(Quick Reference)
The Grails Framework - Reference Documentation
Authors: Graeme Rocher, Peter Ledbrook, Marc Palmer, Jeff Brown, Luke Daley, Burt Beckwith, Lari Hotari
Version: 2.5.5
1 Introduction
Java web development as it stands today is dramatically more complicated than it needs to be. Most modern web frameworks in the Java space are over complicated and don't embrace the Don't Repeat Yourself (DRY) principles.
Dynamic frameworks like Rails, Django and TurboGears helped pave the way to a more modern way of thinking about web applications. Grails builds on these concepts and dramatically reduces the complexity of building web applications on the Java platform. What makes it different, however, is that it does so by building on already established Java technologies like Spring and Hibernate.
Grails is a full stack framework and attempts to solve as many pieces of the web development puzzle through the core technology and its associated plugins. Included out the box are things like:
- An easy to use Object Relational Mapping (ORM) layer built on Hibernate
- An expressive view technology called Groovy Server Pages (GSP)
- A controller layer built on Spring MVC
- A command line scripting environment built on the Groovy-powered Gant
- An embedded Tomcat container which is configured for on the fly reloading
- Dependency injection with the inbuilt Spring container
- Support for internationalization (i18n) built on Spring's core MessageSource concept
- A transactional service layer built on Spring's transaction abstraction
All of these are made easy to use through the power of the
Groovy language and the extensive use of Domain Specific Languages (DSLs)
This documentation will take you through getting started with Grails and building web applications with the Grails framework.
1.1 What's new in Grails 2.5?
Grails 2.5.x is a small release that brings Grails up-to-date with the latest third-party dependencies including:
- Groovy 2.4.x
- Spring 4.1.x
- Spring Loaded 1.2.4
In addition Grails 2.5.x includes many small maintenance fixes to the platform.
1.2 What's new in Grails 2.4?
Groovy 2.3
Grails 2.4 comes with Groovy 2.3 which includes many new features and enhancements.
For more information on Groovy 2.3, see the
comprehensive release notes.
Spring 4.0
Grails 2.4 comes with Spring 4.0.4 which includes many new features and enhancements. See the
Spring documentation.
Hibernate 4.3
Grails 2.4 now uses Hibernate 4.3.5 by default (Hibernate 3 is still available as an optional install).
Standalone GORM and GSP
GORM and GSP can now be used outside of Grails. See the following guides / examples for more information:
The Asset-Pipeline replaces Resources to serve static assets.
The asset-pipeline provides a new, easier to manage, faster means of managing your JavaScript, CSS, and images, while also bringing compiled client languages in to the fray as first-class citizens (e.g. CoffeeScript, LESS, SASS).
All your assets should now live in the
grails-app/assets
subfolders. Three folders are made for you by default:
- javascript
- stylesheets
- images
Now, defining manifests are done directly in your JavaScript files, or CSS by using require directives!
//= require jquery
//= require_self
//= require file_a
//= require_tree .console.log('some javascript');
Easily add your assets to your GSP files:
<asset:javascript src="application.js"/>
<asset:stylesheet href="application.css"/>
<asset:image src="grails_logo.png" height="60" />
Enjoy developing with on the fly asset processing, asset compiling on WAR, and much more. See the
docs for more info.
Static Compilation
Groovy is a dynamically dispatched, dynamically typed language by default but also has great support for static type checking and static compilation. See
these notes on Groovy static compilation. In general, Grails supports Groovy's static compilation but there are a lot of special situations which are common in a Grails app which cannot be statically compiled. For example, if a method marked with
@CompileStatic
contains code which invokes a GORM dynamic finder the code will not compile because the Groovy compiler cannot verify that the dynamic finder is valid. Grails 2.4 improves on this by allowing code to be statically compiled and still do things like invoke GORM dynamic finders.
The
grails.compiler.GrailsCompileStatic annotation behaves much like the
groovy.transform.CompileStatic annotation and provides special handling to recognize Grails specific constructs.
The following controller is marked with
@GrailsCompileStatic
. All of the code that can be statically compiled will be statically compiled. When the compiler encounters code which can not be statically validated, normally that would result in a compile error. The Grails compiler will allow certain things to be considered valid and dynamically dispatch those instructions.
// grails-app/controllers/com/demo/PersonController.groovy
package com.demoimport grails.compiler.GrailsCompileStatic@GrailsCompileStatic
class PersonController { def showKids() {
def kids = Person.findAllByAgeLessThan(16) // …
}
}
There may be situations where most of the code in a class should be statically compiled but a specific method should be left to dynamic compilation. See the following example.
import grails.compiler.GrailsCompileStatic
import groovy.transform.TypeCheckingMode@GrailsCompileStatic
class SomeClass { def update() {
// this method will be statically compiled
} @GrailsCompileStatic(TypeCheckingMode.SKIP)
def save() {
// this method will not be statically compiled
} def delete() {
// this method will be statically compiled
}
}
The
grails.compiler.GrailsTypeChecked annotation behaves much like the
groovy.transform.TypeChecked annotation and provides special handling to recognize Grails specific constructs.
See the
static compilation and type checking section for more details.
More Advanced Subqueries in GORM
The support for subqueries has been extended. You can now use
in
with nested subqueries:
def results = Person.where {
firstName in where { age < 18 }.firstName
}.list()
Criteria and where queries can be seamlessly mixed:
def results = Person.withCriteria {
notIn "firstName", Person.where { age < 18 }.firstName
}
Subqueries can be used with projections:
def results = Person.where {
age > where { age > 18 }.avg('age')
}
Correlated queries that span two domain classes can be used:
def employees = Employee.where {
region.continent in ['APAC', "EMEA"]
}.id()def results = Sale.where {
employee in employees && total > 100000
}.employee.list()
And support for aliases (cross query references) using simple variable declarations has been added to where queries:
def query = Employee.where {
def em1 = Employee
exists Sale.where {
def s1 = Sale
def em2 = employee
return em2.id == em1.id
}.id()
}
def results = query.list()
GORM for Hibernate in Unit tests
It is no longer necessary to create integration tests in order to test GORM interactions with Hibernate. You can now instead use
HibernateTestMixin
:
import grails.test.mixin.TestMixin
import grails.test.mixin.gorm.Domain
import grails.test.mixin.hibernate.HibernateTestMixin
import spock.lang.Specification
@Domain(Person)
@TestMixin(HibernateTestMixin)
class PersonSpec extends Specification { void "Test count people"() {
expect: "Test execute Hibernate count query"
Person.count() == 0
sessionFactory != null
transactionManager != null
session != null
}
}
This library dependency is required in grails-app/conf/BuildConfig.groovy for adding support for HibernateTestMixin
dependencies {
test "org.grails:grails-datastore-test-support:1.0-grails-2.4"
}
HibernateTestMixin is only supported with hibernate4 plugin versions >= 4.3.5.4 .
plugins {
runtime ':hibernate4:4.3.5.4'
}
Views For Namespaced Controllers
The views for namespaced controllers may now be defined in the
grails-app/views/<namespace name>/<controller name>/
directory. See the
Models And Views section for more details.
Improved Programmatic Transactions
Transaction attributes may now be specified when invoking
withTransaction
.
// the keys in the Map must correspond to properties
// of org.springframework.transaction.support.DefaultTransactionDefinitionAccount.withTransaction([propagationBehavior: TransactionDefinition.PROPAGATION_REQUIRES_NEW,
isolationLevel: TransactionDefinition.ISOLATION_REPEATABLE_READ]) {
// …
}
See the
withTransaction docs for more information.
New Maven Plugin
The Maven plugin has been rewritten to use
Aether for dependency resolution and can now be used with both Grails 2.3.x and Grails 2.4.x without releasing a new version of the plugin.
This means that the Maven plugin version number is no longer tied to the version number of Grails and new releases of the Maven plugin will not come out with each new Grails release. Instead, users can continue to use the 2.4.0 version of the plugin for any version of Grails going forward.
Unit Testing improvements
There is a Grails "unit testing runtime" that is based on the previous TestMixin based solution. It now separates the TestMixin classes and the actual runtime that handles the lifecycle of the Grails unit testing runtime. State of the runtime is not kept in static fields of the TestMixin classes anymore. The Groovy AST transformation behind the TestMixin annotation integrates to JUnit and Spock test classes by adding JUnit Rule fields to the class. In the previous solution, Before/BeforeClass and After/AfterClass annotations on AST added mix-in methods were used for the integration.
Some of the main features:
- The programming model remains the same for unit testing of Grails applications
- Setup/teardown method ordering is now deterministic because the integration is now using a single JUnit Rule field and the test runtime uses eventing internally to setup and teardown resources
- There are doWithSpring and doWithConfig callbacks for unit tests - these callback methods get called before the grailsApplication instance in the unit test runtime gets initialized.
- It's possible to register a Spock Mock as a bean to the application context of the Grails unit test runtime application - you can replace a collaborator bean with a mock
- It's possible to reuse a single application context for several test classes and control that so that tests can be made faster when required
- The Grails unit testing runtime has an event-based plugin architecture. It's possible to add new test runtime "features" with new test runtime plugin classes. The test runtime plugin API is due to change. Changes will be made based on feedback from the Grails community. The main interfaces of the API are currently documented in the javadocs: TestPlugin, TestEventInterceptor and TestEvent. Custom test plugins are currently limited since there isn't a solution for scanning for available test plugins. It's now possible to add custom test plugins in a static initialization block of a test class by calling TestRuntimeFactory.addPluginClass .
See the updated
unit testing chapter in the manual for more information of the new features like doWithSpring and doWithConfig.
Improved Unit Testing Support For allowedMethods
The allowedMethods property is now respected in unit tests.
// grails-app/controllers/com/demo/DemoController.groovypackage com.democlass DemoController { static allowedMethods = [save: 'POST', update: 'PUT', delete: 'DELETE'] def save() {
render 'Save was successful!'
} // …
}
// test/unit/com/demo/DemoControllerSpec.groovy
package com.demoimport grails.test.mixin.TestFor
import spock.lang.Specification
import static javax.servlet.http.HttpServletResponse.*@TestFor(DemoController)
class DemoControllerSpec extends Specification { void "test a valid request method"() {
when:
request.method = 'POST'
controller.save() then:
response.status == SC_OK
response.text == 'Save was successful!'
} void "test an invalid request method"() {
when:
request.method = 'DELETE'
controller.save() then:
response.status == SC_METHOD_NOT_ALLOWED
}
}
1.3 What's new in Grails 2.3?
Improved Dependency Management
The default dependency resolution engine used by Grails has been changed to
Aether, the dependency resolution engine used by Maven. Which engine you use can be configured in
BuildConfig
:
grails.project.dependency.resolver = "maven" // or ivy
Using Aether dependency resolution in Grails results in the same behavior as when using the Maven build tool, meaning improved snapshot handling, understanding of custom packaging types and so on.
In addition, the
dependency-report command has been updated to print the dependency graph of the console, which helps in diagnosing dependency resolution failures. See the chapter on
Dependency Resolution for more information.
Data Binder
Grails 2.3 includes a new data binding mechanism which is more flexible and easier to maintain than the data binder used in previous versions. The new data binder includes numerous enhancements including:
See the
Data Binding section for details.
The legacy data binder may be used by assigning
true
to the
grails.databinding.useSpringBinder
property in
grails-app/conf/Config.groovy
. Note that the legacy binder does not support any of the new features provided by the new data binder.
Binding Request Body To Command Objects
If a request is made to a controller action which accepts a command object and the request includes a body, the body will be parsed and used to do data binding to the command object. This simplifies use cases where a request includes a JSON or XML body (for example) that can be bound to a command object. See the
Command Objects documentation for more details.
Domain Classes As Command Objects
When a domain class is used as a command object and there is an
id
request parameter, the framework will retrieve the instance of the domain class from the database using the
id
request parameter. See the
Command Objects documentation for more details.
Forked Execution
All major commands can now be forked into a separate JVM, thus isolating the build path from the runtime / test paths. Forked execution can be controlled via the
BuildConfig
:
grails.project.fork = [
test: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, daemon:true], // configure settings for the test-app JVM
run: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256], // configure settings for the run-app JVM
war: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256], // configure settings for the run-war JVM
console: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256]// configure settings for the Console UI JVM
]
See the documentation on
Forked Mode for more information.
Test Runner Daemon
To speed up testing when using forked execution a new daemon will start-up in the background to run tests when using interactive mode. You can restart the daemon with the
restart-daemon
command from interactive mode:
Server-Side REST Improvements
Grails'
REST support has been significantly improved with the addition of the following features:
- Rich REST URL Mapping support with supports for resource mappings, singular resource mappings, nested resources, versioning and more
- New extensible response rendering and binding APIs
- Support for HAL, Atom and Hypermedia (HATEAOS)
- Scaffolding for REST controllers
See the
user guide for more information.
New Scaffolding 2.0 Plugin
Grails' Scaffolding feature has been split into a
separate plugin. Version 2.0 of the plugin includes support for generating REST controllers, Async controllers, and Spock unit tests.
URL Mappings May Specify A Redirect
URL Mappings may now specify that a redirect should be triggered when the mapping matches an incoming request:
class UrlMappings {
static mappings = {
"/viewBooks"(redirect: '/books/list')
"/viewAuthors"(redirect: [controller: 'author', action: 'list'])
"/viewPublishers"(redirect: [controller: 'publisher', action: 'list', permanent: true]) // …
}
}
See the
user guide for more information.
Async support
Grails 2.3 features new
Asynchronous Programming APIs that allow for asynchronous processing of requests and integrate seamlessly with GORM. Example:
import static grails.async.Promises.*
…
def index() {
tasks books: Book.async.list(),
totalBooks: Book.async.count(),
otherValue: {
// do hard work
}
}
See
the documentation for further information.
Encoding / Escaping Improvements
Grails 2.3 features dedicated support for
Cross Site Scripting (XSS) prevention, including :
- Defaulting to HTML escaping all GSP expressions and scriptlets
- Context sensitive encoding switching for tags
- Double encoding prevention
- Optional automatic encoding of all data in a GSP page not considered safe
See the documentation on
Cross Site Scripting (XSS) prevention for more information.
Hibernate 3 and 4 support
The GORM for Hibernate 3 support for Grails has been extracted into a separate project, allowing new support for Hibernate 4 as a separate plugin.
Controller Exception Handling
Controllers may define exception handler methods which will automatically be invoked any time an action in that controller throws an exception.
// grails-app/controllers/demo/DemoController.groovy
package democlass DemoController { def someAction() {
// do some work
} def handleSQLException(SQLException e) {
render 'A SQLException Was Handled'
} def handleBatchUpdateException(BatchUpdateException e) {
redirect controller: 'logging', action: 'batchProblem'
} def handleNumberFormatException(NumberFormatException nfe) {
[problemDescription: 'A Number Was Invalid']
}
}
See the
controller exception handling docs for more information.
Namespaced Controllers
Controllers may now be defined in a namespace which allows for multiple controllers to be defined with the same name in different packages.
// grails-app/controllers/com/app/reporting/AdminController.groovy
package com.app.reportingclass AdminController { static namespace = 'reports' // …
}
// grails-app/controllers/com/app/security/AdminController.groovy
package com.app.securityclass AdminController { static namespace = 'users' // …
}
// grails-app/conf/UrlMappings.groovy
class UrlMappings { static mappings = {
'/userAdmin' {
controller = 'admin'
namespace = 'users'
} '/reportAdmin' {
controller = 'admin'
namespace = 'reports'
} "/$namespace/$controller/$action?"()
}
}
<g:link controller="admin" namespace="reports">Click For Report Admin</g:link>
<g:link controller="admin" namespace="users">Click For User Admin</g:link>
See the
namespaced controllers docs for more information.
Command Line
The
create-app
command will now by default generate the command line grailsw wrapper for newly created applications. The
--skip-wrapper
switch may be used to prevent the wrapper from being generated.
grails create-app appname --skip-wrapper
1.4 What's new in Grails 2.2?
Namespace Support
Grails 2.2 includes improved support for managing naming conflicts between artifacts provided by an application and its plugins.
Bean names for Service artifacts provided by a plugin are now prefixed with the plugin name. For example, if a Service named
com.publishing.AuthorService
is provided by
a plugin named
PublishingUtilities
and another Service named
com.bookutils.AuthorService
is provided by a plugin named
BookUtilities
, the bean names for those services
will be
publishingUtilitiesAuthorService
and
bookUtilitiesAuthorService
respectively. If a plugin provides a Service that does not have a name which conflicts with any
other Service, then a bean alias will automatically be created that does not contain the prefix and the alias will refer to the bean referenced by the prefixed name. Service
artifacts provided directly by the application will have no prefix added to the relevant bean name. See the
dependency injection and services docs.
Domain classes provided by a plugin will have their default database table name prefixed with the plugin name if the
grails.gorm.table.prefix.enabled
config property is
set to
true
. For example, if the
PublishingUtilities
plugin provides a domain class named
Book
, the default table name for that domain class will be
PUBLISHING_UTILITIES_BOOK
if the
grails.gorm.table.prefix.enabled
config property is set to
true
.
URL Mappings may now include a
plugin
attribute to indicate that the controller referenced in the mapping is provided by a particular plugin.
static mappings = { // requests to /bookAuthors will be handled by the
// AuthorController provided by the BookUtilities plugin
"/bookAuthors" {
controller = 'author'
plugin = 'bookUtilities'
} // requests to /publishingAuthors will be handled by the
// AuthorController provided by the Publishing plugin
"/publishingAuthors" {
controller = 'author'
plugin = 'publishing'
}
}
See the
namespaced controllers docs for more information.
Controller methods and GSP Tags which accept a controller name as a parameter now support an optional parameter indicating
that the controller is provided by a specific plugin.
<g:link controller="user" plugin="springSecurity">Manage Users</g:link>
class DemoController {
def index() {
redirect controller: 'user', action: 'list', plugin: 'springSecurity'
}
}
Forked Tomcat Execution
Grails 2.2 supports forked JVM execution of the Tomcat container in development mode. This has several benefits including:
- Reduced memory consumption, since the Grails build system can exit
- Isolation of the build classpath from the runtime classpath
- The ability to deploy other Grails/Spring applications in parallel without conflicting dependencies
See the
documentation on using forked mode for more information.
SQL Projections In Criteria Queries
Grails 2.2 adds new functionality to criteria queries to provide access to Hibernate's SQL projection API.
// Use SQL projections to retrieve the perimeter and area of all of the Box instances…
def c = Box.createCriteria()def results = c.list {
projections {
sqlProjection '(2 * (width + height)) as perimeter, (width * height) as area', ['perimeter', 'area'], [INTEGER, INTEGER]
}
}
See the
Criteria section for more information.
Groovy 2
Grails 2.2 ships with Groovy 2.0, which has a
bunch of new features itself.
1.5 What's new in Grails 2.1?
Maven Improvements / Multi Module Build Support
Grails' Maven support has been improved in a number of significant ways. Firstly it is now possible to specify plugins within your
pom.xml
file:
<dependency>
<groupId>org.grails.plugins</groupId>
<artifactId>hibernate</artifactId>
<version>2.1.0</version>
<type>zip</type>
<scope>compile</scope>
</dependency>
The Maven plugin now resolves plugins as well as jar dependencies (previously jar dependencies were resolved by Maven and plugins by Ivy). Ivy is completely disabled leaving all dependency resolution up to Maven ensuring that evictions work as expected.
There is also a new Grails
create-multi-project-build
script which features initial support for Maven (Gradle coming in a future release). This script can be run from a parent directory containing Grails applications and plugins and it will generate a Maven multi-module build.
Enabling Maven in a project has been made easier with the inclusion of the
create-pom
command:
grails create-app myapp
cd myapp
grails create-pom com.mycompany
mvn package
To create a multi-module Maven build follow these steps:
grails create-app myapp
grails create-plugin plugin-a
grails create-plugin plugin-b
grails create-multi-project-build com.mycompany:parent:1.0-SNAPSHOT
mvn install
Grails Wrapper
The Grails Wrapper allows a Grails application to build without having to install Grails and configure a GRAILS_HOME environment variable. The wrapper includes a small shell script and a couple of small bootstrap jar files that typically would be checked in to source code control along with the rest of the project. The first time the wrapper is executed it will download and configure a Grails installation. This wrapper makes it more simple to setup a development environment, configure CI and manage upgrades to future versions of Grails. When the application is upgraded to the next version of Grails, the wrapper is updated and checked in to the source code control system and the next time developers update their workspace and run the wrapper, they will automatically be using the correct version of Grails.
See the
Wrapper Documentation for more details.
Debug Option
The
grails
command now supports a
-debug
option which will startup the remote debug agent. This behavior used to be provided by the
grails-debug
command.
grails-debug
is still available but is deprecated and may be removed from a future release.
Grails Command Aliases
The
alias
command may be used to define aliases for grails commands.
The following command creates an alias named
rit
(short for "run integration tests"):
grails alias rit test-app integration:
See the
alias docs for more info.
Cache Plugin
Grails 2.1 installs the
cache plugin by default. This plugin provides powerful and easy to use cache functionality to applications and plugins. The main plugin provides basic map backed caching support. For more robust caching options one of the implementation plugins should be installed and configured. See the
cache-redis docs and the
cache-ehcache docs for details.
See
the main plugin documentation for details on how to configure and use the plugin.
New GORM Methods
In Grails 2.1.1 domain classes now have static methods named
first
and
last
to retrieve the first and last instances from the datastore. See the
first and
last documentation for details.
1.6 What's new in Grails 2.0?
This section covers the new features that are present in 2.0 and is broken down into sections covering the build system, core APIs, the web tier, persistence enhancements and improvements in testing. Note there are many more small enhancements and improvements, these sections just cover some of the highlights.
1.6.1 Development Environment Features
Interactive Mode and Console Enhancements
Grails 2.0 features brand new console output that is more concise and user friendly to consume. An example of the new output when running tests can be seen below:
In general Grails makes its best effort to display update information on a single line and only present the information that is crucial. This means that while in previous versions of Grails the
war command produced many lines of output, in Grails 2.0 only 1 line of output is produced:
In addition simply typing 'grails' at the command line activates the new interactive mode which features TAB completion, command history and keeps the JVM running to ensure commands execute much quicker than otherwise
For more information on the new features of the console refer to the section of the user guide that covers the
console and interactive mode.
Reloading Agent
Grails 2.0 reloading mechanism no longer uses class loaders, but instead uses a JVM agent to reload changes to class files. This results in greatly improved reliability when reloading changes and also ensures that the class files stored in disk remain consistent with the class files loaded in memory, which reduces the need to run the
clean command.
New Test Report and Documentation Templates
There are new templates for displaying test results that are clearer and more user friendly than the previous reports:
In addition, the Grails documentation engine has received a facelift with a new template for presenting Grails application and plugin documentation:
See the section on the
documentation engine for more usage info.
Use a TOC for Project Docs
The old documentation engine relied on you putting section numbers into the gdoc filenames. Although convenient, this effectively made it difficult to restructure your user guide by inserting new chapters and sections. In addition, any such restructuring or renaming of section titles resulted in breaking changes to the URLs.
You can now use logical names for your gdoc files and define the structure and section titles in a YAML table-of-contents file, as described in the section on the
documentation engine. The logical names appear in the URLs, so as long as you don't change those, your URLs will always remain the same no matter how much restructuring or changing of titles you do.
Grails 2.0 even provides a
migrate-docs command to aid you in migrating existing gdoc user guides.
Enhanced Error Reporting and Diagnosis
Error reporting and problem diagnosis has been greatly improved with a new errors view that analyses stack traces and recursively displays problem areas in your code:
In addition stack trace filtering has been further enhanced to display only relevant trace information:
Line | Method
->> 9 | getValue in Book.groovy
- - - - - - - - - - - - - - - - - - - - - - - - -
| 7 | getBookValue in BookService.groovy
| 886 | runTask . . in ThreadPoolExecutor.java
| 908 | run in ''
^ 662 | run . . . . in Thread.java
H2 Database and Console
Grails 2.0 now uses the H2 database instead of HSQLDB, and enables the H2 database console in development mode (at the URI /dbconsole) so that the in-memory database can be easily queried from the browser:
Plugin Usage Tracking
To enhance community awareness of the most popular plugins an opt-in plugin usage tracking system has been included where users can participate in providing feedback to the plugin community on which plugins are most popular.
This will help drive the roadmap and increase support of key plugins while reducing the need to support older or less popular plugins thus helping plugin development teams focus their efforts.
Dependency Resolution Improvements
There are numerous improvements to dependency resolution handling via Ivy including:
- Grails now makes a best effort to cache the previous resolve and avoid resolving again unless you change
BuildConfig.groovy
.
- Plugins dependencies now appear in the dependency report generated by
grails dependency-report
- Plugins published with the release plugin now publish their transitive plugin dependencies in the generated POM which are later resolved.
- It is now possible to customize the ivy cache directory via
BuildConfig.groovy
grails.project.dependency.resolution = {
cacheDir "target/ivy-cache"
}
- You can change the ivy cache directory for all projects via
settings.groovy
grails.dependency.cache.dir = "${userHome}/.ivy2/cache"
- It is now possible to completely disable resolution from inherited repositories (repositories defined by other plugins):
grails.project.dependency.resolution = { repositories {
inherits false // Whether to inherit repository definitions from plugins
…
}
…
}
- It is now possible to easily disable checksum validation errors:
grails.project.dependency.resolution = {
checksums false // whether to verify checksums or not
}
1.6.2 Core Features
Binary Plugins
Grails plugins can now be packaged as JAR files and published to standard maven repositories. This even works for GSP and static resources (with resources plugin 1.0.1). See the section on
Binary plugins for more information.
Groovy 1.8
Grails 2.0 comes with Groovy 1.8 which includes many new
features and enhancementsSpring 3.1 Profile Support
Grails' existing environment support has been bridged into the Spring 3.1 profile support. For example when running with a custom Grails environment called "production", a Spring profile of "production" is activated so that you can use Spring's bean configuration APIs to configure beans for a specific profile.
1.6.3 Web Features
Controller Actions as Methods
It is now possible to define controller actions as methods instead of using closures as in previous versions of Grails. In fact this is now the preferred way of expressing an action. For example:
// action as a method
def index() {}
// action as a closure
def index = {}
Binding Primitive Method Action Arguments
It is now possible to bind form parameters to action arguments where the name of the form element matches the argument name. For example given the following form:
<g:form name="myForm" action="save">
<input name="name" />
<input name="age" />
</g:form>
You can define an action that declares arguments for each input and automatically converts the parameters to the appropriate type:
def save(String name, int age) {
// remaining
}
Static Resource Abstraction
A new
static resource abstraction is included that allows declarative handling of JavaScript, CSS and image resources including automatic ordering, compression, caching and gzip handling.
Servlet 3.0 Async Features
Grails now supports Servlet 3.0 including the Asynchronous programming model defined by the specification:
def index() {
def ctx = startAsync()
ctx.start {
new Book(title:"The Stand").save()
render template:"books", model:[books:Book.list()]
ctx.complete()
}
}
Link Generation API
A general purpose
LinkGenerator
class is now available that is usable anywhere within a Grails application and not just within the context of a controller. For example if you need to generate links in a service or an asynchronous background job outside the scope of a request:
LinkGenerator grailsLinkGeneratordef generateLink() {
grailsLinkGenerator.link(controller:"book", action:"list")
}
Page Rendering API
Like the
LinkGenerator
the new
PageRenderer
can be used to render GSP pages outside the scope of a web request, such as in a scheduled job or web service. The
PageRenderer
class features a very similar API to the
render
method found within controllers:
grails.gsp.PageRenderer groovyPageRenderervoid welcomeUser(User user) {
def contents = groovyPageRenderer.render(view:"/emails/welcomeLetter", model:[user: user])
sendEmail {
to user.email
body contents
}
}
The
PageRenderer
service also allows you to pre-process GSPs into HTML templates:
new File("/path/to/welcome.html").withWriter { w ->
groovyPageRenderer.renderTo(view:"/page/content", w)
}
Filter Exclusions
Filters may now express controller, action and uri exclusions to offer more options for expressing to which requests a particular filter should be applied.
filter1(actionExclude: 'log*') {
before = {
// …
}
}
filter2(controllerExclude: 'auth') {
before = {
// …
}
}filter3(uriExclude: '/secure*') {
before = {
// …
}
}
Performance Improvements
Performance of GSP page rendering has once again been improved by optimizing the GSP compiler to inline method calls where possible.
HTML5 Scaffolding
There is a new HTML5-based scaffolding UI:
jQuery by Default
The jQuery plugin is now the default JavaScript library installed into a Grails application. For backwards compatibility a
Prototype plugin is available. Refer to the
documentation on the Prototype plugin for installation instructions.
Easy Date Parsing
A new
date
method has been added to the
params
object to allow easy, null-safe parsing of dates:
def val = params.date('myDate', 'dd-MM-yyyy')// or a list for formats
def val = params.date('myDate', ['yyyy-MM-dd', 'yyyyMMdd', 'yyMMdd'])// or the format read from messages.properties via the key 'date.myDate.format'
def val = params.date('myDate')
Customizable URL Formats
The default URL Mapping mechanism supports camel case names in the URLs. The default URL for accessing an action named
addNumbers
in a controller named
MathHelperController
would be something like
/mathHelper/addNumbers
. Grails allows for the customization of this pattern and provides an implementation which replaces the camel case convention with a hyphenated convention that would support URLs like
/math-helper/add-numbers
. To enable hyphenated URLs assign a value of "hyphenated" to the
grails.web.url.converter
property in
grails-app/conf/Config.groovy
.
// grails-app/conf/Config.groovygrails.web.url.converter = 'hyphenated'
Arbitrary strategies may be plugged in by providing a class which implements the
UrlConverter interface and adding an instance of that class to the Spring application context with the bean name of
grails.web.UrlConverter.BEAN_NAME
. If Grails finds a bean in the context with that name, it will be used as the default converter and there is no need to assign a value to the
grails.web.url.converter
config property.
// src/groovy/com/myapplication/MyUrlConverterImpl.groovypackage com.myapplicationclass MyUrlConverterImpl implements grails.web.UrlConverter { String toUrlElement(String propertyOrClassName) {
// return some representation of a property or class name that should be used in URLs…
}
}
// grails-app/conf/spring/resources.groovybeans = {
"${grails.web.UrlConverter.BEAN_NAME}"(com.myapplication.MyUrlConverterImpl)
}
Web Flow input and output
It is now possible to provide input arguments when calling a subflow. Flows can also return output values that can be used in a calling flow.
1.6.4 Persistence Features
The GORM API
The GORM API has been formalized into a set of classes (
GormStaticApi
,
GormInstanceApi
and
GormValidationApi
) that get statically wired into every domain class at the byte code level. The result is better code completion for IDEs, better integration with Java and the potential for more GORM implementations for other types of data stores.
Detached Criteria and Where Queries
Grails 2.0 features support for
DetachedCriteria which are criteria queries that are not associated with any session or connection and thus can be more easily reused and composed:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
def results = criteria.list(max:4, sort:"firstName")
To support the addition of
DetachedCriteria queries and encourage their use a new
where
method and DSL has been introduced to greatly reduce the complexity of criteria queries:
def query = Person.where {
(lastName != "Simpson" && firstName != "Fred") || (firstName == "Bart" && age > 9)
}
def results = query.list(sort:"firstName")
See the documentation on
DetachedCriteria and
Where Queries for more information.
New findOrCreate and findOrSave Methods
Domain classes have support for the findOrCreateWhere, findOrSaveWhere, findOrCreateBy and findOrSaveBy query methods which behave just like findWhere and findBy methods except that they should never return null. If a matching instance cannot be found in the database then a new instance is created, populated with values represented in the query parameters and returned. In the case of findOrSaveWhere and findOrSaveBy, the instance is saved before being returned.
def book = Book.findOrCreateWhere(author: 'Douglas Adams', title: "The Hitchhiker's Guide To The Galaxy")
def book = Book.findOrSaveWhere(author: 'Daniel Suarez', title: 'Daemon')
def book = Book.findOrCreateByAuthorAndTitle('Daniel Suarez', 'Daemon')
def book = Book.findOrSaveByAuthorAndTitle('Daniel Suarez', 'Daemon')
Abstract Inheritance
GORM now supports abstract inheritance trees which means you can define queries and associations linking to abstract classes:
abstract class Media {
String title
…
}
class Book extends Media {
}
class Album extends Media {}
class Account {
static hasMany = [purchasedMedia:Media]
}..def allMedia = Media.list()
Multiple Data Sources Support
It is now possible to define multiple datasources in
DataSource.groovy
and declare one or more datasources a particular domain uses by default:
class ZipCode { String code static mapping = {
datasource 'ZIP_CODES'
}
}
If multiple datasources are specified for a domain then you can use the name of a particular datasource as a namespace in front of any regular GORM method:
def zipCode = ZipCode.auditing.get(42)
For more information see the section on
Multiple Data Sources in the user guide.
Database Migrations
A new
database migration plugin has been designed and built for Grails 2.0 allowing you to apply migrations to your database, rollback changes and diff your domain model with the current state of the database.
Database Reverse Engineering
A new
database reverse engineering plugin has been designed and built for Grails 2.0 that allows you to generate a domain model from an existing database schema.
Hibernate 3.6
Grails 2.0 is now built on Hibernate 3.6
Bag Collections
You can now use Hibernate
Bags for mapped collections to avoid the memory and performance issues of loading large collections to enforce
Set
uniqueness or
List
order.
For more information see the section on
Sets, Lists and Maps in the user guide.
1.6.5 Testing Features
New Unit Testing Console Output
Test output from the test-app command has been improved:
New Unit Testing API
There is a new unit testing API based on mixins that supports JUnit 3, 4 and Spock style tests (with Spock 0.6 and above). Example:
import grails.test.mixin.TestFor@TestFor(SimpleController)
class SimpleControllerTests {
void testIndex() {
controller.home() assert view == "/simple/homePage"
assert model.title == "Hello World"
}
}
The
documentation on testing has also been re-written around this new framework.
Unit Testing GORM
A new in-memory GORM implementation is present that supports many more features of the GORM API making unit testing of criteria queries, named queries and other previously unsupported methods possible.
Faster Unit Testing with Interactive Mode
The new interactive mode (activated by typing 'grails') greatly improves the execution time of running unit and integration tests.
Unit Test Scaffolding
A unit test is now generated for scaffolded controllers
2 Getting Started
2.1 Installation Requirements
Before installing Grails you will need as a minimum a Java Development Kit (JDK) installed version 1.6 or above. Download the appropriate JDK for your operating system, run the installer, and then set up an environment variable called
JAVA_HOME
pointing to the location of this installation. If you're unsure how to do this, we recommend the video installation guides from
grailsexample.net:
These will show you how to install Grails too, not just the JDK.
A JDK is required in your Grails development environment. A JRE is not sufficient.
On some platforms (for example OS X) the Java installation is automatically detected. However in many cases you will want to manually configure the location of Java. For example:
export JAVA_HOME=/Library/Java/Home
export PATH="$PATH:$JAVA_HOME/bin"
if you're using bash or another variant of the Bourne Shell.
2.2 Downloading and Installing
The first step to getting up and running with Grails is to install the distribution. To do so follow these steps:
- Download a binary distribution of Grails and extract the resulting zip file to a location of your choice
- Set the GRAILS_HOME environment variable to the location where you extracted the zip
- On Unix/Linux based systems this is typically a matter of adding something like the following
export GRAILS_HOME=/path/to/grails
to your profile
- On Windows this is typically a matter of setting an environment variable under
My Computer/Advanced/Environment Variables
- Then add the
bin
directory to your PATH
variable:
- On Unix/Linux based systems this can be done by adding
export PATH="$PATH:$GRAILS_HOME/bin"
to your profile
- On Windows this is done by modifying the
Path
environment variable under My Computer/Advanced/Environment Variables
If Grails is working correctly you should now be able to type
grails -version
in the terminal window and see output similar to this:
Grails version: 2.0.0
2.3 Creating an Application
To create a Grails application you first need to familiarize yourself with the usage of the
grails
command which is used in the following manner:
Run
create-app to create an application:
grails create-app helloworld
This will create a new directory inside the current one that contains the project. Navigate to this directory in your console:
2.4 A Hello World Example
Let's now take the new project and turn it into the classic "Hello world!" example. First, change into the "helloworld" directory you just created and start the Grails interactive console:
$ cd helloworld
$ grails
You should see a prompt that looks like this:
What we want is a simple page that just prints the message "Hello World!" to the browser. In Grails, whenever you want a new page you just create a new controller action for it. Since we don't yet have a controller, let's create one now with the
create-controller command:
grails> create-controller hello
Don't forget that in the interactive console, we have auto-completion on command names. So you can type "cre" and then press <tab> to get a list of all
create-*
commands. Type a few more letters of the command name and then <tab> again to finish.
The above command will create a new
controller in the
grails-app/controllers/helloworld
directory called
HelloController.groovy
. Why the extra
helloworld
directory? Because in Java land, it's strongly recommended that all classes are placed into packages, so Grails defaults to the application name if you don't provide one. The reference page for
create-controller provides more detail on this.
We now have a controller so let's add an action to generate the "Hello World!" page. The code looks like this:
package helloworldclass HelloController { def index() {
render "Hello World!"
}
}
The action is simply a method. In this particular case, it calls a special method provided by Grails to
render the page.
Job done. To see your application in action, you just need to start up a server with another command called
run-app:
grails> run-app
This will start an embedded server on port 8080 that hosts your application. You should now be able to access your application at the URL
http://localhost:8080/helloworld/ - try it!
If you see the error "Server failed to start for port 8080: Address already in use", then it means another server is running on that port. You can easily work around this by running your server on a different port using -Dserver.port=9090 run-app
. '9090' is just an example: you can pretty much choose anything within the range 1024 to 49151.
The result will look something like this:
This is the Grails intro page which is rendered by the
grails-app/view/index.gsp
file. It detects the presence of your controllers and provides links to them. You can click on the "HelloController" link to see our custom page containing the text "Hello World!". Voila! You have your first working Grails application.
One final thing: a controller can contain many actions, each of which corresponds to a different page (ignoring AJAX at this point). Each page is accessible via a unique URL that is composed from the controller name and the action name: /<appname>/<controller>/<action>. This means you can access the Hello World page via
/helloworld/hello/index, where 'hello' is the controller name (remove the 'Controller' suffix from the class name and lower-case the first letter) and 'index' is the action name. But you can also access the page via the same URL without the action name: this is because 'index' is the
default action . See the end of the
controllers and actions section of the user guide to find out more on default actions.
2.5 Using Interactive Mode
Grails 2.0 features an interactive mode which makes command execution faster since the JVM doesn't have to be restarted for each command. To use interactive mode simple type 'grails' from the root of any projects and use TAB completion to get a list of available commands. See the screenshot below for an example:
For more information on the capabilities of interactive mode refer to the section on
Interactive Mode in the user guide.
2.6 Getting Set Up in an IDE
IntelliJ IDEA
IntelliJ IDEA and the
JetGroovy plugin offer good support for Groovy and Grails developers. Refer to the section on
Groovy and Grails support on the JetBrains website for a feature overview.
IntelliJ IDEA comes in two flavours; the open source "Community Edition" and the commercial "Ultimate Edition".
Both offers support for Groovy, but only Ultimate Edition offers Grails support.
With Ultimate Edition, there is no need to use the
grails integrate-with --intellij
command, as Ultimate Edition understands Grails projects natively. Just open the project with
File -> New Project -> Create project from existing sources
.
You can still use Community Edition for Grails development, but you will miss out on all the Grails specific features like automatic classpath management, GSP editor and quick access to Grails commands.
To integrate Grails with Community Edition run the following command to generate appropriate project files:
grails integrate-with --intellij
Eclipse
We recommend that users of
Eclipse looking to develop Grails application take a look at
Groovy/Grails Tool Suite, which offers built in support for Grails including automatic classpath management, a GSP editor and quick access to Grails commands. See the
STS Integration page for an overview.
NetBeans
NetBeans provides a Groovy/Grails plugin that automatically recognizes Grails projects and provides the ability to run Grails applications in the IDE, code completion and integration with the Glassfish server. For an overview of features see the
NetBeans Integration guide on the Grails website which was written by the NetBeans team.
TextMate
Since Grails' focus is on simplicity it is often possible to utilize more simple editors and
TextMate on the Mac has an excellent Groovy/Grails bundle available from the
TextMate bundles SVN.
To integrate Grails with TextMate run the following command to generate appropriate project files:
grails integrate-with --textmate
Alternatively TextMate can easily open any project with its command line integration by issuing the following command from the root of your project:
2.7 Convention over Configuration
Grails uses "convention over configuration" to configure itself. This typically means that the name and location of files is used instead of explicit configuration, hence you need to familiarize yourself with the directory structure provided by Grails.
Here is a breakdown and links to the relevant sections:
grails-app
- top level directory for Groovy sources
scripts
- Gant scripts.
src
- Supporting sources
groovy
- Other Groovy sources
java
- Other Java sources
test
- Unit and integration tests.
2.8 Running an Application
Grails applications can be run with the built in Tomcat server using the
run-app command which will load a server on port 8080 by default:
You can specify a different port by using the
server.port
argument:
grails -Dserver.port=8090 run-app
Note that it is better to start up the application in interactive mode since a container restart is much quicker:
$ grails
grails> run-app
| Server running. Browse to http://localhost:8080/helloworld
| Application loaded in interactive mode. Type 'stop-app' to shutdown.
| Downloading: plugins-list.xml
grails> stop-app
| Stopping Grails server
grails> run-app
| Server running. Browse to http://localhost:8080/helloworld
| Application loaded in interactive mode. Type 'stop-app' to shutdown.
| Downloading: plugins-list.xml
More information on the
run-app command can be found in the reference guide.
2.9 Testing an Application
The
create-*
commands in Grails automatically create unit or integration tests for you within the
test/unit
or
test/integration
directory. It is of course up to you to populate these tests with valid test logic, information on which can be found in the section on
Testing.
To execute tests you run the
test-app command as follows:
2.10 Deploying an Application
Grails applications are deployed as Web Application Archives (WAR files), and Grails includes the
war command for performing this task:
This will produce a WAR file under the
target
directory which can then be deployed as per your container's instructions.
Unlike most scripts which default to the
development
environment unless overridden, the
war
command runs in the
production
environment by default. You can override this like any script by specifying the environment name, for example:
NEVER deploy Grails using the run-app command as this command sets Grails up for auto-reloading at runtime which has a severe performance and scalability implications
When deploying Grails you should always run your containers JVM with the
-server
option and with sufficient memory allocation. A good set of VM flags would be:
-server -Xmx512M -XX:MaxPermSize=256m
2.11 Supported Java EE Containers
Grails runs on any container that supports Servlet 2.5 and above and is known to work on the following specific container products:
- Tomcat 7
- Tomcat 6
- SpringSource tc Server
- Eclipse Virgo
- GlassFish 3
- GlassFish 2
- Resin 4
- Resin 3
- JBoss 6
- JBoss 5
- Jetty 8
- Jetty 7
- Jetty 6
- Oracle Weblogic 10.3
- Oracle Weblogic 10
- Oracle Weblogic 9
- IBM WebSphere 8.5
- IBM WebSphere 8.0
- IBM WebSphere 7.0
- IBM WebSphere 6.1
It's required to set "-Xverify:none" in "Application servers > server > Process Definition > Java Virtual Machine > Generic JVM arguments" for older versions of WebSphere. This is no longer needed for WebSphere version 8 or newer.
Some containers have bugs however, which in most cases can be worked around. A
list of known deployment issues can be found on the Grails wiki.
2.12 Creating Artefacts
Grails ships with a few convenience targets such as
create-controller,
create-domain-class and so on that will create
Controllers and different artefact types for you.
These are just for your convenience and you can just as easily use an IDE or your favourite text editor.
For example to create the basis of an application you typically need a
domain model:
grails create-app helloworld
cd helloworld
grails create-domain-class book
This will result in the creation of a domain class at
grails-app/domain/helloworld/Book.groovy
such as:
package helloworldclass Book {
}
There are many such
create-*
commands that can be explored in the command line reference guide.
To decrease the amount of time it takes to run Grails scripts, use the interactive mode.
2.13 Generating an Application
To get started quickly with Grails it is often useful to use a feature called
Scaffolding to generate the skeleton of an application. To do this use one of the
generate-*
commands such as
generate-all, which will generate a
controller (and its unit test) and the associated
views:
grails generate-all helloworld.Book
3 Upgrading from Grails 2.2
A number of changes need to be considered when upgrading your application from Grails 2.2, some of them breaking. Here's a quick list with more detail on each item following after:
- New improved data binding (no Spring property editors)
- Much improved XSS prevention with default HTML encoding
- A new dependency resolution engine
- Must be online to fetch Grails dependencies
- Grails core dependencies rearranged
- Tomcat and Hibernate plugins independently versioned now (breaking!)
- Scaffolding is now a separate plugin
- Spock included by default
- Dependency injection does not work in integration tests by default
- Forked execution for tests
- Reloading in
run-app
won't work by default on upgraded apps
grails-debug
doesn't work for forked execution
New Data Binder
There is a new data binding mechanism written from the ground up to meet Grails' needs. If you wish to continue using Spring for data binding then you must set the
grails.databinding.useSpringBinder
property to
true
in
grails-app/conf/Config.groovy
Encoding / Escaping (XSS) Changes
Grails 2.3 includes new features to help prevent XSS attacks. These are enabled by default for new applications, but older applications will require manual intervention. See the section on
Cross Site Scripting (XSS) prevention for how to appropriately configure XSS prevention.
Dependency Resolution changes
Although dependency resolution using Ivy is still supported, the default for Grails 2.3 is to use Aether and the Ivy support will not be improved upon going forward. You may wish to consider using Aether instead for your existing applications by setting the following in
grails-app/conf/BuildConfig.groovy
:
grails.project.dependency.resolver = "maven" // or ivy
If you need to authenticate to a maven repository, you will want to change the definition of that repository like so:
mavenRepo("http://artifactory.mycompany.com/repo") {
authentication(username: "myusername", password: "secret")
}
Dependency Metadata Changes
In addition, the POM and dependency metadata for Grails 2.3 has been re-arranged and cleaned up so that only direct dependencies are specified for an application and all other dependencies are inherited transitively. This has implications to the upgrade since, for example, Ehcache is now a transitive dependency of the Hibernate plugin, whilst before it was a direct dependency. If get a compilation error related to Ehcache, it is most likely that you don't have the Hibernate plugin installed and need to directly declare the Ehcache dependency:
compile "net.sf.ehcache:ehcache:2.8.1"
In addition, excludes may no longer work and may need adjusting when upgrading due to how the metadata has changed. Run the
dependency-report to see the new dependency metadata and make adjustments accordingly.
A common error that may occur when upgrading is:
| Configuring classpath
:: problems summary ::
:::: WARNINGS
::::::::::::::::::::::::::::::::::::::::::::::
:: UNRESOLVED DEPENDENCIES ::
::::::::::::::::::::::::::::::::::::::::::::::
:: org.springframework#spring-test;3.2.2.RELEASE: configuration not found in org.springframework#spring-test;3.2.2.RELEASE: 'compile'. It was required from org.grails#grails-plugin-testing;2.3.0.BUILD-SNAPSHOT compile
::::::::::::::::::::::::::::::::::::::::::::::
This is caused by a plugin that depends on an old version of
spring-test
(for example the
Mail plugin). To correct this run
grails dependency-report
and search for plugins that have a transitive dependency on
spring-test
and exclude them. For example:
plugins {
compile ':mail:1.0', {
excludes 'spring-test'
}
}
However, longer term to solve problems like this we recommend that users move away from Ivy and use Aether instead for dependency resolution:
grails.project.dependency.resolver="maven"
No initial offline mode with Aether
Aether does not support resolving dependencies from a flat file system. This means that the jars we ship with Grails in GRAILS_HOME/lib are not used for the first resolve, but instead the jars are obtained from Maven central. After they have been obtained from Maven central then Aether operates fine offline.
If however you do not have the necessary jars in your local Maven repository, then the only way to get offline execution is to enable Ivy via BuildConfig (see above).
Changes to Core plugin versioning schemes and the Upgrade command
Core plugins like
tomcat
and
hibernate
are no longer versioned the same as the Grails version, instead they are versioned according to the Tomcat and Hibernate version they target. If you are upgrading from Grails 2.2 you need to manually configure the correct Tomcat and Hibernate plugins in
BuildConfig
. The
upgrade
command will not do this for you!
plugins {
// plugins for the build system only
build ':tomcat:7.0.42' // plugins needed at runtime but not for compilation
runtime ':hibernate:3.6.10.2'
}
Note that the
upgrade
command will be deprecated in 2.3 and replaced with a command named
use-current-grails-version
, which will make no attempts to automatically upgrade Grails applications.
Scaffolding moved to a plugin and rewritten
If you have dynamically scaffolded controllers in your application then you will need to configure the 1.0 version of the
Scaffolding plugin in BuildConfig:
plugins {
compile ':scaffolding:1.0.0'
}
By default for new applications the 2.0 version of the scaffolding plugin is used, which is not backwards compatible with 1.0.
Spock included by default
You no longer need to add the Spock plugin to your projects. Simply create Spock specifications as before and they will be run as unit tests. In fact, don't install the Spock plugin, otherwise your specifications will run twice and potentially fail. This also means that the
spock
test type no longer exists. Specifications and JUnit tests run as the same type now.
Dependency Injection for Integration Tests
In order to support alternate JUnit4 test runners, Grails 2.3 no longer uses a special test runner to run tests and integration tests should no longer extend
GroovyTestCase
.
This change requires that any JUnit integration tests that require dependency injection now need to be annotated with:
@TestMixin(IntegrationTestMixin)
For Spock integration tests, extending
IntegrationSpec
also works.
Forked Execution for Testing
Tests are now by default executed in a forked JVM (although this can be disabled). One implication of this is that tests will be slower to execute when using:
The reason for this is the need to load a separate JVM to execute tests. To mitigate this Grails interactive mode has been updated to load a background JVM that can be resumed. If you do:
$ grails // load interactive mode
$ grails -> test-app
$ grails -> test-app
Test execution will be noticeably faster and is the recommended way to run tests in Grails. On older hardware that does not include multiple cores (to run the separate JVMs) it is recommended you disable forked execution for tests to achieve faster test execution times:
forkConfig = [maxMemory: 1024, minMemory: 64, debug: false, maxPerm: 256]
grails.project.fork = [
test: false, // disable forked execution for test-app
run: forkConfig, // configure settings for the run-app JVM
…
]
Forked Execution and the Reloading Agent
In Grails 2.3 the reloading agent is no longer on the build system path unless you pass the
-reloading
flag to the
grails
command:
grails -reloading run-app
The reason for this is that the default in Grails 2.3 and above is to load Grails application in a forked JVM and enable the agent for the forked JVM. If you do not wish to use forked JVMs then you must ensure that you run Grails with the
-reloading
flag. Alternatively, you can enable forking with the following configuration in
BuildConfig
:
forkConfig = [maxMemory: 1024, minMemory: 64, debug: false, maxPerm: 256]
grails.project.fork = [
test: forkConfig, // configure settings for the test-app JVM
run: forkConfig, // configure settings for the run-app JVM
war: forkConfig, // configure settings for the run-war JVM
console: forkConfig // configure settings for the Swing console JVM
]
Forked Execution and Remote Debugging
The
grails-debug
command will no longer work with Grails for remote debugging sessions. The reason is the command enabled debugging for the build system JVM, but not the JVM used in forked execution. The solution to this is to use the
debug-fork
command line argument:
grails --debug-fork run-app
Alternatively you can set the
debug
setting to
true
in
BuildConfig
and use the regular
grails
command to execute:
forkConfig = [maxMemory: 1024, minMemory: 64, debug: true, maxPerm: 256]
grails.project.fork = [
run: forkConfig, // configure settings for the run-app JVM
...
Forked Execution and Functional Test plugins
Some existing plugins (Cucumber plugin for example) do not work with 2.3.x forked execution because they expect the tests to be running in the same JVM as the application under tests. For example it is not possible to setup fixture / test data using GORM inside a functional test and have that data visible to the application under test since the application under test is in a separate JVM. The solution to this is to provide the necessary fixture data in the
BootStrap
of the application (only for the test environment of course).
4 Upgrading from Grails 2.3
The upgrade Command
The
upgrade
command has been removed from Grails 2.4. The procedure for upgrading to the latest version of Grails will be detailed in the user guide from now on. Below are steps that must be taken to upgrade an application from 2.3.x to 2.4.
The set-grails-version Command
The
set-grails-version command should be run to update the application's metadata to indicate which version of Grails the application is built with.
Update to latest Plugin versions
You should update your application's
BuildConfig
to use the latest plugins compatible with Grails 2.3. Example:
plugins {
// plugins for the build system only
build ':tomcat:7.0.52.1' // plugins for the compile step
compile ':scaffolding:2.1.0'
compile ':cache:1.1.3'
compile ':asset-pipeline:1.8.3' // plugins needed at runtime but not for compilation
runtime ':hibernate4:4.3.5.2' // or ':hibernate:3.6.10.14'
runtime ':database-migration:1.4.0'
runtime ':jquery:1.11.0.2'
…
}
You may get compilation errors or incompatibility problems with older versions of the above plugins installed.
grails-debug Script Has Been Removed
The
grails-debug
and
grails-debug.bat
scripts have been removed. To debug the build system JVM run
grails -debug <command>
and to debug the forked JVM run
grails --debug-fork <command>
.
New Command Object Data Binding Behavior
The data binding behavior for command objects has changed in Grails 2.4. Request parameter names may now be prefixed with the name of the controller action argument name that the request parameter should be bound to. For example, if a request is made to the
buy
action in the controller below a request parameter named
buyer.name
will be bound to the
name
property of the
buyer
argument and a request parameter named
seller.name
will be bound to the
name
property of the
seller
argument. See the
Command Objects documentation for more details.
New Behavior For Domain Class Command Objects
If a command object's type is a domain class and there is no
id
request parameter then
null
will be passed into the controller action unless the HTTP request method is "POST", in which case a new instance of the domain class will be created by invoking the domain class constructor. For all of the cases where the domain class instance is non-null, data binding is only performed if the HTTP request method is "POST", "PUT" or "PATCH". See the
Command Objects documentation for more details.
Nullable Command Object Properties
The behavior in Grails 2.3.x is such that constrained properties in command objects and other classes marked with @Validateable are all configured with
nullable: false
by default. Unconstrained properties were not configured with
nullable: false
. In Grails 2.4 all non-static unconstrained properties in command object classes and other classes marked with @Validateable are all configured with
nullable: false
.
class StoreController {
def buy(Person buyer, Person seller) {
// …
}
}class Person {
String name
}
If you wish to retain the old behavior, you can do so on a per-command object basis by using the @Validateable constraint explicitly and passing the
nullable: true
argument:
@Validateable(nullable = true)
class Person {
String name
}
This will cause all properties to be nullable by default unless a constraint is explicitly added (similar to the behavior prior to Grails 2.4).
See the
Command Objects documentation for more details.
Ajax Tags Have Been Deprecated
The
formRemote,
remoteField,
remoteFunction and
remoteLink Ajax tags have been deprecated and will be removed from a future version of Grails. Applications may provide their own Ajax tags and/or Javascript plugins may provide Ajax tags of their own.
The Spring Data Binder Has Been Deprecated
The
grails.databinding.useSpringBinder
config property may be set to
true
to tell Grails to use the Spring data binder instead of the Grails data binder. The Spring data binder has been deprecated and will be removed from a future version of Grails. It is recommended that when upgrading to Grails 2.4 that the Grails data binder be used.
The resources Plugin
As of Grails 2.4 the
resources
plugin has been replaced with the
asset-pipeline
plugin as the default resource management plugin for newly created applications. See the
static resource abstraction section of the User Guide for more details. When upgrading an application to Grails 2.4 if you choose to continue using the resources plugin you will need to use version 1.2.7 or later as previous versions of the plugin are not compatible with Grails 2.4.
Static Holder Classes
The following deprecated classes have been removed from Grails 2.4.x:
org.codehaus.groovy.grails.commons.ApplicationHolder
org.codehaus.groovy.grails.commons.ConfigurationHolder
org.codehaus.groovy.grails.plugins.PluginManagerHolder
org.codehaus.groovy.grails.web.context.ServletContextHolder
org.codehaus.groovy.grails.compiler.support.GrailsResourceLoaderHolder
If you or any plugins you have installed are using these classes you will get a compilation error. The problem can be rectified by updating to new plugins and using
grails.util.Holders
instead.
If your application uses the jquery plugin you will need to update to version 1.11.0.2 or later as previous versions of the plugin made use of the ApplicationHolder
class. If your application uses the resources plugin you will need to update to version 1.2.7 or later as previous versions of the plugin made use of the ConfigurationHolder
class.
Changes To applicationContext.xml
The
web-app/WEB-INF/applicationContext.xml
file contains a bean definition for a
grailsResourceLoader
bean which is an instance of
org.codehaus.groovy.grails.commons.GrailsResourceLoaderFactoryBean
. That bean definition needs to be removed from the file. The
grailsApplication
bean may have the
grailsResourceLoader
bean injected into it as shown below.
<bean id="grailsApplication" class="org.codehaus.groovy.grails.commons.GrailsApplicationFactoryBean">
<description>Grails application factory bean</description>
<property name="grailsDescriptor" value="/WEB-INF/grails.xml" />
<property name="grailsResourceLoader" ref="grailsResourceLoader" />
</bean>
The
grailsApplication
bean definition should be left in the file but the
grailsResourceLoader
bean reference should be removed as shown below.
<bean id="grailsApplication" class="org.codehaus.groovy.grails.commons.GrailsApplicationFactoryBean">
<description>Grails application factory bean</description>
<property name="grailsDescriptor" value="/WEB-INF/grails.xml" />
</bean>
Changes to web.xml
The Sitemesh servlet filter has been removed and the GSP layout feature is now handled by
GrailsLayoutView.
Applications that are using a customized web.xml should apply the customizations to a web.xml file of Grails 2.4 .
This applies only to applications that have used the "install-templates" to install template files in src/templates folder of the application.
It's recommended to rename src/templates to a different name and use a diff tool to apply the possible application specific customizations to the files created with Grails 2.4 install-templates command.
Data Binding Changes
Prior to Grails 2.4 when data binding was performed with the
params
object in a controller, if the request contained a body the body would be parsed and used for data binding instead of the
params
object. In Grails 2.4 this behavior has changed so that if binding is initiated with
params
, the binding will always be done with the
params
object, without regard to whether or not the request has a body. If binding is done with the
request
object, if the request has a body then the body will be parsed and used for data binding, otherwise the request parameters will be used for data binding.
class SomeController { def someAction() {
// Prior to Grails 2.4 if the request contains a body
// then obj1 will be populated with values parsed from
// the body instead of with values in params. // With Grails 2.4 obj1 will be populated with values
// in params.
def obj1 = new SomeDomainClass(params) // the same is true for the following
def obj2 = new SomeDomainClass()
obj2.properties = params
} def someOtherAction() {
// If the request contains a body then obj1 will be
// populated with values parsed from the body, otherwise
// obj1 will be populated with the request parameters. // This is not a new change in behavior.
def obj1 = new SomeDomainClass()
obj1.properties = request
}
}
There is one release in the 2.3.x chain which has the 2.4 behavior described above and that is Grails 2.3.8. None of the 2.3.x releases before or after 2.3.8 have this behavior.
The allowedMethods Property And Unit Tests
The unit testing environment now respects the
allowedMethods property in controllers. Prior to Grails 2.4 a unit test which accessed a controller action which is supposed to be restricted to certain request methods could have skipped the step of setting the request method in the unit test because the allowedMethods property was ignored by the unit test. As of Grails 2.4 if a controller action is limited to be accessed with certain request methods, the unit test must be constructed to deal with that.
// grails-app/controllers/com/demo/DemoController.groovypackage com.democlass DemoController { static allowedMethods = [save: 'POST', update: 'PUT', delete: 'DELETE'] def save() {
render 'Save was successful!'
} // …
}
// test/unit/com/demo/DemoControllerSpec.groovy
package com.demoimport grails.test.mixin.TestFor
import spock.lang.Specification
import static javax.servlet.http.HttpServletResponse.*@TestFor(DemoController)
class DemoControllerSpec extends Specification { void "test a valid request method"() {
when:
request.method = 'POST'
controller.save() then:
response.status == SC_OK
response.text == 'Save was successful!'
} void "test an invalid request method"() {
when:
request.method = 'DELETE'
controller.save() then:
response.status == SC_METHOD_NOT_ALLOWED
}
}
scanning for JSP taglibs has to be configured, no JSTL default dependency
JSP taglib tld files aren't scanned by default any more. This must be configured with the
grails.gsp.tldScanPattern
setting. It accepts a comma separated String value. Spring's PathMatchingResourcePatternResolver is used to resolve the patterns.
You can get the previous behaviour by adding this setting to Config.groovy:
grails.gsp.tldScanPattern='classpath*:/META-INF/*.tld,/WEB-INF/tld/*.tld'
JSTL standard library is no more added as a dependency. In case you are using JSTL, you should also add these dependencies:
runtime 'javax.servlet:jstl:1.1.2'
runtime 'taglibs:standard:1.1.2'
5 Configuration
It may seem odd that in a framework that embraces "convention-over-configuration" that we tackle this topic now. With Grails' default settings you can actually develop an application without doing any configuration whatsoever, as the quick start demonstrates, but it's important to learn where and how to override the conventions when you need to. Later sections of the user guide will mention what configuration settings you can use, but not how to set them. The assumption is that you have at least read the first section of this chapter!
5.1 Basic Configuration
For general configuration Grails provides two files:
grails-app/conf/BuildConfig.groovy
grails-app/conf/Config.groovy
Both of them use Groovy's
ConfigSlurper syntax. The first,
BuildConfig.groovy
, is for settings that are used when running Grails commands, such as
compile
,
doc
, etc. The second file,
Config.groovy
, is for settings that are used when your application is running. This means that
Config.groovy
is packaged with your application, but
BuildConfig.groovy
is not. Don't worry if you're not clear on the distinction: the guide will tell you which file to put a particular setting in.
The most basic syntax is similar to that of Java properties files with dot notation on the left-hand side:
Note that the value is a Groovy string literal! Those quotes around 'world' are important. In fact, this highlights one of the advantages of the ConfigSlurper syntax over properties files: the property values can be any valid Groovy type, such as strings, integers, or arbitrary objects!
Things become more interesting when you have multiple settings with the same base. For example, you could have the two settings
foo.bar.hello = "world"
foo.bar.good = "bye"
both of which have the same base:
foo.bar
. The above syntax works but it's quite repetitive and verbose. You can remove some of that verbosity by nesting properties at the dots:
foo {
bar {
hello = "world"
good = "bye"
}
}
or by only partially nesting them:
foo {
bar.hello = "world"
bar.good = "bye"
}
However, you can't nest after using the dot notation. In other words, this
won't work:
// Won't work!
foo.bar {
hello = "world"
good = "bye"
}
Within both
BuildConfig.groovy
and
Config.groovy
you can access several implicit variables from configuration values:
Variable | Description |
---|
userHome | Location of the home directory for the account that is running the Grails application. |
grailsHome | Location of the directory where you installed Grails. If the GRAILS_HOME environment variable is set, it is used. |
appName | The application name as it appears in application.properties. |
appVersion | The application version as it appears in application.properties. |
For example:
my.tmp.dir = "${userHome}/.grails/tmp"
In addition,
BuildConfig.groovy
has
Variable | Description |
---|
grailsVersion | The version of Grails used to build the project. |
grailsSettings | An object containing various build related settings, such as baseDir . It's of type BuildSettings . |
and
Config.groovy
has
Those are the basics of adding settings to the configuration file, but how do you access those settings from your own application? That depends on which config you want to read.
The settings in
BuildConfig.groovy
are only available from
command scripts and can be accessed via the
grailsSettings.config
property like so:
target(default: "Example command") {
def maxIterations = grailsSettings.config.myapp.iterations.max
…
}
If you want to read runtime configuration settings, i.e. those defined in
Config.groovy
, use the
grailsApplication
object, which is available as a variable in controllers and tag libraries:
class MyController {
def hello() {
def recipient = grailsApplication.config.foo.bar.hello render "Hello ${recipient}"
}
}
and can be easily injected into services and other Grails artifacts:
class MyService {
def grailsApplication String greeting() {
def recipient = grailsApplication.config.foo.bar.hello
return "Hello ${recipient}"
}
}
As you can see, when accessing configuration settings you use the same dot notation as when you define them.
5.1.1 Built in options
Grails has a set of core settings that are worth knowing about. Their defaults are suitable for most projects, but it's important to understand what they do because you may need one or more of them later.
Build settings
Let's start with some important build settings. Although Grails requires JDK 6 when developing your applications, it is possible to deploy those applications to JDK 5 containers. Simply set the following in
BuildConfig.groovy
:
grails.project.source.level = "1.5"
grails.project.target.level = "1.5"
Note that source and target levels are different to the standard public version of JDKs, so JDK 5 -> 1.5, JDK 6 -> 1.6, and JDK 7 -> 1.7.
In addition, Grails supports Servlet versions 2.5 and above but defaults to 2.5. If you wish to use newer features of the Servlet API (such as 3.0 async support) you should configure the
grails.servlet.version
setting appropriately:
grails.servlet.version = "3.0"
Runtime settings
On the runtime front, i.e.
Config.groovy
, there are quite a few more core settings:
grails.config.locations
- The location of properties files or addition Grails Config files that should be merged with main configuration. See the section on externalised config.
grails.enable.native2ascii
- Set this to false if you do not require native2ascii conversion of Grails i18n properties files (default: true).
grails.views.default.codec
- Sets the default encoding regime for GSPs - can be one of 'none', 'html', or 'base64' (default: 'none'). To reduce risk of XSS attacks, set this to 'html'.
grails.views.gsp.encoding
- The file encoding used for GSP source files (default: 'utf-8').
grails.mime.file.extensions
- Whether to use the file extension to dictate the mime type in Content Negotiation (default: true).
grails.mime.types
- A map of supported mime types used for Content Negotiation.
grails.serverURL
- A string specifying the server URL portion of absolute links, including server name e.g. grails.serverURL="http://my.yourportal.com". See createLink. Also used by redirects.
grails.views.gsp.sitemesh.preprocess
- Determines whether SiteMesh preprocessing happens. Disabling this slows down page rendering, but if you need SiteMesh to parse the generated HTML from a GSP view then disabling it is the right option. Don't worry if you don't understand this advanced property: leave it set to true.
grails.reload.excludes
and grails.reload.includes
- Configuring these directives determines the reload behavior for project specific source files. Each directive takes a list of strings that are the class names for project source files that should be excluded from reloading behavior or included accordingly when running the application in development with the run-app
command. If the grails.reload.includes
directive is configured, then only the classes in that list will be reloaded.
War generation
grails.project.war.file
- Sets the name and location of the WAR file generated by the war command
grails.war.dependencies
- A closure containing Ant builder syntax or a list of JAR filenames. Lets you customise what libraries are included in the WAR file.
grails.war.copyToWebApp
- A closure containing Ant builder syntax that is legal inside an Ant copy, for example "fileset()". Lets you control what gets included in the WAR file from the "web-app" directory.
grails.war.resources
- A closure containing Ant builder syntax. Allows the application to do any other work before building the final WAR file
For more information on using these options, see the section on
deployment
5.1.2 Logging
The Basics
Grails uses its common configuration mechanism to provide the settings for the underlying
Log4j log system, so all you have to do is add a
log4j
setting to the file
grails-app/conf/Config.groovy
.
So what does this
log4j
setting look like? Here's a basic example:
log4j = {
error 'org.codehaus.groovy.grails.web.servlet', // controllers
'org.codehaus.groovy.grails.web.pages' // GSP warn 'org.apache.catalina'
}
This says that for loggers whose name starts with 'org.codehaus.groovy.grails.web.servlet' or 'org.codehaus.groovy.grails.web.pages', only messages logged at 'error' level and above will be shown. Loggers with names starting with 'org.apache.catalina' logger only show messages at the 'warn' level and above. What does that mean? First of all, you have to understand how levels work.
Logging levels
There are several standard logging levels, which are listed here in order of descending priority:
- off
- fatal
- error
- warn
- info
- debug
- trace
- all
When you log a message, you implicitly give that message a level. For example, the method
log.error(msg)
will log a message at the 'error' level. Likewise,
log.debug(msg)
will log it at 'debug'. Each of the above levels apart from 'off' and 'all' have a corresponding log method of the same name.
The logging system uses that
message level combined with the configuration for the logger (see next section) to determine whether the message gets written out. For example, if you have an 'org.example.domain' logger configured like so:
warn 'org.example.domain'
then messages with a level of 'warn', 'error', or 'fatal' will be written out. Messages at other levels will be ignored.
Before we go on to loggers, a quick note about those 'off' and 'all' levels. These are special in that they can only be used in the configuration; you can't log messages at these levels. So if you configure a logger with a level of 'off', then no messages will be written out. A level of 'all' means that you will see all messages. Simple.
Loggers
Loggers are fundamental to the logging system, but they are a source of some confusion. For a start, what are they? Are they shared? How do you configure them?
A logger is the object you log messages to, so in the call
log.debug(msg)
,
log
is a logger instance (of type
Log). These loggers are cached and uniquely identified by name, so if two separate classes use loggers with the same name, those loggers are actually the same instance.
There are two main ways to get hold of a logger:
- use the
log
instance injected into artifacts such as domain classes, controllers and services;
- use the Commons Logging API directly.
If you use the dynamic
log
property, then the name of the logger is 'grails.app.<type>.<className>', where
type
is the type of the artifact, for example 'controllers' or 'services', and
className
is the fully qualified name of the artifact. For example, if you have this service:
package org.exampleclass MyService {
…
}
then the name of the logger will be 'grails.app.services.org.example.MyService'.
For other classes, the typical approach is to store a logger based on the class name in a constant static field:
package org.otherimport org.apache.commons.logging.LogFactoryclass MyClass {
private static final log = LogFactory.getLog(this)
…
}
This will create a logger with the name 'org.other.MyClass' - note the lack of a 'grails.app.' prefix since the class isn't an artifact. You can also pass a name to the
getLog()
method, such as "myLogger", but this is less common because the logging system treats names with dots ('.') in a special way.
Configuring loggers
You have already seen how to configure loggers in Grails:
log4j = {
error 'org.codehaus.groovy.grails.web.servlet'
}
This example configures loggers with names starting with 'org.codehaus.groovy.grails.web.servlet' to ignore any messages sent to them at a level of 'warn' or lower. But is there a logger with this name in the application? No. So why have a configuration for it? Because the above rule applies to any logger whose name
begins with 'org.codehaus.groovy.grails.web.servlet.' as well. For example, the rule applies to both the
org.codehaus.groovy.grails.web.servlet.GrailsDispatcherServlet
class and the
org.codehaus.groovy.grails.web.servlet.mvc.GrailsWebRequest
one.
In other words, loggers are hierarchical. This makes configuring them by package much simpler than it would otherwise be.
The most common things that you will want to capture log output from are your controllers, services, and other artifacts. Use the convention mentioned earlier to do that:
grails.app.<artifactType>.<className> . In particular the class name must be fully qualified, i.e. with the package if there is one:
log4j = {
// Set level for all application artifacts
info "grails.app" // Set for a specific controller in the default package
debug "grails.app.controllers.YourController" // Set for a specific domain class
debug "grails.app.domain.org.example.Book" // Set for all taglibs
info "grails.app.taglib"
}
The standard artifact names used in the logging configuration are:
conf
- For anything under grails-app/conf
such as BootStrap.groovy
(but excluding filters)
filters
- For filters
taglib
- For tag libraries
services
- For service classes
controllers
- For controllers
domain
- For domain entities
Grails itself generates plenty of logging information and it can sometimes be helpful to see that. Here are some useful loggers from Grails internals that you can use, especially when tracking down problems with your application:
org.codehaus.groovy.grails.commons
- Core artifact information such as class loading etc.
org.codehaus.groovy.grails.web
- Grails web request processing
org.codehaus.groovy.grails.web.mapping
- URL mapping debugging
org.codehaus.groovy.grails.plugins
- Log plugin activity
grails.spring
- See what Spring beans Grails and plugins are defining
org.springframework
- See what Spring is doing
org.hibernate
- See what Hibernate is doing
So far, we've only looked at explicit configuration of loggers. But what about all those loggers that
don't have an explicit configuration? Are they simply ignored? The answer lies with the root logger.
The Root Logger
All logger objects inherit their configuration from the root logger, so if no explicit configuration is provided for a given logger, then any messages that go to that logger are subject to the rules defined for the root logger. In other words, the root logger provides the default configuration for the logging system.
Grails automatically configures the root logger to only handle messages at 'error' level and above, and all the messages are directed to the console (stdout for those with a C background). You can customise this behaviour by specifying a 'root' section in your logging configuration like so:
log4j = {
root {
info()
}
…
}
The above example configures the root logger to log messages at 'info' level and above to the default console appender. You can also configure the root logger to log to one or more named appenders (which we'll talk more about shortly):
log4j = {
appenders {
file name:'file', file:'/var/logs/mylog.log'
}
root {
debug 'stdout', 'file'
}
}
In the above example, the root logger will log to two appenders - the default 'stdout' (console) appender and a custom 'file' appender.
For power users there is an alternative syntax for configuring the root logger: the root
org.apache.log4j.Logger
instance is passed as an argument to the log4j closure. This lets you work with the logger directly:
log4j = { root ->
root.level = org.apache.log4j.Level.DEBUG
…
}
For more information on what you can do with this
Logger
instance, refer to the Log4j API documentation.
Those are the basics of logging pretty well covered and they are sufficient if you're happy to only send log messages to the console. But what if you want to send them to a file? How do you make sure that messages from a particular logger go to a file but not the console? These questions and more will be answered as we look into appenders.
Appenders
Loggers are a useful mechanism for filtering messages, but they don't physically write the messages anywhere. That's the job of the appender, of which there are various types. For example, there is the default one that writes messages to the console, another that writes them to a file, and several others. You can even create your own appender implementations!
This diagram shows how they fit into the logging pipeline:
As you can see, a single logger may have several appenders attached to it. In a standard Grails configuration, the console appender named 'stdout' is attached to all loggers through the default root logger configuration. But that's the only one. Adding more appenders can be done within an 'appenders' block:
log4j = {
appenders {
rollingFile name: "myAppender",
maxFileSize: 1024,
file: "/tmp/logs/myApp.log"
}
}
The following appenders are available by default:
Each named argument passed to an appender maps to a property of the underlying
Appender implementation. So the previous example sets the
name
,
maxFileSize
and
file
properties of the
RollingFileAppender
instance.
You can have as many appenders as you like - just make sure that they all have unique names. You can even have multiple instances of the same appender type, for example several file appenders that log to different files.
If you prefer to create the appender programmatically or if you want to use an appender implementation that's not available in the above syntax, simply declare an
appender
entry with an instance of the appender you want:
import org.apache.log4j.*log4j = {
appenders {
appender new RollingFileAppender(
name: "myAppender",
maxFileSize: 1024,
file: "/tmp/logs/myApp.log")
}
}
This approach can be used to configure
JMSAppender
,
SocketAppender
,
SMTPAppender
, and more.
Once you have declared your extra appenders, you can attach them to specific loggers by passing the name as a key to one of the log level methods from the previous section:
error myAppender: "grails.app.controllers.BookController"
This will ensure that the 'grails.app.controllers.BookController' logger sends log messages to 'myAppender' as well as any appenders configured for the root logger. To add more than one appender to the logger, then add them to the same level declaration:
error myAppender: "grails.app.controllers.BookController",
myFileAppender: ["grails.app.controllers.BookController",
"grails.app.services.BookService"],
rollingFile: "grails.app.controllers.BookController"
The above example also shows how you can configure more than one logger at a time for a given appender (
myFileAppender
) by using a list.
Be aware that you can only configure a single level for a logger, so if you tried this code:
error myAppender: "grails.app.controllers.BookController"
debug myFileAppender: "grails.app.controllers.BookController"
fatal rollingFile: "grails.app.controllers.BookController"
you'd find that only 'fatal' level messages get logged for 'grails.app.controllers.BookController'. That's because the last level declared for a given logger wins. What you probably want to do is limit what level of messages an appender writes.
An appender that is attached to a logger configured with the 'all' level will generate a lot of logging information. That may be fine in a file, but it makes working at the console difficult. So we configure the console appender to only write out messages at 'info' level or above:
log4j = {
appenders {
console name: "stdout", threshold: org.apache.log4j.Level.INFO
}
}
The key here is the
threshold
argument which determines the cut-off for log messages. This argument is available for all appenders, but do note that you currently have to specify a
Level
instance - a string such as "info" will not work.
Custom Layouts
By default the Log4j DSL assumes that you want to use a
PatternLayout. However, there are other layouts available including:
xml
- Create an XML log file
html
- Creates an HTML log file
simple
- A simple textual log
pattern
- A Pattern layout
You can specify custom patterns to an appender using the
layout
setting:
log4j = {
appenders {
console name: "customAppender",
layout: pattern(conversionPattern: "%c{2} %m%n")
}
}
This also works for the built-in appender "stdout", which logs to the console:
log4j = {
appenders {
console name: "stdout",
layout: pattern(conversionPattern: "%c{2} %m%n")
}
}
Environment-specific configuration
Since the logging configuration is inside
Config.groovy
, you can put it inside an environment-specific block. However, there is a problem with this approach: you have to provide the full logging configuration each time you define the
log4j
setting. In other words, you cannot selectively override parts of the configuration - it's all or nothing.
To get around this, the logging DSL provides its own environment blocks that you can put anywhere in the configuration:
log4j = {
appenders {
console name: "stdout",
layout: pattern(conversionPattern: "%c{2} %m%n") environments {
production {
rollingFile name: "myAppender", maxFileSize: 1024,
file: "/tmp/logs/myApp.log"
}
}
} root {
//…
} // other shared config
info "grails.app.controller" environments {
production {
// Override previous setting for 'grails.app.controller'
error "grails.app.controllers"
}
}
}
The one place you can't put an environment block is
inside the
root
definition, but you can put the
root
definition inside an environment block.
Full stacktraces
When exceptions occur, there can be an awful lot of noise in the stacktrace from Java and Groovy internals. Grails filters these typically irrelevant details and restricts traces to non-core Grails/Groovy class packages.
When this happens, the full trace is always logged to the
StackTrace
logger, which by default writes its output to a file called
stacktrace.log
. As with other loggers though, you can change its behaviour in the configuration. For example if you prefer full stack traces to go to the console, add this entry:
error stdout: "StackTrace"
This won't stop Grails from attempting to create the stacktrace.log file - it just redirects where stack traces are written to. An alternative approach is to change the location of the 'stacktrace' appender's file:
log4j = {
appenders {
rollingFile name: "stacktrace", maxFileSize: 1024,
file: "/var/tmp/logs/myApp-stacktrace.log"
}
}
or, if you don't want to the 'stacktrace' appender at all, configure it as a 'null' appender:
log4j = {
appenders {
'null' name: "stacktrace"
}
}
You can of course combine this with attaching the 'stdout' appender to the 'StackTrace' logger if you want all the output in the console.
Finally, you can completely disable stacktrace filtering by setting the
grails.full.stacktrace
VM property to
true
:
grails -Dgrails.full.stacktrace=true run-app
Masking Request Parameters From Stacktrace Logs
When Grails logs a stacktrace, the log message may include the names and values of all of the request parameters for the current request. To mask out the values of secure request parameters, specify the parameter names in the
grails.exceptionresolver.params.exclude
config property:
grails.exceptionresolver.params.exclude = ['password', 'creditCard']
Request parameter logging may be turned off altogether by setting the
grails.exceptionresolver.logRequestParameters
config property to
false
. The default value is
true
when the application is running in DEVELOPMENT mode and
false
for all other modes.
grails.exceptionresolver.logRequestParameters=false
Logger inheritance
Earlier, we mentioned that all loggers inherit from the root logger and that loggers are hierarchical based on '.'-separated terms. What this means is that unless you override a parent setting, a logger retains the level and the appenders configured for that parent. So with this configuration:
log4j = {
appenders {
file name:'file', file:'/var/logs/mylog.log'
}
root {
debug 'stdout', 'file'
}
}
all loggers in the application will have a level of 'debug' and will log to both the 'stdout' and 'file' appenders. What if you only want to log to 'stdout' for a particular logger? Change the 'additivity' for a logger in that case.
Additivity simply determines whether a logger inherits the configuration from its parent. If additivity is false, then its not inherited. The default for all loggers is true, i.e. they inherit the configuration. So how do you change this setting? Here's an example:
log4j = {
appenders {
…
}
root {
…
} info additivity: false,
stdout: ["grails.app.controllers.BookController",
"grails.app.services.BookService"]
}
So when you specify a log level, add an 'additivity' named argument. Note that you when you specify the additivity, you must configure the loggers for a named appender. The following syntax will
not work:
info additivity: false, ["grails.app.controllers.BookController",
"grails.app.services.BookService"]
Customizing stack trace printing and filtering
Stacktraces in general and those generated when using Groovy in particular are quite verbose and contain many stack frames that aren't interesting when diagnosing problems. So Grails uses a implementation of the
org.codehaus.groovy.grails.exceptions.StackTraceFilterer
interface to filter out irrelevant stack frames. To customize the approach used for filtering, implement that interface in a class in src/groovy or src/java and register it in
Config.groovy
:
grails.logging.stackTraceFiltererClass =
'com.yourcompany.yourapp.MyStackTraceFilterer'
In addition, Grails customizes the display of the filtered stacktrace to make the information more readable. To customize this, implement the
org.codehaus.groovy.grails.exceptions.StackTracePrinter
interface in a class in src/groovy or src/java and register it in
Config.groovy
:
grails.logging.stackTracePrinterClass =
'com.yourcompany.yourapp.MyStackTracePrinter'
Finally, to render error information in the error GSP, an HTML-generating printer implementation is needed. The default implementation is
org.codehaus.groovy.grails.web.errors.ErrorsViewStackTracePrinter
and it's registered as a Spring bean. To use your own implementation, either implement the
org.codehaus.groovy.grails.exceptions.StackTraceFilterer
directly or subclass
ErrorsViewStackTracePrinter
and register it in
grails-app/conf/spring/resources.groovy
as:
import com.yourcompany.yourapp.MyErrorsViewStackTracePrinterbeans = { errorsViewStackTracePrinter(MyErrorsViewStackTracePrinter,
ref('grailsResourceLocator'))
}
Alternative logging libraries
By default, Grails uses Log4J to do its logging. For most people this is absolutely fine, and many users don't even care what logging library is used. But if you're not one of those and want to use an alternative, such as the
JDK logging package or
logback, you can do so by simply excluding a couple of dependencies from the global set and adding your own:
grails.project.dependency.resolution = {
inherits("global") {
excludes "grails-plugin-logging", "log4j"
}
…
dependencies {
runtime "ch.qos.logback:logback-core:0.9.29"
…
}
…
}
If you do this, you will get unfiltered, standard Java stacktraces in your log files and you won't be able to use the logging configuration DSL that's just been described. Instead, you will have to use the standard configuration mechanism for the library you choose.
5.1.3 GORM
Grails provides the following GORM configuration options:
grails.gorm.failOnError
- If set to true
, causes the save()
method on domain classes to throw a grails.validation.ValidationException
if validation fails during a save. This option may also be assigned a list of Strings representing package names. If the value is a list of Strings then the failOnError behavior will only be applied to domain classes in those packages (including sub-packages). See the save method docs for more information.
For example, to enable failOnError for all domain classes:
grails.gorm.failOnError=true
and to enable failOnError for domain classes by package:
grails.gorm.failOnError = ['com.companyname.somepackage',
'com.companyname.someotherpackage']
grails.gorm.autoFlush
= If set to true
, causes the merge, save and delete methods to flush the session, replacing the need to explicitly flush using save(flush: true)
.
5.2 Environments
Per Environment Configuration
Grails supports the concept of per environment configuration. The
Config.groovy
,
DataSource.groovy
, and
BootStrap.groovy
files in the
grails-app/conf
directory can use per-environment configuration using the syntax provided by
ConfigSlurper. As an example consider the following default
DataSource
definition provided by Grails:
dataSource {
pooled = false
driverClassName = "org.h2.Driver"
username = "sa"
password = ""
}
environments {
development {
dataSource {
dbCreate = "create-drop"
url = "jdbc:h2:mem:devDb"
}
}
test {
dataSource {
dbCreate = "update"
url = "jdbc:h2:mem:testDb"
}
}
production {
dataSource {
dbCreate = "update"
url = "jdbc:h2:prodDb"
}
}
}
Notice how the common configuration is provided at the top level and then an
environments
block specifies per environment settings for the
dbCreate
and
url
properties of the
DataSource
.
Packaging and Running for Different Environments
Grails'
command line has built in capabilities to execute any command within the context of a specific environment. The format is:
grails [environment] [command name]
In addition, there are 3 preset environments known to Grails:
dev
,
prod
, and
test
for
development
,
production
and
test
. For example to create a WAR for the
test
environment you wound run:
To target other environments you can pass a
grails.env
variable to any command:
grails -Dgrails.env=UAT run-app
Programmatic Environment Detection
Within your code, such as in a Gant script or a bootstrap class you can detect the environment using the
Environment class:
import grails.util.Environment...switch (Environment.current) {
case Environment.DEVELOPMENT:
configureForDevelopment()
break
case Environment.PRODUCTION:
configureForProduction()
break
}
Per Environment Bootstrapping
It's often desirable to run code when your application starts up on a per-environment basis. To do so you can use the
grails-app/conf/BootStrap.groovy
file's support for per-environment execution:
def init = { ServletContext ctx ->
environments {
production {
ctx.setAttribute("env", "prod")
}
development {
ctx.setAttribute("env", "dev")
}
}
ctx.setAttribute("foo", "bar")
}
Generic Per Environment Execution
The previous
BootStrap
example uses the
grails.util.Environment
class internally to execute. You can also use this class yourself to execute your own environment specific logic:
Environment.executeForCurrentEnvironment {
production {
// do something in production
}
development {
// do something only in development
}
}
5.3 The DataSource
Since Grails is built on Java technology setting up a data source requires some knowledge of JDBC (the technology that doesn't stand for Java Database Connectivity).
If you use a database other than H2 you need a JDBC driver. For example for MySQL you would need
Connector/JDrivers typically come in the form of a JAR archive. It's best to use the dependency resolution to resolve the jar if it's available in a Maven repository, for example you could add a dependency for the MySQL driver like this:
dependencies {
runtime 'mysql:mysql-connector-java:5.1.29'
}
If you can't use dependency resolution then just put the JAR in your project's
lib
directory.
Once you have the JAR resolved you need to get familiar Grails' DataSource descriptor file located at
grails-app/conf/DataSource.groovy
. This file contains the dataSource definition which includes the following settings:
driverClassName
- The class name of the JDBC driver
username
- The username used to establish a JDBC connection
password
- The password used to establish a JDBC connection
url
- The JDBC URL of the database
dbCreate
- Whether to auto-generate the database from the domain model - one of 'create-drop', 'create', 'update' or 'validate'
pooled
- Whether to use a pool of connections (defaults to true)
logSql
- Enable SQL logging to stdout
formatSql
- Format logged SQL
dialect
- A String or Class that represents the Hibernate dialect used to communicate with the database. See the org.hibernate.dialect package for available dialects.
readOnly
- If true
makes the DataSource read-only, which results in the connection pool calling setReadOnly(true)
on each Connection
transactional
- If false
leaves the DataSource's transactionManager bean outside the chained BE1PC transaction manager implementation. This only applies to additional datasources.
persistenceInterceptor
- The default datasource is automatically wired up to the persistence interceptor, other datasources are not wired up automatically unless this is set to true
properties
- Extra properties to set on the DataSource bean. See the Tomcat Pool documentation. There is also a Javadoc format documentation of the properties.
jmxExport
- If false
, will disable registration of JMX MBeans for all DataSources. By default JMX MBeans are added for DataSources with jmxEnabled = true
in properties.
A typical configuration for MySQL may be something like:
dataSource {
pooled = true
dbCreate = "update"
url = "jdbc:mysql://localhost:3306/my_database"
driverClassName = "com.mysql.jdbc.Driver"
dialect = org.hibernate.dialect.MySQL5InnoDBDialect
username = "username"
password = "password"
properties {
jmxEnabled = true
initialSize = 5
maxActive = 50
minIdle = 5
maxIdle = 25
maxWait = 10000
maxAge = 10 * 60000
timeBetweenEvictionRunsMillis = 5000
minEvictableIdleTimeMillis = 60000
validationQuery = "SELECT 1"
validationQueryTimeout = 3
validationInterval = 15000
testOnBorrow = true
testWhileIdle = true
testOnReturn = false
jdbcInterceptors = "ConnectionState;StatementCache(max=200)"
defaultTransactionIsolation = java.sql.Connection.TRANSACTION_READ_COMMITTED
}
}
When configuring the DataSource do not include the type or the def keyword before any of the configuration settings as Groovy will treat these as local variable definitions and they will not be processed. For example the following is invalid:
dataSource {
boolean pooled = true // type declaration results in ignored local variable
…
}
Example of advanced configuration using extra properties:
dataSource {
pooled = true
dbCreate = "update"
url = "jdbc:mysql://localhost:3306/my_database"
driverClassName = "com.mysql.jdbc.Driver"
dialect = org.hibernate.dialect.MySQL5InnoDBDialect
username = "username"
password = "password"
properties {
// Documentation for Tomcat JDBC Pool
// http://tomcat.apache.org/tomcat-7.0-doc/jdbc-pool.html#Common_Attributes
// https://tomcat.apache.org/tomcat-7.0-doc/api/org/apache/tomcat/jdbc/pool/PoolConfiguration.html
jmxEnabled = true
initialSize = 5
maxActive = 50
minIdle = 5
maxIdle = 25
maxWait = 10000
maxAge = 10 * 60000
timeBetweenEvictionRunsMillis = 5000
minEvictableIdleTimeMillis = 60000
validationQuery = "SELECT 1"
validationQueryTimeout = 3
validationInterval = 15000
testOnBorrow = true
testWhileIdle = true
testOnReturn = false
ignoreExceptionOnPreLoad = true
// http://tomcat.apache.org/tomcat-7.0-doc/jdbc-pool.html#JDBC_interceptors
jdbcInterceptors = "ConnectionState;StatementCache(max=200)"
defaultTransactionIsolation = java.sql.Connection.TRANSACTION_READ_COMMITTED // safe default
// controls for leaked connections
abandonWhenPercentageFull = 100 // settings are active only when pool is full
removeAbandonedTimeout = 120
removeAbandoned = true
// use JMX console to change this setting at runtime
logAbandoned = false // causes stacktrace recording overhead, use only for debugging
// JDBC driver properties
// Mysql as example
dbProperties {
// Mysql specific driver properties
// http://dev.mysql.com/doc/connector-j/en/connector-j-reference-configuration-properties.html
// let Tomcat JDBC Pool handle reconnecting
autoReconnect=false
// truncation behaviour
jdbcCompliantTruncation=false
// mysql 0-date conversion
zeroDateTimeBehavior='convertToNull'
// Tomcat JDBC Pool's StatementCache is used instead, so disable mysql driver's cache
cachePrepStmts=false
cacheCallableStmts=false
// Tomcat JDBC Pool's StatementFinalizer keeps track
dontTrackOpenResources=true
// performance optimization: reduce number of SQLExceptions thrown in mysql driver code
holdResultsOpenOverStatementClose=true
// enable MySQL query cache - using server prep stmts will disable query caching
useServerPrepStmts=false
// metadata caching
cacheServerConfiguration=true
cacheResultSetMetadata=true
metadataCacheSize=100
// timeouts for TCP/IP
connectTimeout=15000
socketTimeout=120000
// timer tuning (disable)
maintainTimeStats=false
enableQueryTimeouts=false
// misc tuning
noDatetimeStringSync=true
}
}
}
More on dbCreate
Hibernate can automatically create the database tables required for your domain model. You have some control over when and how it does this through the
dbCreate
property, which can take these values:
- create - Drops the existing schema and creates the schema on startup, dropping existing tables, indexes, etc. first.
- create-drop - Same as create, but also drops the tables when the application shuts down cleanly.
- update - Creates missing tables and indexes, and updates the current schema without dropping any tables or data. Note that this can't properly handle many schema changes like column renames (you're left with the old column containing the existing data).
- validate - Makes no changes to your database. Compares the configuration with the existing database schema and reports warnings.
- any other value - does nothing
You can also remove the
dbCreate
setting completely, which is recommended once your schema is relatively stable and definitely when your application and database are deployed in production. Database changes are then managed through proper migrations, either with SQL scripts or a migration tool like
Liquibase (the
Database Migration plugin uses Liquibase and is tightly integrated with Grails and GORM).
5.3.1 DataSources and Environments
The previous example configuration assumes you want the same config for all environments: production, test, development etc.
Grails' DataSource definition is "environment aware", however, so you can do:
dataSource {
pooled = true
driverClassName = "com.mysql.jdbc.Driver"
dialect = org.hibernate.dialect.MySQL5InnoDBDialect
// other common settings here
}environments {
production {
dataSource {
url = "jdbc:mysql://liveip.com/liveDb"
// other environment-specific settings here
}
}
}
5.3.2 JNDI DataSources
Referring to a JNDI DataSource
Most Java EE containers supply
DataSource
instances via
Java Naming and Directory Interface (JNDI). Grails supports the definition of JNDI data sources as follows:
dataSource {
jndiName = "java:comp/env/myDataSource"
}
The format on the JNDI name may vary from container to container, but the way you define the
DataSource
in Grails remains the same.
Configuring a Development time JNDI resource
The way in which you configure JNDI data sources at development time is plugin dependent. Using the
Tomcat plugin you can define JNDI resources using the
grails.naming.entries
setting in
grails-app/conf/Config.groovy
:
grails.naming.entries = [
"bean/MyBeanFactory": [
auth: "Container",
type: "com.mycompany.MyBean",
factory: "org.apache.naming.factory.BeanFactory",
bar: "23"
],
"jdbc/EmployeeDB": [
type: "javax.sql.DataSource", //required
auth: "Container", // optional
description: "Data source for Foo", //optional
driverClassName: "org.h2.Driver",
url: "jdbc:h2:mem:database",
username: "dbusername",
password: "dbpassword",
maxActive: "8",
maxIdle: "4"
],
"mail/session": [
type: "javax.mail.Session,
auth: "Container",
"mail.smtp.host": "localhost"
]
]
5.3.3 Automatic Database Migration
The
dbCreate
property of the
DataSource
definition is important as it dictates what Grails should do at runtime with regards to automatically generating the database tables from
GORM classes. The options are described in the
DataSource section:
create
create-drop
update
validate
- no value
In
development mode
dbCreate
is by default set to "create-drop", but at some point in development (and certainly once you go to production) you'll need to stop dropping and re-creating the database every time you start up your server.
It's tempting to switch to
update
so you retain existing data and only update the schema when your code changes, but Hibernate's update support is very conservative. It won't make any changes that could result in data loss, and doesn't detect renamed columns or tables, so you'll be left with the old one and will also have the new one.
Grails supports migrations via the
Database Migration plugin which can be installed by declaring the plugin in
grails-app/conf/BuildConfig.groovy
:
grails.project.dependency.resolution = {
…
plugins {
runtime ':database-migration:1.3.1'
}
}
The plugin uses
Liquibase and provides access to all of its functionality, and also has support for GORM (for example generating a change set by comparing your domain classes to a database).
5.3.4 Transaction-aware DataSource Proxy
The actual
dataSource
bean is wrapped in a transaction-aware proxy so you will be given the connection that's being used by the current transaction or Hibernate
Session
if one is active.
If this were not the case, then retrieving a connection from the
dataSource
would be a new connection, and you wouldn't be able to see changes that haven't been committed yet (assuming you have a sensible transaction isolation setting, e.g.
READ_COMMITTED
or better).
The "real" unproxied
dataSource
is still available to you if you need access to it; its bean name is
dataSourceUnproxied
.
You can access this bean like any other Spring bean, i.e. using dependency injection:
class MyService { def dataSourceUnproxied
…
}
or by pulling it from the
ApplicationContext
:
def dataSourceUnproxied = ctx.dataSourceUnproxied
5.3.5 Database Console
The
H2 database console is a convenient feature of H2 that provides a web-based interface to any database that you have a JDBC driver for, and it's very useful to view the database you're developing against. It's especially useful when running against an in-memory database.
You can access the console by navigating to
http://localhost:8080/appname/dbconsole in a browser. The URI can be configured using the
grails.dbconsole.urlRoot
attribute in Config.groovy and defaults to
'/dbconsole'
.
The console is enabled by default in development mode and can be disabled or enabled in other environments by using the
grails.dbconsole.enabled
attribute in Config.groovy. For example you could enable the console in production using
environments {
production {
grails.serverURL = "http://www.changeme.com"
grails.dbconsole.enabled = true
grails.dbconsole.urlRoot = '/admin/dbconsole'
}
development {
grails.serverURL = "http://localhost:8080/${appName}"
}
test {
grails.serverURL = "http://localhost:8080/${appName}"
}
}
If you enable the console in production be sure to guard access to it using a trusted security framework.
Configuration
By default the console is configured for an H2 database which will work with the default settings if you haven't configured an external database - you just need to change the JDBC URL to
jdbc:h2:mem:devDB
. If you've configured an external database (e.g. MySQL, Oracle, etc.) then you can use the Saved Settings dropdown to choose a settings template and fill in the url and username/password information from your DataSource.groovy.
5.3.6 Multiple Datasources
By default all domain classes share a single
DataSource
and a single database, but you have the option to partition your domain classes into two or more
DataSource
s.
Configuring Additional DataSources
The default
DataSource
configuration in
grails-app/conf/DataSource.groovy
looks something like this:
dataSource {
pooled = true
driverClassName = "org.h2.Driver"
username = "sa"
password = ""
}
hibernate {
cache.use_second_level_cache = true
cache.use_query_cache = true
cache.provider_class = 'net.sf.ehcache.hibernate.EhCacheProvider'
}environments {
development {
dataSource {
dbCreate = "create-drop"
url = "jdbc:h2:mem:devDb"
}
}
test {
dataSource {
dbCreate = "update"
url = "jdbc:h2:mem:testDb"
}
}
production {
dataSource {
dbCreate = "update"
url = "jdbc:h2:prodDb"
}
}
}
This configures a single
DataSource
with the Spring bean named
dataSource
. To configure extra
DataSource
s, add another
dataSource
block (at the top level, in an environment block, or both, just like the standard
DataSource
definition) with a custom name, separated by an underscore. For example, this configuration adds a second
DataSource
, using MySQL in the development environment and Oracle in production:
environments {
development {
dataSource {
dbCreate = "create-drop"
url = "jdbc:h2:mem:devDb"
}
dataSource_lookup {
dialect = org.hibernate.dialect.MySQLInnoDBDialect
driverClassName = 'com.mysql.jdbc.Driver'
username = 'lookup'
password = 'secret'
url = 'jdbc:mysql://localhost/lookup'
dbCreate = 'update'
}
}
test {
dataSource {
dbCreate = "update"
url = "jdbc:h2:mem:testDb"
}
}
production {
dataSource {
dbCreate = "update"
url = "jdbc:h2:prodDb"
}
dataSource_lookup {
dialect = org.hibernate.dialect.Oracle10gDialect
driverClassName = 'oracle.jdbc.driver.OracleDriver'
username = 'lookup'
password = 'secret'
url = 'jdbc:oracle:thin:@localhost:1521:lookup'
dbCreate = 'update'
}
}
}
You can use the same or different databases as long as they're supported by Hibernate.
Configuring Domain Classes
If a domain class has no
DataSource
configuration, it defaults to the standard
'dataSource'
. Set the
datasource
property in the
mapping
block to configure a non-default
DataSource
. For example, if you want to use the
ZipCode
domain to use the
'lookup'
DataSource
, configure it like this;
class ZipCode { String code static mapping = {
datasource 'lookup'
}
}
A domain class can also use two or more
DataSource
s. Use the
datasources
property with a list of names to configure more than one, for example:
class ZipCode { String code static mapping = {
datasources(['lookup', 'auditing'])
}
}
If a domain class uses the default
DataSource
and one or more others, use the special name
'DEFAULT'
to indicate the default
DataSource
:
class ZipCode { String code static mapping = {
datasources(['lookup', 'DEFAULT'])
}
}
If a domain class uses all configured
DataSource
s use the special value
'ALL'
:
class ZipCode { String code static mapping = {
datasource 'ALL'
}
}
Namespaces and GORM Methods
If a domain class uses more than one
DataSource
then you can use the namespace implied by each
DataSource
name to make GORM calls for a particular
DataSource
. For example, consider this class which uses two
DataSource
s:
class ZipCode { String code static mapping = {
datasources(['lookup', 'auditing'])
}
}
The first
DataSource
specified is the default when not using an explicit namespace, so in this case we default to 'lookup'. But you can call GORM methods on the 'auditing'
DataSource
with the
DataSource
name, for example:
def zipCode = ZipCode.auditing.get(42)
…
zipCode.auditing.save()
As you can see, you add the
DataSource
to the method call in both the static case and the instance case.
Hibernate Mapped Domain Classes
You can also partition annotated Java classes into separate datasources. Classes using the default datasource are registered in
grails-app/conf/hibernate/hibernate.cfg.xml
. To specify that an annotated class uses a non-default datasource, create a
hibernate.cfg.xml
file for that datasource with the file name prefixed with the datasource name.
For example if the
Book
class is in the default datasource, you would register that in
grails-app/conf/hibernate/hibernate.cfg.xml
:
<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE hibernate-configuration PUBLIC
'-//Hibernate/Hibernate Configuration DTD 3.0//EN'
'http://hibernate.sourceforge.net/hibernate-configuration-3.0.dtd'>
<hibernate-configuration>
<session-factory>
<mapping class='org.example.Book'/>
</session-factory>
</hibernate-configuration>
and if the
Library
class is in the "ds2" datasource, you would register that in
grails-app/conf/hibernate/ds2_hibernate.cfg.xml
:
<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE hibernate-configuration PUBLIC
'-//Hibernate/Hibernate Configuration DTD 3.0//EN'
'http://hibernate.sourceforge.net/hibernate-configuration-3.0.dtd'>
<hibernate-configuration>
<session-factory>
<mapping class='org.example.Library'/>
</session-factory>
</hibernate-configuration>
The process is the same for classes mapped with hbm.xml files - just list them in the appropriate hibernate.cfg.xml file.
Services
Like Domain classes, by default Services use the default
DataSource
and
PlatformTransactionManager
. To configure a Service to use a different
DataSource
, use the static
datasource
property, for example:
class DataService { static datasource = 'lookup' void someMethod(...) {
…
}
}
A transactional service can only use a single
DataSource
, so be sure to only make changes for domain classes whose
DataSource
is the same as the Service.
Note that the datasource specified in a service has no bearing on which datasources are used for domain classes; that's determined by their declared datasources in the domain classes themselves. It's used to declare which transaction manager to use.
What you'll see is that if you have a Foo domain class in dataSource1 and a Bar domain class in dataSource2, and WahooService uses dataSource1, a service method that saves a new Foo and a new Bar will only be transactional for Foo since they share the datasource. The transaction won't affect the Bar instance. If you want both to be transactional you'd need to use two services and XA datasources for two-phase commit, e.g. with the Atomikos plugin.
Transactions across multiple datasources
Grails uses the Best Efforts 1PC pattern for handling transactions across multiple datasources.
The
Best Efforts 1PC pattern is fairly general but can fail in some circumstances that the developer must be aware of. This is a non-XA pattern that involves a synchronized single-phase commit of a number of resources. Because the
2PC is not used, it can never be as safe as an
XA transaction, but is often good enough if the participants are aware of the compromises.
The basic idea is to delay the commit of all resources as late as possible in a transaction so that the only thing that can go wrong is an infrastructure failure (not a business-processing error). Systems that rely on Best Efforts 1PC reason that infrastructure failures are rare enough that they can afford to take the risk in return for higher throughput. If business-processing services are also designed to be idempotent, then little can go wrong in practice.
The BE1PC implementation was added in Grails 2.3.6. . Before this change additional datasources didn't take part in transactions initiated in Grails. The transactions in additional datasources were basically in auto commit mode. In some cases this might be the wanted behavior. One reason might be performance: on the start of each new transaction, the BE1PC transaction manager creates a new transaction to each datasource. It's possible to leave an additional datasource out of the BE1PC transaction manager by setting
transactional = false
in the respective configuration block of the additional dataSource. Datasources with
readOnly = true
will also be left out of the chained transaction manager (since 2.3.7).
By default, the BE1PC implementation will add all beans implementing the Spring
PlatformTransactionManager
interface to the chained BE1PC transaction manager. For example, a possible
JMSTransactionManager
bean in the Grails application context would be added to the Grails BE1PC transaction manager's chain of transaction managers.
You can exclude transaction manager beans from the BE1PC implementation with the this configuration option:
grails.transaction.chainedTransactionManagerPostProcessor.blacklistPattern = '.*'
The exclude matching is done on the name of the transaction manager bean. The transaction managers of datasources with
transactional = false
or
readOnly = true
will be skipped and using this configuration option is not required in that case.
XA and Two-phase Commit
When the Best Efforts 1PC pattern isn't suitable for handling transactions across multiple transactional resources (not only datasources), there are several options available for adding XA/2PC support to Grails applications.
The
Spring transactions documentation contains information about integrating the JTA/XA transaction manager of different application servers. In this case, you can configure a bean with the name
transactionManager
manually in
resources.groovy
or
resources.xml
file.
There is also
Atomikos plugin available for XA support in Grails applications.
5.4 Externalized Configuration
Some deployments require that configuration be sourced from more than one place and be changeable without requiring a rebuild of the application. In order to support deployment scenarios such as these the configuration can be externalized. To do so, point Grails at the locations of the configuration files that should be used by adding a
grails.config.locations
setting in
Config.groovy
, for example:
grails.config.locations = [
"classpath:${appName}-config.properties",
"classpath:${appName}-config.groovy",
"file:${userHome}/.grails/${appName}-config.properties",
"file:${userHome}/.grails/${appName}-config.groovy" ]
In the above example we're loading configuration files (both Java Properties files and
ConfigSlurper configurations) from different places on the classpath and files located in
USER_HOME
.
It is also possible to load config by specifying a class that is a config script.
grails.config.locations = [com.my.app.MyConfig]
This can be useful in situations where the config is either coming from a plugin or some other part of your application. A typical use for this is re-using configuration provided by plugins across multiple applications.
Ultimately all configuration files get merged into the
config
property of the
GrailsApplication object and are hence obtainable from there.
Values that have the same name as previously defined values will overwrite the existing values, and the pointed to configuration sources are loaded in the order in which they are defined.
Config Defaults
The configuration values contained in the locations described by the
grails.config.locations
property will
override any values defined in your application
Config.groovy
file which may not be what you want. You may want to have a set of
default values be be loaded that can be overridden in either your application's
Config.groovy
file or in a named config location. For this you can use the
grails.config.defaults.locations
property.
This property supports the same values as the
grails.config.locations
property (i.e. paths to config scripts, property files or classes), but the config described by
grails.config.defaults.locations
will be loaded
before all other values and can therefore be overridden. Some plugins use this mechanism to supply one or more sets of default configuration that you can choose to include in your application config.
Grails also supports the concept of property place holders and property override configurers as defined in Spring For more information on these see the section on Grails and Spring
5.5 Versioning
Versioning Basics
Grails has built in support for application versioning. The version of the application is set to
0.1
when you first create an application with the
create-app command. The version is stored in the application meta data file
application.properties
in the root of the project.
To change the version of your application you can edit the file manually, or run the
set-version command:
The version is used in various commands including the
war command which will append the application version to the end of the created WAR file.
Detecting Versions at Runtime
You can detect the application version using Grails' support for application metadata using the
GrailsApplication class. For example within
controllers there is an implicit
grailsApplication variable that can be used:
def version = grailsApplication.metadata['app.version']
You can retrieve the version of Grails that is running with:
def grailsVersion = grailsApplication.metadata['app.grails.version']
or the
GrailsUtil
class:
import grails.util.GrailsUtil
…
def grailsVersion = GrailsUtil.grailsVersion
5.6 Project Documentation
Since Grails 1.2, the documentation engine that powers the creation of this documentation has been available for your own Grails projects.
The documentation engine uses a variation on the
Textile syntax to automatically create project documentation with smart linking, formatting etc.
Creating project documentation
To use the engine you need to follow a few conventions. First, you need to create a
src/docs/guide
directory where your documentation source files will go. Then, you need to create the source docs themselves. Each chapter should have its own gdoc file as should all numbered sub-sections. You will end up with something like:
+ src/docs/guide/introduction.gdoc
+ src/docs/guide/introduction/changes.gdoc
+ src/docs/guide/gettingStarted.gdoc
+ src/docs/guide/configuration.gdoc
+ src/docs/guide/configuration/build.gdoc
+ src/docs/guide/configuration/build/controllers.gdoc
Note that you can have all your gdoc files in the top-level directory if you want, but you can also put sub-sections in sub-directories named after the parent section - as the above example shows.
Once you have your source files, you still need to tell the documentation engine what the structure of your user guide is going to be. To do that, you add a
src/docs/guide/toc.yml
file that contains the structure and titles for each section. This file is in
YAML format and basically represents the structure of the user guide in tree form. For example, the above files could be represented as:
introduction:
title: Introduction
changes: Change Log
gettingStarted: Getting Started
configuration:
title: Configuration
build:
title: Build Config
controllers: Specifying Controllers
The format is pretty straightforward. Any section that has sub-sections is represented with the corresponding filename (minus the .gdoc extension) followed by a colon. The next line should contain
title:
plus the title of the section as seen by the end user. Every sub-section then has its own line after the title. Leaf nodes, i.e. those without any sub-sections, declare their title on the same line as the section name but after the colon.
That's it. You can easily add, remove, and move sections within the
toc.yml
to restructure the generated user guide. You should also make sure that all section names, i.e. the gdoc filenames, should be unique since they are used for creating internal links and for the HTML filenames. Don't worry though, the documentation engine will warn you of duplicate section names.
Creating reference items
Reference items appear in the Quick Reference section of the documentation. Each reference item belongs to a category and a category is a directory located in the
src/docs/ref
directory. For example, suppose you have defined a new controller method called
renderPDF
. That belongs to the
Controllers
category so you would create a gdoc text file at the following location:
+ src/docs/ref/Controllers/renderPDF.gdoc
Configuring Output Properties
There are various properties you can set within your
grails-app/conf/Config.groovy
file that customize the output of the documentation such as:
- grails.doc.title - The title of the documentation
- grails.doc.subtitle - The subtitle of the documentation
- grails.doc.authors - The authors of the documentation
- grails.doc.license - The license of the software
- grails.doc.copyright - The copyright message to display
- grails.doc.footer - The footer to use
Other properties such as the version are pulled from your project itself. If a title is not specified, the application name is used.
You can also customise the look of the documentation and provide images by setting a few other options:
- grails.doc.css - The location of a directory containing custom CSS files (type
java.io.File
)
- grails.doc.js - The location of a directory containing custom JavaScript files (type
java.io.File
)
- grails.doc.style - The location of a directory containing custom HTML templates for the guide (type
java.io.File
)
- grails.doc.images - The location of a directory containing image files for use in the style templates and within the documentation pages themselves (type
java.io.File
)
One of the simplest ways to customise the look of the generated guide is to provide a value for
grails.doc.css
and then put a custom.css file in the corresponding directory. Grails will automatically include this CSS file in the guide. You can also place a custom-pdf.css file in that directory. This allows you to override the styles for the PDF version of the guide.
Generating Documentation
Once you have created some documentation (refer to the syntax guide in the next chapter) you can generate an HTML version of the documentation using the command:
This command will output an
docs/manual/index.html
which can be opened in a browser to view your documentation.
Documentation Syntax
As mentioned the syntax is largely similar to Textile or Confluence style wiki markup. The following sections walk you through the syntax basics.
Basic Formatting
Monospace:
monospace
Italic:
italic
Bold:
bold
Image:
!http://grails.org/images/new/grailslogo_topNav.png!
You can also link to internal images like so:
!someFolder/my_diagram.png!
This will link to an image stored locally within your project. There is currently no default location for doc images, but you can specify one with the
grails.doc.images
setting in Config.groovy like so:
grails.doc.images = new File("src/docs/images")
In this example, you would put the my_diagram.png file in the directory 'src/docs/images/someFolder'.
Linking
There are several ways to create links with the documentation generator. A basic external link can either be defined using confluence or textile style markup:
[Pivotal|http://www.pivotal.io/oss]
or
"Pivotal":http://www.pivotal.io/oss
For links to other sections inside the user guide you can use the
guide:
prefix with the name of the section you want to link to:
[Intro|guide:introduction]
The section name comes from the corresponding gdoc filename. The documentation engine will warn you if any links to sections in your guide break.
To link to reference items you can use a special syntax:
In this case the category of the reference item is on the right hand side of the | and the name of the reference item on the left.
Finally, to link to external APIs you can use the
api:
prefix. For example:
[String|api:java.lang.String]
The documentation engine will automatically create the appropriate javadoc link in this case. To add additional APIs to the engine you can configure them in
grails-app/conf/Config.groovy
. For example:
grails.doc.api.org.hibernate=
"http://docs.jboss.org/hibernate/stable/core/javadocs"
The above example configures classes within the
org.hibernate
package to link to the Hibernate website's API docs.
Lists and Headings
Headings can be created by specifying the letter 'h' followed by a number and then a dot:
h3.<space>Heading3
h4.<space>Heading4
Unordered lists are defined with the use of the * character:
* item 1
** subitem 1
** subitem 2
* item 2
Numbered lists can be defined with the # character:
Tables can be created using the
table
macro:
Name | Number |
---|
Albert | 46 |
Wilma | 1348 |
James | 12 |
{table}
*Name* | *Number*
Albert | 46
Wilma | 1348
James | 12
{table}
Code and Notes
You can define code blocks with the
code
macro:
class Book {
String title
}
{code}
class Book {
String title
}
{code}
The example above provides syntax highlighting for Java and Groovy code, but you can also highlight XML markup:
{code:xml}
<hello>world</hello>
{code}
There are also a couple of macros for displaying notes and warnings:
Note:
This is a note!
{note}
This is a note!
{note}
Warning:
This is a warning!
{warning}
This is a warning!
{warning}
5.7 Dependency Resolution
Grails features a dependency resolution DSL that lets you control how plugins and JAR dependencies are resolved.
You can choose to use Aether (since Grails 2.3) or Apache Ivy as the dependency resolution engine. Aether is the dependency resolution library used by the Maven build tool, so if you are looking for Maven-like behavior then Aether is the better choice. Ivy allows more flexibility if you wish to resolve jars from flat file systems or none HTTP repositories. Aether is the default dependency resolution engine for Grails applications since Grails 2.3.
As of Grails 2.4 the Ivy resolver is considered deprecated and no longer maintained. It is recommended all users switch to using Aether.
To configure which dependency resolution engine to use you can specify the
grails.project.dependency.resolver
setting in
grails-app/conf/BuildConfig.groovy
. The default setting is shown below:
grails.project.dependency.resolver = "maven" // or ivy
You can then specify a
grails.project.dependency.resolution
property inside the
grails-app/conf/BuildConfig.groovy
file that configures how dependencies are resolved:
grails.project.dependency.resolution = {
// config here
}
The default configuration looks like the following:
grails.servlet.version = "3.0" // Change depending on target container compliance (2.5 or 3.0)
grails.project.class.dir = "target/classes"
grails.project.test.class.dir = "target/test-classes"
grails.project.test.reports.dir = "target/test-reports"
grails.project.work.dir = "target/work"
grails.project.target.level = 1.6
grails.project.source.level = 1.6
//grails.project.war.file = "target/${appName}-${appVersion}.war"grails.project.fork = [
// configure settings for compilation JVM, note that if you alter the Groovy version forked compilation is required
// compile: [maxMemory: 256, minMemory: 64, debug: false, maxPerm: 256, daemon:true], // configure settings for the test-app JVM, uses the daemon by default
test: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, daemon:true],
// configure settings for the run-app JVM
run: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, forkReserve:false],
// configure settings for the run-war JVM
war: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, forkReserve:false],
// configure settings for the Console UI JVM
console: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256]
]grails.project.dependency.resolver = "maven" // or ivy
grails.project.dependency.resolution = {
// inherit Grails' default dependencies
inherits("global") {
// specify dependency exclusions here; for example, uncomment this to disable ehcache:
// excludes 'ehcache'
}
log "error" // log level of Ivy resolver, either 'error', 'warn', 'info', 'debug' or 'verbose'
checksums true // Whether to verify checksums on resolve
legacyResolve false // whether to do a secondary resolve on plugin installation, not advised and here for backwards compatibility repositories {
inherits true // Whether to inherit repository definitions from plugins grailsPlugins()
grailsHome()
mavenLocal()
grailsCentral()
mavenCentral()
// uncomment these (or add new ones) to enable remote dependency resolution from public Maven repositories
//mavenRepo "http://repository.codehaus.org"
//mavenRepo "http://download.java.net/maven/2/"
} dependencies {
// specify dependencies here under either 'build', 'compile', 'runtime', 'test' or 'provided' scopes e.g.
runtime 'mysql:mysql-connector-java:5.1.24'
compile 'org.springframework.integration:spring-integration-core:2.2.5.RELEASE'
} plugins {
// plugins for the build system only
}
}
The details of the above will be explained in the next few sections.
5.7.1 Configurations and Dependencies
Grails features five dependency resolution configurations (or 'scopes'):
-
build
: Dependencies for the build system only
-
compile
: Dependencies for the compile step
-
runtime
: Dependencies needed at runtime but not for compilation (see above)
-
test
: Dependencies needed for testing but not at runtime (see above)
-
provided
: Dependencies needed at development time, but not during WAR deployment
-
optional
(Aether only): Dependencies considered optional and not required for the execution of the application or plugin
Within the
dependencies
block you can specify a dependency that falls into one of these configurations by calling the equivalent method. For example if your application requires the MySQL driver to function at
runtime
you can specify that like this:
runtime 'com.mysql:mysql-connector-java:5.1.16'
This uses the string syntax:
group:name:version
.
If you are using Aether as the dependency resolution library, the Maven pattern of:
<groupId>:<artifactId>[:<extension>[:<classifier>]]:<version>
You can also use a Map-based syntax:
runtime group: 'com.mysql',
name: 'mysql-connector-java',
version: '5.1.16'
Possible settings to the map syntax are:
group
- The group / organization (or groupId in Maven terminology)
name
- The dependency name (or artifactId in Maven terminology)
version
- The version of the dependency
extension
(Aether only) - The file extension of the dependency
classifier
- The dependency classifier
branch
(Ivy only) - The branch of the dependency
transitive
(Ivy only) - Whether the dependency has transitive dependencies
As you can see from the list above some dependency configuration settings work only in Aether and some only in Ivy.
Multiple dependencies can be specified by passing multiple arguments:
runtime 'com.mysql:mysql-connector-java:5.1.16',
'net.sf.ehcache:ehcache:1.6.1'// Orruntime(
[group:'com.mysql', name:'mysql-connector-java', version:'5.1.16'],
[group:'net.sf.ehcache', name:'ehcache', version:'1.6.1']
)
Disabling transitive dependency resolution
By default, Grails will not only get the JARs and plugins that you declare, but it will also get their transitive dependencies. This is usually what you want, but there are occasions where you want a dependency without all its baggage. In such cases, you can disable transitive dependency resolution on a case-by-case basis:
runtime('com.mysql:mysql-connector-java:5.1.16',
'net.sf.ehcache:ehcache:1.6.1') {
transitive = false
}// Or
runtime group:'com.mysql',
name:'mysql-connector-java',
version:'5.1.16',
transitive:false
Excluding specific transitive dependencies
A far more common scenario is where you want the transitive dependencies, but some of them cause issues with your own dependencies or are unnecessary. For example, many Apache projects have 'commons-logging' as a transitive dependency, but it shouldn't be included in a Grails project (we use SLF4J). That's where the
excludes
option comes in:
runtime('com.mysql:mysql-connector-java:5.1.16',
'net.sf.ehcache:ehcache:1.6.1') {
excludes "xml-apis", "commons-logging"
}// Or
runtime(group:'com.mysql', name:'mysql-connector-java', version:'5.1.16') {
excludes([ group: 'xml-apis', name: 'xml-apis'],
[ group: 'org.apache.httpcomponents' ],
[ name: 'commons-logging' ])
As you can see, you can either exclude dependencies by their artifact ID (also known as a module name) or any combination of group and artifact IDs (if you use the Map notation). You may also come across
exclude
as well, but that can only accept a single string or Map:
runtime('com.mysql:mysql-connector-java:5.1.16',
'net.sf.ehcache:ehcache:1.6.1') {
exclude "xml-apis"
}
Dependency Management (Aether Only)
If you are using Aether then you can take advantage of Maven's notion of
Dependency Management.
To do so you use a
management
block, for example:
management {
dependency "commons-logging:commons-logging:1.1.3"
}
The above declaration will force all any transitive dependencies on
commons-logging
to use the 1.1.3 version without you having to declare an explicit dependency on
commons-logging
yourself. In addition to the version, you can also control the scope and exclusion rules of a dependency.
Where are the JARs?
With all these declarative dependencies, you may wonder where all the JARs end up. They have to go somewhere after all. By default Grails puts them into a directory, called the dependency cache, that resides on your local file system at
user.home/.grails/ivy-cache
or
user.home/.m2/repository
when using Aether. You can change this either via the
settings.groovy
file:
grails.dependency.cache.dir = "${userHome}/.my-dependency-cache"
or in the dependency DSL:
grails.project.dependency.resolution = {
…
cacheDir "target/ivy-cache"
…
}
The
settings.groovy
option applies to all projects, so it's the preferred approach.
5.7.2 Dependency Repositories
Remote Repositories
Initially your BuildConfig.groovy does not use any remote public Maven repositories. There is a default
grailsHome()
repository that will locate the JAR files Grails needs from your Grails installation. To use a public repository, specify it in the
repositories
block:
repositories {
mavenCentral()
}
In this case the default public Maven repository is specified.
You can also specify a specific Maven repository to use by URL:
repositories {
mavenRepo "http://repository.codehaus.org"
}
and even give it a name:
repositories {
mavenRepo name: "Codehaus", root: "http://repository.codehaus.org"
}
so that you can easily identify it in logs.
Offline Mode
There are times when it is not desirable to connect to any remote repositories (whilst working on the train for example!). In this case you can use the
offline
flag to execute Grails commands and Grails will not connect to any remote repositories:
Note that this command will fail if you do not have the necessary dependencies in your local Maven cache
You can also globally configure offline mode by setting
grails.offline.mode
to
true
in
~/.grails/settings.groovy
or in your project's
BuildConfig.groovy
file:
To specify your local Maven cache (
~/.m2/repository
) as a repository:
repositories {
mavenLocal()
}
Authentication with Aether
To authenticate with Aether you can either define the credentials on the repository definition:
mavenRepo(url:"http://localhost:8082/myrepo") {
auth username: "foo", password: "bar"
}
Or you can specify an
id
on the repository:
mavenRepo(id:'myrepo', url:"http://localhost:8082/myrepo")
And then declare your credentials in
USER_HOME/.grails/settings.groovy
:
grails.project.dependency.authentication = {
credentials {
id = "myrepo"
username = "admin"
password = "password"
}
}
Authentication with Ivy
If your repository requires authentication you can configure this using a
credentials
block:
credentials {
realm = ".."
host = "localhost"
username = "myuser"
password = "mypass"
}
This can be placed in your
USER_HOME/.grails/settings.groovy
file using the
grails.project.ivy.authentication
setting:
grails.project.ivy.authentication = {
credentials {
realm = ".."
host = "localhost"
username = "myuser"
password = "mypass"
}
}
5.7.3 Debugging Resolution
If you are having trouble getting a dependency to resolve you can enable more verbose debugging from the underlying engine using the
log
method:
// log level of the Aether or Ivy resolver, either 'error', 'warn',
// 'info', 'debug' or 'verbose'
log "warn"
A common issue is that the checksums for a dependency don't match the associated JAR file, and so Ivy rejects the dependency. This helps ensure that the dependencies are valid. But for a variety of reasons some dependencies simply don't have valid checksums in the repositories, even if they are valid JARs. To get round this, you can disable Ivy's dependency checks like so:
grails.project.dependency.resolution = {
…
log "warn"
checksums false
…
}
This is a global setting, so only use it if you have to.
5.7.4 Inherited Dependencies
By default every Grails application inherits several framework dependencies. This is done through the line:
Inside the
BuildConfig.groovy
file. To exclude specific inherited dependencies you use the
excludes
method:
inherits("global") {
excludes "oscache", "ehcache"
}
5.7.5 Providing Default Dependencies
Most Grails applications have runtime dependencies on several jar files that are provided by the Grails framework. These include libraries like Spring, Sitemesh, Hibernate etc. When a war file is created, all of these dependencies will be included in it. But, an application may choose to exclude these jar files from the war. This is useful when the jar files will be provided by the container, as would normally be the case if multiple Grails applications are deployed to the same container.
The dependency resolution DSL provides a mechanism to express that all of the default dependencies will be provided by the container. This is done by invoking the
defaultDependenciesProvided
method and passing
true
as an argument:
grails.project.dependency.resolution = { defaultDependenciesProvided true // all of the default dependencies will
// be "provided" by the container inherits "global" // inherit Grails' default dependencies repositories {
grailsHome()
…
}
dependencies {
…
}
}
defaultDependenciesProvided
must come before inherits
, otherwise the Grails dependencies will be included in the war.
5.7.6 Snapshots and Other Changing Dependencies
Configuration Changing dependencies
Typically, dependencies are constant. That is, for a given combination of
group
,
name
and
version
the jar (or plugin) that it refers to will never change. The Grails dependency management system uses this fact to cache dependencies in order to avoid having to download them from the source repository each time. Sometimes this is not desirable. For example, many developers use the convention of a
snapshot (i.e. a dependency with a version number ending in “-SNAPSHOT”) that can change from time to time while still retaining the same version number. We call this a "changing dependency".
Whenever you have a changing dependency, Grails will always check the remote repository for a new version. More specifically, when a changing dependency is encountered during dependency resolution its last modified timestamp in the local cache is compared against the last modified timestamp in the dependency repositories. If the version on the remote server is deemed to be newer than the version in the local cache, the new version will be downloaded and used.
Be sure to read the next section on "Dependency Resolution Caching" in addition to this one as it affects changing dependencies.
All dependencies (jars and plugins) with a version number ending in
-SNAPSHOT
are
implicitly considered to be changing by Grails. You can also explicitly specify that a dependency is changing by setting the changing flag in the dependency DSL (This is only required for Ivy, Aether does not support the 'changing' flag and treats dependencies that end with -SNAPSHOT as changing):
runtime ('org.my:lib:1.2.3') {
changing = true
}
Aether and SNAPSHOT dependencies
The semantics for handling snapshots when using Aether in Grails are the same as those when using the Maven build tool. The default snapshot check policy is to check once a day for a new version of the dependency. This means that if a new snapshot is published during the day to a remote repository you may not see that change unless you manually clear out your local snapshot.
If you wish to change the snapshot update policy you can do so by configuring an
updatePolicy
for the repository where the snapshot was resolved from, for example:
repositories {
mavenCentral {
updatePolicy "interval:1"
}
}
The above example configures an update policy that checks once a minute for changes. Note that that an
updatePolicy
like the above will seriously impact performance of dependency resolution. The possibly configuration values for
updatePolicy
are as follows:
never
- Never check for new snapshots
always
- Always check for new snapshots
daily
- Check once a day for new snapshots (the default)
interval:x
- Check once every x minutes for new snapshots
Ivy and Changing dependencies
For those used to Maven snapshot handling, if you use Aether dependency management you can expect the same semantics as Maven. If you choose to use Ivy there is a caveat to the support for changing dependencies that you should be aware of. Ivy will stop looking for newer versions of a dependency once it finds a remote repository that has the dependency.
Consider the following setup:
grails.project.dependency.resolution = {
repositories {
mavenLocal()
mavenRepo "http://my.org/repo"
}
dependencies {
compile "myorg:mylib:1.0-SNAPSHOT"
}
In this example we are using the local maven repository and a remote network maven repository. Assuming that the local OI dependency and the local Maven cache do not contain the dependency but the remote repository does, when we perform dependency resolution the following actions will occur:
- maven local repository is searched, dependency not found
- maven network repository is searched, dependency is downloaded to the cache and used
Note that the repositories are checked in the order they are defined in the
BuildConfig.groovy
file.
If we perform dependency resolution again without the dependency changing on the remote server, the following will happen:
- maven local repository is searched, dependency not found
- maven network repository is searched, dependency is found to be the same "age" as the version in the cache so will not be updated (i.e. downloaded)
Later on, a new version of
mylib 1.0-SNAPSHOT
is published changing the version on the server. The next time we perform dependency resolution, the following will happen:
- maven local repository is searched, dependency not found
- maven network repository is searched, dependency is found to newer than version in the cache so will be updated (i.e. downloaded to the cache)
So far everything is working well.
Now we want to test some local changes to the
mylib
library. To do this we build it locally and install it to the local Maven cache (how doesn't particularly matter). The next time we perform a dependency resolution, the following will occur:
- maven local repository is searched, dependency is found to newer than version in the cache so will be updated (i.e. downloaded to the cache)
- maven network repository is NOT searched as we've already found the dependency
This is what we wanted to occur.
Later on, a new version of
mylib 1.0-SNAPSHOT
is published changing the version on the server. The next time we perform dependency resolution, the following will happen:
- maven local repository is searched, dependency is found to be the same "age" as the version in the cache so will not be updated (i.e. downloaded)
- maven network repository is NOT searched as we've already found the dependency
This is likely to not be the desired outcome. We are now out of sync with the latest published snapshot and will continue to keep using the version from the local maven repository.
The rule to remember is this: when resolving a dependency, Ivy will stop searching as soon as it finds a repository that has the dependency at the specified version number. It will
not continue searching all repositories trying to find a more recently modified instance.
To remedy this situation (i.e. build against the
newer version of
mylib 1.0-SNAPSHOT
in the remote repository), you can either:
- Delete the version from the local maven repository, or
- Reorder the repositories in the
BuildConfig.groovy
file
Where possible, prefer deleting the version from the local maven repository. In general, when you have finished building against a locally built SNAPSHOT always try to clear it from the local maven repository.
This changing dependency behaviour is an unmodifiable characteristic of the underlying dependency management system Apache Ivy. It is currently not possible to have Ivy search all repositories to look for newer versions (in terms of modification date) of the same dependency (i.e. the same combination of group
, name
and version
). If you want this behavior consider switching to Aether as the dependency manager.
5.7.7 Dependency Reports
As mentioned in the previous section a Grails application consists of dependencies inherited from the framework, the plugins installed and the application dependencies itself.
To obtain a report of an application's dependencies you can run the
dependency-report command:
By default this will generate reports in the
target/dependency-report
directory. You can specify which configuration (scope) you want a report for by passing an argument containing the configuration name:
grails dependency-report runtime
As of Grails 2.3 the
dependency-report
command will also output to the console a graph of the dependencies of an application. Example output it shown below:
compile - Dependencies placed on the classpath for compilation (total: 73)
+--- org.codehaus.groovy:groovy-all:2.0.6
+--- org.grails:grails-plugin-codecs:2.3.0
| --- org.grails:grails-web:2.3.0
| --- commons-fileupload:commons-fileupload:1.2.2
| --- xpp3:xpp3_min:1.1.4c
| --- commons-el:commons-el:1.0
| --- opensymphony:sitemesh:2.4
| --- org.springframework:spring-webmvc:3.1.2.RELEASE
| --- commons-codec:commons-codec:1.5
| --- org.slf4j:slf4j-api:1.7.2
+--- org.grails:grails-plugin-controllers:2.3.0
| --- commons-beanutils:commons-beanutils:1.8.3
| --- org.grails:grails-core:2.3.0
...
5.7.8 Plugin JAR Dependencies
Specifying Plugin JAR dependencies
The way in which you specify dependencies for a
plugin is identical to how you specify dependencies in an application. When a plugin is installed into an application the application automatically inherits the dependencies of the plugin.
To define a dependency that is resolved for use with the plugin but not
exported to the application then you can set the
export
property of the dependency:
compile('org.spockframework:spock-core:0.5-groovy-1.8') {
export = false
}
In this case the Spock dependency will be available only to the plugin and not resolved as an application dependency. Alternatively, if you're using the Map syntax:
compile group: 'org.spockframework', name: 'spock-core',
version: '0.5-groovy-1.8', export: false
You can use exported = false
instead of export = false
, but we recommend the latter because it's consistent with the Map argument.
Overriding Plugin JAR Dependencies in Your Application
If a plugin is using a JAR which conflicts with another plugin, or an application dependency then you can override how a plugin resolves its dependencies inside an application using exclusions. For example:
plugins {
compile(":hibernate:$grailsVersion") {
excludes "javassist"
}
}dependencies {
runtime "javassist:javassist:3.4.GA"
}
In this case the application explicitly declares a dependency on the "hibernate" plugin and specifies an exclusion using the
excludes
method, effectively excluding the javassist library as a dependency.
5.7.9 Maven Integration
When using the Grails Maven plugin with the Maven build tool, Grails' dependency resolution mechanics are disabled as it is assumed that you will manage dependencies with Maven's
pom.xml
file.
However, if you would like to continue using Grails regular commands like
run-app,
test-app and so on then you can tell Grails' command line to load dependencies from the Maven
pom.xml
file instead.
To do so simply add the following line to your
BuildConfig.groovy
:
grails.project.dependency.resolution = {
pom true
..
}
The line
pom true
tells Grails to parse Maven's
pom.xml
and load dependencies from there.
5.7.10 Deploying to a Maven Repository
If you use Maven to build your Grails project, you can use the standard Maven targets
mvn install
and
mvn deploy
.
If not, you can deploy a Grails project or plugin to a Maven repository using the
release plugin.
The plugin provides the ability to publish Grails projects and plugins to local and remote Maven repositories. There are two key additional targets added by the plugin:
- maven-install - Installs a Grails project or plugin into your local Maven cache
- maven-deploy - Deploys a Grails project or plugin to a remote Maven repository
By default this plugin will automatically generate a valid
pom.xml
for you unless a
pom.xml
is already present in the root of the project, in which case this
pom.xml
file will be used.
maven-install
The
maven-install
command will install the Grails project or plugin artifact into your local Maven cache:
In the case of plugins, the plugin zip file will be installed, whilst for application the application WAR file will be installed.
maven-deploy
The
maven-deploy
command will deploy a Grails project or plugin into a remote Maven repository:
It is assumed that you have specified the necessary
<distributionManagement>
configuration within a
pom.xml
or that you specify the
id
of the remote repository to deploy to:
grails maven-deploy --repository=myRepo
The
repository
argument specifies the 'id' for the repository. Configure the details of the repository specified by this 'id' within your
grails-app/conf/BuildConfig.groovy
file or in your
$USER_HOME/.grails/settings.groovy
file:
grails.project.dependency.distribution = {
localRepository = "/path/to/my/local"
remoteRepository(id: "myRepo", url: "http://myserver/path/to/repo")
}
The syntax for configuring remote repositories matches the syntax from the
remoteRepository element in the Ant Maven tasks. For example the following XML:
<remoteRepository id="myRepo" url="scp://localhost/www/repository">
<authentication username="..." privateKey="${user.home}/.ssh/id_dsa"/>
</remoteRepository>
Can be expressed as:
remoteRepository(id: "myRepo", url: "scp://localhost/www/repository") {
authentication username: "...", privateKey: "${userHome}/.ssh/id_dsa"
}
By default the plugin will try to detect the protocol to use from the URL of the repository (e.g. "http" from "http://.." etc.), however to specify a different protocol you can do:
grails maven-deploy --repository=myRepo --protocol=webdav
The available protocols are:
- http
- scp
- scpexe
- ftp
- webdav
Groups, Artifacts and Versions
Maven defines the notion of a 'groupId', 'artifactId' and a 'version'. This plugin pulls this information from the Grails project conventions or plugin descriptor.
Projects
For applications this plugin will use the Grails application name and version provided by Grails when generating the
pom.xml
file. To change the version you can run the
set-version
command:
The Maven
groupId
will be the same as the project name, unless you specify a different one in Config.groovy:
grails.project.groupId="com.mycompany"
Plugins
With a Grails plugin the
groupId
and
version
are taken from the following properties in the
GrailsPlugin.groovy
descriptor:
String groupId = 'myOrg'
String version = '0.1'
The 'artifactId' is taken from the plugin name. For example if you have a plugin called
FeedsGrailsPlugin
the
artifactId
will be "feeds". If your plugin does not specify a
groupId
then this defaults to "org.grails.plugins".
5.7.11 Plugin Dependencies
You can declaratively specify plugins as dependencies via the dependency DSL instead of using the
install-plugin command:
grails.project.dependency.resolution = {
…
repositories {
…
} plugins {
runtime ':hibernate:1.2.1'
} dependencies {
…
}
…
}
If you don't specify a group id the default plugin group id of
org.grails.plugins
is used.
Latest Integration
Only the Ivy dependency manager supports the "latest.integration" version. For Aether you can achieve a similar effect with version ranges.
You can specify to use the latest version of a particular plugin by using "latest.integration" as the version number:
plugins {
runtime ':hibernate:latest.integration'
}
Integration vs. Release
The "latest.integration" version label will also include resolving snapshot versions. To not include snapshot versions then use the "latest.release" label:
plugins {
runtime ':hibernate:latest.release'
}
The "latest.release" label only works with Maven compatible repositories. If you have a regular SVN-based Grails repository then you should use "latest.integration".
And of course if you use a Maven repository with an alternative group id you can specify a group id:
plugins {
runtime 'mycompany:hibernate:latest.integration'
}
Plugin Exclusions
You can control how plugins transitively resolves both plugin and JAR dependencies using exclusions. For example:
plugins {
runtime(':weceem:0.8') {
excludes "searchable"
}
}
Here we have defined a dependency on the "weceem" plugin which transitively depends on the "searchable" plugin. By using the
excludes
method you can tell Grails
not to transitively install the searchable plugin. You can combine this technique to specify an alternative version of a plugin:
plugins {
runtime(':weceem:0.8') {
excludes "searchable" // excludes most recent version
}
runtime ':searchable:0.5.4' // specifies a fixed searchable version
}
You can also completely disable transitive plugin installs, in which case no transitive dependencies will be resolved:
plugins {
runtime(':weceem:0.8') {
transitive = false
}
runtime ':searchable:0.5.4' // specifies a fixed searchable version
}
5.7.12 Caching of Dependency Resolution Results
As a performance optimisation, when using Ivy (this does not apply to Aether), Grails does not resolve dependencies for every command invocation. Even with all the necessary dependencies downloaded and cached, resolution may take a second or two. To minimise this cost, Grails caches the result of dependency resolution (i.e. the location on the local file system of all of the declared dependencies, typically inside the dependency cache) and reuses this result for subsequent commands when it can reasonably expect that nothing has changed.
Grails only performs dependency resolution under the following circumstances:
- The project is clean (i.e. fresh checkout or after
grails clean
)
- The
BuildConfig.groovy
file has changed since the last command was run
- The
--refresh-dependencies
command line switch is provided to the command (any command)
- The
refresh-dependencies
command is the command being executed
Generally, this strategy works well and you can ignore dependency resolution caching. Every time you change your dependencies (i.e. modify
BuildConfig.groovy
) Grails will do the right thing and resolve your new dependencies.
However, when you have
changing or
dynamic dependencies you will have to consider dependency resolution caching.
{info}
A
changing dependency is one whose version number does not change, but its contents do (like a SNAPSHOT). A
dynamic dependency is one that is defined as one of many possible options (like a dependency with a version range, or symbolic version number like
latest.integration
).
{info}
Both
changing and
dynamic dependencies are influenced by the environment. With caching active, any changes to the environment are effectively ignored. For example, your project may not automatically fetch the very latest version of a dependency when using
latest.integration
. Or if you declare a
SNAPSHOT
dependency, you may not automatically get the latest that's available on the server.
To ensure you have the correct version of a
changing or
dynamic dependency in your project, you can:
- clean the project
- run the
refresh-dependencies
command
- run any command with the
--refresh-dependencies
switch; or
- make a change to
BuildConfig.groovy
If you have your CI builds configured to not perform clean builds, it may be worth adding the
--refresh-dependencies
switch to the command you use to build your projects.
6 The Command Line
Grails' command line system is built on
Gant - a simple Groovy wrapper around
Apache Ant.
However, Grails takes it further through the use of convention and the
grails
command. When you type:
Grails searches in the following directories for Gant scripts to execute:
USER_HOME/.grails/scripts
PROJECT_HOME/scripts
PROJECT_HOME/plugins/*/scripts
GRAILS_HOME/scripts
Grails will also convert command names that are in lower case form such as run-app into camel case. So typing
Results in a search for the following files:
USER_HOME/.grails/scripts/RunApp.groovy
PROJECT_HOME/scripts/RunApp.groovy
PLUGINS_HOME/*/scripts/RunApp.groovy
GLOBAL_PLUGINS_HOME/*/scripts/RunApp.groovy
GRAILS_HOME/scripts/RunApp.groovy
If multiple matches are found Grails will give you a choice of which one to execute.
When Grails executes a Gant script, it invokes the "default" target defined in that script. If there is no default, Grails will quit with an error.
To get a list of all commands and some help about the available commands type:
which outputs usage instructions and the list of commands Grails is aware of:
Usage (optionals marked with *):
grails [environment]* [target] [arguments]*Examples:
grails dev run-app
grails create-app booksAvailable Targets (type grails help 'target-name' for more info):
grails bootstrap
grails bug-report
grails clean
grails compile
...
Refer to the Command Line reference in the Quick Reference menu of the reference guide for more information about individual commands
It's often useful to provide custom arguments to the JVM when running Grails commands, in particular with
run-app
where you may for example want to set a higher maximum heap size. The Grails command will use any JVM options provided in the general
JAVA_OPTS
environment variable, but you can also specify a Grails-specific environment variable too:
export GRAILS_OPTS="-Xmx1G -Xms256m -XX:MaxPermSize=256m"
grails run-app
non-interactive mode
When you run a script manually and it prompts you for information, you can answer the questions and continue running the script. But when you run a script as part of an automated process, for example a continuous integration build server, there's no way to "answer" the questions. So you can pass the
--non-interactive
switch to the script command to tell Grails to accept the default answer for any questions, for example whether to install a missing plugin.
For example:
grails war --non-interactive
6.1 Interactive Mode
Interactive mode is the a feature of the Grails command line which keeps the JVM running and allows for quicker execution of commands. To activate interactive mode type 'grails' at the command line and then use TAB completion to get a list of commands:
If you need to open a file whilst within interactive mode you can use the
open
command which will TAB complete file paths:
Even better, the
open
command understands the logical aliases 'test-report' and 'dep-report', which will open the most recent test and dependency reports respectively. In other words, to open the test report in a browser simply execute
open test-report
. You can even open multiple files at once:
open test-report test/unit/MyTests.groovy
will open the HTML test report in your browser and the
MyTests.groovy
source file in your text editor.
TAB completion also works for class names after the
create-*
commands:
If you need to run an external process whilst interactive mode is running you can do so by starting the command with a !:
Note that with ! (bang) commands, you get file path auto completion - ideal for external commands that operate on the file system such as 'ls', 'cat', 'git', etc.
The
stop-app
command will stop an application that has been run with the
run-app
command.
To exit interactive mode enter the
exit
command. Note that if the Grails application has been run with
run-app
normally it will terminate when the interactive mode console exits because the JVM will be terminated. An exception to this would be if the application were running in forked mode which means the application is running in a different JVM. In that case the application will be left running after the interactive mode console terminates. If you want to exit interactive mode and stop an application that is running in forked mode, use the
quit
command. The
quit
command will stop the running application and then close interactive mode.
6.2 Forked Execution
Forked Execution
Since Grails 2.3, the
run-app
,
run-war
,
test-app
and
console
commands are now executed in a forked JVM in order to isolate the build classpath from the runtime classpath.
Forked execution is configured via the
grails-app/conf/BuildConfig.groovy
file. The following is the default configuration:
grails.project.fork = [
test: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, daemon:true], // configure settings for the test-app JVM
run: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256], // configure settings for the run-app JVM
war: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256], // configure settings for the run-war JVM
console: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256]// configure settings for the Console UI JVM
]
The memory requirements of the forked JVM can be tweaked as per the requirements of the application.
Forked Test Execution
When running the
test-app command, a separate JVM is launched to execute this tests. This will have a notable impact on the speed of execution of the tests when running the command directly:
To mitigate this, Grails 2.3 and above include a feature that launches a background JVM on standby to run tests when using interactive mode. In other words, running
test-app
from interactive mode will result in faster test execution times:
$ grails
$ grails> test-app
It is recommended that forked execution is used for tests, however it does require modern hardware due to the use of multiple JVMs. You can therefore disable forked execution by setting the
grails.project.fork.test
setting to
false
:
forkConfig = [maxMemory: 1024, minMemory: 64, debug: false, maxPerm: 256]
grails.project.fork = [
test: false,
…
]
Using the Test Runner Daemon to Speed-up Test Execution
The default configuration for the testing is to activate a daemon to run tests using the
daemon
argument:
grails.project.fork = [
test: [maxMemory: 768, minMemory: 64, debug: false, maxPerm: 256, daemon:true], // configure settings for the test-app JVM
...
This only works in interactive mode, so if you start Grails with the 'grails' command and then using
test-app
the daemon will be used:
$ grails
$ grails> test-app
This has the effect of speeding-up test executions times. You can disable the daemon by setting
daemon
to
false
. If the
daemon
becomes unresponsive you can restart it with
restart-daemon
:
Debugging and Forked Execution (--debug vs --debug-fork)
An important consideration when using forked execution is that the
debug
argument will allow a remote debugger to be attached to the build JVM but not the JVM that your application is running in. To debug your application you should use the
debug-fork
argument:
grails test-app --debug-fork
Or for run-app:
grails run-app --debug-fork
Forked Tomcat Execution
Grails 2.2 and above support forked JVM execution of the Tomcat container in development mode. This has several benefits including:
- Reduced memory consumption, since the Grails build system can exit
- Isolation of the build classpath from the runtime classpath
- The ability to deploy other Grails/Spring applications in parallels without conflicting dependencies
To enable forked execution you can set the
grails.project.fork.run
property to
true
:
grails.project.fork.run=true
Then just us the regular
run-app
command as per normal. Note that in forked mode the
grails
process will exit and leave the container running in the background. To stop the server there is a new
stop-app
command:
To customize the JVM arguments passed to the forked JVM you can specify a list instead:
grails.project.fork.run= [maxMemory:1024, minMemory:64, debug:false, maxPerm:256, jvmArgs: ['-Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=5005']]
Auto-deploying additional WAR files in Forked Mode
Since forked execution isolates classpaths more effectively than embedded execution you can deploy additional WAR files (such as other Grails or Spring applications) to the container.
The easiest way to do so is to drop the WAR files into the
src/autodeploy
directory (if it doesn't exist you can create it).
You can customize the location of the autodeploy directory by specifying an alternative location in
BuildConfig.groovy
:
grails.project.autodeploy.dir="/path/to/my/war/files"
Customizing the Forked Tomcat instance
If you want to programmatically customize the forked
Tomcat instance you can do so by implementing a class named
org.grails.plugins.tomcat.ForkedTomcatCustomizer
which provides a method with the following signature:
void customize(Tomcat tomcat) {
// your code here
}
6.3 Creating Gant Scripts
You can create your own Gant scripts by running the
create-script command from the root of your project. For example the following command:
grails create-script compile-sources
Will create a script called
scripts/CompileSources.groovy
. A Gant script itself is similar to a regular Groovy script except that it supports the concept of "targets" and dependencies between them:
target(default:"The default target is the one that gets executed by Grails") {
depends(clean, compile)
}target(clean:"Clean out things") {
ant.delete(dir:"output")
}target(compile:"Compile some sources") {
ant.mkdir(dir:"mkdir")
ant.javac(srcdir:"src/java", destdir:"output")
}
As demonstrated in the script above, there is an implicit
ant
variable (an instance of
groovy.util.AntBuilder
) that allows access to the
Apache Ant API.
In previous versions of Grails (1.0.3 and below), the variable was Ant
, i.e. with a capital first letter.
You can also "depend" on other targets using the
depends
method demonstrated in the
default
target above.
The default target
In the example above, we specified a target with the explicit name "default". This is one way of defining the default target for a script. An alternative approach is to use the
setDefaultTarget()
method:
target("clean-compile": "Performs a clean compilation on the app source") {
depends(clean, compile)
}target(clean:"Clean out things") {
ant.delete(dir:"output")
}target(compile:"Compile some sources") {
ant.mkdir(dir:"mkdir")
ant.javac(srcdir:"src/java", destdir:"output")
}setDefaultTarget("clean-compile")
This lets you call the default target directly from other scripts if you wish. Also, although we have put the call to
setDefaultTarget()
at the end of the script in this example, it can go anywhere as long as it comes
after the target it refers to ("clean-compile" in this case).
Which approach is better? To be honest, you can use whichever you prefer - there don't seem to be any major advantages in either case. One thing we would say is that if you want to allow other scripts to call your "default" target, you should move it into a shared script that doesn't have a default target at all. We'll talk some more about this in the next section.
6.4 Re-using Grails scripts
Grails ships with a lot of command line functionality out of the box that you may find useful in your own scripts (See the command line reference in the reference guide for info on all the commands). Of particular use are the
compile,
package and
bootstrap scripts.
The
bootstrap script for example lets you bootstrap a Spring
ApplicationContext instance to get access to the data source and so on (the integration tests use this):
includeTargets << grailsScript("_GrailsBootstrap")target ('default': "Database stuff") {
depends(configureProxy, packageApp, classpath, loadApp, configureApp) Connection c
try {
c = appCtx.getBean('dataSource').getConnection()
// do something with connection
}
finally {
c?.close()
}
}
Pulling in targets from other scripts
Gant lets you pull in all targets (except "default") from another Gant script. You can then depend upon or invoke those targets as if they had been defined in the current script. The mechanism for doing this is the
includeTargets
property. Simply "append" a file or class to it using the left-shift operator:
includeTargets << new File("/path/to/my/script.groovy")
includeTargets << gant.tools.Ivy
Don't worry too much about the syntax using a class, it's quite specialised. If you're interested, look into the Gant documentation.
Core Grails targets
As you saw in the example at the beginning of this section, you use neither the File- nor the class-based syntax for
includeTargets
when including core Grails targets. Instead, you should use the special
grailsScript()
method that is provided by the Grails command launcher (note that this is not available in normal Gant scripts, just Grails ones).
The syntax for the
grailsScript()
method is pretty straightforward: simply pass it the name of the Grails script to include, without any path information. Here is a list of Grails scripts that you could reuse:
Script | Description |
---|
_GrailsSettings | You really should include this! Fortunately, it is included automatically by all other Grails scripts except _GrailsProxy, so you usually don't have to include it explicitly. |
_GrailsEvents | Include this to fire events. Adds an event(String eventName, List args) method. Again, included by almost all other Grails scripts. |
_GrailsClasspath | Configures compilation, test, and runtime classpaths. If you want to use or play with them, include this script. Again, included by almost all other Grails scripts. |
_GrailsProxy | If you don't have direct access to the internet and use a proxy, include this script to configure access through your proxy. |
_GrailsArgParsing | Provides a parseArguments target that does what it says on the tin: parses the arguments provided by the user when they run your script. Adds them to the argsMap property. |
_GrailsTest | Contains all the shared test code. Useful if you want to add any extra tests. |
_GrailsRun | Provides all you need to run the application in the configured servlet container, either normally (runApp /runAppHttps ) or from a WAR file (runWar /runWarHttps ). |
There are many more scripts provided by Grails, so it is worth digging into the scripts themselves to find out what kind of targets are available. Anything that starts with an "_" is designed for reuse.
Script architecture
You maybe wondering what those underscores are doing in the names of the Grails scripts. That is Grails' way of determining that a script is
internal , or in other words that it has not corresponding "command". So you can't run "grails _grails-settings" for example. That is also why they don't have a default target.
Internal scripts are all about code sharing and reuse. In fact, we recommend you take a similar approach in your own scripts: put all your targets into an internal script that can be easily shared, and provide simple command scripts that parse any command line arguments and delegate to the targets in the internal script. For example if you have a script that runs some functional tests, you can split it like this:
./scripts/FunctionalTests.groovy:includeTargets << new File("${basedir}/scripts/_FunctionalTests.groovy")target(default: "Runs the functional tests for this project.") {
depends(runFunctionalTests)
}./scripts/_FunctionalTests.groovy:includeTargets << grailsScript("_GrailsTest")target(runFunctionalTests: "Run functional tests.") {
depends(...)
…
}
Here are a few general guidelines on writing scripts:
- Split scripts into a "command" script and an internal one.
- Put the bulk of the implementation in the internal script.
- Put argument parsing into the "command" script.
- To pass arguments to a target, create some script variables and initialise them before calling the target.
- Avoid name clashes by using closures assigned to script variables instead of targets. You can then pass arguments direct to the closures.
6.5 Hooking into Events
Grails provides the ability to hook into scripting events. These are events triggered during execution of Grails target and plugin scripts.
The mechanism is deliberately simple and loosely specified. The list of possible events is not fixed in any way, so it is possible to hook into events triggered by plugin scripts, for which there is no equivalent event in the core target scripts.
Defining event handlers
Event handlers are defined in scripts called
_Events.groovy
. Grails searches for these scripts in the following locations:
USER_HOME/.grails/scripts
- user-specific event handlers
PROJECT_HOME/scripts
- applicaton-specific event handlers
PLUGINS_HOME/*/scripts
- plugin-specific event handlers
GLOBAL_PLUGINS_HOME/*/scripts
- event handlers provided by global plugins
Whenever an event is fired,
all the registered handlers for that event are executed. Note that the registration of handlers is performed automatically by Grails, so you just need to declare them in the relevant
_Events.groovy
file.
Event handlers are blocks defined in
_Events.groovy
, with a name beginning with "event". The following example can be put in your /scripts directory to demonstrate the feature:
eventCreatedArtefact = { type, name ->
println "Created $type $name"
}eventStatusUpdate = { msg ->
println msg
}eventStatusFinal = { msg ->
println msg
}
You can see here the three handlers
eventCreatedArtefact
,
eventStatusUpdate
,
eventStatusFinal
. Grails provides some standard events, which are documented in the command line reference guide. For example the
compile command fires the following events:
CompileStart
- Called when compilation starts, passing the kind of compile - source or tests
CompileEnd
- Called when compilation is finished, passing the kind of compile - source or tests
Triggering events
To trigger an event simply include the Init.groovy script and call the event() closure:
includeTargets << grailsScript("_GrailsEvents")event("StatusFinal", ["Super duper plugin action complete!"])
Common Events
Below is a table of some of the common events that can be leveraged:
Event | Parameters | Description |
---|
StatusUpdate | message | Passed a string indicating current script status/progress |
StatusError | message | Passed a string indicating an error message from the current script |
StatusFinal | message | Passed a string indicating the final script status message, i.e. when completing a target, even if the target does not exit the scripting environment |
CreatedArtefact | artefactType,artefactName | Called when a create-xxxx script has completed and created an artefact |
CreatedFile | fileName | Called whenever a project source filed is created, not including files constantly managed by Grails |
Exiting | returnCode | Called when the scripting environment is about to exit cleanly |
PluginInstalled | pluginName | Called after a plugin has been installed |
CompileStart | kind | Called when compilation starts, passing the kind of compile - source or tests |
CompileEnd | kind | Called when compilation is finished, passing the kind of compile - source or tests |
DocStart | kind | Called when documentation generation is about to start - javadoc or groovydoc |
DocEnd | kind | Called when documentation generation has ended - javadoc or groovydoc |
SetClasspath | rootLoader | Called during classpath initialization so plugins can augment the classpath with rootLoader.addURL(...). Note that this augments the classpath after event scripts are loaded so you cannot use this to load a class that your event script needs to import, although you can do this if you load the class by name. |
PackagingEnd | none | Called at the end of packaging (which is called prior to the Tomcat server being started and after web.xml is generated) |
6.6 Customising the build
Grails is most definitely an opinionated framework and it prefers convention to configuration, but this doesn't mean you
can't configure it. In this section, we look at how you can influence and modify the standard Grails build.
The defaults
The core of the Grails build configuration is the
grails.util.BuildSettings
class, which contains quite a bit of useful information. It controls where classes are compiled to, what dependencies the application has, and other such settings.
Here is a selection of the configuration options and their default values:
Property | Config option | Default value |
---|
grailsWorkDir | grails.work.dir | $USER_HOME/.grails/<grailsVersion> |
projectWorkDir | grails.project.work.dir | <grailsWorkDir>/projects/<baseDirName> |
classesDir | grails.project.class.dir | <projectWorkDir>/classes |
testClassesDir | grails.project.test.class.dir | <projectWorkDir>/test-classes |
testReportsDir | grails.project.test.reports.dir | <projectWorkDir>/test/reports |
resourcesDir | grails.project.resource.dir | <projectWorkDir>/resources |
projectPluginsDir | grails.project.plugins.dir | <projectWorkDir>/plugins |
globalPluginsDir | grails.global.plugins.dir | <grailsWorkDir>/global-plugins |
verboseCompile | grails.project.compile.verbose | false |
The
BuildSettings
class has some other properties too, but they should be treated as read-only:
Property | Description |
---|
baseDir | The location of the project. |
userHome | The user's home directory. |
grailsHome | The location of the Grails installation in use (may be null ). |
grailsVersion | The version of Grails being used by the project. |
grailsEnv | The current Grails environment. |
config | The configuration settings defined in the project's BuildConfig.groovy file. Access properties in the same way as you access runtime settings: grailsSettings.config.foo.bar.hello . |
compileDependencies | A list of compile-time project dependencies as File instances. |
testDependencies | A list of test-time project dependencies as File instances. |
runtimeDependencies | A list of runtime-time project dependencies as File instances. |
Of course, these properties aren't much good if you can't get hold of them. Fortunately that's easy to do: an instance of
BuildSettings
is available to your scripts as the
grailsSettings
script variable. You can also access it from your code by using the
grails.util.BuildSettingsHolder
class, but this isn't recommended.
Overriding the defaults
All of the properties in the first table can be overridden by a system property or a configuration option - simply use the "config option" name. For example, to change the project working directory, you could either run this command:
grails -Dgrails.project.work.dir=work compile
or add this option to your
grails-app/conf/BuildConfig.groovy
file:
grails.project.work.dir = "work"
Note that the default values take account of the property values they depend on, so setting the project working directory like this would also relocate the compiled classes, test classes, resources, and plugins.
What happens if you use both a system property and a configuration option? Then the system property wins because it takes precedence over the
BuildConfig.groovy
file, which in turn takes precedence over the default values.
The
BuildConfig.groovy
file is a sibling of
grails-app/conf/Config.groovy
- the former contains options that only affect the build, whereas the latter contains those that affect the application at runtime. It's not limited to the options in the first table either: you will find build configuration options dotted around the documentation, such as ones for specifying the port that the embedded servlet container runs on or for determining what files get packaged in the WAR file.
Available build settings
Name | Description |
---|
grails.server.port.http | Port to run the embedded servlet container on ("run-app" and "run-war"). Integer. |
grails.server.port.https | Port to run the embedded servlet container on for HTTPS ("run-app --https" and "run-war --https"). Integer. |
grails.config.base.webXml | Path to a custom web.xml file to use for the application (alternative to using the web.xml template). |
grails.compiler.dependencies | Legacy approach to adding extra dependencies to the compiler classpath. Set it to a closure containing "fileset()" entries. These entries will be processed by an AntBuilder so the syntax is the Groovy form of the corresponding XML elements in an Ant build file, e.g. fileset(dir: "$basedir/lib", includes: "**/*.class") . |
grails.testing.patterns | A list of Ant path patterns that let you control which files are included in the tests. The patterns should not include the test case suffix, which is set by the next property. |
grails.testing.nameSuffix | By default, tests are assumed to have a suffix of "Tests". You can change it to anything you like but setting this option. For example, another common suffix is "Test". |
grails.project.war.file | A string containing the file path of the generated WAR file, along with its full name (include extension). For example, "target/my-app.war". |
grails.war.dependencies | A closure containing "fileset()" entries that allows you complete control over what goes in the WAR's "WEB-INF/lib" directory. |
grails.war.copyToWebApp | A closure containing "fileset()" entries that allows you complete control over what goes in the root of the WAR. It overrides the default behaviour of including everything under "web-app". |
grails.war.resources | A closure that takes the location of the staging directory as its first argument. You can use any Ant tasks to do anything you like. It is typically used to remove files from the staging directory before that directory is jar'd up into a WAR. |
grails.project.web.xml | The location to generate Grails' web.xml to |
Reloading Agent Cache Directory
Grails uses an agent based reloading system in the development environment that allows source code changes to be picked up while the application is running. This reloading agent caches information needed to carry out the reloading efficiently. By default this information is stored under
<USER_HOME_DIR>/.grails/.slcache/
. The
GRAILS_AGENT_CACHE_DIR
environment variable may be assigned a value to cause this cache information to be stored somewhere else. Note that this is an operating system environment variable, not a JVM system property or a property which may be defined in
BuildConfig.groovy
. This setting must be defined as an environment variable because the agent cache directory must be configured very early in the JVM startup process, before any Grails code is executed.
6.7 Ant and Maven
If all the other projects in your team or company are built using a standard build tool such as Ant or Maven, you become the black sheep of the family when you use the Grails command line to build your application. Fortunately, you can easily integrate the Grails build system into the main build tools in use today (well, the ones in use in Java projects at least).
Maven Integration
Grails provides integration with
Maven 3 with a Maven plugin.
Preparation
In order to use the Maven plugin, all you need is Maven 3 installed and set up. This is because
you no longer need to install Grails separately to use it with Maven!
The Maven 3 integration for Grails has been designed and tested for Maven 3.1.0 and above. It will not work with earlier versions.
Creating a Grails Maven Project
Using the
create-pom
command you can generate a valid Maven
pom.xml
file for any existing Grails project. The below presents an example:
$ grails create-app myapp
$ cd myapp
$ grails create-pom com.mycompany
The
create-pom
command expects a group id as an argument. The name and the version are taken from the
application.properties
of the application. The Maven plugin will keep the version in the
pom.xml
in sync with the version in
application.properties
.
The following standard Maven commands are then possible:
compile
- Compiles a Grails project
package
- Builds a WAR file from the Grails project.
install
- Builds a WAR file (or plugin zip/jar if a plugin) and installs it into your local Maven cache
test
- Runs the tests of a Grails project
clean
- Cleans the Grails project
Other standard Maven commands will likely work too.
You can also use some of the Grails commands that have been wrapped as Maven goals:
For a complete, up to date list, run
mvn grails:help
Defining Plugin Dependencies
All Grails plugins are published to a standard Maven repository located at
. When using the Maven plugin for Grails you must ensure that this repository is declared in your list of remote repositories:
<repository>
<id>grails-plugins</id>
<name>grails-plugins</name>
<url>http://repo.grails.org/grails/plugins</url>
</repository>
With this done you can declare plugin dependencies within your
pom.xml
file:
<dependency>
<groupId>org.grails.plugins</groupId>
<artifactId>database-migration</artifactId>
<version>1.1</version>
<scope>runtime</scope>
<type>zip</type>
</dependency>
Note that the
type
element must be set to
zip
.
Specifying the Grails Version to Use
The 2.4.0 version of the Maven plugin works with different versions of Grails. By default it tries to auto-detect the version of Grails to use from the
grails-dependencies
dependency definition found in the
pom.xml
:
<dependency>
<groupId>org.grails</groupId>
<artifactId>grails-dependencies</artifactId>
<version>2.4.0</version>
</dependency>
If you change the version of
grails-dependencies
then a different version of Grails will be used. However, you can also explicitly define the Grails version to be used in the plugin
configuration
:
<plugin>
<groupId>org.grails</groupId>
<artifactId>grails-maven-plugin</artifactId>
<version>2.4.0</version>
<configuration>
<grailsVersion>2.4.0</grailsVersion>
</configuration>
<extensions>true</extensions>
</plugin>
Debugging Grails Execution
The Maven plugin will run Grails commands in a separate process, meaning that the Grails process occupies a separate JVM as the Maven process.
To debug the Grails process you need to configure the
forkDebug
option in the plugin's
configuration
tag:
<plugin>
<groupId>org.grails</groupId>
<artifactId>grails-maven-plugin</artifactId>
<version>2.4.0</version>
<configuration>
<forkDebug>true</forkDebug>
</configuration>
<extensions>true</extensions>
</plugin>
With this configuration in place the JVM executed in Maven will load in debug mode.
If you need to customize the memory of the forked process the following elements are available:
forkMaxMemory
- The maximum amount of heap (default 1024)
forkMinMemory
- The minimum amount of heap (default 512)
forkPermGen
- The amount of permgen (default 256)
Multi Module Maven Builds
The Maven plugin can be used to power multi-module Grails builds. The easiest way to set this is up is with the
create-multi-project-build
command:
$ grails create-app myapp
$ grails create-plugin plugin1
$ grails create-plugin plugin2
$ grails create-multi-project-build org.mycompany:parent:1.0
Running
mvn install
will build all projects together. To enable the 'grails' command to read the POMs you can modify
BuildConfig.groovy
to use the POM and resolve dependencies from your Maven local cache:
grails.project.dependency.resolution = {
…
pom true
repositories {
…
mavenLocal()
}
}
By reading the
pom.xml
file you can do an initial
mvn install
from the parent project to build all plugins and install them into your local maven cache and then
cd
into your project and use the regular
grails run-app
command to run your application. All previously built plugins will be resolved from the local Maven cache.
Adding Grails commands to phases
The standard POM created for you by Grails already attaches the appropriate core Grails commands to their corresponding build phases, so "compile" goes in the "compile" phase and "war" goes in the "package" phase. That doesn't help though when you want to attach a plugin's command to a particular phase. The classic example is functional tests. How do you make sure that your functional tests (using which ever plugin you have decided on) are run during the "integration-test" phase?
Fear not: all things are possible. In this case, you can associate the command to a phase using an extra "execution" block:
<plugin>
<groupId>org.grails</groupId>
<artifactId>grails-maven-plugin</artifactId>
<version>2.4.0</version>
<extensions>true</extensions>
<executions>
<execution>
<goals>
…
</goals>
</execution>
<execution>
<id>functional-tests</id>
<phase>integration-test</phase>
<goals>
<goal>exec</goal>
</goals>
<configuration>
<command>functional-tests</command>
</configuration>
</execution>
</executions>
</plugin>
This also demonstrates the
grails:exec
goal, which can be used to run any Grails command. Simply pass the name of the command as the
command
system property, and optionally specify the arguments with the
args
property:
mvn grails:exec -Dcommand=create-webtest -Dargs=Book
Raising issues
If you come across any problems with the Maven integration, please raise a
JIRA issue.
Ant Integration
When you create a Grails application with the
create-app command, Grails doesn't automatically create an Ant
build.xml
file but you can generate one with the
integrate-with command:
grails integrate-with --ant
This creates a
build.xml
file containing the following targets:
clean
- Cleans the Grails application
compile
- Compiles your application's source code
test
- Runs the unit tests
run
- Equivalent to "grails run-app"
war
- Creates a WAR file
deploy
- Empty by default, but can be used to implement automatic deployment
Each of these can be run by Ant, for example:
The build file is configured to use
Apache Ivy for dependency management, which means that it will automatically download all the requisite Grails JAR files and other dependencies on demand. You don't even have to install Grails locally to use it! That makes it particularly useful for continuous integration systems such as
CruiseControl or
Jenkins.
It uses the Grails
Ant task to hook into the existing Grails build system. The task lets you run any Grails script that's available, not just the ones used by the generated build file. To use the task, you must first declare it:
<taskdef name="grailsTask"
classname="grails.ant.GrailsTask"
classpathref="grails.classpath"/>
This raises the question: what should be in "grails.classpath"? The task itself is in the "grails-bootstrap" JAR artifact, so that needs to be on the classpath at least. You should also include the "groovy-all" JAR. With the task defined, you just need to use it! The following table shows you what attributes are available:
Attribute | Description | Required |
---|
home | The location of the Grails installation directory to use for the build. | Yes, unless classpath is specified. |
classpathref | Classpath to load Grails from. Must include the "grails-bootstrap" artifact and should include "grails-scripts". | Yes, unless home is set or you use a classpath element. |
script | The name of the Grails script to run, e.g. "TestApp". | Yes. |
args | The arguments to pass to the script, e.g. "-unit -xml". | No. Defaults to "". |
environment | The Grails environment to run the script in. | No. Defaults to the script default. |
includeRuntimeClasspath | Advanced setting: adds the application's runtime classpath to the build classpath if true. | No. Defaults to true . |
The task also supports the following nested elements, all of which are standard Ant path structures:
classpath
- The build classpath (used to load Gant and the Grails scripts).
compileClasspath
- Classpath used to compile the application's classes.
runtimeClasspath
- Classpath used to run the application and package the WAR. Typically includes everything in @compileClasspath.
testClasspath
- Classpath used to compile and run the tests. Typically includes everything in runtimeClasspath
.
How you populate these paths is up to you. If you use the
home
attribute and put your own dependencies in the
lib
directory, then you don't even need to use any of them. For an example of their use, take a look at the generated Ant build file for new apps.
6.8 Grails Wrapper
The Grails Wrapper allows a Grails application to built without having to install Grails and configure a GRAILS_HOME environment variable. The wrapper includes a small shell script and a couple of small bootstrap jar files that typically would be checked in to source code control along with the rest of the project. The first time the wrapper is executed it will download and configure a Grails installation. This wrapper makes it more simple to setup a development environment, configure CI and manage upgrades to future versions of Grails. When the application is upgraded to the next version of Grails, the wrapper is updated and checked in to the source code control system and the next time developers update their workspace and run the wrapper, they will automatically be using the correct version of Grails.
Generating The Wrapper
The
wrapper command can be used to generate the wrapper shell scripts and supporting jar files. Execute the wrapper command at the top of an existing Grails project.
In order to do this of course Grails must be installed and configured. This is only a requirement for bootstrapping the wrapper. Once the wrapper is generated there is no need to have a Grails installation configured in order to use the wrapper.
See the
wrapper command documentation for details about command line arguments.
By default the wrapper command will generate a
grailsw
shell script and
grailsw.bat
batch file at the top of the project. In addition to those, a
wrapper/
directory (the name of the directory is configurable via command line options) is generated which contains some support files which are necessary to run the wrapper. All of these files should be checked into the source code control system along with the rest of the project. This allows developers to check the project out of source code control and immediately start using the wrapper to execute Grails commands without having to install and configure Grails.
Using The Wrapper
The wrapper script accepts all of the same arguments as the normal grails command.
./grailsw create-domain-class com.demo.Person
./grailsw run-app
./grailsw test-app unit:etc...
7 Object Relational Mapping (GORM)
Domain classes are core to any business application. They hold state about business processes and hopefully also implement behavior. They are linked together through relationships; one-to-one, one-to-many, or many-to-many.
GORM is Grails' object relational mapping (ORM) implementation. Under the hood it uses Hibernate 3 (a very popular and flexible open source ORM solution) and thanks to the dynamic nature of Groovy with its static and dynamic typing, along with the convention of Grails, there is far less configuration involved in creating Grails domain classes.
You can also write Grails domain classes in Java. See the section on Hibernate Integration for how to write domain classes in Java but still use dynamic persistent methods. Below is a preview of GORM in action:
def book = Book.findByTitle("Groovy in Action")book
.addToAuthors(name:"Dierk Koenig")
.addToAuthors(name:"Guillaume LaForge")
.save()
7.1 Quick Start Guide
A domain class can be created with the
create-domain-class command:
grails create-domain-class helloworld.Person
If no package is specified with the create-domain-class script, Grails automatically uses the application name as the package name.
This will create a class at the location
grails-app/domain/helloworld/Person.groovy
such as the one below:
package helloworldclass Person {
}
If you have the dbCreate
property set to "update", "create" or "create-drop" on your DataSource, Grails will automatically generate/modify the database tables for you.
You can customize the class by adding properties:
class Person {
String name
Integer age
Date lastVisit
}
Once you have a domain class try and manipulate it with the
shell or
console by typing:
This loads an interactive GUI where you can run Groovy commands with access to the Spring ApplicationContext, GORM, etc.
7.1.1 Basic CRUD
Try performing some basic CRUD (Create/Read/Update/Delete) operations.
Create
To create a domain class use Map constructor to set its properties and call
save:
def p = new Person(name: "Fred", age: 40, lastVisit: new Date())
p.save()
The
save method will persist your class to the database using the underlying Hibernate ORM layer.
Read
Grails transparently adds an implicit
id
property to your domain class which you can use for retrieval:
def p = Person.get(1)
assert 1 == p.id
This uses the
get method that expects a database identifier to read the
Person
object back from the database.
You can also load an object in a read-only state by using the
read method:
In this case the underlying Hibernate engine will not do any dirty checking and the object will not be persisted. Note that
if you explicitly call the
save method then the object is placed back into a read-write state.
In addition, you can also load a proxy for an instance by using the
load method:
This incurs no database access until a method other than getId() is called. Hibernate then initializes the proxied instance, or
throws an exception if no record is found for the specified id.
Update
To update an instance, change some properties and then call
save again:
def p = Person.get(1)
p.name = "Bob"
p.save()
Delete
To delete an instance use the
delete method:
def p = Person.get(1)
p.delete()
7.2 Domain Modelling in GORM
When building Grails applications you have to consider the problem domain you are trying to solve. For example if you were building an
Amazon-style bookstore you would be thinking about books, authors, customers and publishers to name a few.
These are modeled in GORM as Groovy classes, so a
Book
class may have a title, a release date, an ISBN number and so on. The next few sections show how to model the domain in GORM.
To create a domain class you run the
create-domain-class command as follows:
grails create-domain-class org.bookstore.Book
The result will be a class at
grails-app/domain/org/bookstore/Book.groovy
:
package org.bookstoreclass Book {
}
This class will map automatically to a table in the database called
book
(the same name as the class). This behaviour is customizable through the
ORM Domain Specific LanguageNow that you have a domain class you can define its properties as Java types. For example:
package org.bookstoreclass Book {
String title
Date releaseDate
String ISBN
}
Each property is mapped to a column in the database, where the convention for column names is all lower case separated by underscores. For example
releaseDate
maps onto a column
release_date
. The SQL types are auto-detected from the Java types, but can be customized with
Constraints or the
ORM DSL.
7.2.1 Association in GORM
Relationships define how domain classes interact with each other. Unless specified explicitly at both ends, a relationship exists only in the direction it is defined.
7.2.1.1 Many-to-one and one-to-one
A many-to-one relationship is the simplest kind, and is defined with a property of the type of another domain class. Consider this example:
Example A
In this case we have a unidirectional many-to-one relationship from
Face
to
Nose
. To make this relationship bidirectional define the other side as follows (and see the section on controlling the ends of the association just below):
Example B
class Nose {
static belongsTo = [face:Face]
}
In this case we use the
belongsTo
setting to say that
Nose
"belongs to"
Face
. The result of this is that we can create a
Face
, attach a
Nose
instance to it and when we save or delete the
Face
instance, GORM will save or delete the
Nose
. In other words, saves and deletes will cascade from
Face
to the associated
Nose
:
new Face(nose:new Nose()).save()
The example above will save both face and nose. Note that the inverse
is not true and will result in an error due to a transient
Face
:
new Nose(face:new Face()).save() // will cause an error
Now if we delete the
Face
instance, the
Nose
will go too:
def f = Face.get(1)
f.delete() // both Face and Nose deleted
To make the relationship a true one-to-one, use the
hasOne
property on the owning side, e.g.
Face
:
Example C
class Face {
static hasOne = [nose:Nose]
}
Note that using this property puts the foreign key on the inverse table to the example A, so in this case the foreign key column is stored in the
nose
table inside a column called
face_id
. Also,
hasOne
only works with bidirectional relationships.
Finally, it's a good idea to add a unique constraint on one side of the one-to-one relationship:
class Face {
static hasOne = [nose:Nose] static constraints = {
nose unique: true
}
}
Controlling the ends of the association
Occasionally you may find yourself with domain classes that have multiple properties of the same type. They may even be self-referential, i.e. the association property has the same type as the domain class it's in. Such situations can cause problems because Grails may guess incorrectly the type of the association. Consider this simple class:
class Person {
String name
Person parent static belongsTo = [ supervisor: Person ] static constraints = { supervisor nullable: true }
}
As far as Grails is concerned, the
parent
and
supervisor
properties are two directions of the same association. So when you set the
parent
property on a
Person
instance, Grails will automatically set the
supervisor
property on the other
Person
instance. This may be what you want, but if you look at the class, what we in fact have are two unidirectional relationships.
To guide Grails to the correct mapping, you can tell it that a particular association is unidirectional through the
mappedBy
property:
class Person {
String name
Person parent static belongsTo = [ supervisor: Person ] static mappedBy = [ supervisor: "none", parent: "none" ] static constraints = { supervisor nullable: true }
}
You can also replace "none" with any property name of the target class. And of course this works for normal domain classes too, not just self-referential ones. Nor is the
mappedBy
property limited to many-to-one and one-to-one associations: it also works for one-to-many and many-to-many associations as you'll see in the next section.
If you have a property called "none" on your domain class, this approach won't work currently! The "none" property will be treated as the reverse direction of the association (or the "back reference"). Fortunately, "none" is not a common domain class property name.
7.2.1.2 One-to-many
A one-to-many relationship is when one class, example
Author
, has many instances of another class, example
Book
. With Grails you define such a relationship with the
hasMany
setting:
class Author {
static hasMany = [books: Book] String name
}
class Book {
String title
}
In this case we have a unidirectional one-to-many. Grails will, by default, map this kind of relationship with a join table.
The ORM DSL allows mapping unidirectional relationships using a foreign key association instead
Grails will automatically inject a property of type
java.util.Set
into the domain class based on the
hasMany
setting. This can be used to iterate over the collection:
def a = Author.get(1)for (book in a.books) {
println book.title
}
The default fetch strategy used by Grails is "lazy", which means that the collection will be lazily initialized on first access. This can lead to the n+1 problem if you are not careful.If you need "eager" fetching you can use the ORM DSL or specify eager fetching as part of a query
The default cascading behaviour is to cascade saves and updates, but not deletes unless a
belongsTo
is also specified:
class Author {
static hasMany = [books: Book] String name
}
class Book {
static belongsTo = [author: Author]
String title
}
If you have two properties of the same type on the many side of a one-to-many you have to use
mappedBy
to specify which the collection is mapped:
class Airport {
static hasMany = [flights: Flight]
static mappedBy = [flights: "departureAirport"]
}
class Flight {
Airport departureAirport
Airport destinationAirport
}
This is also true if you have multiple collections that map to different properties on the many side:
class Airport {
static hasMany = [outboundFlights: Flight, inboundFlights: Flight]
static mappedBy = [outboundFlights: "departureAirport",
inboundFlights: "destinationAirport"]
}
class Flight {
Airport departureAirport
Airport destinationAirport
}
7.2.1.3 Many-to-many
Grails supports many-to-many relationships by defining a
hasMany
on both sides of the relationship and having a
belongsTo
on the owned side of the relationship:
class Book {
static belongsTo = Author
static hasMany = [authors:Author]
String title
}
class Author {
static hasMany = [books:Book]
String name
}
Grails maps a many-to-many using a join table at the database level. The owning side of the relationship, in this case
Author
, takes responsibility for persisting the relationship and is the only side that can cascade saves across.
For example this will work and cascade saves:
new Author(name:"Stephen King")
.addToBooks(new Book(title:"The Stand"))
.addToBooks(new Book(title:"The Shining"))
.save()
However this will only save the
Book
and not the authors!
new Book(name:"Groovy in Action")
.addToAuthors(new Author(name:"Dierk Koenig"))
.addToAuthors(new Author(name:"Guillaume Laforge"))
.save()
This is the expected behaviour as, just like Hibernate, only one side of a many-to-many can take responsibility for managing the relationship.
Grails' Scaffolding feature does not currently support many-to-many relationship and hence you must write the code to manage the relationship yourself
7.2.1.4 Basic Collection Types
As well as associations between different domain classes, GORM also supports mapping of basic collection types.
For example, the following class creates a
nicknames
association that is a
Set
of
String
instances:
class Person {
static hasMany = [nicknames: String]
}
GORM will map an association like the above using a join table. You can alter various aspects of how the join table is mapped using the
joinTable
argument:
class Person { static hasMany = [nicknames: String] static mapping = {
hasMany joinTable: [name: 'bunch_o_nicknames',
key: 'person_id',
column: 'nickname',
type: "text"]
}
}
The example above will map to a table that looks like the following:
bunch_o_nicknames Table
---------------------------------------------
| person_id | nickname |
---------------------------------------------
| 1 | Fred |
---------------------------------------------
7.2.2 Composition in GORM
As well as
association, Grails supports the notion of composition. In this case instead of mapping classes onto separate tables a class can be "embedded" within the current table. For example:
class Person {
Address homeAddress
Address workAddress
static embedded = ['homeAddress', 'workAddress']
}class Address {
String number
String code
}
The resulting mapping would looking like this:
If you define the Address
class in a separate Groovy file in the grails-app/domain
directory you will also get an address
table. If you don't want this to happen use Groovy's ability to define multiple classes per file and include the Address
class below the Person
class in the grails-app/domain/Person.groovy
file
7.2.3 Inheritance in GORM
GORM supports inheritance both from abstract base classes and concrete persistent GORM entities. For example:
class Content {
String author
}
class BlogEntry extends Content {
URL url
}
class Book extends Content {
String ISBN
}
class PodCast extends Content {
byte[] audioStream
}
In the above example we have a parent
Content
class and then various child classes with more specific behaviour.
Considerations
At the database level Grails by default uses table-per-hierarchy mapping with a discriminator column called
class
so the parent class (
Content
) and its subclasses (
BlogEntry
,
Book
etc.), share the
same table.
Table-per-hierarchy mapping has a down side in that you
cannot have non-nullable properties with inheritance mapping. An alternative is to use table-per-subclass which can be enabled with the
ORM DSLHowever, excessive use of inheritance and table-per-subclass can result in poor query performance due to the use of outer join queries. In general our advice is if you're going to use inheritance, don't abuse it and don't make your inheritance hierarchy too deep.
Polymorphic Queries
The upshot of inheritance is that you get the ability to polymorphically query. For example using the
list method on the
Content
super class will return all subclasses of
Content
:
def content = Content.list() // list all blog entries, books and podcasts
content = Content.findAllByAuthor('Joe Bloggs') // find all by authordef podCasts = PodCast.list() // list only podcasts
7.2.4 Sets, Lists and Maps
Sets of Objects
By default when you define a relationship with GORM it is a
java.util.Set
which is an unordered collection that cannot contain duplicates. In other words when you have:
class Author {
static hasMany = [books: Book]
}
The books property that GORM injects is a
java.util.Set
. Sets guarantee uniqueness but not order, which may not be what you want. To have custom ordering you configure the Set as a
SortedSet
:
class Author { SortedSet books static hasMany = [books: Book]
}
In this case a
java.util.SortedSet
implementation is used which means you must implement
java.lang.Comparable
in your Book class:
class Book implements Comparable { String title
Date releaseDate = new Date() int compareTo(obj) {
releaseDate.compareTo(obj.releaseDate)
}
}
The result of the above class is that the Book instances in the books collection of the Author class will be ordered by their release date.
Lists of Objects
To keep objects in the order which they were added and to be able to reference them by index like an array you can define your collection type as a
List
:
class Author { List books static hasMany = [books: Book]
}
In this case when you add new elements to the books collection the order is retained in a sequential list indexed from 0 so you can do:
author.books[0] // get the first book
The way this works at the database level is Hibernate creates a
books_idx
column where it saves the index of the elements in the collection to retain this order at the database level.
When using a
List
, elements must be added to the collection before being saved, otherwise Hibernate will throw an exception (
org.hibernate.HibernateException
: null index column for collection):
// This won't work!
def book = new Book(title: 'The Shining')
book.save()
author.addToBooks(book)// Do it this way instead.
def book = new Book(title: 'Misery')
author.addToBooks(book)
author.save()
Bags of Objects
If ordering and uniqueness aren't a concern (or if you manage these explicitly) then you can use the Hibernate
Bag type to represent mapped collections.
The only change required for this is to define the collection type as a
Collection
:
class Author { Collection books static hasMany = [books: Book]
}
Since uniqueness and order aren't managed by Hibernate, adding to or removing from collections mapped as a Bag don't trigger a load of all existing instances from the database, so this approach will perform better and require less memory than using a
Set
or a
List
.
Maps of Objects
If you want a simple map of string/value pairs GORM can map this with the following:
class Author {
Map books // map of ISBN:book names
}def a = new Author()
a.books = ["1590597583":"Grails Book"]
a.save()
In this case the key and value of the map MUST be strings.
If you want a Map of objects then you can do this:
class Book { Map authors static hasMany = [authors: Author]
}def a = new Author(name:"Stephen King")def book = new Book()
book.authors = [stephen:a]
book.save()
The static
hasMany
property defines the type of the elements within the Map. The keys for the map
must be strings.
A Note on Collection Types and Performance
The Java
Set
type doesn't allow duplicates. To ensure uniqueness when adding an entry to a
Set
association Hibernate has to load the entire associations from the database. If you have a large numbers of entries in the association this can be costly in terms of performance.
The same behavior is required for
List
types, since Hibernate needs to load the entire association to maintain order. Therefore it is recommended that if you anticipate a large numbers of records in the association that you make the association bidirectional so that the link can be created on the inverse side. For example consider the following code:
def book = new Book(title:"New Grails Book")
def author = Author.get(1)
book.author = author
book.save()
In this example the association link is being created by the child (Book) and hence it is not necessary to manipulate the collection directly resulting in fewer queries and more efficient code. Given an
Author
with a large number of associated
Book
instances if you were to write code like the following you would see an impact on performance:
def book = new Book(title:"New Grails Book")
def author = Author.get(1)
author.addToBooks(book)
author.save()
You could also model the collection as a Hibernate Bag as described above.
7.3 Persistence Basics
A key thing to remember about Grails is that under the surface Grails is using
Hibernate for persistence. If you are coming from a background of using
ActiveRecord or
iBatis/MyBatis, Hibernate's "session" model may feel a little strange.
Grails automatically binds a Hibernate session to the currently executing request. This lets you use the
save and
delete methods as well as other GORM methods transparently.
Transactional Write-Behind
A useful feature of Hibernate over direct JDBC calls and even other frameworks is that when you call
save or
delete it does not necessarily perform any SQL operations
at that point. Hibernate batches up SQL statements and executes them as late as possible, often at the end of the request when flushing and closing the session. This is typically done for you automatically by Grails, which manages your Hibernate session.
Hibernate caches database updates where possible, only actually pushing the changes when it knows that a flush is required, or when a flush is triggered programmatically. One common case where Hibernate will flush cached updates is when performing queries since the cached information might be included in the query results. But as long as you're doing non-conflicting saves, updates, and deletes, they'll be batched until the session is flushed. This can be a significant performance boost for applications that do a lot of database writes.
Note that flushing is not the same as committing a transaction. If your actions are performed in the context of a transaction, flushing will execute SQL updates but the database will save the changes in its transaction queue and only finalize the updates when the transaction commits.
7.3.1 Saving and Updating
An example of using the
save method can be seen below:
def p = Person.get(1)
p.save()
This save will be not be pushed to the database immediately - it will be pushed when the next flush occurs. But there are occasions when you want to control when those statements are executed or, in Hibernate terminology, when the session is "flushed". To do so you can use the flush argument to the save method:
def p = Person.get(1)
p.save(flush: true)
Note that in this case
all pending SQL statements including previous saves, deletes, etc. will be synchronized with the database. This also lets you catch any exceptions, which is typically useful in highly concurrent scenarios involving
optimistic locking:
def p = Person.get(1)
try {
p.save(flush: true)
}
catch (org.springframework.dao.DataIntegrityViolationException e) {
// deal with exception
}
Another thing to bear in mind is that Grails
validates a domain instance every time you save it. If that validation fails the domain instance will
not be persisted to the database. By default,
save()
will simply return
null
in this case, but if you would prefer it to throw an exception you can use the
failOnError
argument:
def p = Person.get(1)
try {
p.save(failOnError: true)
}
catch (ValidationException e) {
// deal with exception
}
You can even change the default behaviour with a setting in
Config.groovy
, as described in the
section on configuration. Just remember that when you are saving domain instances that have been bound with data provided by the user, the likelihood of validation exceptions is quite high and you won't want those exceptions propagating to the end user.
You can find out more about the subtleties of saving data in
this article - a must read!
7.3.2 Deleting Objects
An example of the
delete method can be seen below:
def p = Person.get(1)
p.delete()
As with saves, Hibernate will use transactional write-behind to perform the delete; to perform the delete in-place you can use the
flush
argument:
def p = Person.get(1)
p.delete(flush: true)
Using the
flush
argument lets you catch any errors that occur during a delete. A common error that may occur is if you violate a database constraint, although this is normally down to a programming or schema error. The following example shows how to catch a
DataIntegrityViolationException
that is thrown when you violate the database constraints:
def p = Person.get(1)try {
p.delete(flush: true)
}
catch (org.springframework.dao.DataIntegrityViolationException e) {
flash.message = "Could not delete person ${p.name}"
redirect(action: "show", id: p.id)
}
Note that Grails does not supply a
deleteAll
method as deleting data is discouraged and can often be avoided through boolean flags/logic.
If you really need to batch delete data you can use the
executeUpdate method to do batch DML statements:
Customer.executeUpdate("delete Customer c where c.name = :oldName",
[oldName: "Fred"])
7.3.3 Understanding Cascading Updates and Deletes
It is critical that you understand how cascading updates and deletes work when using GORM. The key part to remember is the
belongsTo
setting which controls which class "owns" a relationship.
Whether it is a one-to-one, one-to-many or many-to-many, defining
belongsTo
will result in updates cascading from the owning class to its dependant (the other side of the relationship), and for many-/one-to-one and one-to-many relationships deletes will also cascade.
If you
do not define
belongsTo
then no cascades will happen and you will have to manually save each object (except in the case of the one-to-many, in which case saves will cascade automatically if a new instance is in a
hasMany
collection).
Here is an example:
class Airport {
String name
static hasMany = [flights: Flight]
}
class Flight {
String number
static belongsTo = [airport: Airport]
}
If I now create an
Airport
and add some
Flight
s to it I can save the
Airport
and have the updates cascaded down to each flight, hence saving the whole object graph:
new Airport(name: "Gatwick")
.addToFlights(new Flight(number: "BA3430"))
.addToFlights(new Flight(number: "EZ0938"))
.save()
Conversely if I later delete the
Airport
all
Flight
s associated with it will also be deleted:
def airport = Airport.findByName("Gatwick")
airport.delete()
However, if I were to remove
belongsTo
then the above cascading deletion code
would not work. To understand this better take a look at the summaries below that describe the default behaviour of GORM with regards to specific associations. Also read
part 2 of the GORM Gotchas series of articles to get a deeper understanding of relationships and cascading.
Bidirectional one-to-many with belongsTo
class A { static hasMany = [bees: B] }
class B { static belongsTo = [a: A] }
In the case of a bidirectional one-to-many where the many side defines a
belongsTo
then the cascade strategy is set to "ALL" for the one side and "NONE" for the many side.
Unidirectional one-to-many
class A { static hasMany = [bees: B] }
In the case of a unidirectional one-to-many where the many side defines no belongsTo then the cascade strategy is set to "SAVE-UPDATE".
Bidirectional one-to-many, no belongsTo
class A { static hasMany = [bees: B] }
In the case of a bidirectional one-to-many where the many side does not define a
belongsTo
then the cascade strategy is set to "SAVE-UPDATE" for the one side and "NONE" for the many side.
Unidirectional one-to-one with belongsTo
class B { static belongsTo = [a: A] }
In the case of a unidirectional one-to-one association that defines a
belongsTo
then the cascade strategy is set to "ALL" for the owning side of the relationship (A->B) and "NONE" from the side that defines the
belongsTo
(B->A)
Note that if you need further control over cascading behaviour, you can use the
ORM DSL.
7.3.4 Eager and Lazy Fetching
Associations in GORM are by default lazy. This is best explained by example:
class Airport {
String name
static hasMany = [flights: Flight]
}
class Flight {
String number
Location destination
static belongsTo = [airport: Airport]
}
class Location {
String city
String country
}
Given the above domain classes and the following code:
def airport = Airport.findByName("Gatwick")
for (flight in airport.flights) {
println flight.destination.city
}
GORM will execute a single SQL query to fetch the
Airport
instance, another to get its flights, and then 1 extra query for
each iteration over the
flights
association to get the current flight's destination. In other words you get N+1 queries (if you exclude the original one to get the airport).
Configuring Eager Fetching
An alternative approach that avoids the N+1 queries is to use eager fetching, which can be specified as follows:
class Airport {
String name
static hasMany = [flights: Flight]
static mapping = {
flights lazy: false
}
}
In this case the
flights
association will be loaded at the same time as its
Airport
instance, although a second query will be executed to fetch the collection. You can also use
fetch: 'join'
instead of
lazy: false
, in which case GORM will only execute a single query to get the airports and their flights. This works well for single-ended associations, but you need to be careful with one-to-manys. Queries will work as you'd expect right up to the moment you add a limit to the number of results you want. At that point, you will likely end up with fewer results than you were expecting. The reason for this is quite technical but ultimately the problem arises from GORM using a left outer join.
So, the recommendation is currently to use
fetch: 'join'
for single-ended associations and
lazy: false
for one-to-manys.
Be careful how and where you use eager loading because you could load your entire database into memory with too many eager associations. You can find more information on the mapping options in the
section on the ORM DSL.
Using Batch Fetching
Although eager fetching is appropriate for some cases, it is not always desirable. If you made everything eager you could quite possibly load your entire database into memory resulting in performance and memory problems. An alternative to eager fetching is to use batch fetching. You can configure Hibernate to lazily fetch results in "batches". For example:
class Airport {
String name
static hasMany = [flights: Flight]
static mapping = {
flights batchSize: 10
}
}
In this case, due to the
batchSize
argument, when you iterate over the
flights
association, Hibernate will fetch results in batches of 10. For example if you had an
Airport
that had 30 flights, if you didn't configure batch fetching you would get 1 query to fetch the
Airport
and then
30
queries to fetch each flight. With batch fetching you get 1 query to fetch the
Airport
and 3 queries to fetch each
Flight
in batches of 10. In other words, batch fetching is an optimization of the lazy fetching strategy. Batch fetching can also be configured at the class level as follows:
class Flight {
…
static mapping = {
batchSize 10
}
}
Check out
part 3 of the GORM Gotchas series for more in-depth coverage of this tricky topic.
7.3.5 Pessimistic and Optimistic Locking
Optimistic Locking
By default GORM classes are configured for optimistic locking. Optimistic locking is a feature of Hibernate which involves storing a version value in a special
version
column in the database that is incremented after each update.
The
version
column gets read into a
version
property that contains the current versioned state of persistent instance which you can access:
def airport = Airport.get(10)println airport.version
When you perform updates Hibernate will automatically check the version property against the version column in the database and if they differ will throw a
StaleObjectException. This will roll back the transaction if one is active.
This is useful as it allows a certain level of atomicity without resorting to pessimistic locking that has an inherit performance penalty. The downside is that you have to deal with this exception if you have highly concurrent writes. This requires flushing the session:
def airport = Airport.get(10)try {
airport.name = "Heathrow"
airport.save(flush: true)
}
catch (org.springframework.dao.OptimisticLockingFailureException e) {
// deal with exception
}
The way you deal with the exception depends on the application. You could attempt a programmatic merge of the data or go back to the user and ask them to resolve the conflict.
Alternatively, if it becomes a problem you can resort to pessimistic locking.
The version
will only be updated after flushing the session.
Pessimistic Locking
Pessimistic locking is equivalent to doing a SQL "SELECT * FOR UPDATE" statement and locking a row in the database. This has the implication that other read operations will be blocking until the lock is released.
In Grails pessimistic locking is performed on an existing instance with the
lock method:
def airport = Airport.get(10)
airport.lock() // lock for update
airport.name = "Heathrow"
airport.save()
Grails will automatically deal with releasing the lock for you once the transaction has been committed. However, in the above case what we are doing is "upgrading" from a regular SELECT to a SELECT..FOR UPDATE and another thread could still have updated the record in between the call to
get()
and the call to
lock()
.
To get around this problem you can use the static
lock method that takes an id just like
get:
def airport = Airport.lock(10) // lock for update
airport.name = "Heathrow"
airport.save()
In this case only SELECT..FOR UPDATE is issued.
As well as the
lock method you can also obtain a pessimistic locking using queries. For example using a dynamic finder:
def airport = Airport.findByName("Heathrow", [lock: true])
Or using criteria:
def airport = Airport.createCriteria().get {
eq('name', 'Heathrow')
lock true
}
7.3.6 Modification Checking
Once you have loaded and possibly modified a persistent domain class instance, it isn't straightforward to retrieve the original values. If you try to reload the instance using
get Hibernate will return the current modified instance from its Session cache. Reloading using another query would trigger a flush which could cause problems if your data isn't ready to be flushed yet. So GORM provides some methods to retrieve the original values that Hibernate caches when it loads the instance (which it uses for dirty checking).
isDirty
You can use the
isDirty method to check if any field has been modified:
def airport = Airport.get(10)
assert !airport.isDirty()airport.properties = params
if (airport.isDirty()) {
// do something based on changed state
}
isDirty()
does not currently check collection associations, but it does check all other persistent properties and associations.
You can also check if individual fields have been modified:
def airport = Airport.get(10)
assert !airport.isDirty()airport.properties = params
if (airport.isDirty('name')) {
// do something based on changed name
}
getDirtyPropertyNames
You can use the
getDirtyPropertyNames method to retrieve the names of modified fields; this may be empty but will not be null:
def airport = Airport.get(10)
assert !airport.isDirty()airport.properties = params
def modifiedFieldNames = airport.getDirtyPropertyNames()
for (fieldName in modifiedFieldNames) {
// do something based on changed value
}
getPersistentValue
You can use the
getPersistentValue method to retrieve the value of a modified field:
def airport = Airport.get(10)
assert !airport.isDirty()airport.properties = params
def modifiedFieldNames = airport.getDirtyPropertyNames()
for (fieldName in modifiedFieldNames) {
def currentValue = airport."$fieldName"
def originalValue = airport.getPersistentValue(fieldName)
if (currentValue != originalValue) {
// do something based on changed value
}
}
7.4 Querying with GORM
GORM supports a number of powerful ways to query from dynamic finders, to criteria to Hibernate's object oriented query language HQL. Depending on the complexity of the query you have the following options in order of flexibility and power:
- Dynamic Finders
- Where Queries
- Criteria Queries
- Hibernate Query Language (HQL)
In addition, Groovy's ability to manipulate collections with
GPath and methods like sort, findAll and so on combined with GORM results in a powerful combination.
However, let's start with the basics.
Listing instances
Use the
list method to obtain all instances of a given class:
The
list method supports arguments to perform pagination:
def books = Book.list(offset:10, max:20)
as well as sorting:
def books = Book.list(sort:"title", order:"asc")
Here, the
sort
argument is the name of the domain class property that you wish to sort on, and the
order
argument is either
asc
for
ascending or
desc
for
descending.
Retrieval by Database Identifier
The second basic form of retrieval is by database identifier using the
get method:
You can also obtain a list of instances for a set of identifiers using
getAll:
def books = Book.getAll(23, 93, 81)
7.4.1 Dynamic Finders
GORM supports the concept of
dynamic finders. A dynamic finder looks like a static method invocation, but the methods themselves don't actually exist in any form at the code level.
Instead, a method is auto-magically generated using code synthesis at runtime, based on the properties of a given class. Take for example the
Book
class:
class Book {
String title
Date releaseDate
Author author
}
class Author {
String name
}
The
Book
class has properties such as
title
,
releaseDate
and
author
. These can be used by the
findBy and
findAllBy methods in the form of "method expressions":
def book = Book.findByTitle("The Stand")book = Book.findByTitleLike("Harry Pot%")book = Book.findByReleaseDateBetween(firstDate, secondDate)book = Book.findByReleaseDateGreaterThan(someDate)book = Book.findByTitleLikeOrReleaseDateLessThan("%Something%", someDate)
Method Expressions
A method expression in GORM is made up of the prefix such as
findBy followed by an expression that combines one or more properties. The basic form is:
Book.findBy([Property][Comparator][Boolean Operator])?[Property][Comparator]
The tokens marked with a '?' are optional. Each comparator changes the nature of the query. For example:
def book = Book.findByTitle("The Stand")book = Book.findByTitleLike("Harry Pot%")
In the above example the first query is equivalent to equality whilst the latter, due to the
Like
comparator, is equivalent to a SQL
like
expression.
The possible comparators include:
InList
- In the list of given values
LessThan
- less than a given value
LessThanEquals
- less than or equal a give value
GreaterThan
- greater than a given value
GreaterThanEquals
- greater than or equal a given value
Like
- Equivalent to a SQL like expression
Ilike
- Similar to a Like
, except case insensitive
NotEqual
- Negates equality
InRange
- Between the from
and to
values of a Groovy Range
Rlike
- Performs a Regexp LIKE in MySQL or Oracle otherwise falls back to Like
Between
- Between two values (requires two arguments)
IsNotNull
- Not a null value (doesn't take an argument)
IsNull
- Is a null value (doesn't take an argument)
Notice that the last three require different numbers of method arguments compared to the rest, as demonstrated in the following example:
def now = new Date()
def lastWeek = now - 7
def book = Book.findByReleaseDateBetween(lastWeek, now)books = Book.findAllByReleaseDateIsNull()
books = Book.findAllByReleaseDateIsNotNull()
Boolean logic (AND/OR)
Method expressions can also use a boolean operator to combine two or more criteria:
def books = Book.findAllByTitleLikeAndReleaseDateGreaterThan(
"%Java%", new Date() - 30)
In this case we're using
And
in the middle of the query to make sure both conditions are satisfied, but you could equally use
Or
:
def books = Book.findAllByTitleLikeOrReleaseDateGreaterThan(
"%Java%", new Date() - 30)
You can combine as many criteria as you like, but they must all be combined with
And
or all
Or
. If you need to combine
And
and
Or
or if the number of criteria creates a very long method name, just convert the query to a
Criteria or
HQL query.
Querying Associations
Associations can also be used within queries:
def author = Author.findByName("Stephen King")def books = author ? Book.findAllByAuthor(author) : []
In this case if the
Author
instance is not null we use it in a query to obtain all the
Book
instances for the given
Author
.
Pagination and Sorting
The same pagination and sorting parameters available on the
list method can also be used with dynamic finders by supplying a map as the final parameter:
def books = Book.findAllByTitleLike("Harry Pot%",
[max: 3, offset: 2, sort: "title", order: "desc"])
7.4.2 Where Queries
The
where
method, introduced in Grails 2.0, builds on the support for
Detached Criteria by providing an enhanced, compile-time checked query DSL for common queries. The
where
method is more flexible than dynamic finders, less verbose than criteria and provides a powerful mechanism to compose queries.
Basic Querying
The
where
method accepts a closure that looks very similar to Groovy's regular collection methods. The closure should define the logical criteria in regular Groovy syntax, for example:
def query = Person.where {
firstName == "Bart"
}
Person bart = query.find()
The returned object is a
DetachedCriteria
instance, which means it is not associated with any particular database connection or session. This means you can use the
where
method to define common queries at the class level:
class Person {
static simpsons = where {
lastName == "Simpson"
}
…
}
…
Person.simpsons.each {
println it.firstname
}
Query execution is lazy and only happens upon usage of the
DetachedCriteria instance. If you want to execute a where-style query immediately there are variations of the
findAll
and
find
methods to accomplish this:
def results = Person.findAll {
lastName == "Simpson"
}
def results = Person.findAll(sort:"firstName") {
lastName == "Simpson"
}
Person p = Person.find { firstName == "Bart" }
Each Groovy operator maps onto a regular criteria method. The following table provides a map of Groovy operators to methods:
Operator | Criteria Method | Description |
---|
== | eq | Equal to |
!= | ne | Not equal to |
> | gt | Greater than |
< | lt | Less than |
>= | ge | Greater than or equal to |
<= | le | Less than or equal to |
in | inList | Contained within the given list |
==~ | like | Like a given string |
=~ | ilike | Case insensitive like |
It is possible use regular Groovy comparison operators and logic to formulate complex queries:
def query = Person.where {
(lastName != "Simpson" && firstName != "Fred") || (firstName == "Bart" && age > 9)
}
def results = query.list(sort:"firstName")
The Groovy regex matching operators map onto like and ilike queries unless the expression on the right hand side is a
Pattern
object, in which case they map onto an
rlike
query:
def query = Person.where {
firstName ==~ ~/B.+/
}
Note that rlike
queries are only supported if the underlying database supports regular expressions
A
between
criteria query can be done by combining the
in
keyword with a range:
def query = Person.where {
age in 18..65
}
Finally, you can do
isNull
and
isNotNull
style queries by using
null
with regular comparison operators:
def query = Person.where {
middleName == null
}
Query Composition
Since the return value of the
where
method is a
DetachedCriteria instance you can compose new queries from the original query:
def query = Person.where {
lastName == "Simpson"
}
def bartQuery = query.where {
firstName == "Bart"
}
Person p = bartQuery.find()
Note that you cannot pass a closure defined as a variable into the
where
method unless it has been explicitly cast to a
DetachedCriteria
instance. In other words the following will produce an error:
def callable = {
lastName == "Simpson"
}
def query = Person.where(callable)
The above must be written as follows:
import grails.gorm.DetachedCriteriadef callable = {
lastName == "Simpson"
} as DetachedCriteria<Person>
def query = Person.where(callable)
As you can see the closure definition is cast (using the Groovy
as
keyword) to a
DetachedCriteria instance targeted at the
Person
class.
Conjunction, Disjunction and Negation
As mentioned previously you can combine regular Groovy logical operators (
||
and
&&
) to form conjunctions and disjunctions:
def query = Person.where {
(lastName != "Simpson" && firstName != "Fred") || (firstName == "Bart" && age > 9)
}
You can also negate a logical comparison using
!
:
def query = Person.where {
firstName == "Fred" && !(lastName == 'Simpson')
}
Property Comparison Queries
If you use a property name on both the left hand and right side of a comparison expression then the appropriate property comparison criteria is automatically used:
def query = Person.where {
firstName == lastName
}
The following table described how each comparison operator maps onto each criteria property comparison method:
Operator | Criteria Method | Description |
---|
== | eqProperty | Equal to |
!= | neProperty | Not equal to |
> | gtProperty | Greater than |
< | ltProperty | Less than |
>= | geProperty | Greater than or equal to |
<= | leProperty | Less than or equal to |
Querying Associations
Associations can be queried by using the dot operator to specify the property name of the association to be queried:
def query = Pet.where {
owner.firstName == "Joe" || owner.firstName == "Fred"
}
You can group multiple criterion inside a closure method call where the name of the method matches the association name:
def query = Person.where {
pets { name == "Jack" || name == "Joe" }
}
This technique can be combined with other top-level criteria:
def query = Person.where {
pets { name == "Jack" } || firstName == "Ed"
}
For collection associations it is possible to apply queries to the size of the collection:
def query = Person.where {
pets.size() == 2
}
The following table shows which operator maps onto which criteria method for each size() comparison:
Operator | Criteria Method | Description |
---|
== | sizeEq | The collection size is equal to |
!= | sizeNe | The collection size is not equal to |
> | sizeGt | The collection size is greater than |
< | sizeLt | The collection size is less than |
>= | sizeGe | The collection size is greater than or equal to |
<= | sizeLe | The collection size is less than or equal to |
Subqueries
It is possible to execute subqueries within where queries. For example to find all the people older than the average age the following query can be used:
final query = Person.where {
age > avg(age)
}
The following table lists the possible subqueries:
Method | Description |
---|
avg | The average of all values |
sum | The sum of all values |
max | The maximum value |
min | The minimum value |
count | The count of all values |
property | Retrieves a property of the resulting entities |
You can apply additional criteria to any subquery by using the
of
method and passing in a closure containing the criteria:
def query = Person.where {
age > avg(age).of { lastName == "Simpson" } && firstName == "Homer"
}
Since the
property
subquery returns multiple results, the criterion used compares all results. For example the following query will find all people younger than people with the surname "Simpson":
Person.where {
age < property(age).of { lastName == "Simpson" }
}
Other Functions
There are several functions available to you within the context of a query. These are summarized in the table below:
Method | Description |
---|
second | The second of a date property |
minute | The minute of a date property |
hour | The hour of a date property |
day | The day of the month of a date property |
month | The month of a date property |
year | The year of a date property |
lower | Converts a string property to upper case |
upper | Converts a string property to lower case |
length | The length of a string property |
trim | Trims a string property |
Currently functions can only be applied to properties or associations of domain classes. You cannot, for example, use a function on a result of a subquery.
For example the following query can be used to find all pet's born in 2011:
def query = Pet.where {
year(birthDate) == 2011
}
You can also apply functions to associations:
def query = Person.where {
year(pets.birthDate) == 2009
}
Batch Updates and Deletes
Since each
where
method call returns a
DetachedCriteria instance, you can use
where
queries to execute batch operations such as batch updates and deletes. For example, the following query will update all people with the surname "Simpson" to have the surname "Bloggs":
def query = Person.where {
lastName == 'Simpson'
}
int total = query.updateAll(lastName:"Bloggs")
Note that one limitation with regards to batch operations is that join queries (queries that query associations) are not allowed.
To batch delete records you can use the
deleteAll
method:
def query = Person.where {
lastName == 'Simpson'
}
int total = query.deleteAll()
7.4.3 Criteria
Criteria is an advanced way to query that uses a Groovy builder to construct potentially complex queries. It is a much better approach than building up query strings using a
StringBuffer
.
Criteria can be used either with the
createCriteria or
withCriteria methods. The builder uses Hibernate's Criteria API. The nodes on this builder map the static methods found in the
Restrictions class of the Hibernate Criteria API. For example:
def c = Account.createCriteria()
def results = c {
between("balance", 500, 1000)
eq("branch", "London")
or {
like("holderFirstName", "Fred%")
like("holderFirstName", "Barney%")
}
maxResults(10)
order("holderLastName", "desc")
}
This criteria will select up to 10
Account
objects in a List matching the following criteria:
balance
is between 500 and 1000
branch
is 'London'
holderFirstName
starts with 'Fred' or 'Barney'
The results will be sorted in descending order by
holderLastName
.
If no records are found with the above criteria, an empty List is returned.
Conjunctions and Disjunctions
As demonstrated in the previous example you can group criteria in a logical OR using an
or { }
block:
or {
between("balance", 500, 1000)
eq("branch", "London")
}
This also works with logical AND:
and {
between("balance", 500, 1000)
eq("branch", "London")
}
And you can also negate using logical NOT:
not {
between("balance", 500, 1000)
eq("branch", "London")
}
All top level conditions are implied to be AND'd together.
Querying Associations
Associations can be queried by having a node that matches the property name. For example say the
Account
class had many
Transaction
objects:
class Account {
…
static hasMany = [transactions: Transaction]
…
}
We can query this association by using the property name
transactions
as a builder node:
def c = Account.createCriteria()
def now = new Date()
def results = c.list {
transactions {
between('date', now - 10, now)
}
}
The above code will find all the
Account
instances that have performed
transactions
within the last 10 days.
You can also nest such association queries within logical blocks:
def c = Account.createCriteria()
def now = new Date()
def results = c.list {
or {
between('created', now - 10, now)
transactions {
between('date', now - 10, now)
}
}
}
Here we find all accounts that have either performed transactions in the last 10 days OR have been recently created in the last 10 days.
Querying with Projections
Projections may be used to customise the results. Define a "projections" node within the criteria builder tree to use projections. There are equivalent methods within the projections node to the methods found in the Hibernate
Projections class:
def c = Account.createCriteria()def numberOfBranches = c.get {
projections {
countDistinct('branch')
}
}
When multiple fields are specified in the projection, a List of values will be returned. A single value will be returned otherwise.
SQL Projections
The criteria DSL provides access to Hibernate's SQL projection API.
// Box is a domain class…
class Box {
int width
int height
}
// Use SQL projections to retrieve the perimeter and area of all of the Box instances…
def c = Box.createCriteria()def results = c.list {
projections {
sqlProjection '(2 * (width + height)) as perimeter, (width * height) as area', ['perimeter', 'area'], [INTEGER, INTEGER]
}
}
The first argument to the
sqlProjection
method is the SQL which defines the projections. The second argument is a list of
Strings which represent column aliases corresponding to the projected values expressed in the SQL. The third argument
is a list of
org.hibernate.type.Type
instances which correspond to the projected values expressed in the SQL. The API
supports all
org.hibernate.type.Type
objects but constants like INTEGER, LONG, FLOAT etc. are provided by the DSL which
correspond to all of the types defined in
org.hibernate.type.StandardBasicTypes
.
Consider that the following table represents the data in the
BOX
table.
The query above would return results like this:
[[18, 14], [20, 16], [22, 18], [26, 36]]
Each of the inner lists contains the 2 projected values for each
Box
, perimeter and area.
Note that if there are other references in scope wherever your criteria query is expressed that have names that conflict
with any of the type constants described above, the code in your criteria will refer to those references, not the type
constants provided by the DSL. In the unlikely event of that happening you can disambiguate the conflict by referring
to the fully qualified Hibernate type. For example StandardBasicTypes.INTEGER
instead of INTEGER
.
If only 1 value is being projected, the alias and the type do not need to be included in a list.
def results = c.list {
projections {
sqlProjection 'sum(width * height) as totalArea', 'totalArea', INTEGER
}
}
That query would return a single result with the value of 84 as the total area of all of the
Box
instances.
The DSL supports grouped projections with the
sqlGroupProjection
method.
def results = c.list {
projections {
sqlGroupProjection 'width, sum(height) as combinedHeightsForThisWidth', 'width', ['width', 'combinedHeightsForThisWidth'], [INTEGER, INTEGER]
}
}
The first argument to the
sqlGroupProjection
method is the SQL which defines the projections. The second argument represents the
group by clause that should be part of the query. That string may be single column name or a comma separated list of column
names. The third argument is a list of
Strings which represent column aliases corresponding to the projected values expressed in the SQL. The fourth argument
is a list of
org.hibernate.type.Type
instances which correspond to the projected values expressed in the SQL.
The query above is projecting the combined heights of boxes grouped by width and would return results that look like this:
Each of the inner lists contains 2 values. The first value is a box width and the second value is the sum of the heights
of all of the boxes which have that width.
Using SQL Restrictions
You can access Hibernate's SQL Restrictions capabilities.
def c = Person.createCriteria()def peopleWithShortFirstNames = c.list {
sqlRestriction "char_length(first_name) <= 4"
}
SQL Restrictions may be parameterized to deal with SQL injection vulnerabilities related to dynamic restrictions.
def c = Person.createCriteria()def peopleWithShortFirstNames = c.list {
sqlRestriction "char_length(first_name) < ? AND char_length(first_name) > ?", [maxValue, minValue]
}
Note that the parameter there is SQL. The first_name
attribute referenced in the example refers to the persistence model, not the object model like in HQL queries. The Person
property named firstName
is mapped to the first_name
column in the database and you must refer to that in the sqlRestriction
string.Also note that the SQL used here is not necessarily portable across databases.
Using Scrollable Results
You can use Hibernate's
ScrollableResults feature by calling the scroll method:
def results = crit.scroll {
maxResults(10)
}
def f = results.first()
def l = results.last()
def n = results.next()
def p = results.previous()def future = results.scroll(10)
def accountNumber = results.getLong('number')
To quote the documentation of Hibernate ScrollableResults:
A result iterator that allows moving around within the results by arbitrary increments. The Query / ScrollableResults pattern is very similar to the JDBC PreparedStatement / ResultSet pattern and the semantics of methods of this interface are similar to the similarly named methods on ResultSet.
Contrary to JDBC, columns of results are numbered from zero.
Setting properties in the Criteria instance
If a node within the builder tree doesn't match a particular criterion it will attempt to set a property on the Criteria object itself. This allows full access to all the properties in this class. This example calls
setMaxResults
and
setFirstResult
on the
Criteria instance:
import org.hibernate.FetchMode as FM
…
def results = c.list {
maxResults(10)
firstResult(50)
fetchMode("aRelationship", FM.JOIN)
}
Querying with Eager Fetching
In the section on
Eager and Lazy Fetching we discussed how to declaratively specify fetching to avoid the N+1 SELECT problem. However, this can also be achieved using a criteria query:
def criteria = Task.createCriteria()
def tasks = criteria.list{
eq "assignee.id", task.assignee.id
join 'assignee'
join 'project'
order 'priority', 'asc'
}
Notice the usage of the
join
method: it tells the criteria API to use a
JOIN
to fetch the named associations with the
Task
instances. It's probably best not to use this for one-to-many associations though, because you will most likely end up with duplicate results. Instead, use the 'select' fetch mode:
import org.hibernate.FetchMode as FM
…
def results = Airport.withCriteria {
eq "region", "EMEA"
fetchMode "flights", FM.SELECT
}
Although this approach triggers a second query to get the
flights
association, you will get reliable results - even with the
maxResults
option.
fetchMode
and join
are general settings of the query and can only be specified at the top-level, i.e. you cannot use them inside projections or association constraints.
An important point to bear in mind is that if you include associations in the query constraints, those associations will automatically be eagerly loaded. For example, in this query:
def results = Airport.withCriteria {
eq "region", "EMEA"
flights {
like "number", "BA%"
}
}
the
flights
collection would be loaded eagerly via a join even though the fetch mode has not been explicitly set.
Method Reference
If you invoke the builder with no method name such as:
The build defaults to listing all the results and hence the above is equivalent to:
Method | Description |
---|
list | This is the default method. It returns all matching rows. |
get | Returns a unique result set, i.e. just one row. The criteria has to be formed that way, that it only queries one row. This method is not to be confused with a limit to just the first row. |
scroll | Returns a scrollable result set. |
listDistinct | If subqueries or associations are used, one may end up with the same row multiple times in the result set, this allows listing only distinct entities and is equivalent to DISTINCT_ROOT_ENTITY of the CriteriaSpecification class. |
count | Returns the number of matching rows. |
Combining Criteria
You can combine multiple criteria closures in the following way:
def emeaCriteria = {
eq "region", "EMEA"
}def results = Airport.withCriteria {
emeaCriteria.delegate = delegate
emeaCriteria()
flights {
like "number", "BA%"
}
}
This technique requires that each criteria must refer to the same domain class (i.e.
Airport
).
A more flexible approach is to use Detached Criteria, as described in the following section.
7.4.4 Detached Criteria
Detached Criteria are criteria queries that are not associated with any given database session/connection. Supported since Grails 2.0, Detached Criteria queries have many uses including allowing you to create common reusable criteria queries, execute subqueries and execute batch updates/deletes.
Building Detached Criteria Queries
The primary point of entry for using the Detached Criteria is the
grails.gorm.DetachedCriteria
class which accepts a domain class as the only argument to its constructor:
import grails.gorm.*
…
def criteria = new DetachedCriteria(Person)
Once you have obtained a reference to a detached criteria instance you can execute
where queries or criteria queries to build up the appropriate query. To build a normal criteria query you can use the
build
method:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
Note that methods on the
DetachedCriteria
instance
do not mutate the original object but instead return a new query. In other words, you have to use the return value of the
build
method to obtain the mutated criteria object:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
def bartQuery = criteria.build {
eq 'firstName', 'Bart'
}
Executing Detached Criteria Queries
Unlike regular criteria, Detached Criteria are lazy, in that no query is executed at the point of definition. Once a Detached Criteria query has been constructed then there are a number of useful query methods which are summarized in the table below:
Method | Description |
---|
list | List all matching entities |
get | Return a single matching result |
count | Count all matching records |
exists | Return true if any matching records exist |
deleteAll | Delete all matching records |
updateAll(Map) | Update all matching records with the given properties |
As an example the following code will list the first 4 matching records sorted by the
firstName
property:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
def results = criteria.list(max:4, sort:"firstName")
You can also supply additional criteria to the list method:
def results = criteria.list(max:4, sort:"firstName") {
gt 'age', 30
}
To retrieve a single result you can use the
get
or
find
methods (which are synonyms):
Person p = criteria.find() // or criteria.get()
The
DetachedCriteria
class itself also implements the
Iterable
interface which means that it can be treated like a list:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
criteria.each {
println it.firstName
}
In this case the query is only executed when the
each
method is called. The same applies to all other Groovy collection iteration methods.
You can also execute dynamic finders on
DetachedCriteria
just like on domain classes. For example:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
def bart = criteria.findByFirstName("Bart")
Using Detached Criteria for Subqueries
Within the context of a regular criteria query you can use
DetachedCriteria
to execute subquery. For example if you want to find all people who are older than the average age the following query will accomplish that:
def results = Person.withCriteria {
gt "age", new DetachedCriteria(Person).build {
projections {
avg "age"
}
}
order "firstName"
}
Notice that in this case the subquery class is the same as the original criteria query class (i.e.
Person
) and hence the query can be shortened to:
def results = Person.withCriteria {
gt "age", {
projections {
avg "age"
}
}
order "firstName"
}
If the subquery class differs from the original criteria query then you will have to use the original syntax.
In the previous example the projection ensured that only a single result was returned (the average age). If your subquery returns multiple results then there are different criteria methods that need to be used to compare the result. For example to find all the people older than the ages 18 to 65 a
gtAll
query can be used:
def results = Person.withCriteria {
gtAll "age", {
projections {
property "age"
}
between 'age', 18, 65
} order "firstName"
}
The following table summarizes criteria methods for operating on subqueries that return multiple results:
Method | Description |
---|
gtAll | greater than all subquery results |
geAll | greater than or equal to all subquery results |
ltAll | less than all subquery results |
leAll | less than or equal to all subquery results |
eqAll | equal to all subquery results |
neAll | not equal to all subquery results |
Batch Operations with Detached Criteria
The
DetachedCriteria
class can be used to execute batch operations such as batch updates and deletes. For example, the following query will update all people with the surname "Simpson" to have the surname "Bloggs":
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
int total = criteria.updateAll(lastName:"Bloggs")
Note that one limitation with regards to batch operations is that join queries (queries that query associations) are not allowed within the DetachedCriteria
instance.
To batch delete records you can use the
deleteAll
method:
def criteria = new DetachedCriteria(Person).build {
eq 'lastName', 'Simpson'
}
int total = criteria.deleteAll()
7.4.5 Hibernate Query Language (HQL)
GORM classes also support Hibernate's query language HQL, a very complete reference for which can be found
in the Hibernate documentation of the Hibernate documentation.
GORM provides a number of methods that work with HQL including
find,
findAll and
executeQuery. An example of a query can be seen below:
def results =
Book.findAll("from Book as b where b.title like 'Lord of the%'")
Positional and Named Parameters
In this case the value passed to the query is hard coded, however you can equally use positional parameters:
def results =
Book.findAll("from Book as b where b.title like ?", ["The Shi%"])
def author = Author.findByName("Stephen King")
def books = Book.findAll("from Book as book where book.author = ?",
[author])
Or even named parameters:
def results =
Book.findAll("from Book as b " +
"where b.title like :search or b.author like :search",
[search: "The Shi%"])
def author = Author.findByName("Stephen King")
def books = Book.findAll("from Book as book where book.author = :author",
[author: author])
Multiline Queries
Use the line continuation character to separate the query across multiple lines:
def results = Book.findAll("\
from Book as b, \
Author as a \
where b.author = a and a.surname = ?", ['Smith'])
Triple-quoted Groovy multiline Strings will NOT work with HQL queries.
Pagination and Sorting
You can also perform pagination and sorting whilst using HQL queries. To do so simply specify the pagination options as a Map at the end of the method call and include an "ORDER BY" clause in the HQL:
def results =
Book.findAll("from Book as b where " +
"b.title like 'Lord of the%' " +
"order by b.title asc",
[max: 10, offset: 20])
7.5 Advanced GORM Features
The following sections cover more advanced usages of GORM including caching, custom mapping and events.
7.5.1 Events and Auto Timestamping
GORM supports the registration of events as methods that get fired when certain events occurs such as deletes, inserts and updates. The following is a list of supported events:
beforeInsert
- Executed before an object is initially persisted to the database. If you return false, the insert will be cancelled.
beforeUpdate
- Executed before an object is updated. If you return false, the update will be cancelled.
beforeDelete
- Executed before an object is deleted. If you return false, the delete will be cancelled.
beforeValidate
- Executed before an object is validated
afterInsert
- Executed after an object is persisted to the database
afterUpdate
- Executed after an object has been updated
afterDelete
- Executed after an object has been deleted
onLoad
- Executed when an object is loaded from the database
To add an event simply register the relevant method with your domain class.
Do not attempt to flush the session within an event (such as with obj.save(flush:true)). Since events are fired during flushing this will cause a StackOverflowError.
Event types
The beforeInsert event
Fired before an object is saved to the database
class Person {
private static final Date NULL_DATE = new Date(0) String firstName
String lastName
Date signupDate = NULL_DATE def beforeInsert() {
if (signupDate == NULL_DATE) {
signupDate = new Date()
}
}
}
The beforeUpdate event
Fired before an existing object is updated
class Person { def securityService String firstName
String lastName
String lastUpdatedBy static constraints = {
lastUpdatedBy nullable: true
} def beforeUpdate() {
lastUpdatedBy = securityService.currentAuthenticatedUsername()
}
}
The beforeDelete event
Fired before an object is deleted.
class Person {
String name def beforeDelete() {
ActivityTrace.withNewSession {
new ActivityTrace(eventName: "Person Deleted", data: name).save()
}
}
}
Notice the usage of
withNewSession
method above. Since events are triggered whilst Hibernate is flushing using persistence methods like
save()
and
delete()
won't result in objects being saved unless you run your operations with a new
Session
.
Fortunately the
withNewSession
method lets you share the same transactional JDBC connection even though you're using a different underlying
Session
.
The beforeValidate event
Fired before an object is validated.
class Person {
String name static constraints = {
name size: 5..45
} def beforeValidate() {
name = name?.trim()
}
}
The
beforeValidate
method is run before any validators are run.
Validation may run more often than you think. It is triggered by the validate()
and save()
methods as you'd expect, but it is also typically triggered just before the view is rendered as well. So when writing beforeValidate()
implementations, make sure that they can handle being called multiple times with the same property values.
GORM supports an overloaded version of
beforeValidate
which accepts a
List
parameter which may include
the names of the properties which are about to be validated. This version of
beforeValidate
will be called
when the
validate
method has been invoked and passed a
List
of property names as an argument.
class Person {
String name
String town
Integer age static constraints = {
name size: 5..45
age range: 4..99
} def beforeValidate(List propertiesBeingValidated) {
// do pre validation work based on propertiesBeingValidated
}
}def p = new Person(name: 'Jacob Brown', age: 10)
p.validate(['age', 'name'])
Note that when validate
is triggered indirectly because of a call to the save
method that
the validate
method is being invoked with no arguments, not a List
that includes all of
the property names.
Either or both versions of
beforeValidate
may be defined in a domain class. GORM will
prefer the
List
version if a
List
is passed to
validate
but will fall back on the
no-arg version if the
List
version does not exist. Likewise, GORM will prefer the
no-arg version if no arguments are passed to
validate
but will fall back on the
List
version if the no-arg version does not exist. In that case,
null
is passed to
beforeValidate
.
The onLoad/beforeLoad event
Fired immediately before an object is loaded from the database:
class Person {
String name
Date dateCreated
Date lastUpdated def onLoad() {
log.debug "Loading ${id}"
}
}
beforeLoad()
is effectively a synonym for
onLoad()
, so only declare one or the other.
The afterLoad event
Fired immediately after an object is loaded from the database:
class Person {
String name
Date dateCreated
Date lastUpdated def afterLoad() {
name = "I'm loaded"
}
}
Custom Event Listeners
As of Grails 2.0 there is a new API for plugins and applications to register and listen for persistence events. This API is not tied to Hibernate and also works for other persistence plugins such as the
MongoDB plugin for GORM.
To use this API you need to subclass
AbstractPersistenceEventListener
(in package
org.grails.datastore.mapping.engine.event ) and implement the methods
onPersistenceEvent
and
supportsEventType
. You also must provide a reference to the datastore to the listener. The simplest possible implementation can be seen below:
public MyPersistenceListener(final Datastore datastore) {
super(datastore)
}@Override
protected void onPersistenceEvent(final AbstractPersistenceEvent event) {
switch(event.eventType) {
case PreInsert:
println "PRE INSERT ${event.entityObject}"
break
case PostInsert:
println "POST INSERT ${event.entityObject}"
break
case PreUpdate:
println "PRE UPDATE ${event.entityObject}"
break;
case PostUpdate:
println "POST UPDATE ${event.entityObject}"
break;
case PreDelete:
println "PRE DELETE ${event.entityObject}"
break;
case PostDelete:
println "POST DELETE ${event.entityObject}"
break;
case PreLoad:
println "PRE LOAD ${event.entityObject}"
break;
case PostLoad:
println "POST LOAD ${event.entityObject}"
break;
}
}@Override
public boolean supportsEventType(Class<? extends ApplicationEvent> eventType) {
return true
}
The
AbstractPersistenceEvent
class has many subclasses (
PreInsertEvent
,
PostInsertEvent
etc.) that provide further information specific to the event. A
cancel()
method is also provided on the event which allows you to veto an insert, update or delete operation.
Once you have created your event listener you need to register it with the
ApplicationContext
. This can be done in
BootStrap.groovy
:
def init = {
application.mainContext.eventTriggeringInterceptor.datastores.each { k, datastore ->
applicationContext.addApplicationListener new MyPersistenceListener(datastore)
}
}
or use this in a plugin:
def doWithApplicationContext = { applicationContext ->
application.mainContext.eventTriggeringInterceptor.datastores.each { k, datastore ->
applicationContext.addApplicationListener new MyPersistenceListener(datastore)
}
}
Hibernate Events
It is generally encouraged to use the non-Hibernate specific API described above, but if you need access to more detailed Hibernate events then you can define custom Hibernate-specific event listeners.
You can also register event handler classes in an application's
grails-app/conf/spring/resources.groovy
or in the
doWithSpring
closure in a plugin descriptor by registering a Spring bean named
hibernateEventListeners
. This bean has one property,
listenerMap
which specifies the listeners to register for various Hibernate events.
The values of the Map are instances of classes that implement one or more Hibernate listener interfaces. You can use one class that implements all of the required interfaces, or one concrete class per interface, or any combination. The valid Map keys and corresponding interfaces are listed here:
For example, you could register a class
AuditEventListener
which implements
PostInsertEventListener
,
PostUpdateEventListener
, and
PostDeleteEventListener
using the following in an application:
beans = { auditListener(AuditEventListener) hibernateEventListeners(HibernateEventListeners) {
listenerMap = ['post-insert': auditListener,
'post-update': auditListener,
'post-delete': auditListener]
}
}
or use this in a plugin:
def doWithSpring = { auditListener(AuditEventListener) hibernateEventListeners(HibernateEventListeners) {
listenerMap = ['post-insert': auditListener,
'post-update': auditListener,
'post-delete': auditListener]
}
}
Automatic timestamping
If you define a
dateCreated
property it will be set to the current date for you when you create new instances. Likewise, if you define a
lastUpdated
property it will be automatically be updated for you when you change persistent instances.
If this is not the behaviour you want you can disable this feature with:
class Person {
Date dateCreated
Date lastUpdated
static mapping = {
autoTimestamp false
}
}
If you have nullable: false
constraints on either dateCreated
or lastUpdated
, your domain instances will fail validation - probably not what you want. Omit constraints from these properties unless you disable automatic timestamping.
7.5.2 Custom ORM Mapping
Grails domain classes can be mapped onto many legacy schemas with an Object Relational Mapping DSL (domain specific language). The following sections takes you through what is possible with the ORM DSL.
None of this is necessary if you are happy to stick to the conventions defined by GORM for table names, column names and so on. You only needs this functionality if you need to tailor the way GORM maps onto legacy schemas or configures caching
Custom mappings are defined using a static
mapping
block defined within your domain class:
class Person {
…
static mapping = {
version false
autoTimestamp false
}
}
You can also configure global mappings in Config.groovy (or an external config file) using this setting:
grails.gorm.default.mapping = {
version false
autoTimestamp false
}
It has the same syntax as the standard
mapping
block but it applies to all your domain classes! You can then override these defaults within the
mapping
block of a domain class.
7.5.2.1 Table and Column Names
Table names
The database table name which the class maps to can be customized using the
table
method:
class Person {
…
static mapping = {
table 'people'
}
}
In this case the class would be mapped to a table called
people
instead of the default name of
person
.
Column names
It is also possible to customize the mapping for individual columns onto the database. For example to change the name you can do:
class Person { String firstName static mapping = {
table 'people'
firstName column: 'First_Name'
}
}
Here
firstName
is a dynamic method within the
mapping
Closure that has a single Map parameter. Since its name corresponds to a domain class persistent field, the parameter values (in this case just
"column"
) are used to configure the mapping for that property.
Column type
GORM supports configuration of Hibernate types with the DSL using the type attribute. This includes specifying user types that implement the Hibernate
org.hibernate.usertype.UserType interface, which allows complete customization of how a type is persisted. As an example if you had a
PostCodeType
you could use it as follows:
class Address { String number
String postCode static mapping = {
postCode type: PostCodeType
}
}
Alternatively if you just wanted to map it to one of Hibernate's basic types other than the default chosen by Grails you could use:
class Address { String number
String postCode static mapping = {
postCode type: 'text'
}
}
This would make the
postCode
column map to the default large-text type for the database you're using (for example TEXT or CLOB).
See the Hibernate documentation regarding
Basic Types for further information.
Many-to-One/One-to-One Mappings
In the case of associations it is also possible to configure the foreign keys used to map associations. In the case of a many-to-one or one-to-one association this is exactly the same as any regular column. For example consider the following:
class Person { String firstName
Address address static mapping = {
table 'people'
firstName column: 'First_Name'
address column: 'Person_Address_Id'
}
}
By default the
address
association would map to a foreign key column called
address_id
. By using the above mapping we have changed the name of the foreign key column to
Person_Adress_Id
.
One-to-Many Mapping
With a bidirectional one-to-many you can change the foreign key column used by changing the column name on the many side of the association as per the example in the previous section on one-to-one associations. However, with unidirectional associations the foreign key needs to be specified on the association itself. For example given a unidirectional one-to-many relationship between
Person
and
Address
the following code will change the foreign key in the
address
table:
class Person { String firstName static hasMany = [addresses: Address] static mapping = {
table 'people'
firstName column: 'First_Name'
addresses column: 'Person_Address_Id'
}
}
If you don't want the column to be in the
address
table, but instead some intermediate join table you can use the
joinTable
parameter:
class Person { String firstName static hasMany = [addresses: Address] static mapping = {
table 'people'
firstName column: 'First_Name'
addresses joinTable: [name: 'Person_Addresses',
key: 'Person_Id',
column: 'Address_Id']
}
}
Many-to-Many Mapping
Grails, by default maps a many-to-many association using a join table. For example consider this many-to-many association:
class Group {
…
static hasMany = [people: Person]
}
class Person {
…
static belongsTo = Group
static hasMany = [groups: Group]
}
In this case Grails will create a join table called
group_person
containing foreign keys called
person_id
and
group_id
referencing the
person
and
group
tables. To change the column names you can specify a column within the mappings for each class.
class Group {
…
static mapping = {
people column: 'Group_Person_Id'
}
}
class Person {
…
static mapping = {
groups column: 'Group_Group_Id'
}
}
You can also specify the name of the join table to use:
class Group {
…
static mapping = {
people column: 'Group_Person_Id',
joinTable: 'PERSON_GROUP_ASSOCIATIONS'
}
}
class Person {
…
static mapping = {
groups column: 'Group_Group_Id',
joinTable: 'PERSON_GROUP_ASSOCIATIONS'
}
}
7.5.2.2 Caching Strategy
Setting up caching
Hibernate features a second-level cache with a customizable cache provider. This needs to be configured in the
grails-app/conf/DataSource.groovy
file as follows:
hibernate {
cache.use_second_level_cache=true
cache.use_query_cache=true
cache.provider_class='org.hibernate.cache.EhCacheProvider'
}
You can customize any of these settings, for example to use a distributed caching mechanism.
For further reading on caching and in particular Hibernate's second-level cache, refer to the Hibernate documentation on the subject.
Caching instances
Call the
cache
method in your mapping block to enable caching with the default settings:
class Person {
…
static mapping = {
table 'people'
cache true
}
}
This will configure a 'read-write' cache that includes both lazy and non-lazy properties. You can customize this further:
class Person {
…
static mapping = {
table 'people'
cache usage: 'read-only', include: 'non-lazy'
}
}
Caching associations
As well as the ability to use Hibernate's second level cache to cache instances you can also cache collections (associations) of objects. For example:
class Person { String firstName static hasMany = [addresses: Address] static mapping = {
table 'people'
version false
addresses column: 'Address', cache: true
}
}
class Address {
String number
String postCode
}
This will enable a 'read-write' caching mechanism on the
addresses
collection. You can also use:
cache: 'read-write' // or 'read-only' or 'transactional'
to further configure the cache usage.
Caching Queries
You can cache queries such as dynamic finders and criteria. To do so using a dynamic finder you can pass the
cache
argument:
def person = Person.findByFirstName("Fred", [cache: true])
In order for the results of the query to be cached, you must enable caching in your mapping as discussed in the previous section.
You can also cache criteria queries:
def people = Person.withCriteria {
like('firstName', 'Fr%')
cache true
}
Cache usages
Below is a description of the different cache settings and their usages:
read-only
- If your application needs to read but never modify instances of a persistent class, a read-only cache may be used.
read-write
- If the application needs to update data, a read-write cache might be appropriate.
nonstrict-read-write
- If the application only occasionally needs to update data (i.e. if it is very unlikely that two transactions would try to update the same item simultaneously) and strict transaction isolation is not required, a nonstrict-read-write
cache might be appropriate.
transactional
- The transactional
cache strategy provides support for fully transactional cache providers such as JBoss TreeCache. Such a cache may only be used in a JTA environment and you must specify hibernate.transaction.manager_lookup_class
in the grails-app/conf/DataSource.groovy
file's hibernate
config.
7.5.2.3 Inheritance Strategies
By default GORM classes use
table-per-hierarchy
inheritance mapping. This has the disadvantage that columns cannot have a
NOT-NULL
constraint applied to them at the database level. If you would prefer to use a
table-per-subclass
inheritance strategy you can do so as follows:
class Payment {
Integer amount static mapping = {
tablePerHierarchy false
}
}class CreditCardPayment extends Payment {
String cardNumber
}
The mapping of the root
Payment
class specifies that it will not be using
table-per-hierarchy
mapping for all child classes.
7.5.2.4 Custom Database Identity
You can customize how GORM generates identifiers for the database using the DSL. By default GORM relies on the native database mechanism for generating ids. This is by far the best approach, but there are still many schemas that have different approaches to identity.
To deal with this Hibernate defines the concept of an id generator. You can customize the id generator and the column it maps to as follows:
class Person {
…
static mapping = {
table 'people'
version false
id generator: 'hilo',
params: [table: 'hi_value',
column: 'next_value',
max_lo: 100]
}
}
In this case we're using one of Hibernate's built in 'hilo' generators that uses a separate table to generate ids.
For more information on the different Hibernate generators refer to the Hibernate reference documentation
Although you don't typically specify the
id
field (Grails adds it for you) you can still configure its mapping like the other properties. For example to customise the column for the id property you can do:
class Person {
…
static mapping = {
table 'people'
version false
id column: 'person_id'
}
}
7.5.2.5 Composite Primary Keys
GORM supports the concept of composite identifiers (identifiers composed from 2 or more properties). It is not an approach we recommend, but is available to you if you need it:
import org.apache.commons.lang.builder.HashCodeBuilderclass Person implements Serializable { String firstName
String lastName boolean equals(other) {
if (!(other instanceof Person)) {
return false
} other.firstName == firstName && other.lastName == lastName
} int hashCode() {
def builder = new HashCodeBuilder()
builder.append firstName
builder.append lastName
builder.toHashCode()
} static mapping = {
id composite: ['firstName', 'lastName']
}
}
The above will create a composite id of the
firstName
and
lastName
properties of the Person class. To retrieve an instance by id you use a prototype of the object itself:
def p = Person.get(new Person(firstName: "Fred", lastName: "Flintstone"))
println p.firstName
Domain classes mapped with composite primary keys must implement the
Serializable
interface and override the
equals
and
hashCode
methods, using the properties in the composite key for the calculations. The example above uses a
HashCodeBuilder
for convenience but it's fine to implement it yourself.
Another important consideration when using composite primary keys is associations. If for example you have a many-to-one association where the foreign keys are stored in the associated table then 2 columns will be present in the associated table.
For example consider the following domain class:
class Address {
Person person
}
In this case the
address
table will have an additional two columns called
person_first_name
and
person_last_name
. If you wish the change the mapping of these columns then you can do so using the following technique:
class Address {
Person person
static mapping = {
columns {
person {
column name: "FirstName"
column name: "LastName"
}
}
}
}
7.5.2.6 Database Indices
To get the best performance out of your queries it is often necessary to tailor the table index definitions. How you tailor them is domain specific and a matter of monitoring usage patterns of your queries. With GORM's DSL you can specify which columns are used in which indexes:
class Person {
String firstName
String address
static mapping = {
table 'people'
version false
id column: 'person_id'
firstName column: 'First_Name', index: 'Name_Idx'
address column: 'Address', index: 'Name_Idx,Address_Index'
}
}
Note that you cannot have any spaces in the value of the
index
attribute; in this example
index:'Name_Idx, Address_Index'
will cause an error.
7.5.2.7 Optimistic Locking and Versioning
As discussed in the section on
Optimistic and Pessimistic Locking, by default GORM uses optimistic locking and automatically injects a
version
property into every class which is in turn mapped to a
version
column at the database level.
If you're mapping to a legacy schema that doesn't have version columns (or there's some other reason why you don't want/need this feature) you can disable this with the
version
method:
class Person {
…
static mapping = {
table 'people'
version false
}
}
If you disable optimistic locking you are essentially on your own with regards to concurrent updates and are open to the risk of users losing data (due to data overriding) unless you use pessimistic locking
Version columns types
By default Grails maps the
version
property as a
Long
that gets incremented by one each time an instance is updated. But Hibernate also supports using a
Timestamp
, for example:
import java.sql.Timestampclass Person { …
Timestamp version static mapping = {
table 'people'
}
}
There's a slight risk that two updates occurring at nearly the same time on a fast server can end up with the same timestamp value but this risk is very low. One benefit of using a
Timestamp
instead of a
Long
is that you combine the optimistic locking and last-updated semantics into a single column.
7.5.2.8 Eager and Lazy Fetching
Lazy Collections
As discussed in the section on
Eager and Lazy fetching, GORM collections are lazily loaded by default but you can change this behaviour with the ORM DSL. There are several options available to you, but the most common ones are:
- lazy: false
- fetch: 'join'
and they're used like this:
class Person { String firstName
Pet pet static hasMany = [addresses: Address] static mapping = {
addresses lazy: false
pet fetch: 'join'
}
}
class Address {
String street
String postCode
}
class Pet {
String name
}
The first option,
lazy: false
, ensures that when a
Person
instance is loaded, its
addresses
collection is loaded at the same time with a second SELECT. The second option is basically the same, except the collection is loaded with a JOIN rather than another SELECT. Typically you want to reduce the number of queries, so
fetch: 'join'
is the more appropriate option. On the other hand, it could feasibly be the more expensive approach if your domain model and data result in more and larger results than would otherwise be necessary.
For more advanced users, the other settings available are:
- batchSize: N
- lazy: false, batchSize: N
where N is an integer. These let you fetch results in batches, with one query per batch. As a simple example, consider this mapping for
Person
:
class Person { String firstName
Pet pet static mapping = {
pet batchSize: 5
}
}
If a query returns multiple
Person
instances, then when we access the first
pet
property, Hibernate will fetch that
Pet
plus the four next ones. You can get the same behaviour with eager loading by combining
batchSize
with the
lazy: false
option. You can find out more about these options in the
Hibernate user guide and this
primer on fetching strategies. Note that ORM DSL does not currently support the "subselect" fetching strategy.
Lazy Single-Ended Associations
In GORM, one-to-one and many-to-one associations are by default lazy. Non-lazy single ended associations can be problematic when you load many entities because each non-lazy association will result in an extra SELECT statement. If the associated entities also have non-lazy associations, the number of queries grows significantly!
Use the same technique as for lazy collections to make a one-to-one or many-to-one association non-lazy/eager:
class Person {
String firstName
}
class Address { String street
String postCode static belongsTo = [person: Person] static mapping = {
person lazy: false
}
}
Here we configure GORM to load the associated
Person
instance (through the
person
property) whenever an
Address
is loaded.
Lazy Single-Ended Associations and Proxies
Hibernate uses runtime-generated proxies to facilitate single-ended lazy associations; Hibernate dynamically subclasses the entity class to create the proxy.
Consider the previous example but with a lazily-loaded
person
association: Hibernate will set the
person
property to a proxy that is a subclass of
Person
. When you call any of the getters (except for the
id
property) or setters on that proxy, Hibernate will load the entity from the database.
Unfortunately this technique can produce surprising results. Consider the following example classes:
class Pet {
String name
}
class Dog extends Pet {
}
class Person {
String name
Pet pet
}
and assume that we have a single
Person
instance with a
Dog
as the
pet
. The following code will work as you would expect:
def person = Person.get(1)
assert person.pet instanceof Dog
assert Pet.get(person.petId) instanceof Dog
But this won't:
def person = Person.get(1)
assert person.pet instanceof Dog
assert Pet.list()[0] instanceof Dog
The second assertion fails, and to add to the confusion, this will work:
assert Pet.list()[0] instanceof Dog
What's going on here? It's down to a combination of how proxies work and the guarantees that the Hibernate session makes. When you load the
Person
instance, Hibernate creates a proxy for its
pet
relation and attaches it to the session. Once that happens, whenever you retrieve that
Pet
instance with a query, a
get()
, or the
pet
relation
within the same session , Hibernate gives you the proxy.
Fortunately for us, GORM automatically unwraps the proxy when you use
get()
and
findBy*()
, or when you directly access the relation. That means you don't have to worry at all about proxies in the majority of cases. But GORM doesn't do that for objects returned with a query that returns a list, such as
list()
and
findAllBy*()
. However, if Hibernate hasn't attached the proxy to the session, those queries will return the real instances - hence why the last example works.
You can protect yourself to a degree from this problem by using the
instanceOf
method by GORM:
def person = Person.get(1)
assert Pet.list()[0].instanceOf(Dog)
However, it won't help here if casting is involved. For example, the following code will throw a
ClassCastException
because the first pet in the list is a proxy instance with a class that is neither
Dog
nor a sub-class of
Dog
:
def person = Person.get(1)
Dog pet = Pet.list()[0]
Of course, it's best not to use static types in this situation. If you use an untyped variable for the pet instead, you can access any
Dog
properties or methods on the instance without any problems.
These days it's rare that you will come across this issue, but it's best to be aware of it just in case. At least you will know why such an error occurs and be able to work around it.
7.5.2.9 Custom Cascade Behaviour
As described in the section on
cascading updates, the primary mechanism to control the way updates and deletes cascade from one association to another is the static
belongsTo property.
However, the ORM DSL gives you complete access to Hibernate's
transitive persistence capabilities using the
cascade
attribute.
Valid settings for the cascade attribute include:
merge
- merges the state of a detached association
save-update
- cascades only saves and updates to an association
delete
- cascades only deletes to an association
lock
- useful if a pessimistic lock should be cascaded to its associations
refresh
- cascades refreshes to an association
evict
- cascades evictions (equivalent to discard()
in GORM) to associations if set
all
- cascade all operations to associations
all-delete-orphan
- Applies only to one-to-many associations and indicates that when a child is removed from an association then it should be automatically deleted. Children are also deleted when the parent is.
It is advisable to read the section in the Hibernate documentation on transitive persistence to obtain a better understanding of the different cascade styles and recommendations for their usage
To specify the cascade attribute simply define one or more (comma-separated) of the aforementioned settings as its value:
class Person { String firstName static hasMany = [addresses: Address] static mapping = {
addresses cascade: "all-delete-orphan"
}
}
class Address {
String street
String postCode
}
7.5.2.10 Custom Hibernate Types
You saw in an earlier section that you can use composition (with the
embedded
property) to break a table into multiple objects. You can achieve a similar effect with Hibernate's custom user types. These are not domain classes themselves, but plain Java or Groovy classes. Each of these types also has a corresponding "meta-type" class that implements
org.hibernate.usertype.UserType.
The
Hibernate reference manual has some information on custom types, but here we will focus on how to map them in Grails. Let's start by taking a look at a simple domain class that uses an old-fashioned (pre-Java 1.5) type-safe enum class:
class Book { String title
String author
Rating rating static mapping = {
rating type: RatingUserType
}
}
All we have done is declare the
rating
field the enum type and set the property's type in the custom mapping to the corresponding
UserType
implementation. That's all you have to do to start using your custom type. If you want, you can also use the other column settings such as "column" to change the column name and "index" to add it to an index.
Custom types aren't limited to just a single column - they can be mapped to as many columns as you want. In such cases you explicitly define in the mapping what columns to use, since Hibernate can only use the property name for a single column. Fortunately, Grails lets you map multiple columns to a property using this syntax:
class Book { String title
Name author
Rating rating static mapping = {
author type: NameUserType, {
column name: "first_name"
column name: "last_name"
}
rating type: RatingUserType
}
}
The above example will create "first_name" and "last_name" columns for the
author
property. You'll be pleased to know that you can also use some of the normal column/property mapping attributes in the column definitions. For example:
column name: "first_name", index: "my_idx", unique: true
The column definitions do
not support the following attributes:
type
,
cascade
,
lazy
,
cache
, and
joinTable
.
One thing to bear in mind with custom types is that they define the
SQL types for the corresponding database columns. That helps take the burden of configuring them yourself, but what happens if you have a legacy database that uses a different SQL type for one of the columns? In that case, override the column's SQL type using the
sqlType
attribute:
class Book { String title
Name author
Rating rating static mapping = {
author type: NameUserType, {
column name: "first_name", sqlType: "text"
column name: "last_name", sqlType: "text"
}
rating type: RatingUserType, sqlType: "text"
}
}
Mind you, the SQL type you specify needs to still work with the custom type. So overriding a default of "varchar" with "text" is fine, but overriding "text" with "yes_no" isn't going to work.
7.5.2.11 Derived Properties
A derived property is one that takes its value from a SQL expression, often but not necessarily based on the value of one or more other persistent properties. Consider a Product class like this:
class Product {
Float price
Float taxRate
Float tax
}
If the
tax
property is derived based on the value of
price
and
taxRate
properties then is probably no need to persist the
tax
property. The SQL used to derive the value of a derived property may be expressed in the ORM DSL like this:
class Product {
Float price
Float taxRate
Float tax static mapping = {
tax formula: 'PRICE * TAX_RATE'
}
}
Note that the formula expressed in the ORM DSL is SQL so references to other properties should relate to the persistence model not the object model, which is why the example refers to
PRICE
and
TAX_RATE
instead of
price
and
taxRate
.
With that in place, when a Product is retrieved with something like
Product.get(42)
, the SQL that is generated to support that will look something like this:
select
product0_.id as id1_0_,
product0_.version as version1_0_,
product0_.price as price1_0_,
product0_.tax_rate as tax4_1_0_,
product0_.PRICE * product0_.TAX_RATE as formula1_0_
from
product product0_
where
product0_.id=?
Since the
tax
property is derived at runtime and not stored in the database it might seem that the same effect could be achieved by adding a method like
getTax()
to the
Product
class that simply returns the product of the
taxRate
and
price
properties. With an approach like that you would give up the ability query the database based on the value of the
tax
property. Using a derived property allows exactly that. To retrieve all
Product
objects that have a
tax
value greater than 21.12 you could execute a query like this:
Product.findAllByTaxGreaterThan(21.12)
Derived properties may be referenced in the Criteria API:
Product.withCriteria {
gt 'tax', 21.12f
}
The SQL that is generated to support either of those would look something like this:
select
this_.id as id1_0_,
this_.version as version1_0_,
this_.price as price1_0_,
this_.tax_rate as tax4_1_0_,
this_.PRICE * this_.TAX_RATE as formula1_0_
from
product this_
where
this_.PRICE * this_.TAX_RATE>?
Because the value of a derived property is generated in the database and depends on the execution of SQL code, derived properties may not have GORM constraints applied to them. If constraints are specified for a derived property, they will be ignored.
7.5.2.12 Custom Naming Strategy
By default Grails uses Hibernate's
ImprovedNamingStrategy
to convert domain class Class and field names to SQL table and column names by converting from camel-cased Strings to ones that use underscores as word separators. You can customize these on a per-class basis in the
mapping
closure but if there's a consistent pattern you can specify a different
NamingStrategy
class to use.
Configure the class name to be used in
grails-app/conf/DataSource.groovy
in the
hibernate
section, e.g.
dataSource {
pooled = true
dbCreate = "create-drop"
…
}hibernate {
cache.use_second_level_cache = true
…
naming_strategy = com.myco.myproj.CustomNamingStrategy
}
You can also specify the name of the class and it will be loaded for you:
hibernate {
…
naming_strategy = 'com.myco.myproj.CustomNamingStrategy'
}
A third option is to provide an instance if there is some configuration required beyond calling the default constructor:
hibernate {
…
def strategy = new com.myco.myproj.CustomNamingStrategy()
// configure as needed
naming_strategy = strategy
}
You can use an existing class or write your own, for example one that prefixes table names and column names:
package com.myco.myprojimport org.hibernate.cfg.ImprovedNamingStrategy
import org.hibernate.util.StringHelperclass CustomNamingStrategy extends ImprovedNamingStrategy { String classToTableName(String className) {
"table_" + StringHelper.unqualify(className)
} String propertyToColumnName(String propertyName) {
"col_" + StringHelper.unqualify(propertyName)
}
}
7.5.3 Default Sort Order
You can sort objects using query arguments such as those found in the
list method:
def airports = Airport.list(sort:'name')
However, you can also declare the default sort order for a collection in the mapping:
class Airport {
…
static mapping = {
sort "name"
}
}
The above means that all collections of
Airport
instances will by default be sorted by the airport name. If you also want to change the sort
order , use this syntax:
class Airport {
…
static mapping = {
sort name: "desc"
}
}
Finally, you can configure sorting at the association level:
class Airport {
…
static hasMany = [flights: Flight] static mapping = {
flights sort: 'number', order: 'desc'
}
}
In this case, the
flights
collection will always be sorted in descending order of flight number.
These mappings will not work for default unidirectional one-to-many or many-to-many relationships because they involve a join table. See this issue for more details. Consider using a SortedSet
or queries with sort parameters to fetch the data you need.
7.6 Programmatic Transactions
Grails is built on Spring and uses Spring's Transaction abstraction for dealing with programmatic transactions. However, GORM classes have been enhanced to make this simpler with the
withTransaction method. This method has a single parameter, a Closure, which has a single parameter which is a Spring
TransactionStatus instance.
Here's an example of using
withTransaction
in a controller methods:
def transferFunds() {
Account.withTransaction { status ->
def source = Account.get(params.from)
def dest = Account.get(params.to) def amount = params.amount.toInteger()
if (source.active) {
if (dest.active) {
source.balance -= amount
dest.amount += amount
}
else {
status.setRollbackOnly()
}
}
}
}
In this example we rollback the transaction if the destination account is not active. Also, if an unchecked
Exception
or
Error
(but not a checked
Exception
, even though Groovy doesn't require that you catch checked exceptions) is thrown during the process the transaction will automatically be rolled back.
You can also use "save points" to rollback a transaction to a particular point in time if you don't want to rollback the entire transaction. This can be achieved through the use of Spring's
SavePointManager interface.
The
withTransaction
method deals with the begin/commit/rollback logic for you within the scope of the block.
7.7 GORM and Constraints
Although constraints are covered in the
Validation section, it is important to mention them here as some of the constraints can affect the way in which the database schema is generated.
Where feasible, Grails uses a domain class's constraints to influence the database columns generated for the corresponding domain class properties.
Consider the following example. Suppose we have a domain model with the following properties:
String name
String description
By default, in MySQL, Grails would define these columns as
Column | Data Type |
---|
name | varchar(255) |
description | varchar(255) |
But perhaps the business rules for this domain class state that a description can be up to 1000 characters in length. If that were the case, we would likely define the column as follows
if we were creating the table with an SQL script.
Column | Data Type |
---|
description | TEXT |
Chances are we would also want to have some application-based validation to make sure we don't exceed that 1000 character limit
before we persist any records. In Grails, we achieve this validation with
constraints. We would add the following constraint declaration to the domain class.
static constraints = {
description maxSize: 1000
}
This constraint would provide both the application-based validation we want and it would also cause the schema to be generated as shown above. Below is a description of the other constraints that influence schema generation.
Constraints Affecting String Properties
If either the
maxSize
or the
size
constraint is defined, Grails sets the maximum column length based on the constraint value.
In general, it's not advisable to use both constraints on the same domain class property. However, if both the
maxSize
constraint and the
size
constraint are defined, then Grails sets the column length to the minimum of the
maxSize
constraint and the upper bound of the size constraint. (Grails uses the minimum of the two, because any length that exceeds that minimum will result in a validation error.)
If the
inList
constraint is defined (and the
maxSize
and the
size
constraints are not defined), then Grails sets the maximum column length based on the length of the longest string in the list of valid values. For example, given a list including values "Java", "Groovy", and "C++", Grails would set the column length to 6 (i.e., the number of characters in the string "Groovy").
Constraints Affecting Numeric Properties
If the
max
,
min
, or
range
constraint is defined, Grails attempts to set the column precision based on the constraint value. (The success of this attempted influence is largely dependent on how Hibernate interacts with the underlying DBMS.)
In general, it's not advisable to combine the pair
min
/
max
and
range
constraints together on the same domain class property. However, if both of these constraints is defined, then Grails uses the minimum precision value from the constraints. (Grails uses the minimum of the two, because any length that exceeds that minimum precision will result in a validation error.)
If the scale constraint is defined, then Grails attempts to set the column
scale based on the constraint value. This rule only applies to floating point numbers (i.e.,
java.lang.Float
,
java.Lang.Double
,
java.lang.BigDecimal
, or subclasses of
java.lang.BigDecimal
). The success of this attempted influence is largely dependent on how Hibernate interacts with the underlying DBMS.
The constraints define the minimum/maximum numeric values, and Grails derives the maximum number of digits for use in the precision. Keep in mind that specifying only one of
min
/
max
constraints will not affect schema generation (since there could be large negative value of property with max:100, for example), unless the specified constraint value requires more digits than default Hibernate column precision is (19 at the moment). For example:
someFloatValue max: 1000000, scale: 3
would yield:
someFloatValue DECIMAL(19, 3) // precision is default
but
someFloatValue max: 12345678901234567890, scale: 5
would yield:
someFloatValue DECIMAL(25, 5) // precision = digits in max + scale
and
someFloatValue max: 100, min: -100000
would yield:
someFloatValue DECIMAL(8, 2) // precision = digits in min + default scale
8 The Web Layer
8.1 Controllers
A controller handles requests and creates or prepares the response. A controller can generate the response directly or delegate to a view. To create a controller, simply create a class whose name ends with
Controller
in the
grails-app/controllers
directory (in a subdirectory if it's in a package).
The default
URL Mapping configuration ensures that the first part of your controller name is mapped to a URI and each action defined within your controller maps to URIs within the controller name URI.
8.1.1 Understanding Controllers and Actions
Creating a controller
Controllers can be created with the
create-controller or
generate-controller command. For example try running the following command from the root of a Grails project:
grails create-controller book
The command will create a controller at the location
grails-app/controllers/myapp/BookController.groovy
:
package myappclass BookController { def index() { }
}
where "myapp" will be the name of your application, the default package name if one isn't specified.
BookController
by default maps to the /book URI (relative to your application root).
The create-controller
and generate-controller
commands are just for convenience and you can just as easily create controllers using your favorite text editor or IDE
Creating Actions
A controller can have multiple public action methods; each one maps to a URI:
class BookController { def list() { // do controller logic
// create model return model
}
}
This example maps to the
/book/list
URI by default thanks to the property being named
list
.
Public Methods as Actions
In earlier versions of Grails actions were implemented with Closures. This is still supported, but the preferred approach is to use methods.
Leveraging methods instead of Closure properties has some advantages:
- Memory efficient
- Allow use of stateless controllers (
singleton
scope)
- You can override actions from subclasses and call the overridden superclass method with
super.actionName()
- Methods can be intercepted with standard proxying mechanisms, something that is complicated to do with Closures since they're fields.
If you prefer the Closure syntax or have older controller classes created in earlier versions of Grails and still want the advantages of using methods, you can set the
grails.compile.artefacts.closures.convert
property to true in
BuildConfig.groovy
:
grails.compile.artefacts.closures.convert = true
and a compile-time AST transformation will convert your Closures to methods in the generated bytecode.
If a controller class extends some other class which is not defined under the grails-app/controllers/
directory, methods inherited from that class are not converted to controller actions. If the intent is to expose those inherited methods as controller actions the methods may be overridden in the subclass and the subclass method may invoke the method in the super class.
The Default Action
A controller has the concept of a default URI that maps to the root URI of the controller, for example
/book
for
BookController
. The action that is called when the default URI is requested is dictated by the following rules:
- If there is only one action, it's the default
- If you have an action named
index
, it's the default
- Alternatively you can set it explicitly with the
defaultAction
property:
static defaultAction = "list"
8.1.2 Controllers and Scopes
Available Scopes
Scopes are hash-like objects where you can store variables. The following scopes are available to controllers:
- servletContext - Also known as application scope, this scope lets you share state across the entire web application. The servletContext is an instance of ServletContext
- session - The session allows associating state with a given user and typically uses cookies to associate a session with a client. The session object is an instance of HttpSession
- request - The request object allows the storage of objects for the current request only. The request object is an instance of HttpServletRequest
- params - Mutable map of incoming request query string or POST parameters
- flash - See below
Accessing Scopes
Scopes can be accessed using the variable names above in combination with Groovy's array index operator, even on classes provided by the Servlet API such as the
HttpServletRequest:
class BookController {
def find() {
def findBy = params["findBy"]
def appContext = request["foo"]
def loggedUser = session["logged_user"]
}
}
You can also access values within scopes using the de-reference operator, making the syntax even more clear:
class BookController {
def find() {
def findBy = params.findBy
def appContext = request.foo
def loggedUser = session.logged_user
}
}
This is one of the ways that Grails unifies access to the different scopes.
Using Flash Scope
Grails supports the concept of
flash scope as a temporary store to make attributes available for this request and the next request only. Afterwards the attributes are cleared. This is useful for setting a message directly before redirecting, for example:
def delete() {
def b = Book.get(params.id)
if (!b) {
flash.message = "User not found for id ${params.id}"
redirect(action:list)
}
… // remaining code
}
When the
list
action is requested, the
message
value will be in scope and can be used to display an information message. It will be removed from the
flash
scope after this second request.
Note that the attribute name can be anything you want, and the values are often strings used to display messages, but can be any object type.
Scoped Controllers
Supported controller scopes are:
prototype
(default) - A new controller will be created for each request (recommended for actions as Closure properties)
session
- One controller is created for the scope of a user session
singleton
- Only one instance of the controller ever exists (recommended for actions as methods)
To enable one of the scopes, add a static
scope
property to your class with one of the valid scope values listed above, for example
static scope = "singleton"
You can define the default strategy under in
Config.groovy
with the
grails.controllers.defaultScope
key, for example:
grails.controllers.defaultScope = "singleton"
Newly created applications have the
grails.controllers.defaultScope
property set in
grails-app/conf/Config.groovy
with a value of "singleton". You may change this value to any
of the supported scopes listed above. If the property is not assigned a value at all, controllers will default to "prototype" scope.
Use scoped controllers wisely. For instance, we don't recommend having any properties in a singleton-scoped controller since they will be shared for all requests.
8.1.3 Models and Views
Returning the Model
A model is a Map that the view uses when rendering. The keys within that Map correspond to variable names accessible by the view. There are a couple of ways to return a model. First, you can explicitly return a Map instance:
def show() {
[book: Book.get(params.id)]
}
The above does not reflect what you should use with the scaffolding views - see the scaffolding section for more details.
A more advanced approach is to return an instance of the Spring
ModelAndView class:
import org.springframework.web.servlet.ModelAndViewdef index() {
// get some books just for the index page, perhaps your favorites
def favoriteBooks = ... // forward to the list view to show them
return new ModelAndView("/book/list", [ bookList : favoriteBooks ])
}
One thing to bear in mind is that certain variable names can not be used in your model:
Currently, no error will be reported if you do use them, but this will hopefully change in a future version of Grails.
Selecting the View
In both of the previous two examples there was no code that specified which
view to render. So how does Grails know which one to pick? The answer lies in the conventions. Grails will look for a view at the location
grails-app/views/book/show.gsp
for this
show
action:
class BookController {
def show() {
[book: Book.get(params.id)]
}
}
To render a different view, use the
render method:
def show() {
def map = [book: Book.get(params.id)]
render(view: "display", model: map)
}
In this case Grails will attempt to render a view at the location
grails-app/views/book/display.gsp
. Notice that Grails automatically qualifies the view location with the
book
directory of the
grails-app/views
directory. This is convenient, but to access shared views you need instead you can use an absolute path instead of a relative one:
def show() {
def map = [book: Book.get(params.id)]
render(view: "/shared/display", model: map)
}
In this case Grails will attempt to render a view at the location
grails-app/views/shared/display.gsp
.
Grails also supports JSPs as views, so if a GSP isn't found in the expected location but a JSP is, it will be used instead.
Selecting Views For Namespaced Controllers
If a controller defines a namespace for itself with the
namespace property that will affect the root directory in which Grails will look for views which are specified with a relative path. The default root directory for views rendered by a namespaced controller is
grails-app/views/<namespace name>/<controller name>/
. If the view is not found in the namespaced directory then Grails will fallback to looking for the view in the non-namespaced directory.
See the example below.
class ReportingController {
static namespace = 'business' def humanResources() {
// This will render grails-app/views/business/reporting/humanResources.gsp
// if it exists. // If grails-app/views/business/reporting/humanResources.gsp does not
// exist the fallback will be grails-app/views/reporting/humanResources.gsp. // The namespaced GSP will take precedence over the non-namespaced GSP. [numberOfEmployees: 9]
}
def accountsReceivable() {
// This will render grails-app/views/business/reporting/accounting.gsp
// if it exists. // If grails-app/views/business/reporting/accounting.gsp does not
// exist the fallback will be grails-app/views/reporting/accounting.gsp. // The namespaced GSP will take precedence over the non-namespaced GSP. render view: 'numberCrunch', model: [numberOfEmployees: 13]
}
}
Rendering a Response
Sometimes it's easier (for example with Ajax applications) to render snippets of text or code to the response directly from the controller. For this, the highly flexible
render
method can be used:
The above code writes the text "Hello World!" to the response. Other examples include:
// write some markup
render {
for (b in books) {
div(id: b.id, b.title)
}
}
// render a specific view
render(view: 'show')
// render a template for each item in a collection
render(template: 'book_template', collection: Book.list())
// render some text with encoding and content type
render(text: "<xml>some xml</xml>", contentType: "text/xml", encoding: "UTF-8")
If you plan on using Groovy's
MarkupBuilder
to generate HTML for use with the
render
method be careful of naming clashes between HTML elements and Grails tags, for example:
import groovy.xml.MarkupBuilder
…
def login() {
def writer = new StringWriter()
def builder = new MarkupBuilder(writer)
builder.html {
head {
title 'Log in'
}
body {
h1 'Hello'
form {
}
}
} def html = writer.toString()
render html
}
This will actually
call the form tag (which will return some text that will be ignored by the
MarkupBuilder
). To correctly output a
<form>
element, use the following:
def login() {
// …
body {
h1 'Hello'
builder.form {
}
}
// …
}
8.1.4 Redirects and Chaining
Redirects
Actions can be redirected using the
redirect controller method:
class OverviewController { def login() {} def find() {
if (!session.user)
redirect(action: 'login')
return
}
…
}
}
Internally the
redirect method uses the
HttpServletResponse object's
sendRedirect
method.
The
redirect
method expects one of:
- Another closure within the same controller class:
// Call the login action within the same class
redirect(action: login)
- The name of an action (and controller name if the redirect isn't to an action in the current controller):
// Also redirects to the index action in the home controller
redirect(controller: 'home', action: 'index')
- A URI for a resource relative the application context path:
// Redirect to an explicit URI
redirect(uri: "/login.html")
// Redirect to a URL
redirect(url: "http://grails.org")
Parameters can optionally be passed from one action to the next using the
params
argument of the method:
redirect(action: 'myaction', params: [myparam: "myvalue"])
These parameters are made available through the
params dynamic property that accesses request parameters. If a parameter is specified with the same name as a request parameter, the request parameter is overridden and the controller parameter is used.
Since the
params
object is a Map, you can use it to pass the current request parameters from one action to the next:
redirect(action: "next", params: params)
Finally, you can also include a fragment in the target URI:
redirect(controller: "test", action: "show", fragment: "profile")
which will (depending on the URL mappings) redirect to something like "/myapp/test/show#profile".
Chaining
Actions can also be chained. Chaining allows the model to be retained from one action to the next. For example calling the
first
action in this action:
class ExampleChainController { def first() {
chain(action: second, model: [one: 1])
} def second () {
chain(action: third, model: [two: 2])
} def third() {
[three: 3])
}
}
results in the model:
[one: 1, two: 2, three: 3]
The model can be accessed in subsequent controller actions in the chain using the
chainModel
map. This dynamic property only exists in actions following the call to the
chain
method:
class ChainController { def nextInChain() {
def model = chainModel.myModel
…
}
}
Like the
redirect
method you can also pass parameters to the
chain
method:
chain(action: "action1", model: [one: 1], params: [myparam: "param1"])
8.1.5 Controller Interceptors
Often it is useful to intercept processing based on either request, session or application state. This can be achieved with action interceptors. There are currently two types of interceptors: before and after.
If your interceptor is likely to apply to more than one controller, you are almost certainly better off writing a Filter. Filters can be applied to multiple controllers or URIs without the need to change the logic of each controller
Before Interception
The
beforeInterceptor
intercepts processing before the action is executed. If it returns
false
then the intercepted action will not be executed. The interceptor can be defined for all actions in a controller as follows:
def beforeInterceptor = {
println "Tracing action ${actionUri}"
}
The above is declared inside the body of the controller definition. It will be executed before all actions and does not interfere with processing. A common use case is very simplistic authentication:
def beforeInterceptor = [action: this.&auth, except: 'login']// defined with private scope, so it's not considered an action
private auth() {
if (!session.user) {
redirect(action: 'login')
return false
}
}def login() {
// display login page
}
The above code defines a method called
auth
. A private method is used so that it is not exposed as an action to the outside world. The
beforeInterceptor
then defines an interceptor that is used on all actions
except the login action and it executes the
auth
method. The
auth
method is referenced using Groovy's method pointer syntax. Within the method it detects whether there is a user in the session, and if not it redirects to the
login
action and returns
false
, causing the intercepted action to not be processed.
After Interception
Use the
afterInterceptor
property to define an interceptor that is executed after an action:
def afterInterceptor = { model ->
println "Tracing action ${actionUri}"
}
The after interceptor takes the resulting model as an argument and can hence manipulate the model or response.
An after interceptor may also modify the Spring MVC
ModelAndView object prior to rendering. In this case, the above example becomes:
def afterInterceptor = { model, modelAndView ->
println "Current view is ${modelAndView.viewName}"
if (model.someVar) modelAndView.viewName = "/mycontroller/someotherview"
println "View is now ${modelAndView.viewName}"
}
This allows the view to be changed based on the model returned by the current action. Note that the
modelAndView
may be
null
if the action being intercepted called
redirect
or
render
.
Interception Conditions
Rails users will be familiar with the authentication example and how the 'except' condition was used when executing the interceptor (interceptors are called 'filters' in Rails; this terminology conflicts with Servlet filter terminology in Java):
def beforeInterceptor = [action: this.&auth, except: 'login']
This executes the interceptor for all actions except the specified action. A list of actions can also be defined as follows:
def beforeInterceptor = [action: this.&auth, except: ['login', 'register']]
The other supported condition is 'only', this executes the interceptor for only the specified action(s):
def beforeInterceptor = [action: this.&auth, only: ['secure']]
8.1.6 Data Binding
Data binding is the act of "binding" incoming request parameters onto the properties of an object or an entire graph of objects. Data binding should deal with all necessary type conversion since request parameters, which are typically delivered by a form submission, are always strings whilst the properties of a Groovy or Java object may well not be.
Map Based Binding
The data binder is capable of converting and assigning values in a Map to properties of an object. The binder will associate entries in the Map to properties of the object using the keys in the Map that have values which correspond to property names on the object. The following code demonstrates the basics:
// grails-app/domain/Person.groovy
class Person {
String firstName
String lastName
Integer age
}
def bindingMap = [firstName: 'Peter', lastName: 'Gabriel', age: 63]def person = new Person(bindingMap)assert person.firstName == 'Peter'
assert person.lastName == 'Gabriel'
assert person.age == 63
To update properties of a domain object you may assign a Map to the
properties
property of the domain class:
def bindingMap = [firstName: 'Peter', lastName: 'Gabriel', age: 63]def person = Person.get(someId)
person.properties = bindingMapassert person.firstName == 'Peter'
assert person.lastName == 'Gabriel'
assert person.age == 63
The binder can populate a full graph of objects using Maps of Maps.
class Person {
String firstName
String lastName
Integer age
Address homeAddress
}class Address {
String county
String country
}
def bindingMap = [firstName: 'Peter', lastName: 'Gabriel', age: 63, homeAddress: [county: 'Surrey', country: 'England'] ]def person = new Person(bindingMap)assert person.firstName == 'Peter'
assert person.lastName == 'Gabriel'
assert person.age == 63
assert person.homeAddress.county == 'Surrey'
assert person.homeAddress.country == 'England'
Binding To Collections And Maps
The data binder can populate and update Collections and Maps. The following code shows a simple example of populating a
List
of objects in a domain class:
class Band {
String name
static hasMany = [albums: Album]
List albums
}class Album {
String title
Integer numberOfTracks
}
def bindingMap = [name: 'Genesis',
'albums[0]': [title: 'Foxtrot', numberOfTracks: 6],
'albums[1]': [title: 'Nursery Cryme', numberOfTracks: 7]]def band = new Band(bindingMap)assert band.name == 'Genesis'
assert band.albums.size() == 2
assert band.albums[0].title == 'Foxtrot'
assert band.albums[0].numberOfTracks == 6
assert band.albums[1].title == 'Nursery Cryme'
assert band.albums[1].numberOfTracks == 7
That code would work in the same way if
albums
were an array instead of a
List
.
Note that when binding to a
Set
the structure of the
Map
being bound to the
Set
is the same as that of a
Map
being bound to a
List
but since a
Set
is unordered, the indexes don't necessarily correspond to the order of elements in the
Set
. In the code example above, if
albums
were a
Set
instead of a
List
, the
bindingMap
could look exactly the same but 'Foxtrot' might be the first album in the
Set
or it might be the second. When updating existing elements in a
Set
the
Map
being assigned to the
Set
must have
id
elements in it which represent the element in the
Set
being updated, as in the following example:
/*
* The value of the indexes 0 and 1 in albums[0] and albums[1] are arbitrary
* values that can be anything as long as they are unique within the Map.
* They do not correspond to the order of elements in albums because albums
* is a Set.
*/
def bindingMap = ['albums[0]': [id: 9, title: 'The Lamb Lies Down On Broadway']
'albums[1]': [id: 4, title: 'Selling England By The Pound']]def band = Band.get(someBandId)/*
* This will find the Album in albums that has an id of 9 and will set its title
* to 'The Lamb Lies Down On Broadway' and will find the Album in albums that has
* an id of 4 and set its title to 'Selling England By The Pound'. In both
* cases if the Album cannot be found in albums then the album will be retrieved
* from the database by id, the Album will be added to albums and will be updated
* with the values described above. If a Album with the specified id cannot be
* found in the database, then a binding error will be created and associated
* with the band object. More on binding errors later.
*/
band.properties = bindingMap
When binding to a
Map
the structure of the binding
Map
is the same as the structure of a
Map
used for binding to a
List
or a
Set
and the index inside of square brackets corresponds to the key in the
Map
being bound to. See the following code:
class Album {
String title
static hasMany = [players: Player]
Map players
}class Player {
String name
}
def bindingMap = [title: 'The Lamb Lies Down On Broadway',
'players[guitar]': [name: 'Steve Hackett'],
'players[vocals]': [name: 'Peter Gabriel'],
'players[keyboards]': [name: 'Tony Banks']]def album = new Album(bindingMap)assert album.title == 'The Lamb Lies Down On Broadway'
assert album.players.size() == 3
assert album.players.guitar.name == 'Steve Hackett'
assert album.players.vocals.name == 'Peter Gabriel'
assert album.players.keyboards.name == 'Tony Banks'
When updating an existing
Map
, if the key specified in the binding
Map
does not exist in the
Map
being bound to then a new value will be created and added to the
Map
with the specified key as in the following example:
def bindingMap = [title: 'The Lamb Lies Down On Broadway',
'players[guitar]': [name: 'Steve Hackett'],
'players[vocals]': [name: 'Peter Gabriel']
'players[keyboards]': [name: 'Tony Banks']]def album = new Album(bindingMap)assert album.title == 'The Lamb Lies Down On Broadway'
assert album.players.size() == 3
assert album.players.guitar == 'Steve Hackett'
assert album.players.vocals == 'Peter Gabriel'
assert album.players.keyboards == 'Tony Banks'def updatedBindingMap = ['players[drums]': [name: 'Phil Collins'],
'players[keyboards]': [name: 'Anthony George Banks']]album.properties = updatedBindingMapassert album.title == 'The Lamb Lies Down On Broadway'
assert album.players.size() == 4
assert album.players.guitar.name == 'Steve Hackett'
assert album.players.vocals.name == 'Peter Gabriel'
assert album.players.keyboards.name == 'Anthony George Banks'
assert album.players.drums.name == 'Phil Collins'
Binding Request Data to the Model
The
params object that is available in a controller has special behavior that helps convert dotted request parameter names into nested Maps that the data binder can work with. For example, if a request includes request parameters named
person.homeAddress.country
and
person.homeAddress.city
with values 'USA' and 'St. Louis' respectively,
params
would include entries like these:
[person: [homeAddress: [country: 'USA', city: 'St. Louis']]]
There are two ways to bind request parameters onto the properties of a domain class. The first involves using a domain classes' Map constructor:
def save() {
def b = new Book(params)
b.save()
}
The data binding happens within the code
new Book(params)
. By passing the
params object to the domain class constructor Grails automatically recognizes that you are trying to bind from request parameters. So if we had an incoming request like:
/book/save?title=The%20Stand&author=Stephen%20King
Then the
title
and
author
request parameters would automatically be set on the domain class. You can use the
properties property to perform data binding onto an existing instance:
def save() {
def b = Book.get(params.id)
b.properties = params
b.save()
}
This has the same effect as using the implicit constructor.
When binding an empty String (a String with no characters in it, not even spaces), the data binder will convert the empty String to null. This simplifies the most common case where the intent is to treat an empty form field as having the value null since there isn't a way to actually submit a null as a request parameter. When this behavior is not desirable the application may assign the value directly.
The mass property binding mechanism will by default automatically trim all Strings at binding time. To disable this behavior set the
grails.databinding.trimStrings
property to false in
grails-app/conf/Config.groovy
.
// the default value is true
grails.databinding.trimStrings = false// ...
The mass property binding mechanism will by default automatically convert all empty Strings to null at binding time. To disable this behavior set the
grails.databinding.convertEmptyStringsToNull
property to false in
grials-app/conf/Config.groovy
.
// the default value is true
grails.databinding.convertEmptyStringsToNull = false// ...
The order of events is that the String trimming happens and then null conversion happens so if
trimStrings
is
true
and
convertEmptyStringsToNull
is
true
, not only will empty Strings be converted to null but also blank Strings. A blank String is any String such that the
trim()
method returns an empty String.
These forms of data binding in Grails are very convenient, but also indiscriminate. In other words, they will bind all non-transient, typed instance properties of the target object, including ones that you may not want bound. Just because the form in your UI doesn't submit all the properties, an attacker can still send malign data via a raw HTTP request. Fortunately, Grails also makes it easy to protect against such attacks - see the section titled "Data Binding and Security concerns" for more information.
Data binding and Single-ended Associations
If you have a
one-to-one
or
many-to-one
association you can use Grails' data binding capability to update these relationships too. For example if you have an incoming request such as:
Grails will automatically detect the
.id
suffix on the request parameter and look up the
Author
instance for the given id when doing data binding such as:
An association property can be set to
null
by passing the literal
String
"null". For example:
/book/save?author.id=null
Data Binding and Many-ended Associations
If you have a one-to-many or many-to-many association there are different techniques for data binding depending of the association type.
If you have a
Set
based association (the default for a
hasMany
) then the simplest way to populate an association is to send a list of identifiers. For example consider the usage of
<g:select>
below:
<g:select name="books"
from="${Book.list()}"
size="5" multiple="yes" optionKey="id"
value="${author?.books}" />
This produces a select box that lets you select multiple values. In this case if you submit the form Grails will automatically use the identifiers from the select box to populate the
books
association.
However, if you have a scenario where you want to update the properties of the associated objects the this technique won't work. Instead you use the subscript operator:
<g:textField name="books[0].title" value="the Stand" />
<g:textField name="books[1].title" value="the Shining" />
However, with
Set
based association it is critical that you render the mark-up in the same order that you plan to do the update in. This is because a
Set
has no concept of order, so although we're referring to
books0
and
books1
it is not guaranteed that the order of the association will be correct on the server side unless you apply some explicit sorting yourself.
This is not a problem if you use
List
based associations, since a
List
has a defined order and an index you can refer to. This is also true of
Map
based associations.
Note also that if the association you are binding to has a size of two and you refer to an element that is outside the size of association:
<g:textField name="books[0].title" value="the Stand" />
<g:textField name="books[1].title" value="the Shining" />
<g:textField name="books[2].title" value="Red Madder" />
Then Grails will automatically create a new instance for you at the defined position.
You can bind existing instances of the associated type to a
List
using the same
.id
syntax as you would use with a single-ended association. For example:
<g:select name="books[0].id" from="${bookList}"
value="${author?.books[0]?.id}" /><g:select name="books[1].id" from="${bookList}"
value="${author?.books[1]?.id}" /><g:select name="books[2].id" from="${bookList}"
value="${author?.books[2]?.id}" />
Would allow individual entries in the
books List
to be selected separately.
Entries at particular indexes can be removed in the same way too. For example:
<g:select name="books[0].id"
from="${Book.list()}"
value="${author?.books[0]?.id}"
noSelection="['null': '']"/>
Will render a select box that will remove the association at
books0
if the empty option is chosen.
Binding to a
Map
property works the same way except that the list index in the parameter name is replaced by the map key:
<g:select name="images[cover].id"
from="${Image.list()}"
value="${book?.images[cover]?.id}"
noSelection="['null': '']"/>
This would bind the selected image into the
Map
property
images
under a key of
"cover"
.
When binding to Maps, Arrays and Collections the data binder will automatically grow the size of the collections as necessary. The default limit to how large the binder will grow a collection is 256. If the data binder encounters an entry that requires the collection be grown beyond that limit, the entry is ignored. The limit may be configured by assigning a value to the
grails.databinding.autoGrowCollectionLimit
property in
Config.groovy
.
// grails-app/conf/Config.groovy// the default value is 256
grails.databinding.autoGrowCollectionLimit = 128// ...
Data binding with Multiple domain classes
It is possible to bind data to multiple domain objects from the
params object.
For example so you have an incoming request to:
/book/save?book.title=The%20Stand&author.name=Stephen%20King
You'll notice the difference with the above request is that each parameter has a prefix such as
author.
or
book.
which is used to isolate which parameters belong to which type. Grails'
params
object is like a multi-dimensional hash and you can index into it to isolate only a subset of the parameters to bind.
def b = new Book(params.book)
Notice how we use the prefix before the first dot of the
book.title
parameter to isolate only parameters below this level to bind. We could do the same with an
Author
domain class:
def a = new Author(params.author)
Data Binding and Action Arguments
Controller action arguments are subject to request parameter data binding. There are 2 categories of controller action arguments. The first category is command objects. Complex types are treated as command objects. See the
Command Objects section of the user guide for details. The other category is basic object types. Supported types are the 8 primitives, their corresponding type wrappers and
java.lang.String. The default behavior is to map request parameters to action arguments by name:
class AccountingController { // accountNumber will be initialized with the value of params.accountNumber
// accountType will be initialized with params.accountType
def displayInvoice(String accountNumber, int accountType) {
// …
}
}
For primitive arguments and arguments which are instances of any of the primitive type wrapper classes a type conversion has to be carried out before the request parameter value can be bound to the action argument. The type conversion happens automatically. In a case like the example shown above, the
params.accountType
request parameter has to be converted to an
int
. If type conversion fails for any reason, the argument will have its default value per normal Java behavior (null for type wrapper references, false for booleans and zero for numbers) and a corresponding error will be added to the
errors
property of the defining controller.
/accounting/displayInvoice?accountNumber=B59786&accountType=bogusValue
Since "bogusValue" cannot be converted to type int, the value of accountType will be zero, the controller's
errors.hasErrors()
will be true, the controller's
errors.errorCount
will be equal to 1 and the controller's
errors.getFieldError('accountType')
will contain the corresponding error.
If the argument name does not match the name of the request parameter then the
@grails.web.RequestParameter
annotation may be applied to an argument to express the name of the request parameter which should be bound to that argument:
import grails.web.RequestParameterclass AccountingController { // mainAccountNumber will be initialized with the value of params.accountNumber
// accountType will be initialized with params.accountType
def displayInvoice(@RequestParameter('accountNumber') String mainAccountNumber, int accountType) {
// …
}
}
Data binding and type conversion errors
Sometimes when performing data binding it is not possible to convert a particular String into a particular target type. This results in a type conversion error. Grails will retain type conversion errors inside the
errors property of a Grails domain class. For example:
class Book {
…
URL publisherURL
}
Here we have a domain class
Book
that uses the
java.net.URL
class to represent URLs. Given an incoming request such as:
/book/save?publisherURL=a-bad-url
it is not possible to bind the string
a-bad-url
to the
publisherURL
property as a type mismatch error occurs. You can check for these like this:
def b = new Book(params)if (b.hasErrors()) {
println "The value ${b.errors.getFieldError('publisherURL').rejectedValue}" +
" is not a valid URL!"
}
Although we have not yet covered error codes (for more information see the section on
Validation), for type conversion errors you would want a message from the
grails-app/i18n/messages.properties
file to use for the error. You can use a generic error message handler such as:
typeMismatch.java.net.URL=The field {0} is not a valid URL
Or a more specific one:
typeMismatch.Book.publisherURL=The publisher URL you specified is not a valid URL
The BindUsing Annotation
The
BindUsing annotation may be used to define a custom binding mechanism for a particular field in a class. Any time data binding is being applied to the field the closure value of the annotation will be invoked with 2 arguments. The first argument is the object that data binding is being applied to and the second argument is
DataBindingSource which is the data source for the data binding. The value returned from the closure will be bound to the property. The following example would result in the upper case version of the
name
value in the source being applied to the
name
field during data binding.
import org.grails.databinding.BindUsingclass SomeClass {
@BindUsing({obj, source -> //source is DataSourceBinding which is similar to a Map
//and defines getAt operation but source.name cannot be used here.
//In order to get name from source use getAt instead as shown below. source['name']?.toUpperCase()
})
String name
}
Note that data binding is only possible when the name of the request parameter matches with the field name in the class.
Here, name
from request parameters matches with name
from SomeClass
.
The
BindUsing annotation may be used to define a custom binding mechanism for all of the fields on a particular class. When the annotation is applied to a class, the value assigned to the annotation should be a class which implements the
BindingHelper interface. An instance of that class will be used any time a value is bound to a property in the class that this annotation has been applied to.
@BindUsing(SomeClassWhichImplementsBindingHelper)
class SomeClass {
String someProperty
Integer someOtherProperty
}
Custom Data Converters
The binder will do a lot of type conversion automatically. Some applications may want to define their own mechanism for converting values and a simple way to do this is to write a class which implements
ValueConverter and register an instance of that class as a bean in the Spring application context.
package com.myapp.convertersimport org.grails.databinding.converters.ValueConverter/**
* A custom converter which will convert String of the
* form 'city:state' into an Address object.
*/
class AddressValueConverter implements ValueConverter { boolean canConvert(value) {
value instanceof String
} def convert(value) {
def pieces = value.split(':')
new com.myapp.Address(city: pieces[0], state: pieces[1])
} Class<?> getTargetType() {
com.myapp.Address
}
}
An instance of that class needs to be registered as a bean in the Spring application context. The bean name is not important. All beans that implemented ValueConverter will be automatically plugged in to the data binding process.
// grails-app/conf/spring/resources.groovybeans = { addressConverter com.myapp.converters.AddressValueConverter // ...}
class Person {
String firstName
Address homeAddress
}class Address {
String city
String state
}def person = new Person()
person.properties = [firstName: 'Jeff', homeAddress: "O'Fallon:Missouri"]
assert person.firstName == 'Jeff'
assert person.homeAddress.city = "O'Fallon"
assert person.homeAddress.state = 'Missouri'
Date Formats For Data Binding
A custom date format may be specified to be used when binding a String to a Date value by applying the
BindingFormat annotation to a Date field.
import org.grails.databinding.BindingFormatclass Person {
@BindingFormat('MMddyyyy')
Date birthDate
}
A global setting may be configured in
Config.groovy
to define date formats which will be used application wide when binding to Date.
// grails-app/conf/Config.groovygrails.databinding.dateFormats = ['MMddyyyy', 'yyyy-MM-dd HH:mm:ss.S', "yyyy-MM-dd'T'hh:mm:ss'Z'"]
The formats specified in
grails.databinding.dateFormats
will be attempted in the order in which they are included in the List. If a property is marked with @BindingFormat, the @BindingFormat will take precedence over the values specified in
grails.databinding.dateFormats
.
The default formats that are used are "yyyy-MM-dd HH:mm:ss.S", "yyyy-MM-dd'T'hh:mm:ss'Z'" and "yyyy-MM-dd HH:mm:ss.S z".
Custom Formatted Converters
You may supply your own handler for the
BindingFormat annotation by writing a class which implements the
FormattedValueConverter interface and registering an instance of that class as a bean in the Spring application context. Below is an example of a trivial custom String formatter that might convert the case of a String based on the value assigned to the BindingFormat annotation.
package com.myapp.convertersimport org.grails.databinding.converters.FormattedValueConverterclass FormattedStringValueConverter implements FormattedValueConverter {
def convert(value, String format) {
if('UPPERCASE' == format) {
value = value.toUpperCase()
} else if('LOWERCASE' == format) {
value = value.toLowerCase()
}
value
} Class getTargetType() {
// specifies the type to which this converter may be applied
String
}
}
An instance of that class needs to be registered as a bean in the Spring application context. The bean name is not important. All beans that implemented FormattedValueConverter will be automatically plugged in to the data binding process.
// grails-app/conf/spring/resources.groovybeans = { formattedStringConverter com.myapp.converters.FormattedStringValueConverter // ...}
With that in place the
BindingFormat
annotation may be applied to String fields to inform the data binder to take advantage of the custom converter.
import org.grails.databinding.BindingFormatclass Person {
@BindingFormat('UPPERCASE')
String someUpperCaseString @BindingFormat('LOWERCASE')
String someLowerCaseString String someOtherString
}
Localized Binding Formats
The
BindingFormat
annotation supports localized format strings by using the optional
code
attribute. If a value is assigned to the code attribute that value will be used as the message code to retrieve the binding format string from the
messageSource
bean in the Spring application context and that lookup will be localized.
import org.grails.databinding.BindingFormatclass Person {
@BindingFormat(code='date.formats.birthdays')
Date birthDate
}
# grails-app/conf/i18n/messages.properties
date.formats.birthdays=MMddyyyy
# grails-app/conf/i18n/messages_es.properties
date.formats.birthdays=ddMMyyyy
Structured Data Binding Editors
A structured data binding editor is a helper class which can bind structured request parameters to a property. The common use case for structured binding is binding to a
Date
object which might be constructed from several smaller pieces of information contained in several request parameters with names like
birthday_month
,
birthday_date
and
birthday_year
. The structured editor would retrieve all of those individual pieces of information and use them to construct a
Date
.
The framework provides a structured editor for binding to
Date
objects. An application may register its own structured editors for whatever types are appropriate. Consider the following classes:
// src/groovy/databinding/Gadget.groovy
package databindingclass Gadget {
Shape expandedShape
Shape compressedShape
}
// src/groovy/databinding/Shape.groovy
package databindingclass Shape {
int area
}
A
Gadget
has 2
Shape
fields. A
Shape
has an
area
property. It may be that the application wants to accept request parameters like
width
and
height
and use those to calculate the
area
of a
Shape
at binding time. A structured binding editor is well suited for that.
The way to register a structured editor with the data binding process is to add an instance of the
org.grails.databinding.TypedStructuredBindingEditor interface to the Spring application context. The easiest way to implement the
TypedStructuredBindingEditor
interface is to extend the
org.grails.databinding.converters.AbstractStructuredBindingEditor abstract class and override the
getPropertyValue
method as shown below:
// src/groovy/databinding/converters/StructuredShapeEditor.groovy
package databinding.convertersimport databinding.Shapeimport org.grails.databinding.converters.AbstractStructuredBindingEditorclass StructuredShapeEditor extends AbstractStructuredBindingEditor<Shape> { public Shape getPropertyValue(Map values) {
// retrieve the individual values from the Map
def width = values.width as int
def height = values.height as int // use the values to calculate the area of the Shape
def area = width * height // create and return a Shape with the appropriate area
new Shape(area: area)
}
}
An instance of that class needs to be registered with the Spring application context:
// grails-app/conf/spring/resources.groovy
beans = {
shapeEditor databinding.converters.StructuredShapeEditor // …
}
When the data binder binds to an instance of the
Gadget
class it will check to see if there are request parameters with names
compressedShape
and
expandedShape
which have a value of "struct" and if they do exist, that will trigger the use of the
StructuredShapeEditor
. The individual components of the structure need to have parameter names of the form propertyName_structuredElementName. In the case of the
Gadget
class above that would mean that the
compressedShape
request parameter should have a value of "struct" and the
compressedShape_width
and
compressedShape_height
parameters should have values which represent the width and the height of the compressed
Shape
. Similarly, the
expandedShape
request parameter should have a value of "struct" and the
expandedShape_width
and
expandedShape_height
parameters should have values which represent the width and the height of the expanded
Shape
.
// grails-app/controllers/demo/DemoController.groovy
class DemoController { def createGadget(Gadget gadget) {
/* /demo/createGadget?expandedShape=struct&expandedShape_width=80&expandedShape_height=30
&compressedShape=struct&compressedShape_width=10&compressedShape_height=3 */ // with the request parameters shown above gadget.expandedShape.area would be 2400
// and gadget.compressedShape.area would be 30 // ... }
}
Typically the request parameters with "struct" as their value would be represented by hidden form fields.
Data Binding Event Listeners
The
DataBindingListener interface provides a mechanism for listeners to be notified of data binding events. The interface looks like this:
package org.grails.databinding.events;import org.grails.databinding.errors.BindingError;public interface DataBindingListener { /**
* @return true if the listener is interested in events for the specified type.
*/
boolean supports(Class<?> clazz); /**
* Called when data binding is about to start.
*
* @param target The object data binding is being imposed upon
* @param errors the Spring Errors instance (a org.springframework.validation.BindingResult)
* @return true if data binding should continue
*/
Boolean beforeBinding(Object target, Object errors); /**
* Called when data binding is about to imposed on a property
*
* @param target The object data binding is being imposed upon
* @param propertyName The name of the property being bound to
* @param value The value of the property being bound
* @param errors the Spring Errors instance (a org.springframework.validation.BindingResult)
* @return true if data binding should continue, otherwise return false
*/
Boolean beforeBinding(Object target, String propertyName, Object value, Object errors); /**
* Called after data binding has been imposed on a property
*
* @param target The object data binding is being imposed upon
* @param propertyName The name of the property that was bound to
* @param errors the Spring Errors instance (a org.springframework.validation.BindingResult)
*/
void afterBinding(Object target, String propertyName, Object errors); /**
* Called after data binding has finished.
*
* @param target The object data binding is being imposed upon
* @param errors the Spring Errors instance (a org.springframework.validation.BindingResult)
*/
void afterBinding(Object target, Object errors); /**
* Called when an error occurs binding to a property
* @param error encapsulates information about the binding error
* @param errors the Spring Errors instance (a org.springframework.validation.BindingResult)
* @see BindingError
*/
void bindingError(BindingError error, Object errors);
}
Any bean in the Spring application context which implements that interface will automatically be registered with the data binder. The
DataBindingListenerAdapter class implements the
DataBindingListener
interface and provides default implementations for all of the methods in the interface so this class is well suited for subclassing so your listener class only needs to provide implementations for the methods your listener is interested in.
The Grails data binder has limited support for the older
BindEventListener style listeners.
BindEventListener
looks like this:
package org.codehaus.groovy.grails.web.binding;import org.springframework.beans.MutablePropertyValues;
import org.springframework.beans.TypeConverter;public interface BindEventListener { /**
* @param target The target to bind to
* @param source The source of the binding, typically a Map
* @param typeConverter The type converter to be used
*/
void doBind(Object target, MutablePropertyValues source, TypeConverter typeConverter);
}
Support for
BindEventListener
is disabled by default. To enable support assign a value of
true
to the
grails.databinding.enableSpringEventAdapter
property in
grails-app/conf/Config.groovy
.
// grails-app/conf/Config.groovy
grails.databinding.enableSpringEventAdapter=true...
With
enableSpringEventAdapter
set to
true
instances of
BindEventListener
which are in the Spring application context will automatically be registered with the data binder. Notice that the
MutablePropertyValues
and
TypeConverter
arguments to the
doBind
method in
BindEventListener
are Spring specific classes and are not relevant to the current data binder. The event adapter will pass
null
values for those arguments. The only real value passed into the
doBind
method will be the object being bound to. This limited support is provided for backward compatibility and will be useful for a subset of scenarios. Developers are encouraged to migrate their
BindEventListener
beans to the newer
DataBindingListener
model.
Using The Data Binder Directly
There are situations where an application may want to use the data binder directly. For example, to do binding in a Service on some arbitrary object which is not a domain class. The following will not work because the
properties
property is read only.
// src/groovy/bindingdemo/Widget.groovy
package bindingdemoclass Widget {
String name
Integer size
}
// grails-app/services/bindingdemo/WidgetService.groovy
package bindingdemoclass WidgetService { def updateWidget(Widget widget, Map data) {
// this will throw an exception because
// properties is read-only
widget.properties = data
}
}
An instance of the data binder is in the Spring application context with a bean name of
grailsWebDataBinder
. That bean implements the
DataBinder interface. The following code demonstrates using the data binder directly.
// grails-app/services/bindingdmeo/WidgetService
package bindingdemoimport org.grails.databinding.SimpleMapDataBindingSourceclass WidgetService { // this bean will be autowired into the service
def grailsWebDataBinder def updateWidget(Widget widget, Map data) {
grailsWebDataBinder.bind widget, data as SimpleMapDataBindingSource
}}
See the
DataBinder documentation for more information about overloaded versions
of the
bind
method.
Data Binding and Security Concerns
When batch updating properties from request parameters you need to be careful not to allow clients to bind malicious data to domain classes and be persisted in the database. You can limit what properties are bound to a given domain class using the subscript operator:
def p = Person.get(1)p.properties['firstName','lastName'] = params
In this case only the
firstName
and
lastName
properties will be bound.
Another way to do this is is to use
Command Objects as the target of data binding instead of domain classes. Alternatively there is also the flexible
bindData method.
The
bindData
method allows the same data binding capability, but to arbitrary objects:
def p = new Person()
bindData(p, params)
The
bindData
method also lets you exclude certain parameters that you don't want updated:
def p = new Person()
bindData(p, params, [exclude: 'dateOfBirth'])
Or include only certain properties:
def p = new Person()
bindData(p, params, [include: ['firstName', 'lastName']])
Note that if an empty List is provided as a value for the include
parameter then all fields will be subject to binding if they are not explicitly excluded.
8.1.7 XML and JSON Responses
Using the render method to output XML
Grails supports a few different ways to produce XML and JSON responses. The first is the
render method.
The
render
method can be passed a block of code to do mark-up building in XML:
def list() { def results = Book.list() render(contentType: "text/xml") {
books {
for (b in results) {
book(title: b.title)
}
}
}
}
The result of this code would be something like:
<books>
<book title="The Stand" />
<book title="The Shining" />
</books>
Be careful to avoid naming conflicts when using mark-up building. For example this code would produce an error:
def list() { def books = Book.list() // naming conflict here render(contentType: "text/xml") {
books {
for (b in results) {
book(title: b.title)
}
}
}
}
This is because there is local variable
books
which Groovy attempts to invoke as a method.
Using the render method to output JSON
The
render
method can also be used to output JSON:
def list() { def results = Book.list() render(contentType: "application/json") {
books = array {
for (b in results) {
book title: b.title
}
}
}
}
In this case the result would be something along the lines of:
[
{"title":"The Stand"},
{"title":"The Shining"}
]
The same dangers with naming conflicts described above for XML also apply to JSON building.
Automatic XML Marshalling
Grails also supports automatic marshalling of
domain classes to XML using special converters.
To start off with, import the
grails.converters
package into your controller:
import grails.converters.*
Now you can use the following highly readable syntax to automatically convert domain classes to XML:
render Book.list() as XML
The resulting output would look something like the following::
<?xml version="1.0" encoding="ISO-8859-1"?>
<list>
<book id="1">
<author>Stephen King</author>
<title>The Stand</title>
</book>
<book id="2">
<author>Stephen King</author>
<title>The Shining</title>
</book>
</list>
For more information on XML marshalling see the section on
RESTAutomatic JSON Marshalling
Grails also supports automatic marshalling to JSON using the same mechanism. Simply substitute
XML
with
JSON
:
render Book.list() as JSON
The resulting output would look something like the following:
[
{"id":1,
"class":"Book",
"author":"Stephen King",
"title":"The Stand"},
{"id":2,
"class":"Book",
"author":"Stephen King",
"releaseDate":new Date(1194127343161),
"title":"The Shining"}
]
8.1.8 More on JSONBuilder
The previous section on on XML and JSON responses covered simplistic examples of rendering XML and JSON responses. Whilst the XML builder used by Grails is the standard
XmlSlurper found in Groovy, the JSON builder is a custom implementation specific to Grails.
JSONBuilder and Grails versions
JSONBuilder behaves different depending on the version of Grails you use. For version below 1.2 the deprecated
grails.web.JSONBuilder class is used. This section covers the usage of the Grails 1.2 JSONBuilder
For backwards compatibility the old
JSONBuilder
class is used with the
render
method for older applications; to use the newer/better
JSONBuilder
class set the following in
Config.groovy
:
grails.json.legacy.builder = false
Rendering Simple Objects
To render a simple JSON object just set properties within the context of the Closure:
render(contentType: "application/json") {
hello = "world"
}
The above will produce the JSON:
Rendering JSON Arrays
To render a list of objects simple assign a list:
render(contentType: "application/json") {
categories = ['a', 'b', 'c']
}
This will produce:
{"categories":["a","b","c"]}
You can also render lists of complex objects, for example:
render(contentType: "application/json") {
categories = [ { a = "A" }, { b = "B" } ]
}
This will produce:
{"categories":[ {"a":"A"} , {"b":"B"}] }
Use the special
element
method to return a list as the root:
render(contentType: "application/json") {
element 1
element 2
element 3
}
The above code produces:
Rendering Complex Objects
Rendering complex objects can be done with Closures. For example:
render(contentType: "application/json") {
categories = ['a', 'b', 'c']
title = "Hello JSON"
information = {
pages = 10
}
}
The above will produce the JSON:
{"categories":["a","b","c"],"title":"Hello JSON","information":{"pages":10}}
Arrays of Complex Objects
As mentioned previously you can nest complex objects within arrays using Closures:
render(contentType: "application/json") {
categories = [ { a = "A" }, { b = "B" } ]
}
You can use the
array
method to build them up dynamically:
def results = Book.list()
render(contentType: "application/json") {
books = array {
for (b in results) {
book title: b.title
}
}
}
Direct JSONBuilder API Access
If you don't have access to the
render
method, but still want to produce JSON you can use the API directly:
def builder = new JSONBuilder()def result = builder.build {
categories = ['a', 'b', 'c']
title = "Hello JSON"
information = {
pages = 10
}
}// prints the JSON text
println result.toString()def sw = new StringWriter()
result.render sw
8.1.9 Uploading Files
Programmatic File Uploads
Grails supports file uploads using Spring's
MultipartHttpServletRequest interface. The first step for file uploading is to create a multipart form like this:
Upload Form: <br />
<g:uploadForm action="upload">
<input type="file" name="myFile" />
<input type="submit" />
</g:uploadForm>
The
uploadForm
tag conveniently adds the
enctype="multipart/form-data"
attribute to the standard
<g:form>
tag.
There are then a number of ways to handle the file upload. One is to work with the Spring
MultipartFile instance directly:
def upload() {
def f = request.getFile('myFile')
if (f.empty) {
flash.message = 'file cannot be empty'
render(view: 'uploadForm')
return
} f.transferTo(new File('/some/local/dir/myfile.txt'))
response.sendError(200, 'Done')
}
This is convenient for doing transfers to other destinations and manipulating the file directly as you can obtain an
InputStream
and so on with the
MultipartFile interface.
File Uploads through Data Binding
File uploads can also be performed using data binding. Consider this
Image
domain class:
class Image {
byte[] myFile static constraints = {
// Limit upload file size to 2MB
myFile maxSize: 1024 * 1024 * 2
}
}
If you create an image using the
params
object in the constructor as in the example below, Grails will automatically bind the file's contents as a
byte
to the
myFile
property:
def img = new Image(params)
It's important that you set the
size or
maxSize constraints, otherwise your database may be created with a small column size that can't handle reasonably sized files. For example, both H2 and MySQL default to a blob size of 255 bytes for
byte
properties.
It is also possible to set the contents of the file as a string by changing the type of the
myFile
property on the image to a String type:
class Image {
String myFile
}
8.1.10 Command Objects
Grails controllers support the concept of command objects. A command object is a class that is used in conjunction with
data binding, usually to allow validation of data that may not fit into an existing domain class.
Note: A class is only considered to be a command object when it is used as a parameter of an action.
Declaring Command Objects
Command object classes are defined just like any other class.
@grails.validation.Validateable
class LoginCommand {
String username
String password static constraints = {
username(blank: false, minSize: 6)
password(blank: false, minSize: 6)
}
}
In this example, the command object is marked with the
Validateable
annotation. The
Validateable
annotation allows the definition of
constraints just like in
domain classes. If the command object is defined in the same source file as the controller that is using it, Grails will automatically mark it as
Validateable
. It is not required that command object classes be validateable.
By default, all
Validateable
object properties are
nullable: false
which matches the behavior of GORM domain objects. If you want a
Validateable
that has
nullable: true
properties by default, you can specify
nullable: true
on the annotation:
@grails.validation.Validateable(nullable=true)
class AuthorSearchCommand {
String name
Integer age
}
In this example, both
name
and
age
will allow null values during validation.
Using Command Objects
To use command objects, controller actions may optionally specify any number of command object parameters. The parameter types must be supplied so that Grails knows what objects to create and initialize.
Before the controller action is executed Grails will automatically create an instance of the command object class and populate its properties by binding the request parameters. If the command object class is marked with
Validateable
then the command object will be validated. For example:
class LoginController { def login(LoginCommand cmd) {
if (cmd.hasErrors()) {
redirect(action: 'loginForm')
return
} // work with the command object data
}
}
If the command object's type is that of a domain class and there is an
id
request parameter then instead of invoking the domain class constructor to create a new instance a call will be made to the static
get
method on the domain class and the value of the
id
parameter will be passed as an argument. Whatever is returned from that call to
get
is what will be passed into the controller action. This means that if there is an
id
request parameter and no corresponding record is found in the database then the value of the command object will be
null
. If an error occurs retrieving the instance from the database then
null
will be passed as an argument to the controller action and an error will be added the controller's
errors
property. If the command object's type is a domain class and there is no
id
request parameter then
null
will be passed into the controller action unless the HTTP request method is "POST", in which case a new instance of the domain class will be created by invoking the domain class constructor. For all of the cases where the domain class instance is non-null, data binding is only performed if the HTTP request method is "POST", "PUT" or "PATCH".
Command Objects And Request Parameter Names
Normally request parameter names will be mapped directly to property names in the command object. Nested parameter names may be used to bind down the object graph in an intuitive way. In the example below a request parameter named
name
will be bound to the
name
property of the
Person
instance and a request parameter named
address.city
will be bound to the
city
property of the
address
property in the
Person
.
class StoreController {
def buy(Person buyer) {
// …
}
}class Person {
String name
Address address
}class Address {
String city
}
A problem may arise if a controller action accepts multiple command objects which happen to contain the same property name. Consider the following example.
class StoreController {
def buy(Person buyer, Product product) {
// …
}
}class Person {
String name
Address address
}class Address {
String city
}class Product {
String name
}
If there is a request parameter named
name
it isn't clear if that should represent the name of the
Product
or the name of the
Person
. Another version of the problem can come up if a controller action accepts 2 command objects of the same type as shown below.
class StoreController {
def buy(Person buyer, Person seller, Product product) {
// …
}
}class Person {
String name
Address address
}class Address {
String city
}class Product {
String name
}
To help deal with this the framework imposes special rules for mapping parameter names to command object types. The command object data binding will treat all parameters that begin with the controller action parameter name as belonging to the corresponding command object. For example, the
product.name
request parameter will be bound to the
name
property in the
product
argument, the
buyer.name
request parameter will be bound to the
name
property in the
buyer
argument the
seller.address.city
request parameter will be bound to the
city
property of the
address
property of the
seller
argument, etc...
Command Objects and Dependency Injection
Command objects can participate in dependency injection. This is useful if your command object has some custom validation logic which uses a Grails
service:
@grails.validation.Validateable
class LoginCommand { def loginService String username
String password static constraints = {
username validator: { val, obj ->
obj.loginService.canLogin(obj.username, obj.password)
}
}
}
In this example the command object interacts with the
loginService
bean which is injected by name from the Spring
ApplicationContext
.
Binding The Request Body To Command Objects
When a request is made to a controller action which accepts a command object and the request contains a body, Grails will attempt to parse the body of the request based on the request content type and use the body to do data binding on the command object. See the following example.
// grails-app/controllers/bindingdemo/DemoController.groovy
package bindingdemoclass DemoController { def createWidget(Widget w) {
render "Name: ${w?.name}, Size: ${w?.size}"
}
}class Widget {
String name
Integer size
}
$ curl -H "Content-Type: application/json" -d '{"name":"Some Widget","size":"42"}' localhost:8080/myapp/demo/createWidget
Name: Some Widget, Size: 42
~ $
$ curl -H "Content-Type: application/xml" -d '<widget><name>Some Other Widget</name><size>2112</size></widget>' localhost:8080/bodybind/demo/createWidget
Name: Some Other Widget, Size: 2112
~ $
Note that the body of the request is being parsed to make that work. Any attempt to read the body of the request after that will fail since the corresponding input stream will be empty. The controller action can either use a command object or it can parse the body of the request on its own (either directly, or by referring to something like request.JSON), but cannot do both.
// grails-app/controllers/bindingdemo/DemoController.groovy
package bindingdemoclass DemoController { def createWidget(Widget w) {
// this will fail because it requires reading the body,
// which has already been read.
def json = request.JSON // ... }
}
Grails has built-in support for handling duplicate form submissions using the "Synchronizer Token Pattern". To get started you define a token on the
form tag:
<g:form useToken="true" ...>
Then in your controller code you can use the
withForm method to handle valid and invalid requests:
withForm {
// good request
}.invalidToken {
// bad request
}
If you only provide the
withForm method and not the chained
invalidToken
method then by default Grails will store the invalid token in a
flash.invalidToken
variable and redirect the request back to the original page. This can then be checked in the view:
<g:if test="${flash.invalidToken}">
Don't click the button twice!
</g:if>
The withForm tag makes use of the session and hence requires session affinity or clustered sessions if used in a cluster.
8.1.12 Simple Type Converters
Type Conversion Methods
If you prefer to avoid the overhead of
Data Binding and simply want to convert incoming parameters (typically Strings) into another more appropriate type the
params object has a number of convenience methods for each type:
def total = params.int('total')
The above example uses the
int
method, and there are also methods for
boolean
,
long
,
char
,
short
and so on. Each of these methods is null-safe and safe from any parsing errors, so you don't have to perform any additional checks on the parameters.
Each of the conversion methods allows a default value to be passed as an optional second argument. The default value will be returned if a corresponding entry cannot be found in the map or if an error occurs during the conversion. Example:
def total = params.int('total', 42)
These same type conversion methods are also available on the
attrs
parameter of GSP tags.
Handling Multi Parameters
A common use case is dealing with multiple request parameters of the same name. For example you could get a query string such as
?name=Bob&name=Judy
.
In this case dealing with one parameter and dealing with many has different semantics since Groovy's iteration mechanics for
String
iterate over each character. To avoid this problem the
params object provides a
list
method that always returns a list:
for (name in params.list('name')) {
println name
}
8.1.13 Declarative Controller Exception Handling
Grails controllers support a simple mechanism for declarative exception handling. If a controller declares a method that accepts a single argument and the argument type is
java.lang.Exception
or some subclass of
java.lang.Exception
, that method will be invoked any time an action in that controller throws an exception of that type. See the following example.
// grails-app/controllers/demo/DemoController.groovy
package democlass DemoController { def someAction() {
// do some work
} def handleSQLException(SQLException e) {
render 'A SQLException Was Handled'
} def handleBatchUpdateException(BatchUpdateException e) {
redirect controller: 'logging', action: 'batchProblem'
} def handleNumberFormatException(NumberFormatException nfe) {
[problemDescription: 'A Number Was Invalid']
}
}
That controller will behave as if it were written something like this...
// grails-app/controllers/demo/DemoController.groovy
package democlass DemoController { def someAction() {
try {
// do some work
} catch (BatchUpdateException e) {
return handleBatchUpdateException(e)
} catch (SQLException e) {
return handleSQLException(e)
} catch (NumberFormatException e) {
return handleNumberFormatException(e)
}
} def handleSQLException(SQLException e) {
render 'A SQLException Was Handled'
} def handleBatchUpdateException(BatchUpdateException e) {
redirect controller: 'logging', action: 'batchProblem'
} def handleNumberFormatException(NumberFormatException nfe) {
[problemDescription: 'A Number Was Invalid']
}
}
The exception handler method names can be any valid method name. The name is not what makes the method an exception handler, the
Exception
argument type is the important part.
The exception handler methods can do anything that a controller action can do including invoking
render
,
redirect
, returning a model, etc.
One way to share exception handler methods across multiple controllers is to use inheritance. Exception handler methods are inherited into subclasses so an application could define the exception handlers in an abstract class that multiple controllers extend from. Another way to share exception handler methods across multiple controllers is to use a trait, as shown below...
// src/groovy/com/demo/DatabaseExceptionHandler.groovy
package com.demotrait DatabaseExceptionHandler {
def handleSQLException(SQLException e) {
// handle SQLException
} def handleBatchUpdateException(BatchUpdateException e) {
// handle BatchUpdateException
}
}
// grails-app/controllers/com/demo/DemoController.groovy
package com.democlass DemoController implements DatabaseExceptionHandler { // all of the exception handler methods defined
// in DatabaseExceptionHandler will be added to
// this class at compile time
}
Exception handler methods must be present at compile time. Specifically, exception handler methods which are runtime metaprogrammed onto a controller class are not supported.
8.2 Groovy Server Pages
Groovy Servers Pages (or GSP for short) is Grails' view technology. It is designed to be familiar for users of technologies such as ASP and JSP, but to be far more flexible and intuitive.
GSPs live in the
grails-app/views
directory and are typically rendered automatically (by convention) or with the
render method such as:
A GSP is typically a mix of mark-up and GSP tags which aid in view rendering.
Although it is possible to have Groovy logic embedded in your GSP and doing this will be covered in this document, the practice is strongly discouraged. Mixing mark-up and code is a bad thing and most GSP pages contain no code and needn't do so.
A GSP typically has a "model" which is a set of variables that are used for view rendering. The model is passed to the GSP view from a controller. For example consider the following controller action:
def show() {
[book: Book.get(params.id)]
}
This action will look up a
Book
instance and create a model that contains a key called
book
. This key can then be referenced within the GSP view using the name
book
:
Embedding data received from user input has the risk of making your application vulnerable to an Cross Site Scripting (XSS) attack. Please read the documentation on XSS prevention for information on how to prevent XSS attacks.
8.2.1 GSP Basics
In the next view sections we'll go through the basics of GSP and what is available to you. First off let's cover some basic syntax that users of JSP and ASP should be familiar with.
GSP supports the usage of
<% %>
scriptlet blocks to embed Groovy code (again this is discouraged):
<html>
<body>
<% out << "Hello GSP!" %>
</body>
</html>
You can also use the
<%= %>
syntax to output values:
<html>
<body>
<%="Hello GSP!" %>
</body>
</html>
GSP also supports JSP-style server-side comments (which are not rendered in the HTML response) as the following example demonstrates:
<html>
<body>
<%-- This is my comment --%>
<%="Hello GSP!" %>
</body>
</html>
Embedding data received from user input has the risk of making your application vulnerable to an Cross Site Scripting (XSS) attack. Please read the documentation on XSS prevention for information on how to prevent XSS attacks.
8.2.1.1 Variables and Scopes
Within the
<% %>
brackets you can declare variables:
and then access those variables later in the page:
Within the scope of a GSP there are a number of pre-defined variables, including:
8.2.1.2 Logic and Iteration
Using the
<% %>
syntax you can embed loops and so on using this syntax:
<html>
<body>
<% [1,2,3,4].each { num -> %>
<p><%="Hello ${num}!" %></p>
<%}%>
</body>
</html>
As well as logical branching:
<html>
<body>
<% if (params.hello == 'true')%>
<%="Hello!"%>
<% else %>
<%="Goodbye!"%>
</body>
</html>
8.2.1.3 Page Directives
GSP also supports a few JSP-style page directives.
The import directive lets you import classes into the page. However, it is rarely needed due to Groovy's default imports and
GSP Tags:
<%@ page import="java.awt.*" %>
GSP also supports the contentType directive:
<%@ page contentType="application/json" %>
The contentType directive allows using GSP to render other formats.
8.2.1.4 Expressions
In GSP the
<%= %>
syntax introduced earlier is rarely used due to the support for GSP expressions. A GSP expression is similar to a JSP EL expression or a Groovy GString and takes the form
${expr}
:
<html>
<body>
Hello ${params.name}
</body>
</html>
However, unlike JSP EL you can have any Groovy expression within the
${..}
block.
Embedding data received from user input has the risk of making your application vulnerable to an Cross Site Scripting (XSS) attack. Please read the documentation on XSS prevention for information on how to prevent XSS attacks.
Now that the less attractive JSP heritage has been set aside, the following sections cover GSP's built-in tags, which are the preferred way to define GSP pages.
The section on Tag Libraries covers how to add your own custom tag libraries.
All built-in GSP tags start with the prefix
g:
. Unlike JSP, you don't specify any tag library imports. If a tag starts with
g:
it is automatically assumed to be a GSP tag. An example GSP tag would look like:
GSP tags can also have a body such as:
<g:example>
Hello world
</g:example>
Expressions can be passed into GSP tag attributes, if an expression is not used it will be assumed to be a String value:
<g:example attr="${new Date()}">
Hello world
</g:example>
Maps can also be passed into GSP tag attributes, which are often used for a named parameter style syntax:
<g:example attr="${new Date()}" attr2="[one:1, two:2, three:3]">
Hello world
</g:example>
Note that within the values of attributes you must use single quotes for Strings:
<g:example attr="${new Date()}" attr2="[one:'one', two:'two']">
Hello world
</g:example>
With the basic syntax out the way, the next sections look at the tags that are built into Grails by default.
8.2.2.1 Variables and Scopes
Variables can be defined within a GSP using the
set tag:
<g:set var="now" value="${new Date()}" />
Here we assign a variable called
now
to the result of a GSP expression (which simply constructs a new
java.util.Date
instance). You can also use the body of the
<g:set>
tag to define a variable:
<g:set var="myHTML">
Some re-usable code on: ${new Date()}
</g:set>
The assigned value can also be a bean from the applicationContext:
<g:set var="bookService" bean="bookService" />
Variables can also be placed in one of the following scopes:
page
- Scoped to the current page (default)
request
- Scoped to the current request
flash
- Placed within flash scope and hence available for the next request
session
- Scoped for the user session
application
- Application-wide scope.
To specify the scope, use the
scope
attribute:
<g:set var="now" value="${new Date()}" scope="request" />
8.2.2.2 Logic and Iteration
GSP also supports logical and iterative tags out of the box. For logic there are
if,
else and
elseif tags for use with branching:
<g:if test="${session.role == 'admin'}">
<%-- show administrative functions --%>
</g:if>
<g:else>
<%-- show basic functions --%>
</g:else>
Use the
each and
while tags for iteration:
<g:each in="${[1,2,3]}" var="num">
<p>Number ${num}</p>
</g:each><g:set var="num" value="${1}" />
<g:while test="${num < 5 }">
<p>Number ${num++}</p>
</g:while>
8.2.2.3 Search and Filtering
If you have collections of objects you often need to sort and filter them. Use the
findAll and
grep tags for these tasks:
Stephen King's Books:
<g:findAll in="${books}" expr="it.author == 'Stephen King'">
<p>Title: ${it.title}</p>
</g:findAll>
The
expr
attribute contains a Groovy expression that can be used as a filter. The
grep tag does a similar job, for example filtering by class:
<g:grep in="${books}" filter="NonFictionBooks.class">
<p>Title: ${it.title}</p>
</g:grep>
Or using a regular expression:
<g:grep in="${books.title}" filter="~/.*?Groovy.*?/">
<p>Title: ${it}</p>
</g:grep>
The above example is also interesting due to its usage of GPath. GPath is an XPath-like language in Groovy. The
books
variable is a collection of
Book
instances. Since each
Book
has a
title
, you can obtain a list of Book titles using the expression
books.title
. Groovy will auto-magically iterate the collection, obtain each title, and return a new list!
8.2.2.4 Links and Resources
GSP also features tags to help you manage linking to controllers and actions. The
link tag lets you specify controller and action name pairing and it will automatically work out the link based on the
URL Mappings, even if you change them! For example:
<g:link action="show" id="1">Book 1</g:link><g:link action="show" id="${currentBook.id}">${currentBook.name}</g:link><g:link controller="book">Book Home</g:link><g:link controller="book" action="list">Book List</g:link><g:link url="[action: 'list', controller: 'book']">Book List</g:link><g:link params="[sort: 'title', order: 'asc', author: currentBook.author]"
action="list">Book List</g:link>
Form Basics
GSP supports many different tags for working with HTML forms and fields, the most basic of which is the
form tag. This is a controller/action aware version of the regular HTML form tag. The
url
attribute lets you specify which controller and action to map to:
<g:form name="myForm" url="[controller:'book',action:'list']">...</g:form>
In this case we create a form called
myForm
that submits to the
BookController
's
list
action. Beyond that all of the usual HTML attributes apply.
Form Fields
In addition to easy construction of forms, GSP supports custom tags for dealing with different types of fields, including:
- textField - For input fields of type 'text'
- passwordField - For input fields of type 'password'
- checkBox - For input fields of type 'checkbox'
- radio - For input fields of type 'radio'
- hiddenField - For input fields of type 'hidden'
- select - For dealing with HTML select boxes
Each of these allows GSP expressions for the value:
<g:textField name="myField" value="${myValue}" />
GSP also contains extended helper versions of the above tags such as
radioGroup (for creating groups of
radio tags),
localeSelect,
currencySelect and
timeZoneSelect (for selecting locales, currencies and time zones respectively).
Multiple Submit Buttons
The age old problem of dealing with multiple submit buttons is also handled elegantly with Grails using the
actionSubmit tag. It is just like a regular submit, but lets you specify an alternative action to submit to:
<g:actionSubmit value="Some update label" action="update" />
One major different between GSP tags and other tagging technologies is that GSP tags can be called as either regular tags or as method calls from
controllers,
tag libraries or GSP views.
Tags as method calls from GSPs
Tags return their results as a String-like object (a
StreamCharBuffer
which has all of the same methods as String) instead of writing directly to the response when called as methods. For example:
Static Resource: ${createLinkTo(dir: "images", file: "logo.jpg")}
This is particularly useful for using a tag within an attribute:
<img src="${createLinkTo(dir: 'images', file: 'logo.jpg')}" />
In view technologies that don't support this feature you have to nest tags within tags, which becomes messy quickly and often has an adverse effect of WYSIWYG tools such as Dreamweaver that attempt to render the mark-up as it is not well-formed:
<img src="<g:createLinkTo dir="images" file="logo.jpg" />" />
Tags as method calls from Controllers and Tag Libraries
You can also invoke tags from controllers and tag libraries. Tags within the default
g:
namespace can be invoked without the prefix and a
StreamCharBuffer
result is returned:
def imageLocation = createLinkTo(dir:"images", file:"logo.jpg").toString()
Prefix the namespace to avoid naming conflicts:
def imageLocation = g.createLinkTo(dir:"images", file:"logo.jpg").toString()
For tags that use a
custom namespace, use that prefix for the method call. For example (from the
FCK Editor plugin):
def editor = fckeditor.editor(name: "text", width: "100%", height: "400")
8.2.3 Views and Templates
Grails also has the concept of templates. These are useful for partitioning your views into maintainable chunks, and combined with
Layouts provide a highly re-usable mechanism for structured views.
Template Basics
Grails uses the convention of placing an underscore before the name of a view to identify it as a template. For example, you might have a template that renders Books located at
grails-app/views/book/_bookTemplate.gsp
:
<div class="book" id="${book?.id}">
<div>Title: ${book?.title}</div>
<div>Author: ${book?.author?.name}</div>
</div>
Use the
render tag to render this template from one of the views in
grails-app/views/book
:
<g:render template="bookTemplate" model="[book: myBook]" />
Notice how we pass into a model to use using the
model
attribute of the
render
tag. If you have multiple
Book
instances you can also render the template for each
Book
using the render tag with a
collection
attribute:
<g:render template="bookTemplate" var="book" collection="${bookList}" />
Shared Templates
In the previous example we had a template that was specific to the
BookController
and its views at
grails-app/views/book
. However, you may want to share templates across your application.
In this case you can place them in the root views directory at grails-app/views or any subdirectory below that location, and then with the template attribute use an absolute location starting with
/
instead of a relative location. For example if you had a template called
grails-app/views/shared/_mySharedTemplate.gsp
, you would reference it as:
<g:render template="/shared/mySharedTemplate" />
You can also use this technique to reference templates in any directory from any view or controller:
<g:render template="/book/bookTemplate" model="[book: myBook]" />
The Template Namespace
Since templates are used so frequently there is template namespace, called
tmpl
, available that makes using templates easier. Consider for example the following usage pattern:
<g:render template="bookTemplate" model="[book:myBook]" />
This can be expressed with the
tmpl
namespace as follows:
<tmpl:bookTemplate book="${myBook}" />
Templates in Controllers and Tag Libraries
You can also render templates from controllers using the
render controller method. This is useful for
Ajax applications where you generate small HTML or data responses to partially update the current page instead of performing new request:
def bookData() {
def b = Book.get(params.id)
render(template:"bookTemplate", model:[book:b])
}
The
render controller method writes directly to the response, which is the most common behaviour. To instead obtain the result of template as a String you can use the
render tag:
def bookData() {
def b = Book.get(params.id)
String content = g.render(template:"bookTemplate", model:[book:b])
render content
}
Notice the usage of the
g
namespace which tells Grails we want to use the
tag as method call instead of the
render method.
8.2.4 Layouts with Sitemesh
Creating Layouts
Grails leverages
Sitemesh, a decorator engine, to support view layouts. Layouts are located in the
grails-app/views/layouts
directory. A typical layout can be seen below:
<html>
<head>
<title><g:layoutTitle default="An example decorator" /></title>
<g:layoutHead />
</head>
<body onload="${pageProperty(name:'body.onload')}">
<div class="menu"></menu>
<div class="body">
<g:layoutBody />
</div>
</div>
</body>
</html>
The key elements are the
layoutHead,
layoutTitle and
layoutBody tag invocations:
layoutTitle
- outputs the target page's title
layoutHead
- outputs the target page's head tag contents
layoutBody
- outputs the target page's body tag contents
The previous example also demonstrates the
pageProperty tag which can be used to inspect and return aspects of the target page.
Triggering Layouts
There are a few ways to trigger a layout. The simplest is to add a meta tag to the view:
<html>
<head>
<title>An Example Page</title>
<meta name="layout" content="main" />
</head>
<body>This is my content!</body>
</html>
In this case a layout called
grails-app/views/layouts/main.gsp
will be used to layout the page. If we were to use the layout from the previous section the output would resemble this:
<html>
<head>
<title>An Example Page</title>
</head>
<body onload="">
<div class="menu"></div>
<div class="body">
This is my content!
</div>
</body>
</html>
Specifying A Layout In A Controller
Another way to specify a layout is to specify the name of the layout by assigning a value to the "layout" property in a controller. For example, if you have a controller such as:
class BookController {
static layout = 'customer' def list() { … }
}
You can create a layout called
grails-app/views/layouts/customer.gsp
which will be applied to all views that the
BookController
delegates to. The value of the "layout" property may contain a directory structure relative to the
grails-app/views/layouts/
directory. For example:
class BookController {
static layout = 'custom/customer' def list() { … }
}
Views rendered from that controller would be decorated with the
grails-app/views/layouts/custom/customer.gsp
template.
Layout by Convention
Another way to associate layouts is to use "layout by convention". For example, if you have this controller:
class BookController {
def list() { … }
}
You can create a layout called
grails-app/views/layouts/book.gsp
, which will be applied to all views that the
BookController
delegates to.
Alternatively, you can create a layout called
grails-app/views/layouts/book/list.gsp
which will only be applied to the
list
action within the
BookController
.
If you have both the above mentioned layouts in place the layout specific to the action will take precedence when the list action is executed.
If a layout may not be located using any of those conventions, the convention of last resort is to look for the application default layout which
is
grails-app/views/layouts/application.gsp
. The name of the application default layout may be changed by defining a property
in
grails-app/conf/Config.groovy
as follows:
grails.sitemesh.default.layout = 'myLayoutName'
With that property in place, the application default layout will be
grails-app/views/layouts/myLayoutName.gsp
.
Inline Layouts
Grails' also supports Sitemesh's concept of inline layouts with the
applyLayout tag. This can be used to apply a layout to a template, URL or arbitrary section of content. This lets you even further modularize your view structure by "decorating" your template includes.
Some examples of usage can be seen below:
<g:applyLayout name="myLayout" template="bookTemplate" collection="${books}" /><g:applyLayout name="myLayout" url="http://www.google.com" /><g:applyLayout name="myLayout">
The content to apply a layout to
</g:applyLayout>
Server-Side Includes
While the
applyLayout tag is useful for applying layouts to external content, if you simply want to include external content in the current page you use the
include tag:
<g:include controller="book" action="list" />
You can even combine the
include tag and the
applyLayout tag for added flexibility:
<g:applyLayout name="myLayout">
<g:include controller="book" action="list" />
</g:applyLayout>
Finally, you can also call the
include tag from a controller or tag library as a method:
def content = include(controller:"book", action:"list")
The resulting content will be provided via the return value of the
include tag.
8.2.5 Static Resources
Grails 2.0 integrates with the
Asset Pipeline plugin to provide sophisticated static asset management. This plugin is installed by default in new Grails applications.
The basic way to include a link to a static asset in your application is to use the
resource tag. This simple approach creates a URI pointing to the file.
However modern applications with dependencies on multiple JavaScript and CSS libraries and frameworks (as well as dependencies on multiple Grails plugins) require something more powerful.
The issues that the Asset-Pipeline plugin tackles are:
- Reduced Dependence - The plugin has compression, minification, and cache-digests built in.
- Easy Debugging - Makes for easy debugging by keeping files separate in development mode.
- Asset Bundling using require directives.
- Web application performance tuning is difficult.
- The need for a standard way to expose static assets in plugins and applications.
- The need for extensible processing to make languages like LESS or Coffee first class citizens.
The asset-pipeline allows you to define your javascript or css requirements right at the top of the file and they get compiled on War creation.
Take a look at the
documentation for the asset-pipeline to get started.
Pulling in resources with r:require
To use resources, your GSP page must indicate which resource modules it requires. For example with the
jQuery plugin, which exposes a "jquery" resource module, to use jQuery in any page on your site you simply add:
<html>
<head>
<r:require module="jquery"/>
<r:layoutResources/>
</head>
<body>
…
<r:layoutResources/>
</body>
</html>
This will automatically include all resources needed for jQuery, including them at the correct locations in the page. By default the plugin sets the disposition to be "head", so they load early in the page.
You can call
r:require
multiple times in a GSP page, and you use the "modules" attribute to provide a list of modules:
<html>
<head>
<r:require modules="jquery, main, blueprint, charting"/>
<r:layoutResources/>
</head>
<body>
…
<r:layoutResources/>
</body>
</html>
The above may result in many JavaScript and CSS files being included, in the correct order, with some JavaScript files loading at the end of the body to improve the apparent page load time.
However you cannot use r:require in isolation - as per the examples you must have the <r:layoutResources/> tag to actually perform the render.
Rendering the links to resources with r:layoutResources
When you have declared the resource modules that your GSP page requires, the framework needs to render the links to those resources at the correct time.
To achieve this correctly, you must include the r:layoutResources tag twice in your page, or more commonly, in your GSP layout:
<html>
<head>
<g:layoutTitle/>
<r:layoutResources/>
</head>
<body>
<g:layoutBody/>
<r:layoutResources/>
</body>
</html>
This represents the simplest Sitemesh layout you can have that supports Resources.
The Resources framework has the concept of a "disposition" for every resource. This is an indication of where in the page the resource should be included.
The default disposition applied depends on the type of resource. All CSS must be rendered in <head> in HTML, so "head" is the default for all CSS, and will be rendered by the first r:layoutResources. Page load times are improved when JavaScript is loaded after the page content, so the default for JavaScript files is "defer", which means it is rendered when the second r:layoutResources is invoked.
Note that both your GSP page and your Sitemesh layout (as well as any GSP template fragments) can call r:require to depend on resources. The only limitation is that you must call r:require before the r:layoutResources that should render it.
Adding page-specific JavaScript code with r:script
Grails has the
javascript tag which is adapted to defer to Resources plugin if installed, but it is recommended that you call
r:script
directly when you need to include fragments of JavaScript code.
This lets you write some "inline" JavaScript which is actually
not rendered inline, but either in the <head> or at the end of the body, based on the disposition.
Given a Sitemesh layout like this:
<html>
<head>
<g:layoutTitle/>
<r:layoutResources/>
</head>
<body>
<g:layoutBody/>
<r:layoutResources/>
</body>
</html>
...in your GSP you can inject some JavaScript code into the head or deferred regions of the page like this:
<html>
<head>
<title>Testing r:script magic!</title>
</head>
<body>
<r:script disposition="head">
window.alert('This is at the end of <head>');
</r:script>
<r:script disposition="defer">
window.alert('This is at the end of the body, and the page has loaded.');
</r:script>
</body>
</html>
The default disposition is "defer", so the disposition in the latter r:script is purely included for demonstration.
Note that such r:script code fragments
always load after any modules that you have used, to ensure that any required libraries have loaded.
Linking to images with r:img
This tag is used to render
<img>
markup, using the Resources framework to process the resource on the fly (if configured to do so - e.g. make it eternally cacheable).
This includes any extra attributes on the
<img>
tag if the resource has been previously declared in a module.
With this mechanism you can specify the width, height and any other attributes in the resource declaration in the module, and they will be pulled in as necessary.
Example:
<html>
<head>
<title>Testing r:img</title>
</head>
<body>
<r:img uri="/images/logo.png"/>
</body>
</html>
Note that Grails has a built-in
g:img
tag as a shortcut for rendering
<img>
tags that refer to a static resource. The Grails
img tag is Resources-aware and will delegate to
r:img
if found. However it is recommended that you use
r:img
directly if using the Resources plugin.
Alongside the regular Grails
resource tag attributes, this also supports the "uri" attribute for increased brevity.
See
r:resource documentation for full details.
r:resource
This is equivalent to the Grails
resource tag, returning a link to the processed static resource. Grails' own
g:resource
tag delegates to this implementation if found, but if your code requires the Resources plugin, you should use
r:resource
directly.
Alongside the regular Grails
resource tag attributes, this also supports the "uri" attribute for increased brevity.
See
r:resource documentation for full details.
r:external
This is a resource-aware version of Grails
external tag which renders the HTML markup necessary to include an external file resource such as CSS, JS or a favicon.
See
r:resource documentation for full details.
8.2.5.3 Declaring resources
A DSL is provided for declaring resources and modules. This can go either in your
Config.groovy
in the case of application-specific resources, or more commonly in a resources artefact in
grails-app/conf
.
Note that you do not need to declare all your static resources, especially images. However you must to establish dependencies or other resources-specific attributes. Any resource that is not declared is called "ad-hoc" and will still be processed using defaults for that resource type.
Consider this example resource configuration file,
grails-app/conf/MyAppResources.groovy
:
modules = {
core {
dependsOn 'jquery, utils' resource url: '/js/core.js', disposition: 'head'
resource url: '/js/ui.js'
resource url: '/css/main.css',
resource url: '/css/branding.css'
resource url: '/css/print.css', attrs: [media: 'print']
} utils {
dependsOn 'jquery' resource url: '/js/utils.js'
} forms {
dependsOn 'core,utils' resource url: '/css/forms.css'
resource url: '/js/forms.js'
}
}
This defines three resource modules; 'core', 'utils' and 'forms'. The resources in these modules will be automatically bundled out of the box according to the module name, resulting in fewer files. You can override this with
bundle:'someOtherName'
on each resource, or call
defaultBundle
on the module (see
resources plugin documentation).
It declares dependencies between them using
dependsOn
, which controls the load order of the resources.
When you include an
<r:require module="forms"/>
in your GSP, it will pull in all the resources from 'core' and 'utils' as well as 'jquery', all in the correct order.
You'll also notice the
disposition:'head'
on the
core.js
file. This tells Resources that while it can defer all the other JS files to the end of the body, this one must go into the
<head>
.
The CSS file for print styling adds custom attributes using the
attrs
map option, and these are passed through to the
r:external
tag when the engine renders the link to the resource, so you can customize the HTML attributes of the generated link.
There is no limit to the number of modules or xxxResources.groovy artefacts you can provide, and plugins can supply them to expose modules to applications, which is exactly how the jQuery plugin works.
To define modules like this in your application's Config.groovy, you simply assign the DSL closure to the
grails.resources.modules
Config variable.
For full details of the resource DSL please see the
resources plugin documentation.
8.2.5.4 Overriding plugin resources
Because a resource module can define the bundle groupings and other attributes of resources, you may find that the settings provided are not correct for your application.
For example, you may wish to bundle jQuery and some other libraries all together in one file. There is a load-time and caching trade-off here, but often it is the case that you'd like to override some of these settings.
To do this, the DSL supports an "overrides" clause, within which you can change the
defaultBundle
setting for a module, or attributes of individual resources that have been declared with a unique id:
modules = {
core {
dependsOn 'jquery, utils'
defaultBundle 'monolith' resource url: '/js/core.js', disposition: 'head'
resource url: '/js/ui.js'
resource url: '/css/main.css',
resource url: '/css/branding.css'
resource url: '/css/print.css', attrs: [media: 'print']
} utils {
dependsOn 'jquery'
defaultBundle 'monolith' resource url: '/js/utils.js'
} forms {
dependsOn 'core,utils'
defaultBundle 'monolith' resource url: '/css/forms.css'
resource url: '/js/forms.js'
} overrides {
jquery {
defaultBundle 'monolith'
}
}
}
This will put all code into a single bundle named 'monolith'. Note that this can still result in multiple files, as separate bundles are required for head and defer dispositions, and JavaScript and CSS files are bundled separately.
Note that overriding individual resources requires the original declaration to have included a unique id for the resource.
For full details of the resource DSL please see the
resources plugin documentation.
8.2.5.5 Optimizing your resources
The Resources framework uses "mappers" to mutate the resources into the final format served to the user.
The resource mappers are applied to each static resource once, in a specific order. You can create your own resource mappers, and several plugins provide some already for zipping, caching and minifying.
Out of the box, the Resources plugin provides bundling of resources into fewer files, which is achieved with a few mappers that also perform CSS re-writing to handle when your CSS files are moved into a bundle.
Bundling multiple resources into fewer files
The 'bundle' mapper operates by default on any resource with a "bundle" defined - or inherited from a
defaultBundle
clause on the module. Modules have an implicit default bundle name the same as the name of the module.
Files of the same kind will be aggregated into this bundle file. Bundles operate across module boundaries:
modules = {
core {
dependsOn 'jquery, utils'
defaultBundle 'common' resource url: '/js/core.js', disposition: 'head'
resource url: '/js/ui.js', bundle: 'ui'
resource url: '/css/main.css', bundle: 'theme'
resource url: '/css/branding.css'
resource url: '/css/print.css', attrs: [media: 'print']
} utils {
dependsOn 'jquery' resource url: '/js/utils.js', bundle: 'common'
} forms {
dependsOn 'core,utils' resource url: '/css/forms.css', bundle: 'ui'
resource url: '/js/forms.js', bundle: 'ui'
}
}
Here you see that resources are grouped into bundles; 'common', 'ui' and 'theme' - across module boundaries.
Note that auto-bundling by module does
not occur if there is only one resource in the module.
Making resources cache "eternally" in the client browser
Caching resources "eternally" in the client is only viable if the resource has a unique name that changes whenever the contents change, and requires caching headers to be set on the response.
The
cached-resources plugin provides a mapper that achieves this by hashing your files and renaming them based on this hash. It also sets the caching headers on every response for those resources. To use, simply install the cached-resources plugin.
Note that the caching headers can only be set if your resources are being served by your application. If you have another server serving the static content from your app (e.g. Apache HTTPD), configure it to send caching headers. Alternatively you can configure it to request and proxy the resources from your container.
Zipping resources
Returning gzipped resources is another way to reduce page load times and reduce bandwidth.
The
zipped-resources plugin provides a mapper that automatically compresses your content, excluding by default already compressed formats such as gif, jpeg and png.
Simply install the zipped-resources plugin and it works.
Minifying
There are a number of CSS and JavaScript minifiers available to obfuscate and reduce the size of your code. At the time of writing none are publicly released but releases are imminent.
8.2.5.6 Debugging
When your resources are being moved around, renamed and otherwise mutated, it can be hard to debug client-side issues. Modern browsers, especially Safari, Chrome and Firefox have excellent tools that let you view all the resources requested by a page, including the headers and other information about them.
There are several debugging features built in to the Resources framework.
X-Grails-Resources-Original-Src Header
Every resource served in development mode will have the X-Grails-Resources-Original-Src: header added, indicating the original source file(s) that make up the response.
Adding the debug flag
If you add a query parameter
_debugResources=y to your URL and request the page, Resources will bypass any processing so that you can see your original source files.
This also adds a unique timestamp to all your resource URLs, to defeat any caching that browsers may use. This means that you should always see your very latest code when you reload the page.
Turning on debug all the time
You can turn on the aforementioned debug mechanism without requiring a query parameter, but turning it on in Config.groovy:
grails.resources.debug = true
You can of course set this per-environment.
8.2.5.7 Preventing processing of resources
Sometimes you do not want a resource to be processed in a particular way, or even at all. Occasionally you may also want to disable all resource mapping.
Preventing the application of a specific mapper to an individual resource
All resource declarations support a convention of noXXXX:true where XXXX is a mapper name.
So for example to prevent the "hashandcache" mapper from being applied to a resource (which renames and moves it, potentially breaking relative links written in JavaScript code), you would do this:
modules = {
forms {
resource url: '/css/forms.css', nohashandcache: true
resource url: '/js/forms.js', nohashandcache: true
}
}
Excluding/including paths and file types from specific mappers
Mappers have includes/excludes Ant patterns to control whether they apply to a given resource. Mappers set sensible defaults for these based on their activity, for example the zipped-resources plugin's "zip" mapper is set to exclude images by default.
You can configure this in your
Config.groovy
using the mapper name e.g:
// We wouldn't link to .exe files using Resources but for the sake of example:
grails.resources.zip.excludes = ['**/*.zip', '**/*.exe']// Perhaps for some reason we want to prevent bundling on "less" CSS files:
grails.resources.bundle.excludes = ['**/*.less']
There is also an "includes" inverse. Note that settings these replaces the default includes/excludes for that mapper - it is not additive.
Controlling what is treated as an "ad-hoc" (legacy) resource
Ad-hoc resources are those undeclared, but linked to directly in your application
without using the Grails or Resources linking tags (resource, img or external).
These may occur with some legacy plugins or code with hardcoded paths in.
There is a Config.groovy setting
grails.resources.adhoc.patterns which defines a list of Servlet API compliant filter URI mappings, which the Resources filter will use to detect such "ad-hoc resource" requests.
By default this is set to:
grails.resources.adhoc.patterns = ['images/*', '*.js', '*.css']
8.2.5.8 Other Resources-aware plugins
At the time of writing, the following plugins include support for the Resources framework:
8.2.6 Sitemesh Content Blocks
Although it is useful to decorate an entire page sometimes you may find the need to decorate independent sections of your site. To do this you can use content blocks. To get started, partition the page to be decorated using the
<content>
tag:
<content tag="navbar">
… draw the navbar here…
</content><content tag="header">
… draw the header here…
</content><content tag="footer">
… draw the footer here…
</content><content tag="body">
… draw the body here…
</content>
Then within the layout you can reference these components and apply individual layouts to each:
<html>
<body>
<div id="header">
<g:applyLayout name="headerLayout">
<g:pageProperty name="page.header" />
</g:applyLayout>
</div>
<div id="nav">
<g:applyLayout name="navLayout">
<g:pageProperty name="page.navbar" />
</g:applyLayout>
</div>
<div id="body">
<g:applyLayout name="bodyLayout">
<g:pageProperty name="page.body" />
</g:applyLayout>
</div>
<div id="footer">
<g:applyLayout name="footerLayout">
<g:pageProperty name="page.footer" />
</g:applyLayout>
</div>
</body>
</html>
8.2.7 Making Changes to a Deployed Application
One of the main issues with deploying a Grails application (or typically any servlet-based one) is that any change to the views requires that you redeploy your whole application. If all you want to do is fix a typo on a page, or change an image link, it can seem like a lot of unnecessary work. For such simple requirements, Grails does have a solution: the
grails.gsp.view.dir
configuration setting.
How does this work? The first step is to decide where the GSP files should go. Let's say we want to keep them unpacked in a
/var/www/grails/my-app
directory. We add these two lines to
grails-app/conf/Config.groovy
:
grails.gsp.enable.reload = true
grails.gsp.view.dir = "/var/www/grails/my-app/"
The first line tells Grails that modified GSP files should be reloaded at runtime. If you don't have this setting, you can make as many changes as you like but they won't be reflected in the running application until you restart. The second line tells Grails where to load the views and layouts from.
The trailing slash on the grails.gsp.view.dir
value is important! Without it, Grails will look for views in the parent directory.
Setting "grails.gsp.view.dir" is optional. If it's not specified, you can update files directly to the application server's deployed war directory. Depending on the application server, these files might get overwritten when the server is restarted. Most application servers support "exploded war deployment" which is recommended in this case.
With those settings in place, all you need to do is copy the views from your web application to the external directory. On a Unix-like system, this would look something like this:
mkdir -p /var/www/grails/my-app/grails-app/views
cp -R grails-app/views/* /var/www/grails/my-app/grails-app/views
The key point here is that you must retain the view directory structure, including the
grails-app/views
bit. So you end up with the path
/var/www/grails/my-app/grails-app/views/...
.
One thing to bear in mind with this technique is that every time you modify a GSP, it uses up permgen space. So at some point you will eventually hit "out of permgen space" errors unless you restart the server. So this technique is not recommended for frequent or large changes to the views.
There are also some System properties to control GSP reloading:
Name | Description | Default |
---|
grails.gsp.enable.reload | alternative system property for enabling the GSP reload mode without changing Config.groovy | |
grails.gsp.reload.interval | interval between checking the lastmodified time of the gsp source file, unit is milliseconds | 5000 |
grails.gsp.reload.granularity | the number of milliseconds leeway to give before deciding a file is out of date. this is needed because different roundings usually cause a 1000ms difference in lastmodified times | 1000 |
GSP reloading is supported for precompiled GSPs since Grails 1.3.5 .
8.2.8 GSP Debugging
Viewing the generated source code
- Adding "?showSource=true" or "&showSource=true" to the url shows the generated Groovy source code for the view instead of rendering it. It won't show the source code of included templates. This only works in development mode
- The saving of all generated source code can be activated by setting the property "grails.views.gsp.keepgenerateddir" (in Config.groovy) . It must point to a directory that exists and is writable.
- During "grails war" gsp pre-compilation, the generated source code is stored in grails.project.work.dir/gspcompile (usually in ~/.grails/(grails_version)/projects/(project name)/gspcompile).
Debugging GSP code with a debugger
Viewing information about templates used to render a single url
GSP templates are reused in large web applications by using the
g:render
taglib. Several small templates can be used to render a single page.
It might be hard to find out what GSP template actually renders the html seen in the result.
The debug templates -feature adds html comments to the output. The comments contain debug information about gsp templates used to render the page.
Usage is simple: append "?debugTemplates" or "&debugTemplates" to the url and view the source of the result in your browser.
"debugTemplates" is restricted to development mode. It won't work in production.
Here is an example of comments added by debugTemplates :
<!-- GSP #2 START template: /home/.../views/_carousel.gsp
precompiled: false lastmodified: … -->
.
.
.
<!-- GSP #2 END template: /home/.../views/_carousel.gsp
rendering time: 115 ms -->
Each comment block has a unique id so that you can find the start & end of each template call.
8.3 Tag Libraries
Like
Java Server Pages (JSP), GSP supports the concept of custom tag libraries. Unlike JSP, Grails' tag library mechanism is simple, elegant and completely reloadable at runtime.
Quite simply, to create a tag library create a Groovy class that ends with the convention
TagLib
and place it within the
grails-app/taglib
directory:
Now to create a tag create a Closure property that takes two arguments: the tag attributes and the body content:
class SimpleTagLib {
def simple = { attrs, body -> }
}
The
attrs
argument is a Map of the attributes of the tag, whilst the
body
argument is a Closure that returns the body content when invoked:
class SimpleTagLib {
def emoticon = { attrs, body ->
out << body() << (attrs.happy == 'true' ? " :-)" : " :-(")
}
}
As demonstrated above there is an implicit
out
variable that refers to the output
Writer
which you can use to append content to the response. Then you can reference the tag inside your GSP; no imports are necessary:
<g:emoticon happy="true">Hi John</g:emoticon>
To help IDEs like Spring Tool Suite (STS) and others autocomplete tag attributes, you should add Javadoc comments to your tag closures with @attr
descriptions. Since taglibs use Groovy code it can be difficult to reliably detect all usable attributes.For example:class SimpleTagLib { /**
* Renders the body with an emoticon.
*
* @attr happy whether to show a happy emoticon ('true') or
* a sad emoticon ('false')
*/
def emoticon = { attrs, body ->
out << body() << (attrs.happy == 'true' ? " :-)" : " :-(")
}
}
and any mandatory attributes should include the REQUIRED keyword, e.g.class SimpleTagLib { /**
* Creates a new password field.
*
* @attr name REQUIRED the field name
* @attr value the field value
*/
def passwordField = { attrs ->
attrs.type = "password"
attrs.tagName = "passwordField"
fieldImpl(out, attrs)
}
}
8.3.1 Variables and Scopes
Within the scope of a tag library there are a number of pre-defined variables including:
actionName
- The currently executing action name
controllerName
- The currently executing controller name
flash
- The flash object
grailsApplication
- The GrailsApplication instance
out
- The response writer for writing to the output stream
pageScope
- A reference to the pageScope object used for GSP rendering (i.e. the binding)
params
- The params object for retrieving request parameters
pluginContextPath
- The context path to the plugin that contains the tag library
request
- The HttpServletRequest instance
response
- The HttpServletResponse instance
servletContext
- The javax.servlet.ServletContext instance
session
- The HttpSession instance
As demonstrated in the previous example it is easy to write simple tags that have no body and just output content. Another example is a
dateFormat
style tag:
def dateFormat = { attrs, body ->
out << new java.text.SimpleDateFormat(attrs.format).format(attrs.date)
}
The above uses Java's
SimpleDateFormat
class to format a date and then write it to the response. The tag can then be used within a GSP as follows:
<g:dateFormat format="dd-MM-yyyy" date="${new Date()}" />
With simple tags sometimes you need to write HTML mark-up to the response. One approach would be to embed the content directly:
def formatBook = { attrs, body ->
out << "<div id="${attrs.book.id}">"
out << "Title : ${attrs.book.title}"
out << "</div>"
}
Although this approach may be tempting it is not very clean. A better approach would be to reuse the
render tag:
def formatBook = { attrs, body ->
out << render(template: "bookTemplate", model: [book: attrs.book])
}
And then have a separate GSP template that does the actual rendering.
You can also create logical tags where the body of the tag is only output once a set of conditions have been met. An example of this may be a set of security tags:
def isAdmin = { attrs, body ->
def user = attrs.user
if (user && checkUserPrivs(user)) {
out << body()
}
}
The tag above checks if the user is an administrator and only outputs the body content if he/she has the correct set of access privileges:
<g:isAdmin user="${myUser}">
// some restricted content
</g:isAdmin>
Iterative tags are easy too, since you can invoke the body multiple times:
def repeat = { attrs, body ->
attrs.times?.toInteger()?.times { num ->
out << body(num)
}
}
In this example we check for a
times
attribute and if it exists convert it to a number, then use Groovy's
times
method to iterate the specified number of times:
<g:repeat times="3">
<p>Repeat this 3 times! Current repeat = ${it}</p>
</g:repeat>
Notice how in this example we use the implicit
it
variable to refer to the current number. This works because when we invoked the body we passed in the current value inside the iteration:
That value is then passed as the default variable
it
to the tag. However, if you have nested tags this can lead to conflicts, so you should instead name the variables that the body uses:
def repeat = { attrs, body ->
def var = attrs.var ?: "num"
attrs.times?.toInteger()?.times { num ->
out << body((var):num)
}
}
Here we check if there is a
var
attribute and if there is use that as the name to pass into the body invocation on this line:
Note the usage of the parenthesis around the variable name. If you omit these Groovy assumes you are using a String key and not referring to the variable itself.
Now we can change the usage of the tag as follows:
<g:repeat times="3" var="j">
<p>Repeat this 3 times! Current repeat = ${j}</p>
</g:repeat>
Notice how we use the
var
attribute to define the name of the variable
j
and then we are able to reference that variable within the body of the tag.
8.3.5 Tag Namespaces
By default, tags are added to the default Grails namespace and are used with the
g:
prefix in GSP pages. However, you can specify a different namespace by adding a static property to your
TagLib
class:
class SimpleTagLib {
static namespace = "my" def example = { attrs ->
…
}
}
Here we have specified a
namespace
of
my
and hence the tags in this tag lib must then be referenced from GSP pages like this:
<my:example name="..." />
where the prefix is the same as the value of the static
namespace
property. Namespaces are particularly useful for plugins.
Tags within namespaces can be invoked as methods using the namespace as a prefix to the method call:
out << my.example(name:"foo")
This works from GSP, controllers or tag libraries
8.3.6 Using JSP Tag Libraries
In addition to the simplified tag library mechanism provided by GSP, you can also use JSP tags from GSP. To do so simply declare the JSP to use with the
taglib
directive:
<%@ taglib prefix="fmt" uri="http://java.sun.com/jsp/jstl/fmt" %>
Besides this you have to configure Grails to scan for the JSP tld files.
This is configured with the
grails.gsp.tldScanPattern
setting. It accepts a comma separated String value. Spring's PathMatchingResourcePatternResolver is used to resolve the patterns.
For example you could scan for all available tld files by adding this to Config.groovy:
grails.gsp.tldScanPattern='classpath*:/META-INF/*.tld,/WEB-INF/tld/*.tld'
JSTL standard library is no more added as a dependency by default. In case you are using JSTL, you should also add these dependencies to BuildConfig.groovy:
runtime 'javax.servlet:jstl:1.1.2'
runtime 'taglibs:standard:1.1.2'
Then you can use JSP tags like any other tag:
<fmt:formatNumber value="${10}" pattern=".00"/>
With the added bonus that you can invoke JSP tags like methods:
${fmt.formatNumber(value:10, pattern:".00")}
8.3.7 Tag return value
A taglib can be used in a GSP as an ordinary tag or it might be used as a function in other taglibs or GSP expressions.
Internally Grails intercepts calls to taglib closures.
The "out" that is available in a taglib is mapped to a
java.io.Writer
implementation that writes to a buffer
that "captures" the output of the taglib call. This buffer is the return value of a tag library call when it's
used as a function.
If the tag is listed in the library's static
returnObjectForTags
array, then its return value will written to
the output when it's used as a normal tag. The return value of the tag lib closure will be returned as-is
if it's used as a function in GSP expressions or other taglibs.
If the tag is not included in the returnObjectForTags array, then its return value will be discarded.
Using "out" to write output in returnObjectForTags is not supported.
Example:
class ObjectReturningTagLib {
static namespace = "cms"
static returnObjectForTags = ['content'] def content = { attrs, body ->
CmsContent.findByCode(attrs.code)?.content
}
}
Given this example cmd.content(code:'something') call in another taglib or GSP expression would return the value "CmsContent.content" directly to the caller without
wrapping the return value in a buffer. It might be worth doing so also because of performance optimization reasons. There is no need to wrap the
tag return value in an output buffer in such cases.
8.4 URL Mappings
Throughout the documentation so far the convention used for URLs has been the default of
/controller/action/id
. However, this convention is not hard wired into Grails and is in fact controlled by a URL Mappings class located at
grails-app/conf/UrlMappings.groovy
.
The
UrlMappings
class contains a single property called
mappings
that has been assigned a block of code:
class UrlMappings {
static mappings = {
}
}
8.4.1 Mapping to Controllers and Actions
To create a simple mapping simply use a relative URL as the method name and specify named parameters for the controller and action to map to:
"/product"(controller: "product", action: "list")
In this case we've mapped the URL
/product
to the
list
action of the
ProductController
. Omit the action definition to map to the default action of the controller:
"/product"(controller: "product")
An alternative syntax is to assign the controller and action to use within a block passed to the method:
"/product" {
controller = "product"
action = "list"
}
Which syntax you use is largely dependent on personal preference.
If you have mappings that all fall under a particular path you can group mappings with the
group
method:
group "/product", {
"/apple"(controller:"product", id:"apple")
"/htc"(controller:"product", id:"htc")
}
To rewrite one URI onto another explicit URI (rather than a controller/action pair) do something like this:
"/hello"(uri: "/hello.dispatch")
Rewriting specific URIs is often useful when integrating with other frameworks.
8.4.2 Mapping to REST resources
Since Grails 2.3, it possible to create RESTful URL mappings that map onto controllers by convention. The syntax to do so is as follows:
"/books"(resources:'book')
You define a base URI and the name of the controller to map to using the
resources
parameter. The above mapping will result in the following URLs:
HTTP Method | URI | Grails Action |
---|
GET | /books | index |
GET | /books/create | create |
POST | /books | save |
GET | /books/${id} | show |
GET | /books/${id}/edit | edit |
PUT | /books/${id} | update |
DELETE | /books/${id} | delete |
If you wish to include or exclude any of the generated URL mappings you can do so with the
includes
or
excludes
parameter, which accepts the name of the Grails action to include or exclude:
"/books"(resources:'book', excludes:['delete', 'update'])or"/books"(resources:'book', includes:['index', 'show'])
Single resources
A single resource is a resource for which there is only one (possibly per user) in the system. You can create a single resource using the
resource
parameter (as oppose to
resources
):
This results in the following URL mappings:
HTTP Method | URI | Grails Action |
---|
GET | /book/create | create |
POST | /book | save |
GET | /book | show |
GET | /book/edit | edit |
PUT | /book | update |
DELETE | /book | delete |
The main difference is that the id is not included in the URL mapping.
Nested Resources
You can nest resource mappings to generate child resources. For example:
"/books"(resources:'book') {
"/authors"(resources:"author")
}
The above will result in the following URL mappings:
HTTP Method | URL | Grails Action |
---|
GET | /books/${bookId}/authors | index |
GET | /books/${bookId}/authors/create | create |
POST | /books/${bookId}/authors | save |
GET | /books/${bookId}/authors/${id} | show |
GET | /books/${bookId}/authors/edit/${id} | edit |
PUT | /books/${bookId}/authors/${id} | update |
DELETE | /books/${bookId}/authors/${id} | delete |
You can also nest regular URL mappings within a resource mapping:
"/books"(resources: "book") {
"/publisher"(controller:"publisher")
}
This will result in the following URL being available:
HTTP Method | URL | Grails Action |
---|
GET | /books/1/publisher | index |
Linking to RESTful Mappings
You can link to any URL mapping created with the
g:link
tag provided by Grails simply by referencing the controller and action to link to:
<g:link controller="book" action="index">My Link</g:link>
As a convenience you can also pass a domain instance to the
resource
attribute of the
link
tag:
<g:link resource="${book}">My Link</g:link>
This will automatically produce the correct link (in this case "/books/1" for an id of "1").
The case of nested resources is a little different as they typically required two identifiers (the id of the resource and the one it is nested within). For example given the nested resources:
"/books"(resources:'book') {
"/authors"(resources:"author")
}
If you wished to link to the
show
action of the
author
controller, you would write:
// Results in /books/1/authors/2
<g:link controller="author" action="show" method="GET" params="[bookId:1]" id="2">The Author</g:link>
However, to make this more concise there is a
resource
attribute to the link tag which can be used instead:
// Results in /books/1/authors/2
<g:link resource="book/author" action="show" bookId="1" id="2">My Link</g:link>
The resource attribute accepts a path to the resource separated by a slash (in this case "book/author"). The attributes of the tag can be used to specify the necessary
bookId
parameter.
8.4.3 Redirects In URL Mappings
Since Grails 2.3, it is possible to define URL mappings which specify a redirect.
When a URL mapping specifies a redirect, any time that mapping matches an incoming
request, a redirect is initiated with information provided by the mapping.
When a URL mapping specifies a redirect the mapping must either supply a String
representing a URI to redirect to or must provide a Map representing the target
of the redirect. That Map is structured just like the Map that may be passed
as an argument to the
redirect
method in a controller.
"/viewBooks"(redirect: '/books/list')
"/viewAuthors"(redirect: [controller: 'author', action: 'list'])
"/viewPublishers"(redirect: [controller: 'publisher', action: 'list', permanent: true])
Request parameters that were part of the original request will be included in the redirect.
8.4.4 Embedded Variables
Simple Variables
The previous section demonstrated how to map simple URLs with concrete "tokens". In URL mapping speak tokens are the sequence of characters between each slash, '/'. A concrete token is one which is well defined such as as
/product
. However, in many circumstances you don't know what the value of a particular token will be until runtime. In this case you can use variable placeholders within the URL for example:
static mappings = {
"/product/$id"(controller: "product")
}
In this case by embedding a $id variable as the second token Grails will automatically map the second token into a parameter (available via the
params object) called
id
. For example given the URL
/product/MacBook
, the following code will render "MacBook" to the response:
class ProductController {
def index() { render params.id }
}
You can of course construct more complex examples of mappings. For example the traditional blog URL format could be mapped as follows:
static mappings = {
"/$blog/$year/$month/$day/$id"(controller: "blog", action: "show")
}
The above mapping would let you do things like:
/graemerocher/2007/01/10/my_funky_blog_entry
The individual tokens in the URL would again be mapped into the
params object with values available for
year
,
month
,
day
,
id
and so on.
Dynamic Controller and Action Names
Variables can also be used to dynamically construct the controller and action name. In fact the default Grails URL mappings use this technique:
static mappings = {
"/$controller/$action?/$id?"()
}
Here the name of the controller, action and id are implicitly obtained from the variables
controller
,
action
and
id
embedded within the URL.
You can also resolve the controller name and action name to execute dynamically using a closure:
static mappings = {
"/$controller" {
action = { params.goHere }
}
}
Optional Variables
Another characteristic of the default mapping is the ability to append a ? at the end of a variable to make it an optional token. In a further example this technique could be applied to the blog URL mapping to have more flexible linking:
static mappings = {
"/$blog/$year?/$month?/$day?/$id?"(controller:"blog", action:"show")
}
With this mapping all of these URLs would match with only the relevant parameters being populated in the
params object:
/graemerocher/2007/01/10/my_funky_blog_entry
/graemerocher/2007/01/10
/graemerocher/2007/01
/graemerocher/2007
/graemerocher
Optional File Extensions
If you wish to capture the extension of a particular path, then a special case mapping exists:
"/$controller/$action?/$id?(.$format)?"()
By adding the
(.$format)?
mapping you can access the file extension using the
response.format
property in a controller:
def index() {
render "extension is ${response.format}"
}
Arbitrary Variables
You can also pass arbitrary parameters from the URL mapping into the controller by just setting them in the block passed to the mapping:
"/holiday/win" {
id = "Marrakech"
year = 2007
}
This variables will be available within the
params object passed to the controller.
Dynamically Resolved Variables
The hard coded arbitrary variables are useful, but sometimes you need to calculate the name of the variable based on runtime factors. This is also possible by assigning a block to the variable name:
"/holiday/win" {
id = { params.id }
isEligible = { session.user != null } // must be logged in
}
In the above case the code within the blocks is resolved when the URL is actually matched and hence can be used in combination with all sorts of logic.
8.4.5 Mapping to Views
You can resolve a URL to a view without a controller or action involved. For example to map the root URL
/
to a GSP at the location
grails-app/views/index.gsp
you could use:
static mappings = {
"/"(view: "/index") // map the root URL
}
Alternatively if you need a view that is specific to a given controller you could use:
static mappings = {
"/help"(controller: "site", view: "help") // to a view for a controller
}
8.4.6 Mapping to Response Codes
Grails also lets you map HTTP response codes to controllers, actions or views. Just use a method name that matches the response code you are interested in:
static mappings = {
"403"(controller: "errors", action: "forbidden")
"404"(controller: "errors", action: "notFound")
"500"(controller: "errors", action: "serverError")
}
Or you can specify custom error pages:
static mappings = {
"403"(view: "/errors/forbidden")
"404"(view: "/errors/notFound")
"500"(view: "/errors/serverError")
}
Declarative Error Handling
In addition you can configure handlers for individual exceptions:
static mappings = {
"403"(view: "/errors/forbidden")
"404"(view: "/errors/notFound")
"500"(controller: "errors", action: "illegalArgument",
exception: IllegalArgumentException)
"500"(controller: "errors", action: "nullPointer",
exception: NullPointerException)
"500"(controller: "errors", action: "customException",
exception: MyException)
"500"(view: "/errors/serverError")
}
With this configuration, an
IllegalArgumentException
will be handled by the
illegalArgument
action in
ErrorsController
, a
NullPointerException
will be handled by the
nullPointer
action, and a
MyException
will be handled by the
customException
action. Other exceptions will be handled by the catch-all rule and use the
/errors/serverError
view.
You can access the exception from your custom error handing view or controller action using the request's
exception
attribute like so:
class ErrorController {
def handleError() {
def exception = request.exception
// perform desired processing to handle the exception
}
}
If your error-handling controller action throws an exception as well, you'll end up with a StackOverflowException
.
8.4.7 Mapping to HTTP methods
URL mappings can also be configured to map based on the HTTP method (GET, POST, PUT or DELETE). This is very useful for RESTful APIs and for restricting mappings based on HTTP method.
As an example the following mappings provide a RESTful API URL mappings for the
ProductController
:
static mappings = {
"/product/$id"(controller:"product", action: "update", method: "PUT")
}
8.4.8 Mapping Wildcards
Grails' URL mappings mechanism also supports wildcard mappings. For example consider the following mapping:
static mappings = {
"/images/*.jpg"(controller: "image")
}
This mapping will match all paths to images such as
/image/logo.jpg
. Of course you can achieve the same effect with a variable:
static mappings = {
"/images/$name.jpg"(controller: "image")
}
However, you can also use double wildcards to match more than one level below:
static mappings = {
"/images/**.jpg"(controller: "image")
}
In this cases the mapping will match
/image/logo.jpg
as well as
/image/other/logo.jpg
. Even better you can use a double wildcard variable:
static mappings = {
// will match /image/logo.jpg and /image/other/logo.jpg
"/images/$name**.jpg"(controller: "image")
}
In this case it will store the path matched by the wildcard inside a
name
parameter obtainable from the
params object:
def name = params.name
println name // prints "logo" or "other/logo"
If you use wildcard URL mappings then you may want to exclude certain URIs from Grails' URL mapping process. To do this you can provide an
excludes
setting inside the
UrlMappings.groovy
class:
class UrlMappings {
static excludes = ["/images/*", "/css/*"]
static mappings = {
…
}
}
In this case Grails won't attempt to match any URIs that start with
/images
or
/css
.
8.4.9 Automatic Link Re-Writing
Another great feature of URL mappings is that they automatically customize the behaviour of the
link tag so that changing the mappings don't require you to go and change all of your links.
This is done through a URL re-writing technique that reverse engineers the links from the URL mappings. So given a mapping such as the blog one from an earlier section:
static mappings = {
"/$blog/$year?/$month?/$day?/$id?"(controller:"blog", action:"show")
}
If you use the link tag as follows:
<g:link controller="blog" action="show"
params="[blog:'fred', year:2007]">
My Blog
</g:link><g:link controller="blog" action="show"
params="[blog:'fred', year:2007, month:10]">
My Blog - October 2007 Posts
</g:link>
Grails will automatically re-write the URL in the correct format:
<a href="/fred/2007">My Blog</a>
<a href="/fred/2007/10">My Blog - October 2007 Posts</a>
8.4.10 Applying Constraints
URL Mappings also support Grails' unified
validation constraints mechanism, which lets you further "constrain" how a URL is matched. For example, if we revisit the blog sample code from earlier, the mapping currently looks like this:
static mappings = {
"/$blog/$year?/$month?/$day?/$id?"(controller:"blog", action:"show")
}
This allows URLs such as:
/graemerocher/2007/01/10/my_funky_blog_entry
However, it would also allow:
/graemerocher/not_a_year/not_a_month/not_a_day/my_funky_blog_entry
This is problematic as it forces you to do some clever parsing in the controller code. Luckily, URL Mappings can be constrained to further validate the URL tokens:
"/$blog/$year?/$month?/$day?/$id?" {
controller = "blog"
action = "show"
constraints {
year(matches:/\d{4}/)
month(matches:/\d{2}/)
day(matches:/\d{2}/)
}
}
In this case the constraints ensure that the
year
,
month
and
day
parameters match a particular valid pattern thus relieving you of that burden later on.
8.4.11 Named URL Mappings
URL Mappings also support named mappings, that is mappings which have a name associated with them. The name may be used to refer to a specific mapping when links are generated.
The syntax for defining a named mapping is as follows:
static mappings = {
name <mapping name>: <url pattern> {
// …
}
}
For example:
static mappings = {
name personList: "/showPeople" {
controller = 'person'
action = 'list'
}
name accountDetails: "/details/$acctNumber" {
controller = 'product'
action = 'accountDetails'
}
}
The mapping may be referenced in a link tag in a GSP.
<g:link mapping="personList">List People</g:link>
That would result in:
<a href="/showPeople">List People</a>
Parameters may be specified using the params attribute.
<g:link mapping="accountDetails" params="[acctNumber:'8675309']">
Show Account
</g:link>
That would result in:
<a href="/details/8675309">Show Account</a>
Alternatively you may reference a named mapping using the link namespace.
<link:personList>List People</link:personList>
That would result in:
<a href="/showPeople">List People</a>
The link namespace approach allows parameters to be specified as attributes.
<link:accountDetails acctNumber="8675309">Show Account</link:accountDetails>
That would result in:
<a href="/details/8675309">Show Account</a>
To specify attributes that should be applied to the generated
href
, specify a
Map
value to the
attrs
attribute. These attributes will be applied directly to the href, not passed through to be used as request parameters.
<link:accountDetails attrs="[class: 'fancy']" acctNumber="8675309">
Show Account
</link:accountDetails>
That would result in:
<a href="/details/8675309" class="fancy">Show Account</a>
The default URL Mapping mechanism supports camel case names in the URLs. The default URL for accessing an action named
addNumbers
in a controller named
MathHelperController
would be something like
/mathHelper/addNumbers
. Grails allows for the customization of this pattern and provides an implementation which replaces the camel case convention with a hyphenated convention that would support URLs like
/math-helper/add-numbers
. To enable hyphenated URLs assign a value of "hyphenated" to the
grails.web.url.converter
property in
grails-app/conf/Config.groovy
.
// grails-app/conf/Config.groovygrails.web.url.converter = 'hyphenated'
Arbitrary strategies may be plugged in by providing a class which implements the
UrlConverter interface and adding an instance of that class to the Spring application context with the bean name of
grails.web.UrlConverter.BEAN_NAME
. If Grails finds a bean in the context with that name, it will be used as the default converter and there is no need to assign a value to the
grails.web.url.converter
config property.
// src/groovy/com/myapplication/MyUrlConverterImpl.groovypackage com.myapplicationclass MyUrlConverterImpl implements grails.web.UrlConverter { String toUrlElement(String propertyOrClassName) {
// return some representation of a property or class name that should be used in URLs…
}
}
// grails-app/conf/spring/resources.groovybeans = {
"${grails.web.UrlConverter.BEAN_NAME}"(com.myapplication.MyUrlConverterImpl)
}
8.4.13 Namespaced Controllers
If an application defines multiple controllers with the same name
in different packages, the controllers must be defined in a
namespace. The way to define a namespace for a controller is to
define a static property named
namespace
in the controller and
assign a String to the property that represents the namespace.
// grails-app/controllers/com/app/reporting/AdminController.groovy
package com.app.reportingclass AdminController { static namespace = 'reports' // …
}
// grails-app/controllers/com/app/security/AdminController.groovy
package com.app.securityclass AdminController { static namespace = 'users' // …
}
When defining url mappings which should be associated with a namespaced
controller, the
namespace
variable needs to be part of the URL mapping.
// grails-app/conf/UrlMappings.groovy
class UrlMappings { static mappings = {
'/userAdmin' {
controller = 'admin'
namespace = 'users'
} '/reportAdmin' {
controller = 'admin'
namespace = 'reports'
} "/$namespace/$controller/$action?"()
}
}
Reverse URL mappings also require that the
namespace
be specified.
<g:link controller="admin" namespace="reports">Click For Report Admin</g:link>
<g:link controller="admin" namespace="users">Click For User Admin</g:link>
When resolving a URL mapping (forward or reverse) to a namespaced controller,
a mapping will only match if the
namespace
has been provided. If
the application provides several controllers with the same name in different
packages, at most 1 of them may be defined without a
namespace
property. If
there are multiple controllers with the same name that do not define a
namespace
property, the framework will not know how to distinguish between
them for forward or reverse mapping resolutions.
It is allowed for an application to use a plugin which provides a controller
with the same name as a controller provided by the application and for neither
of the controllers to define a
namespace
property as long as the
controllers are in separate packages. For example, an application
may include a controller named
com.accounting.ReportingController
and the application may use a plugin which provides a controller
named
com.humanresources.ReportingController
. The only issue
with that is the URL mapping for the controller provided by the
plugin needs to be explicit in specifying that the mapping applies
to the
ReportingController
which is provided by the plugin.
See the following example.
static mappings = {
"/accountingReports" {
controller = "reporting"
}
"/humanResourceReports" {
controller = "reporting"
plugin = "humanResources"
}
}
With that mapping in place, a request to
/accountingReports
will
be handled by the
ReportingController
which is defined in the
application. A request to
/humanResourceReports
will be handled
by the
ReportingController
which is provided by the
humanResources
plugin.
There could be any number of
ReportingController
controllers provided
by any number of plugins but no plugin may provide more than one
ReportingController
even if they are defined in separate packages.
Assigning a value to the
plugin
variable in the mapping is only
required if there are multiple controllers with the same name
available at runtime provided by the application and/or plugins.
If the
humanResources
plugin provides a
ReportingController
and
there is no other
ReportingController
available at runtime, the
following mapping would work.
static mappings = {
"/humanResourceReports" {
controller = "reporting"
}
}
It is best practice to be explicit about the fact that the controller
is being provided by a plugin.
8.5 Filters
Although Grails
controllers support fine grained interceptors, these are only really useful when applied to a few controllers and become difficult to manage with larger applications. Filters on the other hand can be applied across a whole group of controllers, a URI space or to a specific action. Filters are far easier to plugin and maintain completely separately to your main controller logic and are useful for all sorts of cross cutting concerns such as security, logging, and so on.
8.5.1 Applying Filters
To create a filter create a class that ends with the convention
Filters
in the
grails-app/conf
directory. Within this class define a code block called
filters
that contains the filter definitions:
class ExampleFilters {
def filters = {
// your filters here
}
}
Each filter you define within the
filters
block has a name and a scope. The name is the method name and the scope is defined using named arguments. For example to define a filter that applies to all controllers and all actions you can use wildcards:
sampleFilter(controller:'*', action:'*') {
// interceptor definitions
}
The scope of the filter can be one of the following things:
- A controller and/or action name pairing with optional wildcards
- A URI, with Ant path matching syntax
Filter rule attributes:
controller
- controller matching pattern, by default * is replaced with .* and a regex is compiled
controllerExclude
- controller exclusion pattern, by default * is replaced with .* and a regex is compiled
action
- action matching pattern, by default * is replaced with .* and a regex is compiled
actionExclude
- action exclusion pattern, by default * is replaced with .* and a regex is compiled
regex
(true
/false
) - use regex syntax (don't replace '*' with '.*')
uri
- a uri to match, expressed with as Ant style path (e.g. /book/**)
uriExclude
- a uri pattern to exclude, expressed with as Ant style path (e.g. /book/**)
find
(true
/false
) - rule matches with partial match (see java.util.regex.Matcher.find()
)
invert
(true
/false
) - invert the rule (NOT rule)
Some examples of filters include:
- All controllers and actions
all(controller: '*', action: '*') {}
- Only for the
BookController
justBook(controller: 'book', action: '*') {}
- All controllers except the
BookController
notBook(controller: 'book', invert: true) {}
- All actions containing 'save' in the action name
saveInActionName(action: '*save*', find: true) {}
- All actions starting with the letter 'b' except for actions beginning with the phrase 'bad*'
actionBeginningWithBButNotBad(action: 'b*', actionExclude: 'bad*', find: true) {}
someURIs(uri: '/book/**') {}
In addition, the order in which you define the filters within the
filters
code block dictates the order in which they are executed. To control the order of execution between
Filters
classes, you can use the
dependsOn
property discussed in
filter dependencies section.
Note: When exclude patterns are used they take precedence over the matching patterns. For example, if action is 'b*' and actionExclude is 'bad*' then actions like 'best' and 'bien' will have that filter applied but actions like 'bad' and 'badlands' will not.
8.5.2 Filter Types
Within the body of the filter you can then define one or several of the following interceptor types for the filter:
before
- Executed before the action. Return false
to indicate that the response has been handled that that all future filters and the action should not execute
after
- Executed after an action. Takes a first argument as the view model to allow modification of the model before rendering the view
afterView
- Executed after view rendering. Takes an Exception as an argument which will be non-null
if an exception occurs during processing. Note: this Closure is called before the layout is applied.
For example to fulfill the common simplistic authentication use case you could define a filter as follows:
class SecurityFilters {
def filters = {
loginCheck(controller: '*', action: '*') {
before = {
if (!session.user && !actionName.equals('login')) {
redirect(action: 'login')
return false
}
}
}
}
}
Here the
loginCheck
filter uses a
before
interceptor to execute a block of code that checks if a user is in the session and if not redirects to the login action. Note how returning false ensure that the action itself is not executed.
Here's a more involved example that demonstrates all three filter types:
import java.util.concurrent.atomic.AtomicLongclass LoggingFilters { private static final AtomicLong REQUEST_NUMBER_COUNTER = new AtomicLong()
private static final String START_TIME_ATTRIBUTE = 'Controller__START_TIME__'
private static final String REQUEST_NUMBER_ATTRIBUTE = 'Controller__REQUEST_NUMBER__' def filters = { logFilter(controller: '*', action: '*') { before = {
if (!log.debugEnabled) return true long start = System.currentTimeMillis()
long currentRequestNumber = REQUEST_NUMBER_COUNTER.incrementAndGet() request[START_TIME_ATTRIBUTE] = start
request[REQUEST_NUMBER_ATTRIBUTE] = currentRequestNumber log.debug "preHandle request #$currentRequestNumber : " +
"'$request.servletPath'/'$request.forwardURI', " +
"from $request.remoteHost ($request.remoteAddr) " +
" at ${new Date()}, Ajax: $request.xhr, controller: $controllerName, " +
"action: $actionName, params: ${new TreeMap(params)}" return true
} after = { Map model -> if (!log.debugEnabled) return true long start = request[START_TIME_ATTRIBUTE]
long end = System.currentTimeMillis()
long requestNumber = request[REQUEST_NUMBER_ATTRIBUTE] def msg = "postHandle request #$requestNumber: end ${new Date()}, " +
"controller total time ${end - start}ms"
if (log.traceEnabled) {
log.trace msg + "; model: $model"
}
else {
log.debug msg
}
} afterView = { Exception e -> if (!log.debugEnabled) return true long start = request[START_TIME_ATTRIBUTE]
long end = System.currentTimeMillis()
long requestNumber = request[REQUEST_NUMBER_ATTRIBUTE] def msg = "afterCompletion request #$requestNumber: " +
"end ${new Date()}, total time ${end - start}ms"
if (e) {
log.debug "$msg \n\texception: $e.message", e
}
else {
log.debug msg
}
}
}
}
}
In this logging example we just log various request information, but note that the
model
map in the
after
filter is mutable. If you need to add or remove items from the model map you can do that in the
after
filter.
8.5.3 Variables and Scopes
Filters support all the common properties available to
controllers and
tag libraries, plus the application context:
However, filters only support a subset of the methods available to controllers and tag libraries. These include:
- redirect - For redirects to other controllers and actions
- render - For rendering custom responses
8.5.4 Filter Dependencies
In a
Filters
class, you can specify any other
Filters
classes that should first be executed using the
dependsOn
property. This is used when a
Filters
class depends on the behavior of another
Filters
class (e.g. setting up the environment, modifying the request/session, etc.) and is defined as an array of
Filters
classes.
Take the following example
Filters
classes:
class MyFilters {
def dependsOn = [MyOtherFilters] def filters = {
checkAwesome(uri: "/*") {
before = {
if (request.isAwesome) { // do something awesome }
}
} checkAwesome2(uri: "/*") {
before = {
if (request.isAwesome) { // do something else awesome }
}
}
}
}
class MyOtherFilters {
def filters = {
makeAwesome(uri: "/*") {
before = {
request.isAwesome = true
}
}
doNothing(uri: "/*") {
before = {
// do nothing
}
}
}
}
MyFilters specifically
dependsOn
MyOtherFilters. This will cause all the filters in MyOtherFilters whose scope matches the current request to be executed before those in MyFilters. For a request of "/test", which will match the scope of every filter in the example, the execution order would be as follows:
- MyOtherFilters - makeAwesome
- MyOtherFilters - doNothing
- MyFilters - checkAwesome
- MyFilters - checkAwesome2
The filters within the MyOtherFilters class are processed in order first, followed by the filters in the MyFilters class. Execution order between
Filters
classes are enabled and the execution order of filters within each
Filters
class are preserved.
If any cyclical dependencies are detected, the filters with cyclical dependencies will be added to the end of the filter chain and processing will continue. Information about any cyclical dependencies that are detected will be written to the logs. Ensure that your root logging level is set to at least WARN or configure an appender for the Grails Filters Plugin (
org.codehaus.groovy.grails.plugins.web.filters.FiltersGrailsPlugin
) when debugging filter dependency issues.
8.6 Ajax
Ajax is the driving force behind the shift to richer web applications. These types of applications in general are better suited to agile, dynamic frameworks written in languages like
Groovy and
Ruby Grails provides support for building Ajax applications through its Ajax tag library. For a full list of these see the Tag Library Reference.
Note: JavaScript examples use the jQuery library.
8.6.1 Ajax Support
By default Grails ships with the
jQuery library, but through the
Plugin system provides support for other frameworks such as
Prototype,
,
and the
Google Web Toolkit.
This section covers Grails' support for Ajax in general. To get started, add this line to the
<head>
tag of your page:
<g:javascript library="jquery" />
You can replace
jQuery
with any other library supplied by a plugin you have installed. This works because of Grails' support for adaptive tag libraries. Thanks to Grails' plugin system there is support for a number of different Ajax libraries including (but not limited to):
- jQuery
- Prototype
- Dojo
- YUI
- MooTools
8.6.1.1 Remoting Linking
Remote content can be loaded in a number of ways, the most commons way is through the
remoteLink tag. This tag allows the creation of HTML anchor tags that perform an asynchronous request and optionally set the response in an element. The simplest way to create a remote link is as follows:
<g:remoteLink action="delete" id="1">Delete Book</g:remoteLink>
The above link sends an asynchronous request to the
delete
action of the current controller with an id of
1
.
8.6.1.2 Updating Content
This is great, but usually you provide feedback to the user about what happened:
def delete() {
def b = Book.get(params.id)
b.delete()
render "Book ${b.id} was deleted"
}
GSP code:
<div id="message"></div>
<g:remoteLink action="delete" id="1" update="message">
Delete Book
</g:remoteLink>
The above example will call the action and set the contents of the
message
div
to the response in this case
"Book 1 was deleted"
. This is done by the
update
attribute on the tag, which can also take a Map to indicate what should be updated on failure:
<div id="message"></div>
<div id="error"></div>
<g:remoteLink update="[success: 'message', failure: 'error']"
action="delete" id="1">
Delete Book
</g:remoteLink>
Here the
error
div will be updated if the request failed.
An HTML form can also be submitted asynchronously in one of two ways. Firstly using the
formRemote tag which expects similar attributes to those for the
remoteLink tag:
<g:formRemote url="[controller: 'book', action: 'delete']"
update="[success: 'message', failure: 'error']">
<input type="hidden" name="id" value="1" />
<input type="submit" value="Delete Book!" />
</g:formRemote >
Or alternatively you can use the
submitToRemote tag to create a submit button. This allows some buttons to submit remotely and some not depending on the action:
<form action="delete">
<input type="hidden" name="id" value="1" />
<g:submitToRemote action="delete"
update="[success: 'message', failure: 'error']" />
</form>
8.6.1.4 Ajax Events
Specific JavaScript can be called if certain events occur, all the events start with the "on" prefix and let you give feedback to the user where appropriate, or take other action:
<g:remoteLink action="show"
id="1"
update="success"
onLoading="showProgress()"
onComplete="hideProgress()">Show Book 1</g:remoteLink>
The above code will execute the "showProgress()" function which may show a progress bar or whatever is appropriate. Other events include:
onSuccess
- The JavaScript function to call if successful
onFailure
- The JavaScript function to call if the call failed
onERROR_CODE
- The JavaScript function to call to handle specified error codes (e.g. on404="alert('not found!')"
)
onUninitialized
- The JavaScript function to call the a Ajax engine failed to initialise
onLoading
- The JavaScript function to call when the remote function is loading the response
onLoaded
- The JavaScript function to call when the remote function is completed loading the response
onComplete
- The JavaScript function to call when the remote function is complete, including any updates
You can simply refer to the
XMLHttpRequest
variable to obtain the request:
<g:javascript>
function fireMe(event) {
alert("XmlHttpRequest = " + event)
}
}
</g:javascript>
<g:remoteLink action="example"
update="success"
onFailure="fireMe(XMLHttpRequest)">Ajax Link</g:remoteLink>
8.6.2 Ajax with Prototype
Grails features an external plugin to add
Prototype support to Grails. To install the plugin, list it in BuildConfig.groovy:
runtime ":prototype:latest.release"
This will download the current supported version of the Prototype plugin and install it into your Grails project. With that done you can add the following reference to the top of your page:
<g:javascript library="prototype" />
If you require
Scriptaculous too you can do the following instead:
<g:javascript library="scriptaculous" />
Now all of Grails tags such as
remoteLink,
formRemote and
submitToRemote work with Prototype remoting.
8.6.3 Ajax with Dojo
Grails features an external plugin to add
Dojo support to Grails. To install the plugin, list it in BuildConfig.groovy:
compile ":dojo:latest.release"
This will download the current supported version of Dojo and install it into your Grails project. With that done you can add the following reference to the top of your page:
<g:javascript library="dojo" />
Now all of Grails tags such as
remoteLink,
formRemote and
submitToRemote work with Dojo remoting.
8.6.4 Ajax with GWT
Grails also features support for the
Google Web Toolkit through a plugin. There is comprehensive
documentation available on the Grails wiki.
8.6.5 Ajax on the Server
There are a number of different ways to implement Ajax which are typically broken down into:
- Content Centric Ajax - Where you just use the HTML result of a remote call to update the page
- Data Centric Ajax - Where you actually send an XML or JSON response from the server and programmatically update the page
- Script Centric Ajax - Where the server sends down a stream of JavaScript to be evaluated on the fly
Most of the examples in the
Ajax section cover Content Centric Ajax where you are updating the page, but you may also want to use Data Centric or Script Centric. This guide covers the different styles of Ajax.
Content Centric Ajax
Just to re-cap, content centric Ajax involves sending some HTML back from the server and is typically done by rendering a template with the
render method:
def showBook() {
def b = Book.get(params.id) render(template: "bookTemplate", model: [book: b])
}
Calling this on the client involves using the
remoteLink tag:
<g:remoteLink action="showBook" id="${book.id}"
update="book${book.id}">Update Book</g:remoteLink><div id="book${book.id}">
</div>
Data Centric Ajax with JSON
Data Centric Ajax typically involves evaluating the response on the client and updating programmatically. For a JSON response with Grails you would typically use Grails'
JSON marshalling capability:
import grails.converters.JSONdef showBook() {
def b = Book.get(params.id) render b as JSON
}
And then on the client parse the incoming JSON request using an Ajax event handler:
<g:javascript>
function updateBook(data) {
$("#book" + data.id + "_title").html( data.title );
}
</g:javascript>
<g:remoteLink action="showBook" id="${book.id}" onSuccess="updateBook(data)">
Update Book
</g:remoteLink>
<g:set var="bookId">book${book.id}</g:set>
<div id="${bookId}">
<div id="${bookId}_title">The Stand</div>
</div>
Data Centric Ajax with XML
On the server side using XML is equally simple:
import grails.converters.XMLdef showBook() {
def b = Book.get(params.id) render b as XML
}
However, since DOM is involved the client gets more complicated:
<g:javascript>
function updateBook(data) {
var id = $(data).find("book").attr("id");
$("#book" + id + "_title").html( $(data).find("title").text() );
}
</g:javascript>
<g:remoteLink action="showBook" id="${book.id}" onSuccess="updateBook(data)">
Update Book
</g:remoteLink>
<g:set var="bookId">book${book.id}</g:set>
<div id="${bookId}">
<div id="${bookId}_title">The Stand</div>
</div>
Script Centric Ajax with JavaScript
Script centric Ajax involves actually sending JavaScript back that gets evaluated on the client. An example of this can be seen below:
def showBook() {
def b = Book.get(params.id) response.contentType = "text/javascript"
String title = b.title.encodeAsJavaScript()
render "$('#book${b.id}_title').html('${title}');"
}
The important thing to remember is to set the
contentType
to
text/javascript
. If you use Prototype on the client the returned JavaScript will automatically be evaluated due to this
contentType
setting.
Obviously in this case it is critical that you have an agreed client-side API as you don't want changes on the client breaking the server. This is one of the reasons Rails has something like RJS. Although Grails does not currently have a feature such as RJS there is a
Dynamic JavaScript Plugin that offers similar capabilities.
Responding to both Ajax and non-Ajax requests
It's straightforward to have the same Grails controller action handle both Ajax and non-Ajax requests. Grails adds the
isXhr()
method to
HttpServletRequest
which can be used to identify Ajax requests. For example you could render a page fragment using a template for Ajax requests or the full page for regular HTTP requests:
def listBooks() {
def books = Book.list(params)
if (request.xhr) {
render template: "bookTable", model: [books: books]
} else {
render view: "list", model: [books: books]
}
}
8.7 Content Negotiation
Grails has built in support for
Content negotiation using either the HTTP
Accept
header, an explicit format request parameter or the extension of a mapped URI.
Configuring Mime Types
Before you can start dealing with content negotiation you need to tell Grails what content types you wish to support. By default Grails comes configured with a number of different content types within
grails-app/conf/Config.groovy
using the
grails.mime.types
setting:
grails.mime.types = [ // the first one is the default format
all: '*/*', // 'all' maps to '*' or the first available format in withFormat
atom: 'application/atom+xml',
css: 'text/css',
csv: 'text/csv',
form: 'application/x-www-form-urlencoded',
html: ['text/html','application/xhtml+xml'],
js: 'text/javascript',
json: ['application/json', 'text/json'],
multipartForm: 'multipart/form-data',
rss: 'application/rss+xml',
text: 'text/plain',
hal: ['application/hal+json','application/hal+xml'],
xml: ['text/xml', 'application/xml']
]
The above bit of configuration allows Grails to detect to format of a request containing either the 'text/xml' or 'application/xml' media types as simply 'xml'. You can add your own types by simply adding new entries into the map.
The first one is the default format.
Content Negotiation using the format parameter
Let's say a controller action can return a resource in a variety of formats: HTML, XML, and JSON. What format will the client get? The easiest and most reliable way for the client to control this is through a
format
URL parameter.
So if you, as a browser or some other client, want a resource as XML, you can use a URL like this:
http://my.domain.org/books?format=xml
The result of this on the server side is a
format
property on the
response
object with the value
xml
. You could code your controller action to return XML based on this property, but you can also make use of the controller-specific
withFormat()
method:
import grails.converters.JSON
import grails.converters.XMLclass BookController { def list() {
def books = Book.list() withFormat {
html bookList: books
json { render books as JSON }
xml { render books as XML }
'*' { render books as JSON }
}
}
}
In this example, Grails will only execute the block inside
withFormat()
that matches the requested content type. So if the preferred format is
html
then Grails will execute the
html()
call only. Each 'block' can either be a map model for the corresponding view (as we are doing for 'html' in the above example) or a closure. The closure can contain any standard action code, for example it can return a model or render content directly.
When no format matches explicitly, a
(wildcard) block can be used to handle all other formats.
There is a special format, "all", that is handled differently from the explicit formats. If "all" is specified (normally this happens through the Accept header - see below), then the first block of
withFormat()
is executed when there isn't a
(wildcard) block available.
You should not add an explicit "all" block. In this example, a format of "all" will trigger the
html
handler (
html
is the first block and there is no
*
block).
withFormat {
html bookList: books
json { render books as JSON }
xml { render books as XML }
}
When using withFormat make sure it is the last call in your controller action as the return value of the withFormat
method is used by the action to dictate what happens next.
Using the Accept header
Every incoming HTTP request has a special
Accept header that defines what media types (or mime types) a client can "accept". In older browsers this is typically:
which simply means anything. However, newer browsers send more interesting values such as this one sent by Firefox:
text/xml, application/xml, application/xhtml+xml, text/html;q=0.9,
text/plain;q=0.8, image/png, */*;q=0.5
This particular accept header is unhelpful because it indicates that XML is the preferred response format whereas the user is really expecting HTML. That's why Grails ignores the accept header by default for browsers. However, non-browser clients are typically more specific in their requirements and can send accept headers such as
As mentioned the default configuration in Grails is to ignore the accept header for browsers. This is done by the configuration setting
grails.mime.disable.accept.header.userAgents
, which is configured to detect the major rendering engines and ignore their ACCEPT headers. This allows Grails' content negotiation to continue to work for non-browser clients:
grails.mime.disable.accept.header.userAgents = ['Gecko', 'WebKit', 'Presto', 'Trident']
For example, if it sees the accept header above ('application/json') it will set
format
to
json
as you'd expect. And of course this works with the
withFormat()
method in just the same way as when the
format
URL parameter is set (although the URL parameter takes precedence).
An accept header of '*/*' results in a value of
all
for the
format
property.
If the accept header is used but contains no registered content types, Grails will assume a broken browser is making the request and will set the HTML format - note that this is different from how the other content negotiation modes work as those would activate the "all" format!
Request format vs. Response format
As of Grails 2.0, there is a separate notion of the
request format and the
response format. The request format is dictated by the
CONTENT_TYPE
header and is typically used to detect if the incoming request can be parsed into XML or JSON, whilst the response format uses the file extension, format parameter or ACCEPT header to attempt to deliver an appropriate response to the client.
The
withFormat available on controllers deals specifically with the response format. If you wish to add logic that deals with the request format then you can do so using a separate
withFormat
method available on the request:
request.withFormat {
xml {
// read XML
}
json {
// read JSON
}
}
Content Negotiation with the format Request Parameter
If fiddling with request headers if not your favorite activity you can override the format used by specifying a
format
request parameter:
You can also define this parameter in the
URL Mappings definition:
"/book/list"(controller:"book", action:"list") {
format = "xml"
}
Content Negotiation with URI Extensions
Grails also supports content negotiation using URI extensions. For example given the following URI:
This works as a result of the default URL Mapping definition which is:
"/$controller/$action?/$id?(.$format)?"{
Note the inclusion of the
format
variable in the path. If you do not wish to use content negotiation via the file extension then simply remove this part of the URL mapping:
"/$controller/$action?/$id?"{
Testing Content Negotiation
To test content negotiation in a unit or integration test (see the section on
Testing) you can either manipulate the incoming request headers:
void testJavascriptOutput() {
def controller = new TestController()
controller.request.addHeader "Accept",
"text/javascript, text/html, application/xml, text/xml, */*" controller.testAction()
assertEquals "alert('hello')", controller.response.contentAsString
}
Or you can set the format parameter to achieve a similar effect:
void testJavascriptOutput() {
def controller = new TestController()
controller.params.format = 'js' controller.testAction()
assertEquals "alert('hello')", controller.response.contentAsString
}
9 Web Services
Web Services are all about providing a web API onto your web application and are typically implemented in either
REST or
SOAP
9.1 REST
REST is not really a technology in itself, but more an architectural pattern. REST is very simple and just involves using plain XML or JSON as a communication medium, combined with URL patterns that are "representational" of the underlying system, and HTTP methods such as GET, PUT, POST and DELETE.
Each HTTP method maps to an action type. For example GET for retrieving data, POST for creating data, PUT for updating and so on.
Grails includes flexible features that make it easy to create RESTful APIs. Creating a RESTful resource can be as simple as one line of code, as demonstrated in the next section.
9.1.1 Domain classes as REST resources
The easiest way to create a RESTful API in Grails is to expose a domain class as a REST resource. This can be done by adding the
grails.rest.Resource
transformation to any domain class:
import grails.rest.*@Resource(uri='/books')
class Book { String title static constraints = {
title blank:false
}
}
Simply by adding the
Resource
transformation and specifying a URI, your domain class will automatically be available as a REST resource in either XML or JSON formats. The transformation will automatically register the necessary
RESTful URL mapping and create a controller called
BookController
.
You can try it out by adding some test data to
BootStrap.groovy
:
def init = { servletContext -> new Book(title:"The Stand").save()
new Book(title:"The Shining").save()
}
And then hitting the URL http://localhost:8080/myapp/books/1, which will render the response like:
<?xml version="1.0" encoding="UTF-8"?>
<book id="1">
<title>The Stand</title>
</book>
If you change the URL to
http://localhost:8080/myapp/books/1.json
you will get a JSON response such as:
{"id":1,"title":"The Stand"}
If you wish to change the default to return JSON instead of XML, you can do this by setting the
formats
attribute of the
Resource
transformation:
import grails.rest.*@Resource(uri='/books', formats=['json', 'xml'])
class Book {
…
}
With the above example JSON will be prioritized. The list that is passed should contain the names of the formats that the resource should expose. The names of formats are defined in the
grails.mime.types
setting of
Config.groovy
:
grails.mime.types = [
…
json: ['application/json', 'text/json'],
…
xml: ['text/xml', 'application/xml']
]
See the section on
Configuring Mime Types in the user guide for more information.
Instead of using the file extension in the URI, you can also obtain a JSON response using the ACCEPT header. Here's an example using the Unix
curl
tool:
$ curl -i -H "Accept: application/json" localhost:8080/myapp/books/1
{"id":1,"title":"The Stand"}
This works thanks to Grails'
Content Negotiation features.
You can create a new resource by issuing a
POST
request:
$ curl -i -X POST -H "Content-Type: application/json" -d '{"title":"Along Came A Spider"}' localhost:8080/myapp/books
HTTP/1.1 201 Created
Server: Apache-Coyote/1.1
...
Updating can be done with a
PUT
request:
$ curl -i -X PUT -H "Content-Type: application/json" -d '{"title":"Along Came A Spider"}' localhost:8080/myapp/books/1
HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
...
Finally a resource can be deleted with
DELETE
request:
$ curl -i -X DELETE localhost:8080/myapp/books/1
HTTP/1.1 204 No Content
Server: Apache-Coyote/1.1
...
As you can see, the
Resource
transformation enables all of the HTTP method verbs on the resource. You can enable only read-only capabilities by setting the
readOnly
attribute to true:
import grails.rest.*@Resource(uri='/books', readOnly=true)
class Book {
…
}
In this case POST, PUT and DELETE requests will be forbidden.
9.1.2 Mapping to REST resources
If you prefer to keep the declaration of the URL mapping in your
UrlMappings.groovy
file then simply removing the
uri
attribute of the
Resource
transformation and adding the following line to
UrlMappings.groovy
will suffice:
"/books"(resources:"book")
Extending your API to include more end points then becomes trivial:
"/books"(resources:"book") {
"/publisher"(controller:"publisher", method:"GET")
}
The above example will expose the URI
/books/1/publisher
.
A more detailed explanation on
creating RESTful URL mappings can be found in the
URL Mappings section of the user guide.
9.1.3 Linking to REST resources
The
link
tag offers an easy way to link to any domain class resource:
<g:link resource="${book}">My Link</g:link>
However, currently you cannot use g:link to link to the DELETE action and most browsers do not support sending the DELETE method directly.
The best way to accomplish this is to use a form submit:
<form action="/book/2" method="post">
<input type="hidden" name="_method" value="DELETE"/>
</form>
Grails supports overriding the request method via the hidden _method parameter. This is for browser compatibility purposes. This is useful when using restful resource mappings to create powerful web interfaces.
To make a link fire this type of event, perhaps capture all click events for links with a `data-method` attribute and issue a form submit via javascript.
9.1.4 Versioning REST resources
A common requirement with a REST API is to expose different versions at the same time. There are a few ways this can be achieved in Grails.
Versioning using the URI
A common approach is to use the URI to version APIs (although this approach is discouraged in favour of Hypermedia). For example, you can define the following URL mappings:
"/books/v1"(resources:"book", namespace:'v1')
"/books/v2"(resources:"book", namespace:'v2')
That will match the following controllers:
package myapp.v1class BookController {
static namespace = 'v1'
}package myapp.v2class BookController {
static namespace = 'v2'
}
This approach has the disadvantage of requiring two different URI namespaces for your API.
Versioning with the Accept-Version header
As an alternative Grails supports the passing of an
Accept-Version
header from clients. For example you can define the following URL mappings:
"/books"(version:'1.0', resources:"book", namespace:'v1')
"/books"(version:'2.0', resources:"book", namespace:'v2')
Then in the client simply pass which version you need using the
Accept-Version
header:
$ curl -i -H "Accept-Version: 1.0" -X GET http://localhost:8080/myapp/books
Versioning using Hypermedia / Mime Types
Another approach to versioning is to use Mime Type definitions to declare the version of your custom media types (see the section on "Hypermedia as the Engine of Application State" for more information about Hypermedia concepts). For example, in
Config.groovy
you can declare a custom Mime Type for your resource that includes a version parameter (the 'v' parameter):
grails.mime.types = [
all: '*/*',
book: "application/vnd.books.org.book+json;v=1.0",
bookv2: "application/vnd.books.org.book+json;v=2.0",
…
}
It is critical that place your new mime types after the 'all' Mime Type because if the Content Type of the request cannot be established then the first entry in the map is used for the response. If you have your new Mime Type at the top then Grails will always try and send back your new Mime Type if the requested Mime Type cannot be established.
Then override the renderer (see the section on "Customizing Response Rendering" for more information on custom renderers) to send back the custom Mime Type in
grails-app/conf/spring/resourses.groovy
:
import grails.rest.render.json.*
import org.codehaus.groovy.grails.web.mime.*beans = {
bookRendererV1(JsonRenderer, myapp.v1.Book, new MimeType("application/vnd.books.org.book+json", [v:"1.0"]))
bookRendererV2(JsonRenderer, myapp.v2.Book, new MimeType("application/vnd.books.org.book+json", [v:"2.0"]))
}
Then using the
Accept
header you can specify which version you need using the Mime Type:
$ curl -i -H "Accept: application/vnd.books.org.book+json;v=1.0" -X GET http://localhost:8080/myapp/books
9.1.5 Implementing REST controllers
The
Resource
transformation is a quick way to get started, but typically you'll want to customize the controller logic, the rendering of the response or extend the API to include additional actions.
9.1.5.1 Extending the RestfulController super class
The easiest way to get started doing so is to create a new controller for your resource that extends the
grails.rest.RestfulController
super class. For example:
class BookController extends RestfulController {
static responseFormats = ['json', 'xml']
BookController() {
super(Book)
}
}
To customize any logic you can just override the appropriate action. The following table provides the names of the action names and the URIs they map to:
HTTP Method | URI | Controller Action |
---|
GET | /books | index |
GET | /books/create | create |
POST | /books | save |
GET | /books/${id} | show |
GET | /books/${id}/edit | edit |
PUT | /books/${id} | update |
DELETE | /books/${id} | delete |
Note that the create
and edit
actions are only needed if the controller exposes an HTML interface.
As an example, if you have a
nested resource then you would typically want to query both the parent and the child identifiers. For example, given the following URL mapping:
"/authors"(resources:'author') {
"/books"(resources:'book')
}
You could implement the nested controller as follows:
class BookController extends RestfulController {
static responseFormats = ['json', 'xml']
BookController() {
super(Book)
} @Override
protected Book queryForResource(Serializable id) {
Book.where {
id == id && author.id = params.authorId
}.find()
}}
The example above subclasses
RestfulController
and overrides the protected
queryForResource
method to customize the query for the resource to take into account the parent resource.
Customizing Data Binding In A RestfulController Subclass
The RestfulController class contains code which does data binding for actions like
save
and
update
. The class defines a
getObjectToBind()
method which returns a value which will be used as the source for data binding. For example, the update action does something like this...
class RestfulController<T> { def update() {
T instance = // retrieve instance from the database... instance.properties = getObjectToBind() // …
} // …
}
By default the
getObjectToBind()
method returns the
request object. When the
request
object is used as the binding source, if the request has a body then the body will be parsed and its contents will be used to do the data binding, otherwise the request parameters will be used to do the data binding. Subclasses of RestfulController may override the
getObjectToBind()
method and return anything that is a valid binding source, including a
Map or a
DataBindingSource. For most use cases binding the request is appropriate but the
getObjectToBind()
method allows for changing that behavior where desired.
Using custom subclass of RestfulController with Resource annotation
You can also customize the behaviour of the controller that backs the Resource annotation.
The class must provide a constructor that takes a domain class as it's argument. The second constructor is required for supporting Resource annotation with readOnly=true.
This is a template that can be used for subclassed RestfulController classes used in Resource annotations:
class SubclassRestfulController<T> extends RestfulController<T> {
SubclassRestfulController(Class<T> domainClass) {
this(domainClass, false)
} SubclassRestfulController(Class<T> domainClass, boolean readOnly) {
super(domainClass, readOnly)
}
}
You can specify the super class of the controller that backs the Resource annotation with the
superClass
attribute.
import grails.rest.*@Resource(uri='/books', superClass=SubclassRestfulController)
class Book { String title static constraints = {
title blank:false
}
}
If you don't want to take advantage of the features provided by the
RestfulController
super class, then you can implement each HTTP verb yourself manually. The first step is to create a controller:
$ grails create-controller book
Then add some useful imports and enable readOnly by default:
import grails.transaction.*
import static org.springframework.http.HttpStatus.*
import static org.springframework.http.HttpMethod.*@Transactional(readOnly = true)
class BookController {
…
}
Recall that each HTTP verb matches a particular Grails action according to the following conventions:
HTTP Method | URI | Controller Action |
---|
GET | /books | index |
GET | /books/${id} | show |
GET | /books/create | create |
GET | /books/${id}/edit | edit |
POST | /books | save |
PUT | /books/${id} | update |
DELETE | /books/${id} | delete |
The 'create' and 'edit' actions are already required if you plan to implement an HTML interface for the REST resource. They are there in order to render appropriate HTML forms to create and edit a resource. If this is not a requirement they can be discarded.
The key to implementing REST actions is the
respond method introduced in Grails 2.3. The
respond
method tries to produce the most appropriate response for the requested content type (JSON, XML, HTML etc.)
Implementing the 'index' action
For example, to implement the
index
action, simply call the
respond
method passing the list of objects to respond with:
def index(Integer max) {
params.max = Math.min(max ?: 10, 100)
respond Book.list(params), model:[bookCount: Book.count()]
}
Note that in the above example we also use the
model
argument of the
respond
method to supply the total count. This is only required if you plan to support pagination via some user interface.
The
respond
method will, using
Content Negotiation, attempt to reply with the most appropriate response given the content type requested by the client (via the ACCEPT header or file extension).
If the content type is established to be HTML then a model will be produced such that the action above would be the equivalent of writing:
def index(Integer max) {
params.max = Math.min(max ?: 10, 100)
[bookList: Book.list(params), bookCount: Book.count()]
}
By providing an
index.gsp
file you can render an appropriate view for the given model. If the content type is something other than HTML then the
respond
method will attempt to lookup an appropriate
grails.rest.render.Renderer
instance that is capable of rendering the passed object. This is done by inspecting the
grails.rest.render.RendererRegistry
.
By default there are already renderers configured for JSON and XML, to find out how to register a custom renderer see the section on "Customizing Response Rendering".
Implementing the 'show' action
The
show
action, which is used to display and individual resource by id, can be implemented in one line of Groovy code (excluding the method signature):
def show(Book book) {
respond book
}
By specifying the domain instance as a parameter to the action Grails will automatically attempt to lookup the domain instance using the
id
parameter of the request. If the domain instance doesn't exist, then
null
will be passed into the action. The
respond
method will return a 404 error if null is passed otherwise once again it will attempt to render an appropriate response. If the format is HTML then an appropriate model will produced. The following action is functionally equivalent to the above action:
def show(Book book) {
if(book == null) {
render status:404
}
else {
return [book: book]
}
}
Implementing the 'save' action
The
save
action creates new resource representations. To start off, simply define an action that accepts a resource as the first argument and mark it as
Transactional
with the
grails.transaction.Transactional
transform:
@Transactional
def save(Book book) {
…
}
Then the first thing to do is check whether the resource has any
validation errors and if so respond with the errors:
if(book.hasErrors()) {
respond book.errors, view:'create'
}
else {
…
}
In the case of HTML the 'create' view will be rendered again so the user can correct the invalid input. In the case of other formats (JSON, XML etc.), the errors object itself will be rendered in the appropriate format and a status code of 422 (UNPROCESSABLE_ENTITY) returned.
If there are no errors then the resource can be saved and an appropriate response sent:
book.save flush:true
withFormat {
html {
flash.message = message(code: 'default.created.message', args: [message(code: 'book.label', default: 'Book'), book.id])
redirect book
}
'*' { render status: CREATED }
}
In the case of HTML a redirect is issued to the originating resource and for other formats a status code of 201 (CREATED) is returned.
Implementing the 'update' action
The
update
action updates an existing resource representations and is largely similar to the
save
action. First define the method signature:
@Transactional
def update(Book book) {
…
}
If the resource exists then Grails will load the resource, otherwise null we passed. In the case of null, you should return a 404:
if(book == null) {
render status: NOT_FOUND
}
else {
…
}
Then once again check for errors
validation errors and if so respond with the errors:
if(book.hasErrors()) {
respond book.errors, view:'edit'
}
else {
…
}
In the case of HTML the 'edit' view will be rendered again so the user can correct the invalid input. In the case of other formats (JSON, XML etc.) the errors object itself will be rendered in the appropriate format and a status code of 422 (UNPROCESSABLE_ENTITY) returned.
If there are no errors then the resource can be saved and an appropriate response sent:
book.save flush:true
withFormat {
html {
flash.message = message(code: 'default.updated.message', args: [message(code: 'book.label', default: 'Book'), book.id])
redirect book
}
'*' { render status: OK }
}
In the case of HTML a redirect is issued to the originating resource and for other formats a status code of 200 (OK) is returned.
Implementing the 'delete' action
The
delete
action deletes an existing resource. The implementation is largely similar to the
update
action, expect the
delete()
method is called instead:
book.delete flush:true
withFormat {
html {
flash.message = message(code: 'default.deleted.message', args: [message(code: 'Book.label', default: 'Book'), book.id])
redirect action:"index", method:"GET"
}
'*'{ render status: NO_CONTENT }
}
Notice that for an HTML response a redirect is issued back to the
index
action, whilst for other content types a response code 204 (NO_CONTENT) is returned.
9.1.5.3 Generating a REST controller using scaffolding
To see some of these concepts in action and help you get going the
Scaffolding plugin, version 2.0 and above, can generate a REST ready controller for you, simply run the command:
$ grails generate-controller [Domain Class Name]
9.1.6 Customizing Response Rendering
There are several ways to customize response rendering in Grails.
9.1.6.1 Customizing the Default Renderers
The default renderers for XML and JSON can be found in the
grails.rest.render.xml
and
grails.rest.render.json
packages respectively. These use the Grails converters (
grails.converters.XML
and
grails.converters.JSON
) by default for response rendering.
You can easily customize response rendering using these default renderers. A common change you may want to make is to include or exclude certain properties from rendering.
Including or Excluding Properties from Rendering
As mentioned previously, Grails maintains a registry of
grails.rest.render.Renderer
instances. There are some default configured renderers and the ability to register or override renderers for a given domain class or even for a collection of domain classes. To include a particular property from rendering you need to register a custom renderer by defining a bean in
grails-app/conf/spring/resources.groovy
:
import grails.rest.render.xml.*beans = {
bookRenderer(XmlRenderer, Book) {
includes = ['title']
}
}
The bean name is not important (Grails will scan the application context for all registered renderer beans), but for organizational and readability purposes it is recommended you name it something meaningful.
To exclude a property, the
excludes
property of the
XmlRenderer
class can be used:
import grails.rest.render.xml.*beans = {
bookRenderer(XmlRenderer, Book) {
excludes = ['isbn']
}
}
Customizing the Converters
As mentioned previously, the default renders use the
grails.converters
package under the covers. In other words, under the covers they essentially do the following:
import grails.converters.*…
render book as XML// or render book as JSON
Why the separation between converters and renderers? Well a renderer has more flexibility to use whatever rendering technology you chose. When implementing a custom renderer you could use
Jackson,
Gson or any Java library to implement the renderer. Converters on the other hand are very much tied to Grails' own marshalling implementation.
9.1.6.2 Registering Custom Objects Marshallers
Grails' Converters feature the notion of an
ObjectMarshaller and each type can have a registered
ObjectMarshaller
. You can register custom
ObjectMarshaller
instances to completely customize response rendering. For example, you can define the following in
BootStrap.init
:
XML.registerObjectMarshaller Book, { Book book, XML xml ->
xml.attribute 'id', book.id
xml.build {
title(book.title)
}
}
You can customize the formatting of an individual value this way too. For example the
JodaTime plugin does the following to support rendering of JodaTime dates in JSON output:
JSON.registerObjectMarshaller(DateTime) {
return it?.toString("yyyy-MM-dd'T'HH:mm:ss'Z'")
}
In the case of JSON it's often simple to use a map to customize output:
JSON.registerObjectMarshaller(Book) {
def map= [:]
map['titl'] = it.title
map['auth'] = it.author
return map
}
Registering Custom Marshallers via Spring
Note that if you have many custom marshallers it is recommended you split the registration of these into a separate class:
class CustomMarshallerRegistrar { @javax.annotation.PostConstruct
void registerMarshallers() {
JSON.registerObjectMarshaller(DateTime) {
return it?.toString("yyyy-MM-dd'T'HH:mm:ss'Z'")
}
}
}
Then define this class as Spring bean in
grails-app/conf/spring/resources.groovy
:
beans = {
myCustomMarshallerRegistrar(CustomMarshallerRegistrar)
}
The
PostConstruct
annotation will get triggered on startup of your application.
9.1.6.3 Using Named Configurations for Object Marshallers
It is also possible to register named configurations. For example:
XML.createNamedConfig('publicApi') {
it.registerObjectMarshaller(Book) { Book book, XML xml ->
// do public API
}
}
XML.createNamedConfig('adminApi') {
it.registerObjectMarshaller(Book) { Book book, XML xml ->
// do admin API
}
}
Then when you use either the
render
or
respond
methods you can wrap the call in a named configuration if necessary to customize rendering per request:
XML.use( isAdmin ? 'adminApi' : 'publicApi') {
render book as XML
}
or
XML.use( isAdmin ? 'adminApi' : 'publicApi') {
respond book
}
9.1.6.4 Implementing the ObjectMarshaller Interface
For more complex marshallers it is recommended you implement the
ObjectMarshaller interface. For example given a domain class:
class Book {
String title
}
By default the output when using:
Would look like:
<book id="1">
<title>The Stand</title>
</book>
To write a custom marshaller you can do the following:
class BookMarshaller implements ObjectMarshaller<XML> { public boolean supports(Object object) {
return object instanceof Book
} public void marshalObject(Object object, XML converter) {
Book book = (Book)object
converter.chars book.title
}
}
And then register the marshaller with:
XML.registerObjectMarshaller(new BookMarshaller())
With the custom
ObjectMarshaller
in place, the output is now:
Customizing the Name of the Root Element
If you wish the customize the name of the surrounding element, you can implement
NameAwareMarshaller:
class BookMarshaller implements ObjectMarshaller<XML>,NameAwareMarshaller { ... String getElementName(Object o) {
return 'custom-book'
}}
With the above change the output would now be:
<custom-book>The Stand</custom-book>
Outputting Markup Using the Converters API or Builder
With the passed Converter object you can explicitly code to the Converters API to stream markup to the response:
public void marshalObject(Object object, XML converter) {
Book book = (Book)object converter.attribute 'id', book.id.toString()
converter.attribute 'date-released', book.dateReleased.toString() converter.startNode 'title'
converter.chars book.title
converter.end()}
The above code results in:
<book id="1" date-released="...">
<title>The Stand</title>
</book>
You can also use a builder notation to achieve a similar result (although the builder notation does not work for
CompileStatic
):
public void marshalObject(Object object, XML converter) {
Book b = (Book)object converter.build {
book(id: b.id) {
title b.title
}
}
}
Using the convertAnother Method to Recursively Convert Objects
To create more complex responses you can use the
convertAnother
method to convert associations and other objects:
public void marshalObject(Object object, XML converter) {
Book book = (Book)object converter.startNode 'title'
converter.chars book.title
converter.end() if (book.authors) {
converter.startNode 'authors'
for(author in book.authors) {
converter.convertAnother author
}
converter.end()
}
}
9.1.6.5 Implementing a Custom Renderer
If you want even more control of the rendering or prefer to use your own marshalling techniques then you can implement your own
Renderer
instance. For example below is a simple implementation that customizes the rendering of the
Book
class:
package myapp
import grails.rest.render.*
import org.codehaus.groovy.grails.web.mime.MimeTypeclass BookXmlRenderer extends AbstractRenderer<Book> {
BookXmlRenderer() {
super(Book, [MimeType.XML,MimeType.TEXT_XML] as MimeType[])
} void render(Book object, RenderContext context) {
context.contentType = MimeType.XML.name def xml = new groovy.xml.MarkupBuilder(context.writer)
xml.book(id: object.id, title:object.title)
}
}
The
AbstractRenderer
super class has a constructor that takes the class that it renders and the
MimeType
(s) that are accepted (via the ACCEPT header or file extension) for the renderer.
To configure this renderer, simply add it is a bean to
grails-app/conf/spring/resources.groovy
:
beans = {
bookRenderer(myapp.BookXmlRenderer)
}
The result will be that all
Book
instances will be rendered in the following format:
<book id="1" title="The Stand"/>
Note that if you change the rendering to a completely different format like the above, then you also need to change the binding if you plan to support POST and PUT requests. Grails will not automatically know how to bind data from a custom XML format to a domain class otherwise. See the section on "Customizing Binding of Resources" for further information.
Container Renderers
A
grails.rest.render.ContainerRenderer
is a renderer that renders responses for containers of objects (lists, maps, collections etc.). The interface is largely the same as the
Renderer
interface except for the addition of the
getComponentType()
method, which should return the "contained" type. For example:
class BookListRenderer implements ContainerRenderer<List, Book> {
Class<List> getTargetType() { List }
Class<Book> getComponentType() { Book }
MimeType[] getMimeTypes() { [ MimeType.XML] as MimeType[] }
void render(List object, RenderContext context) {
....
}
}
9.1.6.6 Using GSP to Customize Rendering
You can also customize rendering on a per action basis using Groovy Server Pages (GSP). For example given the
show
action mentioned previously:
def show(Book book) {
respond book
}
You could supply a
show.xml.gsp
file to customize the rendering of the XML:
<%@page contentType="application/xml"%>
<book id="${book.id}" title="${book.title}"/>
HATEOS, an abbreviation for Hypermedia as the Engine of Application State, is a common pattern applied to REST architectures that uses hypermedia and linking to define the REST API.
Hypermedia (also called Mime or Media Types) are used to describe the state of a REST resource, and links tell clients how to transition to the next state. The format of the response is typically JSON or XML, although standard formats such as
Atom and/or
HAL are frequently used.
9.1.7.1 HAL Support
HAL is a standard exchange format commonly used when developing REST APIs that follow HATEOAS principals. An example HAL document representing a list of orders can be seen below:
{
"_links": {
"self": { "href": "/orders" },
"next": { "href": "/orders?page=2" },
"find": {
"href": "/orders{?id}",
"templated": true
},
"admin": [{
"href": "/admins/2",
"title": "Fred"
}, {
"href": "/admins/5",
"title": "Kate"
}]
},
"currentlyProcessing": 14,
"shippedToday": 20,
"_embedded": {
"order": [{
"_links": {
"self": { "href": "/orders/123" },
"basket": { "href": "/baskets/98712" },
"customer": { "href": "/customers/7809" }
},
"total": 30.00,
"currency": "USD",
"status": "shipped"
}, {
"_links": {
"self": { "href": "/orders/124" },
"basket": { "href": "/baskets/97213" },
"customer": { "href": "/customers/12369" }
},
"total": 20.00,
"currency": "USD",
"status": "processing"
}]
}
}
Exposing Resources Using HAL
To return HAL instead of regular JSON for a resource you can simply override the renderer in
grails-app/conf/spring/resources.groovy
with an instance of
grails.rest.render.hal.HalJsonRenderer
(or
HalXmlRenderer
for the XML variation):
import grails.rest.render.hal.*
beans = {
halBookRenderer(HalJsonRenderer, rest.test.Book)
}
With the bean in place requesting the HAL content type will return HAL:
$ curl -i -H "Accept: application/hal+json" http://localhost:8080/myapp/books/1HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Type: application/hal+json;charset=ISO-8859-1{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/1",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": ""The Stand""
}
To use HAL XML format simply change the renderer:
import grails.rest.render.hal.*
beans = {
halBookRenderer(HalXmlRenderer, rest.test.Book)
}
Rendering Collections Using HAL
To return HAL instead of regular JSON for a list of resources you can simply override the renderer in
grails-app/conf/spring/resources.groovy
with an instance of
grails.rest.render.hal.HalJsonCollectionRenderer
:
import grails.rest.render.hal.*
beans = {
halBookCollectionRenderer(HalJsonCollectionRenderer, rest.test.Book)
}
With the bean in place requesting the HAL content type will return HAL:
$ curl -i -H "Accept: application/hal+json" http://localhost:8080/myapp/books
HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Type: application/hal+json;charset=UTF-8
Transfer-Encoding: chunked
Date: Thu, 17 Oct 2013 02:34:14 GMT{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books",
"hreflang": "en",
"type": "application/hal+json"
}
},
"_embedded": {
"book": [
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/1",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "The Stand"
},
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/2",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "Infinite Jest"
},
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/3",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "Walden"
}
]
}
}
Notice that the key associated with the list of
Book
objects in the rendered JSON is
book
which is derived from the type of objects in the collection, namely
Book
. In order to customize the value of this key assign a value to the
collectionName
property on the
HalJsonCollectionRenderer
bean as shown below:
import grails.rest.render.hal.*
beans = {
halBookCollectionRenderer(HalCollectionJsonRenderer, rest.test.Book) {
collectionName = 'publications'
}
}
With that in place the rendered HAL will look like the following:
$ curl -i -H "Accept: application/hal+json" http://localhost:8080/myapp/books
HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Type: application/hal+json;charset=UTF-8
Transfer-Encoding: chunked
Date: Thu, 17 Oct 2013 02:34:14 GMT{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books",
"hreflang": "en",
"type": "application/hal+json"
}
},
"_embedded": {
"publications": [
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/1",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "The Stand"
},
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/2",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "Infinite Jest"
},
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/3",
"hreflang": "en",
"type": "application/hal+json"
}
},
"title": "Walden"
}
]
}
}
Using Custom Media / Mime Types
If you wish to use a custom Mime Type then you first need to declare the Mime Types in
grails-app/conf/Config.groovy
:
grails.mime.types = [
all: "*/*",
book: "application/vnd.books.org.book+json",
bookList: "application/vnd.books.org.booklist+json",
…
]
It is critical that place your new mime types after the 'all' Mime Type because if the Content Type of the request cannot be established then the first entry in the map is used for the response. If you have your new Mime Type at the top then Grails will always try and send back your new Mime Type if the requested Mime Type cannot be established.
Then override the renderer to return HAL using the custom Mime Types:
import grails.rest.render.hal.*
import org.codehaus.groovy.grails.web.mime.*beans = {
halBookRenderer(HalJsonRenderer, rest.test.Book, new MimeType("application/vnd.books.org.book+json", [v:"1.0"]))
halBookListRenderer(HalJsonCollectionRenderer, rest.test.Book, new MimeType("application/vnd.books.org.booklist+json", [v:"1.0"]))
}
In the above example the first bean defines a HAL renderer for a single book instance that returns a Mime Type of
application/vnd.books.org.book+json
. The second bean defines the Mime Type used to render a collection of books (in this case
application/vnd.books.org.booklist+json
).
With this in place issuing a request for the new Mime Type returns the necessary HAL:
$ curl -i -H "Accept: application/vnd.books.org.book+json" http://localhost:8080/myapp/books/1HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Type: application/vnd.books.org.book+json;charset=ISO-8859-1
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/1",
"hreflang": "en",
"type": "application/vnd.books.org.book+json"
}
},
"title": ""The Stand""
}
Customizing Link Rendering
An important aspect of HATEOAS is the usage of links that describe the transitions the client can use to interact with the REST API. By default the
HalJsonRenderer
will automatically create links for you for associations and to the resource itself (using the "self" relationship).
However you can customize link rendering using the
link
method that is added to all domain classes annotated with
grails.rest.Resource
or any class annotated with
grails.rest.Linkable
. For example, the
show
action can be modified as follows to provide a new link in the resulting output:
def show(Book book) {
book.link rel:'publisher', href: g.link(resource:"publisher", params:[bookId: book.id])
respond book
}
Which will result in output such as:
{
"_links": {
"self": {
"href": "http://localhost:8080/myapp/books/1",
"hreflang": "en",
"type": "application/vnd.books.org.book+json"
}
"publisher": {
"href": "http://localhost:8080/myapp/books/1/publisher",
"hreflang": "en"
}
},
"title": ""The Stand""
}
The
link
method can be passed named arguments that match the properties of the
grails.rest.Link
class.
9.1.7.2 Atom Support
Atom is another standard interchange format used to implement REST APIs. An example of Atom output can be seen below:
<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom"> <title>Example Feed</title>
<link href="http://example.org/"/>
<updated>2003-12-13T18:30:02Z</updated>
<author>
<name>John Doe</name>
</author>
<id>urn:uuid:60a76c80-d399-11d9-b93C-0003939e0af6</id> <entry>
<title>Atom-Powered Robots Run Amok</title>
<link href="http://example.org/2003/12/13/atom03"/>
<id>urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a</id>
<updated>2003-12-13T18:30:02Z</updated>
<summary>Some text.</summary>
</entry></feed>
To use Atom rendering again simply define a custom renderer:
import grails.rest.render.atom.*
beans = {
halBookRenderer(AtomRenderer, rest.test.Book)
halBookListRenderer(AtomCollectionRenderer, rest.test.Book)
}
9.1.7.3 Vnd.Error Support
Vnd.Error is a standardised way of expressing an error response.
By default when a validation error occurs when attempting to POST new resources then the errors object will be sent back allow with a 422 respond code:
$ curl -i -H "Accept: application/json" -H "Content-Type: application/json" -X POST -d "" http://localhost:8080/myapp/booksHTTP/1.1 422 Unprocessable Entity
Server: Apache-Coyote/1.1
Content-Type: application/json;charset=ISO-8859-1{"errors":[{"object":"rest.test.Book", "field":"title", "rejected-value":null, "message":"Property [title] of class [class rest.test.Book] cannot be null"}]}
If you wish to change the format to Vnd.Error then simply register
grails.rest.render.errors.VndErrorJsonRenderer
bean in
grails-app/conf/spring/resources.groovy
:
beans = {
vndJsonErrorRenderer(grails.rest.render.errors.VndErrorJsonRenderer)
// for Vnd.Error XML format
vndXmlErrorRenderer(grails.rest.render.errors.VndErrorXmlRenderer)
}
Then if you alter the client request to accept Vnd.Error you get an appropriate response:
$ curl -i -H "Accept: application/vnd.error+json,application/json" -H "Content-Type: application/json" -X POST -d "" http://localhost:8080/myapp/books
HTTP/1.1 200 OK
Server: Apache-Coyote/1.1
Content-Type: application/vnd.error+json;charset=ISO-8859-1[
{
"logref": ""book.nullable"",
"message": "Property [title] of class [class rest.test.Book] cannot be null",
"_links": {
"resource": {
"href": "http://localhost:8080/rest-test/books"
}
}
}
]
9.1.8 Customizing Binding of Resources
The framework provides a sophisticated but simple mechanism for binding REST requests to domain objects and command objects. One way to take advantage of this is to bind the
request
property in a controller the
properties
of a domain class. Given the following XML as the body of the request, the
createBook
action will create a new
Book
and assign "The Stand" to the
title
property and "Stephen King" to the
authorName
property.
<?xml version="1.0" encoding="UTF-8"?>
<book>
<title>The Stand</title>
<authorName>Stephen King</authorName>
</book>
class BookController { def createBook() {
def book = new Book()
book.properties = request // …
}
}
If the root element of the XML document contains an
id
attribute, the
id
value will be used to retrieve the corresponding persistent instance from the database and then the rest of the document will be bound to the instance. If no corresponding record is found in the database, the command object reference will be null.
<?xml version="1.0" encoding="UTF-8"?>
<book>
<title>The Stand</title>
<authorName>Stephen King</authorName>
</book>
Command objects will automatically be bound with the body of the request:
class BookController {
def createBook(BookCommand book) { // …
}
}class BookCommand {
String title
String authorName
}
If the command object type is a domain class and the root element of the XML document contains an
id
attribute, the
id
value will be used to retrieve the corresponding persistent instance from the database and then the rest of the document will be bound to the instance. If no corresponding record is found in the database, the command object reference will be null.
<?xml version="1.0" encoding="UTF-8"?>
<book id="42">
<title>Walden</title>
<authorName>Henry David Thoreau</authorName>
</book>
class BookController {
def updateBook(Book book) {
// The book will have been retrieved from the database and updated
// by doing something like this:
//
// book == Book.get('42')
// if(book != null) {
// book.properties = request
// }
//
// the code above represents what the framework will
// have done. There is no need to write that code. // ... }
}
The data binding depends on an instance of the
DataBindingSource interface created by an instance of the
DataBindingSourceCreator interface. The specific implementation of
DataBindingSourceCreator
will be selected based on the
contentType
of the request. Several implementations are provided to handle common content types. The default implementations will be fine for most use cases. The following table lists the content types which are supported by the core framework and which
DataBindingSourceCreator
implementations are used for each. All of the implementation classes are in the
org.codehaus.groovy.grails.web.binding.bindingsource
package.
Content Type(s) | Bean Name | DataBindingSourceCreator Impl. |
---|
application/xml, text/xml | xmlDataBindingSourceCreator | XmlDataBindingSourceCreator |
application/json, text/json | jsonDataBindingSourceCreator | JsonDataBindingSourceCreator |
application/hal+json | halJsonDataBindingSourceCreator | HalJsonDataBindingSourceCreator |
application/hal+xml | halXmlDataBindingSourceCreator | HalXmlDataBindingSourceCreator |
In order to provide your own
DataBindingSourceCreator
for any of those content types, write a class which implements
DataBindingSourceCreator
and register an instance of that class in the Spring application context. If you
are replacing one of the existing helpers, use the corresponding bean name from above. If you are providing a
helper for a content type other than those accounted for by the core framework, the bean name may be anything that
you like but you should take care not to conflict with one of the bean names above.
The
DataBindingSourceCreator
interface defines just 2 methods:
package org.grails.databinding.bindingsourceimport org.codehaus.groovy.grails.web.mime.MimeType
import org.grails.databinding.DataBindingSource/**
* A factory for DataBindingSource instances
*
* @since 2.3
* @see DataBindingSourceRegistry
* @see DataBindingSource
*
*/
interface DataBindingSourceCreator { /**
* return All of the {
link MimeType} supported by this helper
*/
MimeType[] getMimeTypes() /**
* Creates a DataBindingSource suitable for binding bindingSource to bindingTarget
*
* @param mimeType a mime type
* @param bindingTarget the target of the data binding
* @param bindingSource the value being bound
* @return a DataBindingSource
*/
DataBindingSource createDataBindingSource(MimeType mimeType, Object bindingTarget, Object bindingSource)
}
AbstractRequestBodyDataBindingSourceCreator
is an abstract class designed to be extended to simplify writing custom
DataBindingSourceCreator
classes. Classes which
extend
AbstractRequestbodyDatabindingSourceCreator
need to implement a method named
createBindingSource
which accepts an
InputStream
as an argument and returns a
DataBindingSource
as well as implementing the
getMimeTypes
method described in the
DataBindingSourceCreator
interface above. The
InputStream
argument to
createBindingSource
provides access to the body of the request.
The code below shows a simple implementation.
// MyCustomDataBindingSourceCreator.groovy in
// src/groovy/com/demo/myapp/databinding
package com.demo.myapp.databindingimport org.codehaus.groovy.grails.web.mime.MimeType
import org.grails.databinding.DataBindingSource
import org...databinding.SimpleMapDataBindingSource
import org...databinding.bindingsource.AbstractRequestBodyDataBindingSourceCreator/**
* A custom DataBindingSourceCreator capable of parsing key value pairs out of
* a request body containing a comma separated list of key:value pairs like:
*
* name:Herman,age:99,town:STL
*
*/
class MyCustomDataBindingSourceCreator extends AbstractRequestBodyDataBindingSourceCreator { @Override
public MimeType[] getMimeTypes() {
[new MimeType('text/custom+demo+csv')] as MimeType[]
} @Override
protected DataBindingSource createBindingSource(InputStream inputStream) {
def map = [:] def reader = new InputStreamReader(inputStream) // this is an obviously naive parser and is intended
// for demonstration purposes only. reader.eachLine { line ->
def keyValuePairs = line.split(',')
keyValuePairs.each { keyValuePair ->
if(keyValuePair?.trim()) {
def keyValuePieces = keyValuePair.split(':')
def key = keyValuePieces[0].trim()
def value = keyValuePieces[1].trim()
map[key] = value
}
}
} // create and return a DataBindingSource which contains the parsed data
new SimpleMapDataBindingSource(map)
}
}
An instance of
MyCustomDataSourceCreator
needs to be registered in the spring application context.
// grails-app/conf/spring/resources.groovy
beans = { myCustomCreator com.demo.myapp.databinding.MyCustomDataBindingSourceCreator // …
}
With that in place the framework will use the
myCustomCreator
bean any time a
DataBindingSourceCreator
is needed
to deal with a request which has a
contentType
of "text/custom+demo+csv".
9.2 SOAP
Grails does not feature SOAP support out-of-the-box, but there are several plugins that can help for both producing SOAP servers and calling SOAP web services.
SOAP Clients
To call SOAP web services there are generally 2 approaches taken, one is to use a tool to generate client stubs, the other is to manually construct the SOAP calls. The former can be easier to use, but the latter provides more flexibility / control.
The
CXF client plugin uses the CXF framework, which includes a
wsdl2java
tool for generating a client. There is nothing Groovy/Grails specific here in the generated code as it simply provides a Java API which you can invoke to call SOAP web services.
See the documentation on the
CXF client plugin for further information.
Alternatively, if you prefer more control over your SOAP calls the
WS-Lite library is an excellent choice and features a
Grails plugin. You have more control over the SOAP requests sent, and since Groovy has fantastic support for building and parsing XML it can be very productive approach.
Below is an example of a SOAP call with wslite:
withSoap(serviceURL: 'http://www.holidaywebservice.com/Holidays/US/Dates/USHolidayDates.asmx') {
def response = send {
body {
GetMothersDay(xmlns: 'http://www.27seconds.com/Holidays/US/Dates/') {
year(2011)
}
}
}
println response.GetMothersDayResponse.GetMothersDayResult.text()
}
It is not recommended that you use the
GroovyWS library, it pulls in many dependencies which increases the likelihood of conflicts. The WSlite library provides a far simpler and easier to use solution.
SOAP Servers
Again, Grails does not have direct support for exposing SOAP web services, however if you wish to expose a SOAP service from your application then the
CXF plugin (not to be confused with the cxf-client plugin), provides an easy way to do so.
Typically it involves taking a Grails service and adding 'expose'-style configuration, such as the below:
static expose = EndpointType.JAX_WS_WSDL
//your path (preferred) or url to wsdl
static wsdl = 'org/grails/cxf/test/soap/CustomerService.wsdl'
Please refer to the
documentation of the plugin for more information.
9.3 RSS and Atom
No direct support is provided for RSS or Atom within Grails. You could construct RSS or ATOM feeds with the
render method's XML capability. There is however a
Feeds plugin available for Grails that provides a RSS and Atom builder using the popular
ROME library. An example of its usage can be seen below:
def feed() {
render(feedType: "rss", feedVersion: "2.0") {
title = "My test feed"
link = "http://your.test.server/yourController/feed" for (article in Article.list()) {
entry(article.title) {
link = "http://your.test.server/article/${article.id}"
article.content // return the content
}
}
}
}
10 Asynchronous Programming
With modern hardware featuring multiple cores, many programming languages have been adding asynchronous, parallel programming APIs, Groovy being no exception.
The excellent
GPars project features a whole range of different APIs for asynchronous programming techniques including actors, promises, STM and data flow concurrency.
Added Grails 2.3, the Async features of Grails aim to simplify concurrent programming within the framework and include the concept of Promises and a unified event model.
10.1 Promises
A Promise is a concept being embraced by many concurrency frameworks. They are similar to
java.util.concurrent.Future
instances, but include a more user friendly exception handling model, useful features like chaining and the ability to attach listeners.
Promise Basics
In Grails the
grails.async.Promises
class provides the entry point to the Promise API:
import static grails.async.Promises.*
To create promises you can use the
task
method, which returns an instance of the
grails.async.Promise
interface:
def p1 = task { 2 * 2 }
def p2 = task { 4 * 4 }
def p3 = task { 8 * 8 }
assert [4,16,64] == waitAll(p1, p2, p3)
The
waitAll
method waits synchronously, blocking the current thread, for all of the concurrent tasks to complete and returns the results.
If you prefer not to block the current thread you can use the
onComplete
method:
onComplete([p1,p2,p3]) { List results ->
assert [4,16,64] == results
}
The
waitAll
method will throw an exception if an error occurs executing one of the promises. The originating exception will be thrown. The
onComplete
method, however, will simply not execute the passed closure if an exception occurs. You can register an
onError
listener if you wish to handle exceptions without blocking:
onError([p1,p2,p3]) { Throwable t ->
println "An error occured ${t.message}"
}
If you have just a single long running promise then the
grails.async.Promise
interface provides a similar API on the promise itself. For example:
import static java.util.concurrent.TimeUnit.*
import static grails.async.Promises.*Promise p = task {
// Long running task
}
p.onError { Throwable err ->
println "An error occured ${err.message}"
}
p.onComplete { result ->
println "Promise returned $result"
}
// block until result is called
def result = p.get()
// block for the specified time
def result = p.get(1,MINUTES)
Promise Chaining
It is possible to chain several promises and wait for the chain to complete using the
then
method:
final polish = { … }
final transform = { … }
final save = { … }
final notify = { … }Promise promise = task {
// long running task
}
promise.then polish then transform then save then {
// notify end result
}
If an exception occurs at any point in the chain it will be propagated back to the caller and the next step in the chain will not be called.
Promise Lists and Maps
Grails' async API also features the concept of a promise lists and maps. These are represented by the
grails.async.PromiseList
and
grails.async.PromiseMap
classes respectively.
The easiest way to create a promise list or map is via the
tasks
method of the
Promises
class:
import static grails.async.Promises.*def promiseList = tasks([{ 2 * 2 }, { 4 * 4}, { 8 * 8 }])assert [4,16,64] == promiseList.get()
The
tasks
method, when passed a list of closures, returns a
PromiseList
. You can also construct a
PromiseList
manually:
import grails.async.*def list = new PromiseList()
list << { 2 * 2 }
list << { 4 * 4 }
list << { 8 * 8 }
list.onComplete { List results ->
assert [4,16,64] == results
}
The PromiseList
class does not implement the java.util.List interface, but instead returns a java.util.List from the get() method
Working with
PromiseMap
instances is largely similar. Again you can either use the
tasks
method:
import static grails.async.Promises.*def promiseList = tasks one:{ 2 * 2 },
two:{ 4 * 4},
three:{ 8 * 8 }assert [one:4,two:16,three:64] == promiseList.get()
Or construct a
PromiseMap
manually:
import grails.async.*def map = new PromiseMap()
map['one'] = { 2 * 2 }
map['two'] = { 4 * 4 }
map['three'] = { 8 * 8 }
map.onComplete { Map results ->
assert [one:4,two:16,three:64] == results
}
Promise Factories
The
Promises
class uses a
grails.async.PromiseFactory
instance to create
Promise
instances.
The default implementation uses the GPars concurrency library and is called
org.grails.async.factory.gpars.GparsPromiseFactory
, however it is possible to swap implementations by setting the
Promises.promiseFactory
variable.
One common use case for this is unit testing, typically you do not want promises to execute asynchronously during unit tests, as this makes tests harder to write. For this purpose Grails ships with a
org.grails.async.factory.SynchronousPromiseFactory
instance that makes it easier to test promises:
import org.grails.async.factory.*
import grails.async.*Promises.promiseFactory = new SynchronousPromiseFactory()
Using the
PromiseFactory
mechanism is theoretically possible to plug in other concurrency libraries into the Grails framework.
DelegateAsync Transformation
It is quite common to require both synchronous and asynchronous versions of the same API. Developing both can result in a maintenance problem as typically the asynchronous API would simply delegate to the synchronous version.
The
DelegateAsync
transformation is designed to mitigate this problem by transforming any synchronous API into an asynchronous one.
For example, consider the following service:
class BookService {
List<Book> findBooks(String title) {
// implementation
}
}
The
findBooks
method executes synchronously in the same thread as the caller. To make an asynchronous version of this API you can define another class as follows:
import grails.async.*class AsyncBookService {
@DelegateAsync BookService bookService
}
The
DelegateAsync
transformation will automatically add a new method that looks like the following to the
AsyncBookService
class:
Promise<List<Book>> findBooks(String title) {
Promises.task {
bookService.findBooks(title)
}
}
As you see the transform adds equivalent methods that return a Promise and execute asynchronously.
The
AsyncBookService
can then be injected into other controllers and services and used as follows:
AsyncBookService asyncBookService
def findBooks(String title) {
asyncBookService.findBooks(title)
.onComplete { List results ->
println "Books = ${results}"
}
}
10.2 Asynchronous GORM
Since Grails 2.3, GORM features an asynchronous programming model that works across all supported datastores (Hibernate, MongoDB etc.).
Async Namespace
The Asynchronous GORM API is available on every domain class via the
async
namespace.
For example, the following code listing reads 3 objects from the database asynchronously:
import static grails.async.Promises.*def p1 = Person.async.get(1L)
def p2 = Person.async.get(2L)
def p3 = Person.async.get(3L)
def results = waitAll(p1, p2, p3)
Using the
async
namespace, all the regular GORM methods are available (even dynamic finders), but instead of executing synchronously, the query is run in the background and a
Promise
instance is returned.
The following code listing shows a few common examples of GORM queries executed asynchronously:
import static grails.async.Promises.*Person.async.list().onComplete { List results ->
println "Got people = ${results}"
}
def p = Person.async.getAll(1L, 2L, 3L)
List results = p.get()def p1 = Person.async.findByFirstName("Homer")
def p2 = Person.async.findByFirstName("Bart")
def p3 = Person.async.findByFirstName("Barney")
results = waitAll(p1, p2, p3)
Async and the Session
When using GORM async each promise is executed in a different thread. Since the Hibernate session is not concurrency safe, a new session is bound per thread.
This is an important consideration when using GORM async (particularly with Hibernate as the persistence engine). The objects returned from asynchronous queries will be detached entities.
This means you cannot save objects returned from asynchronous queries without first merging them back into session. For example the following will not work:
def promise = Person.async.findByFirstName("Homer")
def person = promise.get()
person.firstName = "Bart"
person.save()
Instead you need to merge the object with the session bound to the calling thread. The above code needs to be written as:
def promise = Person.async.findByFirstName("Homer")
def person = promise.get()
person.merge()
person.firstName = "Bart"
Note that
merge()
is called first because it may refresh the object from the cache or database, which would result in the change being lost. In general it is not recommended to read and write objects in different threads and you should avoid this technique unless absolutely necessary.
Finally, another issue with detached objects is that association lazy loading
will not work and you will encounter
LazyInitializationException
errors if you do so. If you plan to access the associated objects of those returned from asynchronous queries you should use eager queries (which is recommended anyway to avoid N+1 problems).
Multiple Asynchronous GORM calls
As discussed in the previous section you should avoid reading and writing objects in different threads as merging tends to be inefficient.
However, if you wish to do more complex GORM work asynchronously then the GORM async namespace provides a
task
method that makes this possible. For example:
def promise = Person.async.task {
withTransaction {
def person = findByFirstName("Homer")
person.firstName = "Bart"
person.save(flush:true)
}
}Person updatedPerson = promise.get()
Note that the GORM
task
method differs from the static
Promises.task
method in that it deals with binding a new session to the asynchronous thread for you. If you do not use the GORM version and do asynchronous work with GORM then you need to do this manually. Example:
import static grails.async.Promises.*def promise = task {
Person.withNewSession {
// your logic here
}
}
Async DetachedCriteria
The
DetachedCriteria
class also supports the
async
namespace. For example you can do the following:
DetachedCriteria query = Person.where {
lastName == "Simpson"
}def promise = query.async.list()
10.3 Asynchronous Request Handling
If you are deploying to a Servlet 3.0 container such as Tomcat 7 and above then it is possible to deal with responses asynchronously.
In general for controller actions that execute quickly there is little benefit in handling requests asynchronously. However, for long running controller actions it is extremely beneficial.
The reason being that with an asynchronous / non-blocking response, the one thread == one request == one response relationship is broken. The container can keep a client response open and active, and at the same time return the thread back to the container to deal with another request, improving scalability.
For example, if you have 70 available container threads and an action takes a minute to complete, if the actions are not executed in a non-blocking fashion the likelihood of all 70 threads being occupied and the container not being able to respond is quite high and you should consider asynchronous request processing.
Since Grails 2.3, Grails features a simplified API for creating asynchronous responses built on the
Promise
mechanism discussed previously.
The implementation is based on Servlet 3.0 async so to enable the async features you need to set your servlet target version to 3.0 in BuildConfig.groovy:
grails.servlet.version = "3.0"
Async Models
A typical activity in a Grails controller is to produce a model (a map of key/value pairs) that can be rendered by a view.
If the model takes a while to produce then the server could arrive at a blocking state, impacting scalability. You tell Grails to build the model asynchronously by returning a
grails.async.PromiseMap
via the
Promises.tasks
method:
import static grails.async.Promises.*
…
def index() {
tasks books: Book.async.list(),
totalBooks: Book.async.count(),
otherValue: {
// do hard work
}
}
Grails will handle the response asynchronously, waiting for the promises to complete before rendering the view. The equivalent synchronous action of the above is:
def index() {
def otherValue = …
[ books: Book.list() ,
totalBooks: Book.count(),
otherValue: otherValue ]
}
You can even render different view by passing the
PromiseMap
to the
model
attribute of the
render
method:
import static grails.async.Promises.*
…
def index() {
render view:"myView", model: tasks( one:{ 2 * 2 },
two:{ 3 * 3 } )
}
Async Response Rendering
You can also write to the response asynchronously using promises in Grails 2.3 and above:
import static grails.async.Promises.*
class StockController { def stock(String ticker) {
task {
ticker = ticker ?: 'GOOG'
def url = new URL("http://download.finance.yahoo.com/d/quotes.csv?s=${ticker}&f=nsl1op&e=.csv")
Double price = url.text.split(',')[-1] as Double
render "ticker: $ticker, price: $price"
}
}
}
The above example using Yahoo Finance to query stock prices, executing asynchronously and only rendering the response once the result has been obtained. This is done by returning a
Promise
instance from the controller action.
If the Yahoo URL is unresponsive the original request thread will not be blocked and the container will not become unresponsive.
10.4 Servlet 3.0 Async
In addition to the higher level async features discussed earlier in the section, you can access the raw Servlet 3.0 asynchronous API from a Grails application.
Servlet 3.0 Asynchronous Rendering
You can render content (templates, binary data etc.) in an asynchronous manner by calling the
startAsync
method which returns an instance of the Servlet 3.0
AsyncContext
. Once you have a reference to the
AsyncContext
you can use Grails' regular render method to render content:
def index() {
def ctx = startAsync()
ctx.start {
new Book(title:"The Stand").save()
render template:"books", model:[books:Book.list()]
ctx.complete()
}
}
Note that you must call the
complete()
method to terminate the connection.
Resuming an Async Request
You resume processing of an async request (for example to delegate to view rendering) by using the
dispatch
method of the
AsyncContext
class:
def index() {
def ctx = startAsync()
ctx.start {
// do working
…
// render view
ctx.dispatch()
}
}
11 Validation
Grails validation capability is built on
Spring's Validator API and data binding capabilities. However Grails takes this further and provides a unified way to define validation "constraints" with its constraints mechanism.
Constraints in Grails are a way to declaratively specify validation rules. Most commonly they are applied to
domain classes, however
URL Mappings and
Command Objects also support constraints.
11.1 Declaring Constraints
Within a domain class
constraints are defined with the constraints property that is assigned a code block:
class User {
String login
String password
String email
Integer age static constraints = {
…
}
}
You then use method calls that match the property name for which the constraint applies in combination with named parameters to specify constraints:
class User {
... static constraints = {
login size: 5..15, blank: false, unique: true
password size: 5..15, blank: false
email email: true, blank: false
age min: 18
}
}
In this example we've declared that the
login
property must be between 5 and 15 characters long, it cannot be blank and must be unique. We've also applied other constraints to the
password
,
email
and
age
properties.
By default, all domain class properties are not nullable (i.e. they have an implicit nullable: false
constraint).
A complete reference for the available constraints can be found in the Quick Reference section under the Constraints heading.
Note that constraints are only evaluated once which may be relevant for a constraint that relies on a value like an instance of
java.util.Date
.
class User {
... static constraints = {
// this Date object is created when the constraints are evaluated, not
// each time an instance of the User class is validated.
birthDate max: new Date()
}
}
A word of warning - referencing domain class properties from constraints
It's very easy to attempt to reference instance variables from the static constraints block, but this isn't legal in Groovy (or Java). If you do so, you will get a
MissingPropertyException
for your trouble. For example, you may try
class Response {
Survey survey
Answer answer static constraints = {
survey blank: false
answer blank: false, inList: survey.answers
}
}
See how the
inList
constraint references the instance property
survey
? That won't work. Instead, use a custom
validator:
class Response {
…
static constraints = {
survey blank: false
answer blank: false, validator: { val, obj -> val in obj.survey.answers }
}
}
In this example, the
obj
argument to the custom validator is the domain
instance that is being validated, so we can access its
survey
property and return a boolean to indicate whether the new value for the
answer
property,
val
, is valid.
11.2 Validating Constraints
Validation Basics
Call the
validate method to validate a domain class instance:
def user = new User(params)if (user.validate()) {
// do something with user
}
else {
user.errors.allErrors.each {
println it
}
}
The
errors
property on domain classes is an instance of the Spring
Errors interface. The
Errors
interface provides methods to navigate the validation errors and also retrieve the original values.
Validation Phases
Within Grails there are two phases of validation, the first one being
data binding which occurs when you bind request parameters onto an instance such as:
def user = new User(params)
At this point you may already have errors in the
errors
property due to type conversion (such as converting Strings to Dates). You can check these and obtain the original input value using the
Errors
API:
if (user.hasErrors()) {
if (user.errors.hasFieldErrors("login")) {
println user.errors.getFieldError("login").rejectedValue
}
}
The second phase of validation happens when you call
validate or
save. This is when Grails will validate the bound values against the
constraints you defined. For example, by default the
save method calls
validate
before executing, allowing you to write code like:
if (user.save()) {
return user
}
else {
user.errors.allErrors.each {
println it
}
}
11.3 Sharing Constraints Between Classes
A common pattern in Grails is to use
command objects for validating user-submitted data and then copy the properties of the command object to the relevant domain classes. This often means that your command objects and domain classes share properties and their constraints. You could manually copy and paste the constraints between the two, but that's a very error-prone approach. Instead, make use of Grails' global constraints and import mechanism.
Global Constraints
In addition to defining constraints in domain classes, command objects and
other validateable classes, you can also define them in
grails-app/conf/Config.groovy
:
grails.gorm.default.constraints = {
'*'(nullable: true, size: 1..20)
myShared(nullable: false, blank: false)
}
These constraints are not attached to any particular classes, but they can be easily referenced from any validateable class:
class User {
... static constraints = {
login shared: "myShared"
}
}
Note the use of the
shared
argument, whose value is the name of one of the constraints defined in
grails.gorm.default.constraints
. Despite the name of the configuration setting, you can reference these shared constraints from any validateable class, such as command objects.
The '*' constraint is a special case: it means that the associated constraints ('nullable' and 'size' in the above example) will be applied to all properties in all validateable classes. These defaults can be overridden by the constraints declared in a validateable class.
Importing Constraints
Grails 2 introduced an alternative approach to sharing constraints that allows you to import a set of constraints from one class into another.
Let's say you have a domain class like so:
class User {
String firstName
String lastName
String passwordHash static constraints = {
firstName blank: false, nullable: false
lastName blank: false, nullable: false
passwordHash blank: false, nullable: false
}
}
You then want to create a command object,
UserCommand
, that shares some of the properties of the domain class and the corresponding constraints. You do this with the
importFrom()
method:
class UserCommand {
String firstName
String lastName
String password
String confirmPassword static constraints = {
importFrom User password blank: false, nullable: false
confirmPassword blank: false, nullable: false
}
}
This will import all the constraints from the
User
domain class and apply them to
UserCommand
. The import will ignore any constraints in the source class (
User
) that don't have corresponding properties in the importing class (
UserCommand
). In the above example, only the 'firstName' and 'lastName' constraints will be imported into
UserCommand
because those are the only properties shared by the two classes.
If you want more control over which constraints are imported, use the
include
and
exclude
arguments. Both of these accept a list of simple or regular expression strings that are matched against the property names in the source constraints. So for example, if you only wanted to import the 'lastName' constraint you would use:
…
static constraints = {
importFrom User, include: ["lastName"]
…
}
or if you wanted all constraints that ended with 'Name':
…
static constraints = {
importFrom User, include: [/.*Name/]
…
}
Of course,
exclude
does the reverse, specifying which constraints should
not be imported.
11.4 Validation on the Client
Displaying Errors
Typically if you get a validation error you redirect back to the view for rendering. Once there you need some way of displaying errors. Grails supports a rich set of tags for dealing with errors. To render the errors as a list you can use
renderErrors:
<g:renderErrors bean="${user}" />
If you need more control you can use
hasErrors and
eachError:
<g:hasErrors bean="${user}">
<ul>
<g:eachError var="err" bean="${user}">
<li>${err}</li>
</g:eachError>
</ul>
</g:hasErrors>
Highlighting Errors
It is often useful to highlight using a red box or some indicator when a field has been incorrectly input. This can also be done with the
hasErrors by invoking it as a method. For example:
<div class='value ${hasErrors(bean:user,field:'login','errors')}'>
<input type="text" name="login" value="${fieldValue(bean:user,field:'login')}"/>
</div>
This code checks if the
login
field of the
user
bean has any errors and if so it adds an
errors
CSS class to the
div
, allowing you to use CSS rules to highlight the
div
.
Retrieving Input Values
Each error is actually an instance of the
FieldError class in Spring, which retains the original input value within it. This is useful as you can use the error object to restore the value input by the user using the
fieldValue tag:
<input type="text" name="login" value="${fieldValue(bean:user,field:'login')}"/>
This code will check for an existing
FieldError
in the
User
bean and if there is obtain the originally input value for the
login
field.
11.5 Validation and Internationalization
Another important thing to note about errors in Grails is that error messages are not hard coded anywhere. The
FieldError class in Spring resolves messages from message bundles using Grails'
i18n support.
Constraints and Message Codes
The codes themselves are dictated by a convention. For example consider the constraints we looked at earlier:
package com.mycompany.myappclass User {
... static constraints = {
login size: 5..15, blank: false, unique: true
password size: 5..15, blank: false
email email: true, blank: false
age min: 18
}
}
If a constraint is violated Grails will by convention look for a message code of the form:
[Class Name].[Property Name].[Constraint Code]
In the case of the
blank
constraint this would be
user.login.blank
so you would need a message such as the following in your
grails-app/i18n/messages.properties
file:
user.login.blank=Your login name must be specified!
The class name is looked for both with and without a package, with the packaged version taking precedence. So for example, com.mycompany.myapp.User.login.blank will be used before user.login.blank. This allows for cases where your domain class message codes clash with a plugin's.
For a reference on what codes are for which constraints refer to the reference guide for each constraint.
Displaying Messages
The
renderErrors tag will automatically look up messages for you using the
message tag. If you need more control of rendering you can handle this yourself:
<g:hasErrors bean="${user}">
<ul>
<g:eachError var="err" bean="${user}">
<li><g:message error="${err}" /></li>
</g:eachError>
</ul>
</g:hasErrors>
In this example within the body of the
eachError tag we use the
message tag in combination with its
error
argument to read the message for the given error.
11.6 Applying Validation to Other Classes
Domain classes and
command objects support validation by default. Other classes may be made validateable by defining the static
constraints
property in the class (as described above) and then telling the framework about them. It is important that the application register the validateable classes with the framework. Simply defining the
constraints
property is not sufficient.
The Validateable Annotation
Classes which define the static
constraints
property and are annotated with @Validateable can be made validateable by the framework. Consider this example:
// src/groovy/com/mycompany/myapp/User.groovy
package com.mycompany.myappimport grails.validation.Validateable@Validateable
class User {
... static constraints = {
login size: 5..15, blank: false, unique: true
password size: 5..15, blank: false
email email: true, blank: false
age min: 18
}
}
Registering Validateable Classes
If a class is not marked with
Validateable, it may still be made validateable by the framework. The steps required to do this are to define the static
constraints
property in the class (as described above) and then telling the framework about the class by assigning a value to the
grails.validateable.classes
property in
Config.groovy@:
grails.validateable.classes = [com.mycompany.myapp.User, com.mycompany.dto.Account]
12 The Service Layer
Grails defines the notion of a service layer. The Grails team discourages the embedding of core application logic inside controllers, as it does not promote reuse and a clean separation of concerns.
Services in Grails are the place to put the majority of the logic in your application, leaving controllers responsible for handling request flow with redirects and so on.
Creating a Service
You can create a Grails service by running the
create-service command from the root of your project in a terminal window:
grails create-service helloworld.simple
If no package is specified with the create-service script, Grails automatically uses the application name as the package name.
The above example will create a service at the location
grails-app/services/helloworld/SimpleService.groovy
. A service's name ends with the convention
Service
, other than that a service is a plain Groovy class:
package helloworldclass SimpleService {
}
12.1 Declarative Transactions
Default Declarative Transactions
Services are typically involved with coordinating logic between
domain classes, and hence often involved with persistence that spans large operations. Given the nature of services, they frequently require transactional behaviour. You can use programmatic transactions with the
withTransaction method, however this is repetitive and doesn't fully leverage the power of Spring's underlying transaction abstraction.
Services enable transaction demarcation, which is a declarative way of defining which methods are to be made transactional. All services are transactional by default. To disable this set the
transactional
property to
false
:
class CountryService {
static transactional = false
}
You may also set this property to
true
to make it clear that the service is intentionally transactional.
Warning: dependency injection is the only way that declarative transactions work. You will not get a transactional service if you use the new
operator such as new BookService()
The result is that all methods are wrapped in a transaction and automatic rollback occurs if a method throws a runtime exception (i.e. one that extends
RuntimeException
) or an
Error
. The propagation level of the transaction is by default set to
PROPAGATION_REQUIRED.
Checked exceptions do not roll back transactions. Even though Groovy blurs the distinction between checked and unchecked exceptions, Spring isn't aware of this and its default behaviour is used, so it's important to understand the distinction between checked and unchecked exceptions.
Custom Transaction Configuration
Grails also provides
@Transactional
and
@NotTransactional
annotations for cases where you need more fine-grained control over transactions at a per-method level or need to specify an alternative propagation level. For example, the
@NotTransactional
annotation can be used to mark a particular method to be skipped when a class is annotated with
@Transactional
.
The grails.transaction.Transactional
annotation was first introduced in Grails 2.3. Prior to 2.3, Spring's @Transactional annotation was used.
Annotating a service method with Transactional
disables the default Grails transactional behavior for that service (in the same way that adding transactional=false
does) so if you use any annotations you must annotate all methods that require transactions.
In this example
listBooks
uses a read-only transaction,
updateBook
uses a default read-write transaction, and
deleteBook
is not transactional (probably not a good idea given its name).
import org.springframework.transaction.annotation.Transactionalclass BookService { @Transactional(readOnly = true)
def listBooks() {
Book.list()
} @Transactional
def updateBook() {
// …
} def deleteBook() {
// …
}
}
You can also annotate the class to define the default transaction behavior for the whole service, and then override that default per-method. For example, this service is equivalent to one that has no annotations (since the default is implicitly
transactional=true
):
import org.springframework.transaction.annotation.Transactional@Transactional
class BookService { def listBooks() {
Book.list()
} def updateBook() {
// …
} def deleteBook() {
// …
}
}
This version defaults to all methods being read-write transactional (due to the class-level annotation), but the
listBooks
method overrides this to use a read-only transaction:
import org.springframework.transaction.annotation.Transactional@Transactional
class BookService { @Transactional(readOnly = true)
def listBooks() {
Book.list()
} def updateBook() {
// …
} def deleteBook() {
// …
}
}
Although
updateBook
and
deleteBook
aren't annotated in this example, they inherit the configuration from the class-level annotation.
For more information refer to the section of the Spring user guide on
Using @Transactional.
Unlike Spring you do not need any prior configuration to use
Transactional
; just specify the annotation as needed and Grails will detect them up automatically.
12.1.1 Transactions Rollback and the Session
Understanding Transactions and the Hibernate Session
When using transactions there are important considerations you must take into account with regards to how the underlying persistence session is handled by Hibernate. When a transaction is rolled back the Hibernate session used by GORM is cleared. This means any objects within the session become detached and accessing uninitialized lazy-loaded collections will lead to
LazyInitializationException
s.
To understand why it is important that the Hibernate session is cleared. Consider the following example:
class Author {
String name
Integer age static hasMany = [books: Book]
}
If you were to save two authors using consecutive transactions as follows:
Author.withTransaction { status ->
new Author(name: "Stephen King", age: 40).save()
status.setRollbackOnly()
}Author.withTransaction { status ->
new Author(name: "Stephen King", age: 40).save()
}
Only the second author would be saved since the first transaction rolls back the author
save()
by clearing the Hibernate session. If the Hibernate session were not cleared then both author instances would be persisted and it would lead to very unexpected results.
It can, however, be frustrating to get
LazyInitializationException
s due to the session being cleared.
For example, consider the following example:
class AuthorService { void updateAge(id, int age) {
def author = Author.get(id)
author.age = age
if (author.isTooOld()) {
throw new AuthorException("too old", author)
}
}
}
class AuthorController { def authorService def updateAge() {
try {
authorService.updateAge(params.id, params.int("age"))
}
catch(e) {
render "Author books ${e.author.books}"
}
}
}
In the above example the transaction will be rolled back if the
Author
's age exceeds the maximum value defined in the
isTooOld()
method by throwing an
AuthorException
. The
AuthorException
references the author but when the
books
association is accessed a
LazyInitializationException
will be thrown because the underlying Hibernate session has been cleared.
To solve this problem you have a number of options. One is to ensure you query eagerly to get the data you will need:
class AuthorService {
…
void updateAge(id, int age) {
def author = Author.findById(id, [fetch:[books:"eager"]])
...
In this example the
books
association will be queried when retrieving the
Author
.
This is the optimal solution as it requires fewer queries then the following suggested solutions.
Another solution is to redirect the request after a transaction rollback:
class AuthorController { AuthorService authorService def updateAge() {
try {
authorService.updateAge(params.id, params.int("age"))
}
catch(e) {
flash.message "Can't update age"
redirect action:"show", id:params.id
}
}
}
In this case a new request will deal with retrieving the
Author
again. And, finally a third solution is to retrieve the data for the
Author
again to make sure the session remains in the correct state:
class AuthorController { def authorService def updateAge() {
try {
authorService.updateAge(params.id, params.int("age"))
}
catch(e) {
def author = Author.read(params.id)
render "Author books ${author.books}"
}
}
}
Validation Errors and Rollback
A common use case is to rollback a transaction if there are validation errors. For example consider this service:
import grails.validation.ValidationExceptionclass AuthorService { void updateAge(id, int age) {
def author = Author.get(id)
author.age = age
if (!author.validate()) {
throw new ValidationException("Author is not valid", author.errors)
}
}
}
To re-render the same view that a transaction was rolled back in you can re-associate the errors with a refreshed instance before rendering:
import grails.validation.ValidationExceptionclass AuthorController { def authorService def updateAge() {
try {
authorService.updateAge(params.id, params.int("age"))
}
catch (ValidationException e) {
def author = Author.read(params.id)
author.errors = e.errors
render view: "edit", model: [author:author]
}
}
}
12.2 Scoped Services
By default, access to service methods is not synchronised, so nothing prevents concurrent execution of those methods. In fact, because the service is a singleton and may be used concurrently, you should be very careful about storing state in a service. Or take the easy (and better) road and never store state in a service.
You can change this behaviour by placing a service in a particular scope. The supported scopes are:
prototype
- A new service is created every time it is injected into another class
request
- A new service will be created per request
flash
- A new service will be created for the current and next request only
flow
- In web flows the service will exist for the scope of the flow
conversation
- In web flows the service will exist for the scope of the conversation. ie a root flow and its sub flows
session
- A service is created for the scope of a user session
singleton
(default) - Only one instance of the service ever exists
If your service is flash
, flow
or conversation
scoped it must implement java.io.Serializable
and can only be used in the context of a Web Flow.
To enable one of the scopes, add a static scope property to your class whose value is one of the above, for example
For new Grails apps since 2.3, default controller scope is singleton
, resulting in prototype
scope services that are effectively per-controller singletons. If non-singleton services are required, controller scope should be changed as well.
12.3 Dependency Injection and Services
Dependency Injection Basics
A key aspect of Grails services is the ability to use
Spring Framework's dependency injection features. Grails supports "dependency injection by convention". In other words, you can use the property name representation of the class name of a service to automatically inject them into controllers, tag libraries, and so on.
As an example, given a service called
BookService
, if you define a property called
bookService
in a controller as follows:
class BookController {
def bookService
…
}
In this case, the Spring container will automatically inject an instance of that service based on its configured scope. All dependency injection is done by name. You can also specify the type as follows:
class AuthorService {
BookService bookService
}
NOTE: Normally the property name is generated by lower casing the first letter of the type. For example, an instance of the BookService
class would map to a property named bookService
.To be consistent with standard JavaBean conventions, if the first 2 letters of the class name are upper case, the property name is the same as the class name. For example, the property name of the JDBCHelperService
class would be JDBCHelperService
, not jDBCHelperService
or jdbcHelperService
.See section 8.8 of the JavaBean specification for more information on de-capitalization rules.
Dependency Injection and Services
You can inject services in other services with the same technique. If you had an
AuthorService
that needed to use the
BookService
, declaring the
AuthorService
as follows would allow that:
class AuthorService {
def bookService
}
Dependency Injection and Domain Classes / Tag Libraries
You can even inject services into domain classes and tag libraries, which can aid in the development of rich domain models and views:
class Book {
…
def bookService def buyBook() {
bookService.buyBook(this)
}
}
Service Bean Names
The default bean name which is associated with a service can be problematic if there are multiple services with the same name defined in different packages. For example consider the situation where an application defines a service class named
com.demo.ReportingService
and the application uses a plugin named
ReportingUtilities
and that plugin provides a service class named
com.reporting.util.ReportingService
. The default bean name for each of those would be
reportingService
so they would conflict with each other. Grails manages this by changing the default bean name for services provided by plugins by prefixing the bean name with the plugin name. In the scenario described above the
reportingService
bean would be an instance of the
com.demo.ReportingService
class defined in the application and the
reportingUtilitiesReportingService
bean would be an instance of the
com.reporting.util.ReportingService
class provided by the
ReportingUtilities
plugin. For all service beans provided by plugins, if there are no other services with the same name within the application or other plugins in the application then a bean alias will be created which does not include the plugin name and that alias points to the bean referred to by the name that does include the plugin name prefix. For example, if the
ReportingUtilities
plugin provides a service named
com.reporting.util.AuthorService
and there is no other
AuthorService
in the application or in any of the plugins that the application is using then there will be a bean named
reportingUtilitiesAuthorService
which is an instance of this
com.reporting.util.AuthorService
class and there will be a bean alias defined in the context named
authorService
which points to that same bean.
12.4 Using Services from Java
One of the powerful things about services is that since they encapsulate re-usable logic, you can use them from other classes, including Java classes. There are a couple of ways you can reuse a service from Java. The simplest way is to move your service into a package within the
grails-app/services
directory. The reason this is important is that it is not possible to import classes into Java from the default package (the package used when no package declaration is present). So for example the
BookService
below cannot be used from Java as it stands:
class BookService {
void buyBook(Book book) {
// logic
}
}
However, this can be rectified by placing this class in a package, by moving the class into a sub directory such as
grails-app/services/bookstore
and then modifying the package declaration:
package bookstoreclass BookService {
void buyBook(Book book) {
// logic
}
}
An alternative to packages is to instead have an interface within a package that the service implements:
package bookstoreinterface BookStore {
void buyBook(Book book)
}
And then the service:
class BookService implements bookstore.BookStore {
void buyBook(Book b) {
// logic
}
}
This latter technique is arguably cleaner, as the Java side only has a reference to the interface and not to the implementation class (although it's always a good idea to use packages). Either way, the goal of this exercise to enable Java to statically resolve the class (or interface) to use, at compile time.
Now that this is done you can create a Java class within the
src/java
directory and add a setter that uses the type and the name of the bean in Spring:
// src/java/bookstore/BookConsumer.java
package bookstore;public class BookConsumer { private BookStore store; public void setBookStore(BookStore storeInstance) {
this.store = storeInstance;
}
…
}
Once this is done you can configure the Java class as a Spring bean in
grails-app/conf/spring/resources.xml
(for more information see the section on
Grails and Spring):
<bean id="bookConsumer" class="bookstore.BookConsumer">
<property name="bookStore" ref="bookService" />
</bean>
or in
grails-app/conf/spring/resources.groovy
:
import bookstore.BookConsumerbeans = {
bookConsumer(BookConsumer) {
bookStore = ref("bookService")
}
}
13 Static Type Checking And Compilation
Groovy is a dynamic language and by default Groovy uses a dynamic dispatch mechanism to carry out method calls and property access. This dynamic dispatch mechanism provides a lot of flexibility and power to the language. For example, it is possible to dynamically add methods to classes at runtime and it is possible to dynamically replace existing methods at runtime. Features like these are important and provide a lot of power to the language. However, there are times when you may want to disable this dynamic dispatch in favor of a more static dispatch mechanism and Groovy provides a way to do that. The way to tell the Groovy compiler that a particular class should compiled statically is to mark the class with the
groovy.transform.CompileStatic annotation as shown below.
import groovy.transform.CompileStatic@CompileStatic
class MyClass { // this class will be statically compiled...}
See
these notes on Groovy static compilation for more details on how
CompileStatic
works and why you might want to use it.
One limitation of using
CompileStatic
is that when you use it you give up access to the power and flexibility offered by dynamic dispatch. For example, in Grails you would not be able to invoke a GORM dynamic finder from a class that is marked with
CompileStatic
because the compiler cannot verify that the dynamic finder method exists, because it doesn't exist at compile time. It may be that you want to take advantage of Groovy's static compilation benefits without giving up access to dynamic dispatch for Grails specific things like dynamic finders and this is where
grails.compiler.GrailsCompileStatic comes in.
GrailsCompileStatic
behaves just like
CompileStatic
but is aware of certain Grails features and allows access to those specific features to be accessed dynamically.
13.1 The GrailsCompileStatic Annotation
GrailsCompileStatic
The
GrailsCompileStatic
annotation may be applied to a class or methods within a class.
import grails.compiler.GrailsCompileStatic@GrailsCompileStatic
class SomeClass { // all of the code in this class will be statically compiled def methodOne() {
// …
} def methodTwo() {
// …
} def methodThree() {
// …
}
}
import grails.compiler.GrailsCompileStaticclass SomeClass { // methodOne and methodThree will be statically compiled
// methodTwo will be dynamically compiled @GrailsCompileStatic
def methodOne() {
// …
} def methodTwo() {
// …
} @GrailsCompileStatic
def methodThree() {
// …
}
}
It is possible to mark a class with
GrailsCompileStatic
and exclude specific methods by marking them with
GrailsCompileStatic
and specifying that the type checking should be skipped for that particular method as shown below.
import grails.compiler.GrailsCompileStatic
import groovy.transform.TypeCheckingMode@GrailsCompileStatic
class SomeClass { // methodOne and methodThree will be statically compiled
// methodTwo will be dynamically compiled def methodOne() {
// …
} @GrailsCompileStatic(TypeCheckingMode.SKIP)
def methodTwo() {
// …
} def methodThree() {
// …
}
}
Code that is marked with
GrailsCompileStatic
will all be statically compiled except for Grails specific interactions that cannot be statically compiled but that
GrailsCompileStatic
can identify as permissible for dynamic dispatch. These include things like invoking dynamic finders and DSL code in configuration blocks like constraints and mapping closures in domain classes.
Care must be taken when deciding to statically compile code. There are benefits associated with static compilation but in order to take advantage of those benefits you are giving up the power and flexibility of dynamic dispatch. For example if code is statically compiled it cannot take advantage of runtime metaprogramming enhancements which may be provided by plugins.
13.2 The GrailsTypeChecked Annotation
GrailsTypeChecked
The
grails.compiler.GrailsTypeChecked annotation works a lot like the
GrailsCompileStatic
annotation except that it only enables static type checking, not static compilation. This affords compile time feedback for expressions which cannot be validated statically at compile time while still leaving dynamic dispatch in place for the class.
import grails.compiler.GrailsTypeChecked@GrailsTypeChecked
class SomeClass { // all of the code in this class will be statically type
// checked and will be dynamically dispatched at runtime def methodOne() {
// …
} def methodTwo() {
// …
} def methodThree() {
// …
}
}
14 Testing
Automated testing is a key part of Grails. Hence, Grails provides many ways to making testing easier from low level unit testing to high level functional tests. This section details the different capabilities that Grails offers for testing.
Grails 1.3.x and below used the grails.test.GrailsUnitTestCase
class hierarchy for testing in a JUnit 3 style. Grails 2.0.x and above deprecates these test harnesses in favour of mixins that can be applied to a range of different kinds of tests (JUnit 3, JUnit 4, Spock etc.) without subclassing
The first thing to be aware of is that all of the
create-*
and
generate-*
commands create
unit
or
integration
tests automatically. For example if you run the
create-controller command as follows:
grails create-controller com.acme.app.simple
Grails will create a controller at
grails-app/controllers/com/acme/app/SimpleController.groovy
, and also a unit test at
test/unit/com/acme/app/SimpleControllerTests.groovy
. What Grails won't do however is populate the logic inside the test! That is left up to you.
The default class name suffix is Tests
but as of Grails 1.2.2, the suffix of Test
is also supported.
Running Tests
Tests are run with the
test-app command:
The command will produce output such as:
-------------------------------------------------------
Running Unit Tests…
Running test FooTests...FAILURE
Unit Tests Completed in 464ms …
-------------------------------------------------------Tests failed: 0 errors, 1 failures
whilst showing the reason for each test failure.
You can force a clean before running tests by passing -clean
to the test-app
command.
Grails writes both plain text and HTML test reports to the
target/test-reports
directory, along with the original XML files. The HTML reports are generally the best ones to look at.
Using Grails'
interactive mode confers some distinct advantages when executing tests. First, the tests will execute significantly faster on the second and subsequent runs. Second, a shortcut is available to open the HTML reports in your browser:
You can also run your unit tests from within most IDEs.
Targeting Tests
You can selectively target the test(s) to be run in different ways. To run all tests for a controller named
SimpleController
you would run:
grails test-app SimpleController
This will run any tests for the class named
SimpleController
. Wildcards can be used...
grails test-app *Controller
This will test all classes ending in
Controller
. Package names can optionally be specified...
grails test-app some.org.*Controller
or to run all tests in a package...
grails test-app some.org.*
or to run all tests in a package including subpackages...
grails test-app some.org.**.*
You can also target particular test methods...
grails test-app SimpleController.testLogin
This will run the
testLogin
test in the
SimpleController
tests. You can specify as many patterns in combination as you like...
grails test-app some.org.* SimpleController.testLogin BookController
Targeting Test Types and/or Phases
In addition to targeting certain tests, you can also target test
types and/or
phases by using the
phase:type
syntax.
Grails organises tests by phase and by type. A test phase relates to the state of the Grails application during the tests, and the type relates to the testing mechanism.Grails comes with support for 4 test phases (unit
, integration
, functional
and other
) and JUnit test types for the unit
and integration
phases. These test types have the same name as the phase.Testing plugins may provide new test phases or new test types for existing phases. Refer to the plugin documentation.
To execute the JUnit
integration
tests you can run:
grails test-app integration:integration
Both
phase
and
type
are optional. Their absence acts as a wildcard. The following command will run all test types in the
unit
phase:
The Grails
Spock Plugin is one plugin that adds new test types to Grails. It adds a
spock
test type to the
unit
,
integration
and
functional
phases. To run all spock tests in all phases you would run the following:
To run the all of the spock tests in the
functional
phase you would run...
grails test-app functional:spock
More than one pattern can be specified...
grails test-app unit:spock integration:spock
Targeting Tests in Types and/or Phases
Test and type/phase targetting can be applied at the same time:
grails test-app integration: unit: some.org.**.*
This would run all tests in the
integration
and
unit
phases that are in the package
some.org
or a subpackage.
14.1 Unit Testing
Unit testing are tests at the "unit" level. In other words you are testing individual methods or blocks of code without consideration for surrounding infrastructure. Unit tests are typically run without the presence of physical resources that involve I/O such databases, socket connections or files. This is to ensure they run as quick as possible since quick feedback is important.
The Test Mixins
Since Grails 2.0, a collection of unit testing mixins is provided by Grails that lets you enhance the behavior of a typical JUnit 3, JUnit 4 or Spock test. The following sections cover the usage of these mixins.
The previous JUnit 3-style GrailsUnitTestCase
class hierarchy is still present in Grails for backwards compatibility, but is now deprecated. The previous documentation on the subject can be found in the Grails 1.3.x documentation
You won't normally have to import any of the testing classes because Grails does that for you. But if you find that your IDE for example can't find the classes, here they all are:
grails.test.mixin.TestFor
grails.test.mixin.Mock
grails.test.mixin.TestMixin
grails.test.mixin.support.GrailsUnitTestMixin
grails.test.mixin.domain.DomainClassUnitTestMixin
grails.test.mixin.services.ServiceUnitTestMixin
grails.test.mixin.web.ControllerUnitTestMixin
grails.test.mixin.web.FiltersUnitTestMixin
grails.test.mixin.web.GroovyPageUnitTestMixin
grails.test.mixin.web.UrlMappingsUnitTestMixin
grails.test.mixin.hibernate.HibernateTestMixin
Note that you're only ever likely to use the first two explicitly. The rest are there for reference.
Test Mixin Basics
Most testing can be achieved via the
TestFor
annotation in combination with the
Mock
annotation for mocking collaborators. For example, to test a controller and associated domains you would define the following:
@TestFor(BookController)
@Mock([Book, Author, BookService])
The
TestFor
annotation defines the class under test and will automatically create a field for the type of class under test. For example in the above case a "controller" field will be present, however if
TestFor
was defined for a service a "service" field would be created and so on.
The
Mock
annotation creates mock version of any collaborators. There is an in-memory implementation of GORM that will simulate most interactions with the GORM API. For those interactions that are not automatically mocked you can use the built in support for defining mocks and stubs programmatically. For example:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(BookController)
@Mock(Book)
class BookControllerSpec extends Specification { void "test search"() {
given:
def searchMock = mockFor(SearchService)
searchMock.demand.searchWeb { String q -> ['first result', 'second result'] }
searchMock.demand.static.logResults { List results -> }
controller.searchService = searchMock.createMock() when:
controller.search() then:
controller.response.text.contains "Found 2 results"
}
}
doWithSpring and doWithConfig callback methods, FreshRuntime annotation
The
doWithSpring
callback method can be used to add beans with the BeanBuilder DSL. There is the
doWithConfig
callback method for changing the grailsApplication.config values before the grailsApplication instance of the test runtime gets initialized.
import grails.test.mixin.support.GrailsUnitTestMixinimport org.junit.ClassRule
import org.junit.rules.TestRuleimport spock.lang.Ignore;
import spock.lang.IgnoreRest
import spock.lang.Shared;
import spock.lang.Specification@TestMixin(GrailsUnitTestMixin)
class StaticCallbacksSpec extends Specification {
static doWithSpring = {
myService(MyService)
} static doWithConfig(c) {
c.myConfigValue = 'Hello'
} def "grailsApplication is not null"() {
expect:
grailsApplication != null
} def "doWithSpring callback is executed"() {
expect:
grailsApplication.mainContext.getBean('myService') != null
} def "doWithConfig callback is executed"(){
expect:
config.myConfigValue == 'Hello'
}
}
You can also use these callbacks without "static" together with the
grails.test.runtime.FreshRuntime
annotation.
In this case, a clean application context and grails application instance is initialized for each test method call.
import grails.test.mixin.support.GrailsUnitTestMixin
import grails.test.runtime.FreshRuntime;import org.junit.ClassRule
import org.junit.rules.TestRuleimport spock.lang.Ignore;
import spock.lang.IgnoreRest
import spock.lang.Shared;
import spock.lang.Specification@FreshRuntime
@TestMixin(GrailsUnitTestMixin)
class TestInstanceCallbacksSpec extends Specification {
def doWithSpring = {
myService(MyService)
} def doWithConfig(c) {
c.myConfigValue = 'Hello'
} def "grailsApplication is not null"() {
expect:
grailsApplication != null
} def "doWithSpring callback is executed"() {
expect:
grailsApplication.mainContext.getBean('myService') != null
} def "doWithConfig callback is executed"(){
expect:
config.myConfigValue == 'Hello'
}
}
You can use
org.codehaus.groovy.grails.commons.InstanceFactoryBean
together with doWithSpring and the
FreshRuntime annotation to mock beans in tests.
import grails.test.mixin.support.GrailsUnitTestMixin
import grails.test.runtime.FreshRuntimeimport org.codehaus.groovy.grails.commons.InstanceFactoryBean
import org.junit.ClassRuleimport spock.lang.Shared
import spock.lang.Specification@FreshRuntime
@TestMixin(GrailsUnitTestMixin)
class MockedBeanSpec extends Specification {
def myService=Mock(MyService) def doWithSpring = {
myService(InstanceFactoryBean, myService, MyService)
} def "doWithSpring callback is executed"() {
when:
def myServiceBean=grailsApplication.mainContext.getBean('myService')
myServiceBean.prova()
then:
1 * myService.prova() >> { true }
}
}
The DirtiesRuntime annotation
Test methods may be marked with the
grails.test.runtime.DirtiesRuntime
annotation to indicate that the test modifies the runtime in ways which might be problematic for other tests and as such the runtime should be refreshed after this test method runs.
import grails.test.mixin.TestFor
import spock.lang.Specification
import grails.test.runtime.DirtiesRuntime@TestFor(PersonController)
class PersonControllerSpec extends Specification { @DirtiesRuntime
void "a test method which modifies the runtime"() {
when:
Person.metaClass.someMethod = { … }
// ... then:
// …
} void "a test method which should not be affected by the previous test method"() {
// …
}
}
Sharing test runtime grailsApplication instance and beans for several test classes
It's possible to share a single grailsApplication instance and beans for several test classes.
This feature is activated by the
SharedRuntime
annotation. This annotation takes an optional class parameter
implements
SharedRuntimeConfigurer
interface. All test classes referencing the same SharedRuntimeConfigurer implementation
class will share the same runtime during a single test run.
The value class for SharedRuntimeConfigurer annotation can also implement
TestEventInterceptor
. In this case the instance of the class
will be registered as a test event interceptor for the test runtime.
Loading application beans in unit tests
Adding
static loadExternalBeans = true
field definition to a unit test class makes the Grails unit test runtime load all bean definitions from
grails-app/conf/spring/resources.groovy
and
grails-app/conf/spring/resources.xml
files.
import spock.lang.Issue
import spock.lang.Specification
import grails.test.mixin.support.GrailsUnitTestMixin@TestMixin(GrailsUnitTestMixin)
class LoadExternalBeansSpec extends Specification {
static loadExternalBeans = true void "should load external beans"(){
expect:
applicationContext.getBean('simpleBean') == 'Hello world!'
}
}
14.1.1 Unit Testing Controllers
The Basics
You use the
grails.test.mixin.TestFor
annotation to unit test controllers. Using
TestFor
in this manner activates the
grails.test.mixin.web.ControllerUnitTestMixin
and its associated API. For example:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void "test something"() {
}
}
Adding the
TestFor
annotation to a controller causes a new
controller
field to be automatically created for the controller under test.
The TestFor
annotation will also automatically annotate any public methods starting with "test" with JUnit 4's @Test annotation. If any of your test method don't start with "test" just add this manually
To test the simplest "Hello World"-style example you can do the following:
// Test class
class SimpleController {
def hello() {
render "hello"
}
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void "test hello"() {
when:
controller.hello() then:
response.text == 'hello'
}
}
The
response
object is an instance of
GrailsMockHttpServletResponse
(from the package
org.codehaus.groovy.grails.plugins.testing
) which extends Spring's
MockHttpServletResponse
class and has a number of useful methods for inspecting the state of the response.
For example to test a redirect you can use the
redirectedUrl
property:
class SimpleController {
def index() {
redirect action: 'hello'
}
…
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test index'() {
when:
controller.index() then:
response.redirectedUrl == '/simple/hello'
}
}
Many actions make use of the parameter data associated with the request. For example, the 'sort', 'max', and 'offset' parameters are quite common. Providing these in the test is as simple as adding appropriate values to a special
params
variable:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(PersonController)
class PersonControllerSpec extends Specification { void 'test list'() {
when:
params.sort = 'name'
params.max = 20
params.offset = 0
controller.list() then:
// …
}
}
You can even control what type of request the controller action sees by setting the
method
property of the mock request:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(PersonController)
class PersonControllerSpec extends Specification { void 'test save'() {
when:
request.method = 'POST'
controller.save() then:
// …
}
}
This is particularly important if your actions do different things depending on the type of the request. Finally, you can mark a request as AJAX like so:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(PersonController)
class PersonControllerSpec extends Specification { void 'test list'() {
when:
request.method = 'POST'
request.makeAjaxRequest()
controller.getPage() then:
// …
}
}
You only need to do this though if the code under test uses the
xhr
property on the request.
Testing View Rendering
To test view rendering you can inspect the state of the controller's
modelAndView
property (an instance of
org.springframework.web.servlet.ModelAndView
) or you can use the
view
and
model
properties provided by the mixin:
class SimpleController {
def home() {
render view: "homePage", model: [title: "Hello World"]
}
…
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test home'() {
when:
controller.home() then:
view == '/simple/homePage'
model.title == 'Hello World'
}
}
Note that the view string is the absolute view path, so it starts with a '/' and will include path elements, such as the directory named after the action's controller.
Testing Template Rendering
Unlike view rendering, template rendering will actually attempt to write the template directly to the response rather than returning a
ModelAndView
hence it requires a different approach to testing.
Consider the following controller action:
class SimpleController {
def display() {
render template:"snippet"
}
}
In this example the controller will look for a template in
grails-app/views/simple/_snippet.gsp
. You can test this as follows:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test display'() {
when:
controller.display() then:
response.text == 'contents of the template'
}
}
However, you may not want to render the real template, but just test that is was rendered. In this case you can provide mock Groovy Pages:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test display with mock template'() {
when:
views['/simple/_snippet.gsp'] = 'mock template contents'
controller.display() then:
response.text == 'mock template contents'
}
}
Testing Actions Which Return A Map
When a controller action returns a
java.util.Map
that
Map
may be inspected directly to assert that it contains the expected data:
class SimpleController {
def showBookDetails() {
[title: 'The Nature Of Necessity', author: 'Alvin Plantinga']
}
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test show book details'() {
when:
def model = controller.showBookDetails() then:
model.author == 'Alvin Plantinga'
}
}
Testing XML and JSON Responses
XML and JSON response are also written directly to the response. Grails' mocking capabilities provide some conveniences for testing XML and JSON response. For example consider the following action:
def renderXml() {
render(contentType:"text/xml") {
book(title:"Great")
}
}
This can be tested using the
xml
property of the response:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test render xml'() {
when:
controller.renderXml() then:
response.text == "<book title='Great'/>"
response.xml.@title.text() == 'Great'
}
}
The
xml
property is a parsed result from Groovy's
XmlSlurper class which is very convenient for parsing XML.
Testing JSON responses is pretty similar, instead you use the
json
property:
// controller action
def renderJson() {
render(contentType:"application/json") {
book = "Great"
}
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test render json'() {
when:
controller.renderJson() then:
response.text == '{"book":"Great"}'
response.json.book == 'Great'
}
}
The
json
property is an instance of
org.codehaus.groovy.grails.web.json.JSONElement
which is a map-like structure that is useful for parsing JSON responses.
Testing XML and JSON Requests
Grails provides various convenient ways to automatically parse incoming XML and JSON packets. For example you can bind incoming JSON or XML requests using Grails' data binding:
def consumeBook(Book b) {
render "The title is ${b.title}."
}
To test this Grails provides an easy way to specify an XML or JSON packet via the
xml
or
json
properties. For example the above action can be tested by specifying a String containing the XML:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
@Mock([Book])
class SimpleControllerSpec extends Specification {
void 'test consume book xml'() {
when:
request.xml = '<book><title>Wool</title></book>'
controller.consumeBook() then:
response.text == 'The title is Wool.'
}
}
Or alternatively a domain instance can be specified and it will be auto-converted into the appropriate XML request:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
@Mock([Book])
class SimpleControllerSpec extends Specification { void 'test consume book xml'() {
when:
request.xml = new Book(title: 'Shift')
controller.consumeBook() then:
response.text == 'The title is Shift.'
}
}
The same can be done for JSON requests:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
@Mock([Book])
class SimpleControllerSpec extends Specification { void 'test consume book json'() {
when:
request.json = new Book(title: 'Shift')
controller.consumeBook() then:
response.text == 'The title is Shift.'
}
}
If you prefer not to use Grails' data binding but instead manually parse the incoming XML or JSON that can be tested too. For example consider the controller action below:
def consume() {
request.withFormat {
xml {
render "The XML Title Is ${request.XML.@title}."
}
json {
render "The JSON Title Is ${request.JSON.title}."
}
}
}
To test the XML request you can specify the XML as a string:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test consume xml'() {
when:
request.xml = '<book title="The Stand"/>'
controller.consume() then:
response.text == 'The XML Title Is The Stand.'
} void 'test consume json'() {
when:
request.json = '{title:"The Stand"}'
controller.consume() then:
response.text == 'The JSON Title Is The Stand.'
}
}
Testing Mime Type Handling
You can test mime type handling and the
withFormat
method quite simply by setting the request's
contentType
attribute:
// controller action
def sayHello() {
def data = [Hello:"World"]
request.withFormat {
xml { render data as grails.converters.XML }
json { render data as grails.converters.JSON }
html data
}
}
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test say hello xml'() {
when:
request.contentType = 'application/xml'
controller.sayHello() then:
response.text == '<?xml version="1.0" encoding="UTF-8"?><map><entry key="Hello">World</entry></map>'
} void 'test say hello json'() {
when:
request.contentType = 'application/json'
controller.sayHello() then:
response.text == '{"Hello":"World"}'
}
}
There are constants provided by
ControllerUnitTestMixin
for all of the common common content types as shown below:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test say hello xml'() {
when:
request.contentType = XML_CONTENT_TYPE
controller.sayHello() then:
response.text == '<?xml version="1.0" encoding="UTF-8"?><map><entry key="Hello">World</entry></map>'
} void 'test say hello json'() {
when:
request.contentType = JSON_CONTENT_TYPE
controller.sayHello() then:
response.text == '{"Hello":"World"}'
}
}
The defined constants are listed below:
Constant | Value |
---|
ALL_CONTENT_TYPE | */* |
FORM_CONTENT_TYPE | application/x-www-form-urlencoded |
MULTIPART_FORM_CONTENT_TYPE | multipart/form-data |
HTML_CONTENT_TYPE | text/html |
XHTML_CONTENT_TYPE | application/xhtml+xml |
XML_CONTENT_TYPE | application/xml |
JSON_CONTENT_TYPE | application/json |
TEXT_XML_CONTENT_TYPE | text/xml |
TEXT_JSON_CONTENT_TYPE | text/json |
HAL_JSON_CONTENT_TYPE | application/hal+json |
HAL_XML_CONTENT_TYPE | application/hal+xml |
ATOM_XML_CONTENT_TYPE | application/atom+xml |
Testing Duplicate Form Submissions
Testing duplicate form submissions is a little bit more involved. For example if you have an action that handles a form such as:
def handleForm() {
withForm {
render "Good"
}.invalidToken {
render "Bad"
}
}
you want to verify the logic that is executed on a good form submission and the logic that is executed on a duplicate submission. Testing the bad submission is simple. Just invoke the controller:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test duplicate form submission'() {
when:
controller.handleForm() then:
response.text == 'Bad'
}
}
Testing the successful submission requires providing an appropriate
SynchronizerToken
:
import grails.test.mixin.TestFor
import spock.lang.Specificationimport org.codehaus.groovy.grails.web.servlet.mvc.SynchronizerTokensHolder@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test valid form submission'() {
when:
def tokenHolder = SynchronizerTokensHolder.store(session) params[SynchronizerTokensHolder.TOKEN_URI] = '/controller/handleForm'
params[SynchronizerTokensHolder.TOKEN_KEY] = tokenHolder.generateToken(params[SynchronizerTokensHolder.TOKEN_URI])
controller.handleForm() then:
response.text == 'Good'
}
}
If you test both the valid and the invalid request in the same test be sure to reset the response between executions of the controller:
import grails.test.mixin.TestFor
import spock.lang.Specificationimport org.codehaus.groovy.grails.web.servlet.mvc.SynchronizerTokensHolder@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test form submission'() {
when:
controller.handleForm() then:
response.text == 'Bad' when:
response.reset()
def tokenHolder = SynchronizerTokensHolder.store(session) params[SynchronizerTokensHolder.TOKEN_URI] = '/controller/handleForm'
params[SynchronizerTokensHolder.TOKEN_KEY] = tokenHolder.generateToken(params[SynchronizerTokensHolder.TOKEN_URI])
controller.handleForm() then:
response.text == 'Good'
}
}
Testing File Upload
You use the
GrailsMockMultipartFile
class to test file uploads. For example consider the following controller action:
def uploadFile() {
MultipartFile file = request.getFile("myFile")
file.transferTo(new File("/local/disk/myFile"))
}
To test this action you can register a
GrailsMockMultipartFile
with the request:
import grails.test.mixin.TestFor
import spock.lang.Specificationimport org.codehaus.groovy.grails.plugins.testing.GrailsMockMultipartFile@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test file upload'() {
when:
def file = new GrailsMockMultipartFile('myFile', 'some file contents'.bytes)
request.addFile file
controller.uploadFile() then:
file.targetFileLocation.path == '/local/disk/myFile'
}
}
The
GrailsMockMultipartFile
constructor arguments are the name and contents of the file. It has a mock implementation of the
transferTo
method that simply records the
targetFileLocation
and doesn't write to disk.
Testing Command Objects
Special support exists for testing command object handling with the
mockCommandObject
method. For example consider the following action:
class SimpleController {
def handleCommand(SimpleCommand simple) {
if(simple.hasErrors()) {
render 'Bad'
} else {
render 'Good'
}
}
}class SimpleCommand {
String name static constraints = {
name blank: false
}
}
To test this you mock the command object, populate it and then validate it as follows:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test valid command object'() {
given:
def simpleCommand = new SimpleCommand(name: 'Hugh')
simpleCommand.validate() when:
controller.handleCommand(simpleCommand) then:
response.text == 'Good'
} void 'test invalid command object'() {
given:
def simpleCommand = new SimpleCommand(name: '')
simpleCommand.validate() when:
controller.handleCommand(simpleCommand) then:
response.text == 'Bad'
}
}
The testing framework also supports allowing Grails to create the command object instance automatically. To test this invoke the no-arg version of the controller action method. Grails will create an instance of the command object, perform data binding on it using the request parameters and validate the object just like it does in when the application is running. See the test below.
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test valid command object'() {
when:
params.name = 'Hugh'
controller.handleCommand() then:
response.text == 'Good'
} void 'test invalid command object'() {
when:
params.name = ''
controller.handleCommand() then:
response.text == 'Bad'
}
}
Testing allowedMethods
The unit testing environment respects the
allowedMethods property in controllers. If a controller action is limited to be accessed with certain request methods, the unit test must be constructed to deal with that.
// grails-app/controllers/com/demo/DemoController.groovypackage com.democlass DemoController { static allowedMethods = [save: 'POST', update: 'PUT', delete: 'DELETE'] def save() {
render 'Save was successful!'
} // …
}
// test/unit/com/demo/DemoControllerSpec.groovy
package com.demoimport grails.test.mixin.TestFor
import spock.lang.Specification
import static javax.servlet.http.HttpServletResponse.*@TestFor(DemoController)
class DemoControllerSpec extends Specification { void "test a valid request method"() {
when:
request.method = 'POST'
controller.save() then:
response.status == SC_OK
response.text == 'Save was successful!'
} void "test an invalid request method"() {
when:
request.method = 'DELETE'
controller.save() then:
response.status == SC_METHOD_NOT_ALLOWED
}
}
Testing Calling Tag Libraries
You can test calling tag libraries using
ControllerUnitTestMixin
, although the mechanism for testing the tag called varies from tag to tag. For example to test a call to the
message
tag, add a message to the
messageSource
. Consider the following action:
def showMessage() {
render g.message(code: "foo.bar")
}
This can be tested as follows:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
class SimpleControllerSpec extends Specification { void 'test render message tag'() {
given:
messageSource.addMessage 'foo.bar', request.locale, 'Hello World' when:
controller.showMessage() then:
response.text == 'Hello World'
}
}
See
unit testing tag libraries for more information.
14.1.2 Unit Testing Tag Libraries
The Basics
Tag libraries and GSP pages can be tested with the
grails.test.mixin.web.GroovyPageUnitTestMixin
mixin. To use the mixin declare which tag library is under test with the
TestFor
annotation:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleTagLib)
class SimpleTagLibSpec extends Specification { void "test something"() {
}
}
Adding the
TestFor
annotation to a TagLib class causes a new
tagLib
field to be automatically created for the TagLib class under test.
The tagLib field can be used to test calling tags as function calls. The return value of a function call is either a
StreamCharBuffer instance or
the object returned from the tag closure when
returnObjectForTags feature is used.
Note that if you are testing invocation of a custom tag from a controller you can combine the
ControllerUnitTestMixin
and the
GroovyPageUnitTestMixin
using the
Mock
annotation:
import spock.lang.Specification@TestFor(SimpleController)
@Mock(SimpleTagLib)
class SimpleControllerSpec extends Specification {}
Testing Custom Tags
The core Grails tags don't need to be enabled during testing, however custom tag libraries do. The
GroovyPageUnitTestMixin
class provides a
mockTagLib()
method that you can use to mock a custom tag library. For example consider the following tag library:
class SimpleTagLib { static namespace = 's' def hello = { attrs, body ->
out << "Hello ${attrs.name ?: 'World'}"
} def bye = { attrs, body ->
out << "Bye ${attrs.author.name ?: 'World'}"
}
}
You can test this tag library by using
TestFor
and supplying the name of the tag library:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleTagLib)
class SimpleTagLibSpec extends Specification { void "test hello tag"() {
expect:
applyTemplate('<s:hello />') == 'Hello World'
applyTemplate('<s:hello name="Fred" />') == 'Hello Fred'
applyTemplate('<s:bye author="${author}" />', [author: new Author(name: 'Fred')]) == 'Bye Fred'
} void "test tag calls"() {
expect:
tagLib.hello().toString() == 'Hello World'
tagLib.hello(name: 'Fred').toString() == 'Hello Fred'
tagLib.bye(author: new Author(name: 'Fred')).toString == 'Bye Fred'
}
}
Alternatively, you can use the
TestMixin
annotation and mock multiple tag libraries using the
mockTagLib()
method:
import spock.lang.Specification
import grails.test.mixin.TestMixin
import grails.test.mixin.web.GroovyPageUnitTestMixin@TestMixin(GroovyPageUnitTestMixin)
class MultipleTagLibSpec extends Specification { void "test multiple tags"() {
given:
mockTagLib(SomeTagLib)
mockTagLib(SomeOtherTagLib) expect:
// …
}
}
The
GroovyPageUnitTestMixin
provides convenience methods for asserting that the template output equals or matches an expected value.
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleTagLib)
class SimpleTagLibSpec extends Specification { void "test hello tag"() {
expect:
assertOutputEquals ('Hello World', '<s:hello />')
assertOutputMatches (/.*Fred.*/, '<s:hello name="Fred" />')
}
}
Testing View and Template Rendering
You can test rendering of views and templates in
grails-app/views
via the
render(Map)
method provided by
GroovyPageUnitTestMixin
:
import spock.lang.Specification
import grails.test.mixin.TestMixin
import grails.test.mixin.web.GroovyPageUnitTestMixin@TestMixin(GroovyPageUnitTestMixin)
class RenderingSpec extends Specification { void "test rendering template"() {
when:
def result = render(template: '/simple/hello') then:
result == 'Hello World!'
}
}
This will attempt to render a template found at the location
grails-app/views/simple/_hello.gsp
. Note that if the template depends on any custom tag libraries you need to call
mockTagLib
as described in the previous section.
Some core tags use the active controller and action as input. In GroovyPageUnitTestMixin tests, you can manually set the active controller and action name by setting controllerName and actionName properties on the webRequest object:
webRequest.controllerName = 'simple'
webRequest.actionName = 'hello'
14.1.3 Unit Testing Domains
Overview
Domain class interaction can be tested without involving a real database connection using
DomainClassUnitTestMixin
or by using the
HibernateTestMixin
.
The GORM implementation in DomainClassUnitTestMixin is using a simple in-memory
ConcurrentHashMap
implementation. Note that this has limitations compared to a real GORM implementation.
A large, commonly-used portion of the GORM API can be mocked using
DomainClassUnitTestMixin
including:
- Simple persistence methods like
save()
, delete()
etc.
- Dynamic Finders
- Named Queries
- Query-by-example
- GORM Events
HibernateTestMixin
uses Hibernate 4 and a H2 in-memory database. This makes it possible to use all GORM features also in Grails unit tests.
All features of GORM for Hibernate can be tested within a
HibernateTestMixin
unit test including:
- String-based HQL queries
- composite identifiers
- dirty checking methods
- any direct interaction with Hibernate
The implementation behind
HibernateTestMixin
takes care of setting up the Hibernate with the in-memory H2 database. It only configures the given domain classes for use in a unit test. The @Domain annotation is used to tell which domain classes should be configured.
DomainClassUnitTestMixin Basics
DomainClassUnitTestMixin
is typically used in combination with testing either a controller, service or tag library where the domain is a mock collaborator defined by the
Mock
annotation:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(BookController)
@Mock(Book)
class BookControllerSpec extends Specification {
// …
}
The example above tests the
SimpleController
class and mocks the behavior of the
Simple
domain class as well. For example consider a typical scaffolded
save
controller action:
class BookController {
def save() {
def book = new Book(params)
if (book.save(flush: true)) {
flash.message = message(
code: 'default.created.message',
args: [message(code: 'book.label', default: 'Book'), book.id])
redirect(action: "show", id: book.id)
}
else {
render(view: "create", model: [bookInstance: book])
}
}
}
Tests for this action can be written as follows:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(BookController)
@Mock(Book)
class BookControllerSpec extends Specification {
void "test saving an invalid book"() {
when:
controller.save() then:
model.bookInstance != null
view == '/book/create'
} void "test saving a valid book"() {
when:
params.title = "The Stand"
params.pages = "500" controller.save() then:
response.redirectedUrl == '/book/show/1'
flash.message != null
Book.count() == 1
}
}
Mock
annotation also supports a list of mock collaborators if you have more than one domain to mock:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(BookController)
@Mock([Book, Author])
class BookControllerSpec extends Specification {
// …
}
Alternatively you can also use the
DomainClassUnitTestMixin
directly with the
TestMixin
annotation and then call the
mockDomain
method to mock domains during your test:
import grails.test.mixin.TestFor
import grails.test.mixin.TestMixin
import spock.lang.Specification
import grails.test.mixin.domain.DomainClassUnitTestMixin@TestFor(BookController)
@TestMixin(DomainClassUnitTestMixin)
class BookControllerSpec extends Specification { void setupSpec() {
mockDomain(Book)
} void "test saving an invalid book"() {
when:
controller.save() then:
model.bookInstance != null
view == '/book/create'
} void "test saving a valid book"() {
when:
params.title = "The Stand"
params.pages = "500" controller.save() then:
response.redirectedUrl == '/book/show/1'
flash.message != null
Book.count() == 1
}
}
The
mockDomain
method also includes an additional parameter that lets you pass a Map of Maps to configure a domain, which is useful for fixture-like data:
mockDomain(Book, [
[title: "The Stand", pages: 1000],
[title: "The Shining", pages: 400],
[title: "Along Came a Spider", pages: 300] ])
Testing Constraints
There are 4 types of validateable classes:
- Domain classes
- Classes marked with the
Validateable
annotation
- Command Objects which have been made validateable automatically
- Classes configured to be validateable via the
grails.validateable.classes
property in Config.groovy
The first 3 are easily testable in a unit test with no special configuration necessary as long as the test method is marked with
TestFor
or explicitly applies the
GrailsUnitTestMixin
using
TestMixin
. See the examples below.
// src/groovy/com/demo/MyValidateable.groovy
package com.demo@grails.validation.Validateable
class MyValidateable {
String name
Integer age static constraints = {
name matches: /[A-Z].*/
age range: 1..99
}
}
// grails-app/domain/com/demo/Person.groovy
package com.democlass Person {
String name static constraints = {
name matches: /[A-Z].*/
}
}
// grails-app/controllers/com/demo/DemoController.groovy
package com.democlass DemoController { def addItems(MyCommandObject co) {
if(co.hasErrors()) {
render 'something went wrong'
} else {
render 'items have been added'
}
}
}class MyCommandObject {
Integer numberOfItems static constraints = {
numberOfItems range: 1..10
}
}
// test/unit/com/demo/PersonSpec.groovy
package com.demoimport grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(Person)
class PersonSpec extends Specification { void "Test that name must begin with an upper case letter"() {
when: 'the name begins with a lower letter'
def p = new Person(name: 'jeff') then: 'validation should fail'
!p.validate() when: 'the name begins with an upper case letter'
p = new Person(name: 'Jeff') then: 'validation should pass'
p.validate()
}
}
// test/unit/com/demo/DemoControllerSpec.groovy
package com.demoimport grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(DemoController)
class DemoControllerSpec extends Specification { void 'Test an invalid number of items'() {
when:
params.numberOfItems = 42
controller.addItems() then:
response.text == 'something went wrong'
} void 'Test a valid number of items'() {
when:
params.numberOfItems = 8
controller.addItems() then:
response.text == 'items have been added'
}
}
// test/unit/com/demo/MyValidateableSpec.groovy
package com.demoimport grails.test.mixin.TestMixin
import grails.test.mixin.support.GrailsUnitTestMixin
import spock.lang.Specification
@TestMixin(GrailsUnitTestMixin)
class MyValidateableSpec extends Specification { void 'Test validate can be invoked in a unit test with no special configuration'() {
when: 'an object is valid'
def validateable = new MyValidateable(name: 'Kirk', age: 47) then: 'validate() returns true and there are no errors'
validateable.validate()
!validateable.hasErrors()
validateable.errors.errorCount == 0 when: 'an object is invalid'
validateable.name = 'kirk' then: 'validate() returns false and the appropriate error is created'
!validateable.validate()
validateable.hasErrors()
validateable.errors.errorCount == 1
validateable.errors['name'].code == 'matches.invalid' when: 'the clearErrors() is called'
validateable.clearErrors() then: 'the errors are gone'
!validateable.hasErrors()
validateable.errors.errorCount == 0 when: 'the object is put back in a valid state'
validateable.name = 'Kirk' then: 'validate() returns true and there are no errors'
validateable.validate()
!validateable.hasErrors()
validateable.errors.errorCount == 0
}
}
// test/unit/com/demo/MyCommandObjectSpec.groovy
package com.demoimport grails.test.mixin.TestMixin
import grails.test.mixin.support.GrailsUnitTestMixin
import spock.lang.Specification@TestMixin(GrailsUnitTestMixin)
class MyCommandObjectSpec extends Specification { void 'Test that numberOfItems must be between 1 and 10'() {
when: 'numberOfItems is less than 1'
def co = new MyCommandObject()
co.numberOfItems = 0 then: 'validation fails'
!co.validate()
co.hasErrors()
co.errors['numberOfItems'].code == 'range.toosmall' when: 'numberOfItems is greater than 10'
co.numberOfItems = 11 then: 'validation fails'
!co.validate()
co.hasErrors()
co.errors['numberOfItems'].code == 'range.toobig' when: 'numberOfItems is greater than 1'
co.numberOfItems = 1 then: 'validation succeeds'
co.validate()
!co.hasErrors() when: 'numberOfItems is greater than 10'
co.numberOfItems = 10 then: 'validation succeeds'
co.validate()
!co.hasErrors()
}
}
For validateable classes which are not one of the first 3 types listed above but are configured with the
grails.validateable.classes
property in
Config.groovy
, one additional step is required to test validation.
GrailsUnitTestMixin
provides a method named
mockForConstraintsTests
that will mock validation support for these classes. See the example below.
// src/groovy/com/demo/Book.groovy
package com.democlass Book {
String title
String author static constraints = {
author minSize: 5
}
}
// grails-app/conf/Config.groovy
grails.validateable.classes = [com.demo.Book]// ...
// test/unit/com/demo/BookSpec.groovy
package com.demoimport grails.test.mixin.TestMixin
import grails.test.mixin.support.GrailsUnitTestMixin
import spock.lang.Specification@TestMixin(GrailsUnitTestMixin)
class BookSpec extends Specification { void 'Test validation'() {
given:
mockForConstraintsTests Book when: 'the author name has only 4 characters'
def book = new Book()
book.author = 'Jeff' then: 'validation should fail'
!book.validate()
book.hasErrors()
book.errors['author'] == 'minSize' when: 'the author name has 5 characters'
book.author = 'Jacob' then: 'validation should pass'
book.validate()
!book.hasErrors()
}
}
Note that the
mockForConstraintsTests
method changes the behavior of the errors object such that something like
book.errors'author'
will evaluate to the name of the failed constraint, not a
org.springframework.validation.FieldError
object. This is convenient for unit tests. If your unit test really does want a reference to the
org.springframework.validation.FieldError
object use something like
book.errors.getFieldError('author')
.
That's it for testing constraints. One final thing we would like to say is that testing the constraints in this way catches a common error: typos in the "constraints" property name which is a mistake that is easy to make and equally easy to overlook. A unit test for your constraints will highlight the problem straight away.
HibernateTestMixin Basics
HibernateTestMixin
allows Hibernate 4 to be used in Grails unit tests. It uses a H2 in-memory database.
import grails.test.mixin.TestMixin
import grails.test.mixin.gorm.Domain
import grails.test.mixin.hibernate.HibernateTestMixin
import spock.lang.Specification
@Domain(Person)
@TestMixin(HibernateTestMixin)
class PersonSpec extends Specification { void "Test count people"() {
expect: "Test execute Hibernate count query"
Person.count() == 0
sessionFactory != null
transactionManager != null
session != null
}
}
This library dependency is required in grails-app/conf/BuildConfig.groovy for adding support for
HibernateTestMixin
.
dependencies {
test 'org.grails:grails-datastore-test-support:1.0-grails-2.4'
}
HibernateTestMixin is only supported with hibernate4 plugin versions >= 4.3.5.4 .
plugins {
runtime ':hibernate4:4.3.5.4'
}
Configuring domain classes for HibernateTestMixin tests
The
grails.test.mixin.gorm.Domain
annotation is used to configure the list of domain classes to configure for Hibernate sessionFactory instance that gets configured when the unit test runtime is initialized.
Domain
annotations will be collected from several locations:
- the annotations on the test class
- the package annotations in the package-info.java/package-info.groovy file in the package of the test class
- each super class of the test class and their respective package annotations
- the possible
SharedRuntime
class
Domain
annotations can be shared by adding them as package annotations to package-info.java/package-info.groovy files or by adding them to a
SharedRuntime
class which has been added for the test.
It's not possible to use DomainClassUnitTestMixin's
Mock
annotation in HibernateTestMixin tests. Use the
Domain
annotation in the place of
Mock
in HibernateTestMixin tests.
14.1.4 Unit Testing Filters
Unit testing filters is typically a matter of testing a controller where a filter is a mock collaborator. For example consider the following filters class:
class CancellingFilters {
def filters = {
all(controller:"simple", action:"list") {
before = {
redirect(controller:"book")
return false
}
}
}
}
This filter interceptors the
list
action of the
simple
controller and redirects to the
book
controller. To test this filter you start off with a test that targets the
SimpleController
class and add the
CancellingFilters
as a mock collaborator:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
@Mock(CancellingFilters)
class SimpleControllerSpec extends Specification { // ...}
You can then implement a test that uses the
withFilters
method to wrap the call to an action in filter execution:
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(SimpleController)
@Mock(CancellingFilters)
class SimpleControllerSpec extends Specification { void "test list action is filtered"() {
when:
withFilters(action:"list") {
controller.list()
} then:
response.redirectedUrl == '/book'
}
}
Note that the
action
parameter is required because it is unknown what the action to invoke is until the action is actually called. The
controller
parameter is optional and taken from the controller under test. If it is another controller you are testing then you can specify it:
withFilters(controller:"book",action:"list") {
controller.list()
}
14.1.5 Unit Testing URL Mappings
The Basics
Testing URL mappings can be done with the
TestFor
annotation testing a particular URL mappings class. For example to test the default URL mappings you can do the following:
import com.demo.SimpleController
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(UrlMappings)
@Mock(SimpleController)
class UrlMappingsSpec extends Specification {
// …
}
As you can see, any controller that is the target of a URL mapping that you're testing
must be added to the
@Mock
annotation.
Note that since the default UrlMappings
class is in the default package your test must also be in the default package
With that done there are a number of useful methods that are defined by the
grails.test.mixin.web.UrlMappingsUnitTestMixin
for testing URL mappings. These include:
assertForwardUrlMapping
- Asserts a URL mapping is forwarded for the given controller class (note that controller will need to be defined as a mock collaborate for this to work)
assertReverseUrlMapping
- Asserts that the given URL is produced when reverse mapping a link to a given controller and action
assertUrlMapping
- Asserts a URL mapping is valid for the given URL. This combines the assertForwardUrlMapping
and assertReverseUrlMapping
assertions
Asserting Forward URL Mappings
You use
assertForwardUrlMapping
to assert that a given URL maps to a given controller. For example, consider the following URL mappings:
static mappings = {
"/actionOne"(controller: "simple", action: "action1")
"/actionTwo"(controller: "simple", action: "action2")
}
The following test can be written to assert these URL mappings:
import com.demo.SimpleController
import grails.test.mixin.TestFor
import spock.lang.Specification@TestFor(UrlMappings)
@Mock(SimpleController)
class UrlMappingsSpec extends Specification { void "test forward mappings"() {
expect:
assertForwardUrlMapping("/actionOne", controller: 'simple', action: "action1")
assertForwardUrlMapping("/actionTwo", controller: 'simple', action: "action2")
}
}
Assert Reverse URL Mappings
You use
assertReverseUrlMapping
to check that correct links are produced for your URL mapping when using the
link
tag in GSP views. An example test is largely identical to the previous listing except you use
assertReverseUrlMapping
instead of
assertForwardUrlMapping
. Note that you can combine these 2 assertions with
assertUrlMapping
.
14.1.6 Mocking Collaborators
The Spock Framework manual has a chapter on
Interaction Based Testing which also explains mocking collaborators.
14.1.7 Mocking Codecs
The
GrailsUnitTestMixin
provides a
mockCodec
method for mocking
custom codecs which may be invoked while a unit test is running.
Failing to mock a codec which is invoked while a unit test is running may result in a MissingMethodException.
If runtime metaprogramming needs to be done in a unit test it needs to be done early in the process before the unit testing environment is fully initialized. This should be done when the unit test class is being initialized. For a Spock based test this should be done in the
setupSpec()
method. For a JUnit test this should be done in a method marked with
@BeforeClass
.
package myappimport grails.test.mixin.*
import spock.lang.Specification@TestFor(SomeController)
class SomeControllerSpec extends Specification {
def setupSpec() {
SomeClass.metaClass.someMethod = { ->
// do something here
}
} // …
}
package myappimport grails.test.mixin.*
import org.junit.*@TestFor(SomeController)
class SomeControllerTests { @BeforeClass
static void metaProgramController() {
SomeClass.metaClass.someMethod = { ->
// do something here
}
} // ...}
14.2 Integration Testing
Integration tests differ from unit tests in that you have full access to the Grails environment within the test. Grails uses an in-memory H2 database for integration tests and clears out all the data from the database between tests.
One thing to bear in mind is that logging is enabled for your application classes, but it is different from logging in tests. So if you have something like this:
class MyServiceTests extends GroovyTestCase {
void testSomething() {
log.info "Starting tests"
…
}
}
the "starting tests" message is logged using a different system than the one used by the application. The
log
property in the example above is an instance of
java.util.logging.Logger
(inherited from the base class, not injected by Grails), which doesn't have the same methods as the
log
property injected into your application artifacts. For example, it doesn't have
debug()
or
trace()
methods, and the equivalent of
warn()
is in fact
warning()
.
Transactions
Integration tests run inside a database transaction by default, which is rolled back at the end of the each test. This means that data saved during a test is not persisted to the database. Add a
transactional
property to your test class to check transactional behaviour:
class MyServiceTests extends GroovyTestCase {
static transactional = false void testMyTransactionalServiceMethod() {
…
}
}
Be sure to remove any persisted data from a non-transactional test, for example in the
tearDown
method, so these tests don't interfere with standard transactional tests that expect a clean database.
Testing Controllers
To test controllers you first have to understand the Spring Mock Library.
Grails automatically configures each test with a
MockHttpServletRequest,
MockHttpServletResponse, and
MockHttpSession that you can use in your tests. For example consider the following controller:
class FooController { def text() {
render "bar"
} def someRedirect() {
redirect(action:"bar")
}
}
The tests for this would be:
class FooControllerTests extends GroovyTestCase { void testText() {
def fc = new FooController()
fc.text()
assertEquals "bar", fc.response.contentAsString
} void testSomeRedirect() {
def fc = new FooController()
fc.someRedirect()
assertEquals "/foo/bar", fc.response.redirectedUrl
}
}
In the above case
response
is an instance of
MockHttpServletResponse
which we can use to obtain the generated content with
contentAsString
(when writing to the response) or the redirected URL. These mocked versions of the Servlet API are completely mutable (unlike the real versions) and hence you can set properties on the request such as the
contextPath
and so on.
Grails
does not invoke
interceptors or servlet filters when calling actions during integration testing. You should test interceptors and filters in isolation, using
functional testing if necessary.
Testing Controllers with Services
If your controller references a service (or other Spring beans), you have to explicitly initialise the service from your test.
Given a controller using a service:
class FilmStarsController {
def popularityService def update() {
// do something with popularityService
}
}
The test for this would be:
class FilmStarsTests extends GroovyTestCase {
def popularityService void testInjectedServiceInController () {
def fsc = new FilmStarsController()
fsc.popularityService = popularityService
fsc.update()
}
}
Testing Controller Command Objects
With command objects you just supply parameters to the request and it will automatically do the command object work for you when you call your action with no parameters:
Given a controller using a command object:
class AuthenticationController {
def signup(SignupForm form) {
…
}
}
You can then test it like this:
def controller = new AuthenticationController()
controller.params.login = "marcpalmer"
controller.params.password = "secret"
controller.params.passwordConfirm = "secret"
controller.signup()
Grails auto-magically sees your call to
signup()
as a call to the action and populates the command object from the mocked request parameters. During controller testing, the
params
are mutable with a mocked request supplied by Grails.
Testing Controllers and the render Method
The
render method lets you render a custom view at any point within the body of an action. For instance, consider the example below:
def save() {
def book = Book(params)
if (book.save()) {
// handle
}
else {
render(view:"create", model:[book:book])
}
}
In the above example the result of the model of the action is not available as the return value, but instead is stored within the
modelAndView
property of the controller. The
modelAndView
property is an instance of Spring MVC's
ModelAndView class and you can use it to the test the result of an action:
def bookController = new BookController()
bookController.save()
def model = bookController.modelAndView.model.book
Simulating Request Data
You can use the Spring
MockHttpServletRequest to test an action that requires request data, for example a REST web service. For example consider this action which performs data binding from an incoming request:
def create() {
[book: new Book(params.book)]
}
To simulate the 'book' parameter as an XML request you could do something like the following:
void testCreateWithXML() { def controller = new BookController() controller.request.contentType = 'text/xml'
controller.request.content = '''\
<?xml version="1.0" encoding="ISO-8859-1"?>
<book>
<title>The Stand</title>
…
</book>
'''.stripIndent().getBytes() // note we need the bytes def model = controller.create()
assert model.book
assertEquals "The Stand", model.book.title
}
The same can be achieved with a JSON request:
void testCreateWithJSON() { def controller = new BookController() controller.request.contentType = "application/json"
controller.request.content =
'{"id":1,"class":"Book","title":"The Stand"}'.getBytes() def model = controller.create()
assert model.book
assertEquals "The Stand", model.book.title
}
With JSON don't forget the class
property to specify the name the target type to bind to. In XML this is implicit within the name of the <book>
node, but this property is required as part of the JSON packet.
For more information on the subject of REST web services see the section on
REST.
Testing Tag Libraries
Testing tag libraries is simple because when a tag is invoked as a method it returns its result as a string (technically a
StreamCharBuffer
but this class implements all of the methods of
String
). So for example if you have a tag library like this:
class FooTagLib { def bar = { attrs, body ->
out << "<p>Hello World!</p>"
} def bodyTag = { attrs, body ->
out << "<${attrs.name}>"
out << body()
out << "</${attrs.name}>"
}
}
The tests would look like:
class FooTagLibTests extends GroovyTestCase { void testBarTag() {
assertEquals "<p>Hello World!</p>",
new FooTagLib().bar(null, null).toString()
} void testBodyTag() {
assertEquals "<p>Hello World!</p>",
new FooTagLib().bodyTag(name: "p") {
"Hello World!"
}.toString()
}
}
Notice that for the second example,
testBodyTag
, we pass a block that returns the body of the tag. This is convenient to representing the body as a String.
Testing Tag Libraries with GroovyPagesTestCase
In addition to doing simple testing of tag libraries like in the above examples, you can also use the
grails.test.GroovyPagesTestCase
class to test tag libraries with integration tests.
The
GroovyPagesTestCase
class is a subclass of the standard
GroovyTestCase
class and adds utility methods for testing the output of GSP rendering.
GroovyPagesTestCase
can only be used in an integration test.
For example, consider this date formatting tag library:
import java.text.SimpleDateFormatclass FormatTagLib {
def dateFormat = { attrs, body ->
out << new SimpleDateFormat(attrs.format) << attrs.date
}
}
This can be easily tested as follows:
class FormatTagLibTests extends GroovyPagesTestCase {
void testDateFormat() {
def template =
'<g:dateFormat format="dd-MM-yyyy" date="${myDate}" />' def testDate = … // create the date
assertOutputEquals('01-01-2008', template, [myDate:testDate])
}
}
You can also obtain the result of a GSP using the
applyTemplate
method of the
GroovyPagesTestCase
class:
class FormatTagLibTests extends GroovyPagesTestCase {
void testDateFormat() {
def template =
'<g:dateFormat format="dd-MM-yyyy" date="${myDate}" />' def testDate = … // create the date
def result = applyTemplate(template, [myDate:testDate]) assertEquals '01-01-2008', result
}
}
Testing Domain Classes
Testing domain classes is typically a simple matter of using the
GORM API, but there are a few things to be aware of. Firstly, when testing queries you often need to "flush" to ensure the correct state has been persisted to the database. For example take the following example:
void testQuery() {
def books = [
new Book(title: "The Stand"),
new Book(title: "The Shining")]
books*.save() assertEquals 2, Book.list().size()
}
This test will fail because calling
save does not actually persist the
Book
instances when called. Calling
save
only indicates to Hibernate that at some point in the future these instances should be persisted. To commit changes immediately you "flush" them:
void testQuery() {
def books = [
new Book(title: "The Stand"),
new Book(title: "The Shining")]
books*.save(flush: true) assertEquals 2, Book.list().size()
}
In this case since we're passing the argument
flush
with a value of
true
the updates will be persisted immediately and hence will be available to the query later on.
14.3 Functional Testing
Functional tests involve making HTTP requests against the running application and verifying the resultant behaviour. The functional testing phase differs from the integration phase in that the Grails application is now listening and responding to actual HTTP requests. This is useful for end-to-end testing scenarios, such as making REST calls against a JSON API.
Grails does not ship with any support for writing functional tests directly, but there are several plugins available for this.
Consult the documentation for each plugin for its capabilities.
Common Options
There are options that are common to all plugins that control how the Grails application is launched, if at all.
inline
The
-inline
option specifies that the grails application should be started inline (i.e. like
run-app
).
This option is implicitly set unless the baseUrl
or war
options are setwar
The
-war
option specifies that the grails application should be packaged as a war and started. This is useful as it tests your application in a production-like state, but it has a longer startup time than the
-inline
option. It also runs the war in a forked JVM, meaning that you cannot access any internal application objects.
grails test-app functional: -war
Note that the same build/config options for the
run-war command apply to functional testing against the WAR.
https
The
-https
option results in the application being able to receive https requests as well as http requests. It is compatible with both the
-inline
and
-war
options.
grails test-app functional: -https
Note that this does not change the test
base url to be https, it will still be http unless the
-httpsBaseUrl
option is also given.
httpsBaseUrl
The
-httpsBaseUrl
causes the implicit base url to be used for tests to be a https url.
grails test-app functional: -httpsBaseUrl
This option is ignored if the
-baseUrl
option is specified.
baseUrl
The
baseUrl
option allows the base url for tests to be specified.
grails test-app functional: -baseUrl=http://mycompany.com/grailsapp
This option will prevent the local grails application being started unless
-inline
or
-war
are given as well. To use a custom base url but still test against the local Grails application you
must specify one of either the
-inline
or
-war
options.
15 Internationalization
Grails supports Internationalization (i18n) out of the box by leveraging the underlying Spring MVC internationalization support. With Grails you are able to customize the text that appears in a view based on the user's Locale. To quote the javadoc for the
Locale class:
A Locale object represents a specific geographical, political, or cultural region. An operation that requires a Locale to perform its task is called locale-sensitive and uses the Locale to tailor information for the user. For example, displaying a number is a locale-sensitive operation--the number should be formatted according to the customs/conventions of the user's native country, region, or culture.
A Locale is made up of a
language code and a
country code. For example "en_US" is the code for US English, whilst "en_GB" is the code for British English.
15.1 Understanding Message Bundles
Now that you have an idea of locales, to use them in Grails you create message bundle file containing the different languages that you wish to render. Message bundles in Grails are located inside the
grails-app/i18n
directory and are simple Java properties files.
Each bundle starts with the name
messages
by convention and ends with the locale. Grails ships with several message bundles for a whole range of languages within the
grails-app/i18n
directory. For example:
- messages.properties
- messages_da.properties
- messages_de.properties
- messages_es.properties
- messages_fr.properties
- ...
By default Grails looks in
messages.properties
for messages unless the user has specified a locale. You can create your own message bundle by simply creating a new properties file that ends with the locale you are interested. For example
messages_en_GB.properties
for British English.
15.2 Changing Locales
By default the user locale is detected from the incoming
Accept-Language
header. However, you can provide users the capability to switch locales by simply passing a parameter called
lang
to Grails as a request parameter:
Grails will automatically switch the user's locale and store it in a cookie so subsequent requests will have the new header.
15.3 Reading Messages
Reading Messages in the View
The most common place that you need messages is inside the view. Use the
message tag for this:
<g:message code="my.localized.content" />
As long as you have a key in your
messages.properties
(with appropriate locale suffix) such as the one below then Grails will look up the message:
my.localized.content=Hola, Me llamo John. Hoy es domingo.
Messages can also include arguments, for example:
<g:message code="my.localized.content" args="${ ['Juan', 'lunes'] }" />
The message declaration specifies positional parameters which are dynamically specified:
my.localized.content=Hola, Me llamo {0}. Hoy es {1}.
Reading Messages in Controllers and Tag Libraries
It's simple to read messages in a controller since you can invoke tags as methods:
def show() {
def msg = message(code: "my.localized.content", args: ['Juan', 'lunes'])
}
The same technique can be used in
tag libraries, but if your tag library uses a custom
namespace then you must prefix the call with
g.
:
def myTag = { attrs, body ->
def msg = g.message(code: "my.localized.content", args: ['Juan', 'lunes'])
}
15.4 Scaffolding and i18n
Grails
scaffolding templates for controllers and views are fully i18n-aware. The GSPs use the
message tag for labels, buttons etc. and controller
flash
messages use i18n to resolve locale-specific messages.
The scaffolding includes locale specific labels for domain classes and domain fields. For example, if you have a
Book
domain class with a
title
field:
class Book {
String title
}
The scaffolding will use labels with the following keys:
book.label = Libro
book.title.label = Título del libro
You can use this property pattern if you'd like or come up with one of your own. There is nothing special about the use of the word
label
as part of the key other than it's the convention used by the scaffolding.
16 Security
Grails is no more or less secure than Java Servlets. However, Java servlets (and hence Grails) are extremely secure and largely immune to common buffer overrun and malformed URL exploits due to the nature of the Java Virtual Machine underpinning the code.
Web security problems typically occur due to developer naivety or mistakes, and there is a little Grails can do to avoid common mistakes and make writing secure applications easier to write.
What Grails Automatically Does
Grails has a few built in safety mechanisms by default.
- All standard database access via GORM domain objects is automatically SQL escaped to prevent SQL injection attacks
- The default scaffolding templates HTML escape all data fields when displayed
- Grails link creating tags (link, form, createLink, createLinkTo and others) all use appropriate escaping mechanisms to prevent code injection
- Grails provides codecs to let you trivially escape data when rendered as HTML, JavaScript and URLs to prevent injection attacks here.
16.1 Securing Against Attacks
SQL injection
Hibernate, which is the technology underlying GORM domain classes, automatically escapes data when committing to database so this is not an issue. However it is still possible to write bad dynamic HQL code that uses unchecked request parameters. For example doing the following is vulnerable to HQL injection attacks:
def vulnerable() {
def books = Book.find("from Book as b where b.title ='" + params.title + "'")
}
or the analogous call using a GString:
def vulnerable() {
def books = Book.find("from Book as b where b.title ='${params.title}'")
}
Do
not do this. Use named or positional parameters instead to pass in parameters:
def safe() {
def books = Book.find("from Book as b where b.title = ?",
[params.title])
}
or
def safe() {
def books = Book.find("from Book as b where b.title = :title",
[title: params.title])
}
Phishing
This really a public relations issue in terms of avoiding hijacking of your branding and a declared communication policy with your customers. Customers need to know how to identify valid emails.
XSS - cross-site scripting injection
It is important that your application verifies as much as possible that incoming requests were originated from your application and not from another site. It is also important to ensure that all data values rendered into views are escaped correctly. For example when rendering to HTML or XHTML you must ensure that people cannot maliciously inject JavaScript or other HTML into data or tags viewed by others.
Grails 2.3 and above include special support for automatically encoded data placed into GSP pages. See the documentation on
Cross Site Scripting (XSS) prevention for further information.
You must also avoid the use of request parameters or data fields for determining the next URL to redirect the user to. If you use a
successURL
parameter for example to determine where to redirect a user to after a successful login, attackers can imitate your login procedure using your own site, and then redirect the user back to their own site once logged in, potentially allowing JavaScript code to then exploit the logged-in account on the site.
Cross-site request forgery
CSRF involves unauthorized commands being transmitted from a user that a website trusts. A typical example would be another website embedding a link to perform an action on your website if the user is still authenticated.
The best way to decrease risk against these types of attacks is to use the
useToken
attribute on your forms. See
Handling Duplicate Form Submissions for more information on how to use it. An additional measure would be to not use remember-me cookies.
HTML/URL injection
This is where bad data is supplied such that when it is later used to create a link in a page, clicking it will not cause the expected behaviour, and may redirect to another site or alter request parameters.
HTML/URL injection is easily handled with the
codecs supplied by Grails, and the tag libraries supplied by Grails all use
encodeAsURL where appropriate. If you create your own tags that generate URLs you will need to be mindful of doing this too.
Denial of service
Load balancers and other appliances are more likely to be useful here, but there are also issues relating to excessive queries for example where a link is created by an attacker to set the maximum value of a result set so that a query could exceed the memory limits of the server or slow the system down. The solution here is to always sanitize request parameters before passing them to dynamic finders or other GORM query methods:
int limit = 100
def safeMax = Math.min(params.max?.toInteger() ?: limit, limit) // limit to 100 results
return Book.list(max:safeMax)
Guessable IDs
Many applications use the last part of the URL as an "id" of some object to retrieve from GORM or elsewhere. Especially in the case of GORM these are easily guessable as they are typically sequential integers.
Therefore you must assert that the requesting user is allowed to view the object with the requested id before returning the response to the user.
Not doing this is "security through obscurity" which is inevitably breached, just like having a default password of "letmein" and so on.
You must assume that every unprotected URL is publicly accessible one way or another.
16.2 Cross Site Scripting (XSS) Prevention
Cross Site Scripting (XSS) attacks are a common attack vector for web applications. They typically involve submitting HTML or Javascript code in a form such that when that code is displayed, the browser does something nasty. It could be as simple as popping up an alert box, or it could be much worse. The solution is to escape all untrusted user input when it is displayed in a page. For example,
<script>alert('Got ya!');</script>
will become
<script>alert('Got ya!');</script>
when rendered, nullifying the effects of the malicious input.
By default, Grails plays it safe and escapes all content in
${}
expressions in GSPs. All the standard GSP tags are also safe by default, escaping any relevant attribute values.
So what happens when you want to stop Grails from escaping some content? There are valid use cases for putting HTML into the database and rendering it as-is, as long as that content is
trusted. In such cases, you can tell Grails that the content is safe as should be rendered raw, i.e. without any escaping:
<section>${raw(page.content)}</section>
The
raw()
method you see here is available from controllers, tag libraries and GSP pages.
XSS prevention is hard and requires a lot of developer attentionAlthough Grails plays it safe by default, that is no guarantee that your application will be invulnerable to an XSS-style attack. Such an attack is less likely to succeed than would otherwise be the case, but developers should always be conscious of potential attack vectors and attempt to uncover vulnerabilities in the application during testing. It's also easy to switch to an unsafe default, thereby increasing the risk of a vulnerability being introduced.
There are more details about the XSS in
OWASP - XSS prevention rules and
OWASP - Types of Cross-Site Scripting. Types of XSS are:
Stored XSS,
Reflected XSS and
DOM based XSS.
DOM based XSS prevention is coming more important because of the popularity of Javascript client side templating and Single Page Apps.
Grails codecs are mainly for preventing stored and reflected XSS type of attacks. Grails 2.4 includes HTMLJS codec that assists in preventing some DOM based XSS attacks.
It's difficult to make a solution that works for everyone, and so Grails provides a lot of flexibility with regard to fine-tuning how escaping works, allowing you to keep most of your application safe while switching off default escaping or changing the codec used for pages, tags, page fragments, and more.
Configuration
It is recommended that you review the configuration of a newly created Grails application to garner an understanding of XSS prevention works in Grails.
GSP features the ability to automatically HTML encode GSP expressions, and as of Grails 2.3 this is the default configuration. The default configuration (found in
Config.groovy
) for a newly created Grails application can be seen below:
grails {
views {
gsp {
encoding = 'UTF-8'
htmlcodec = 'xml' // use xml escaping instead of HTML4 escaping
codecs {
expression = 'html' // escapes values inside ${}
scriptlet = 'html' // escapes output from scriptlets in GSPs
taglib = 'none' // escapes output from taglibs
staticparts = 'none' // escapes output from static template parts
}
}
// escapes all not-encoded output at final stage of outputting
// filteringCodecForContentType.'text/html' = 'html'
}
}
}
GSP features several codecs that it uses when writing the page to the response. The codecs are configured in the
codecs
block and are described below:
expression
- The expression codec is used to encode any code found within ${..} expressions. The default for newly created application is html
encoding.
scriptlet
- Used for output from GSP scriplets (<% %>, <%= %> blocks). The default for newly created applications is html
encoding
taglib
- Used to encode output from GSP tag libraries. The default is none
for new applications, as typically it is the responsibility of the tag author to define the encoding of a given tag and by specifying none
Grails remains backwards compatible with older tag libraries.
staticparts
- Used to encode the raw markup output by a GSP page. The default is none
.
Double Encoding Prevention
Versions of Grails prior to 2.3, included the ability to set the default codec to
html
, however enabling this setting sometimes proved problematic when using existing plugins due to encoding being applied twice (once by the
html
codec and then again if the plugin manually called
encodeAsHTML
).
Grails 2.3 includes double encoding prevention so that when an expression is evaluated, it will not encode if the data has already been encoded (Example
${foo.encodeAsHTML()}
).
Raw Output
If you are 100% sure that the value you wish to present on the page has not been received from user input, and you do not wish the value to be encoded then you can use the
raw
method:
The 'raw' method is available in tag libraries, controllers and GSP pages.
Per Plugin Encoding
Grails also features the ability to control the codecs used on a per plugin basis. For example if you have a plugin named
foo
installed, then placing the following configuration in your application's
Config.groovy
will disable encoding for only the
foo
plugin
foo.grails.views.gsp.codecs.expression = "none"
Per Page Encoding
You can also control the various codecs used to render a GSP page on a per page basis, using a page directive:
<%@page expressionCodec="none" %>
Per Tag Library Encoding
Each tag library created has the opportunity to specify a default codec used to encode output from the tag library using the "defaultEncodeAs" property:
static defaultEncodeAs = 'html'
Encoding can also be specified on a per tag basis using "encodeAsForTags":
static encodeAsForTags = [tagName: 'raw']
Context Sensitive Encoding Switching
Certain tags require certain encodings and Grails features the ability to enable a codec only a certain part of a tag's execution using the "withCodec" method. Consider for example the "<g:javascript>"" tag which allows you to embed JavaScript code in the page. This tag requires JavaScript encoding, not HTML coding for the execution of the body of the tag (but not for the markup that is output):
out.println '<script type="text/javascript">'
withCodec("JavaScript") {
out << body()
}
out.println()
out.println '</script>'
Forced Encoding for Tags
If a tag specifies a default encoding that differs from your requirements you can force the encoding for any tag by passing the optional 'encodeAs' attribute:
<g:message code="foo.bar" encodeAs="JavaScript" />
Default Encoding for All Output
The default configuration for new applications is fine for most use cases, and backwards compatible with existing plugins and tag libraries. However, you can also make your application even more secure by configuring Grails to always encode all output at the end of a response. This is done using the
filteringCodecForContentType
configuration in
Config.groovy
:
grails.views.gsp.filteringCodecForContentType.'text/html' = 'html'
Note that, if activated, the
staticparts
codec typically needs to be set to
raw
so that static markup is not encoded:
codecs {
expression = 'html' // escapes values inside ${}
scriptlet = 'html' // escapes output from scriptlets in GSPs
taglib = 'none' // escapes output from taglibs
staticparts = 'raw' // escapes output from static template parts
}
16.3 Encoding and Decoding Objects
Grails supports the concept of dynamic encode/decode methods. A set of standard codecs are bundled with Grails. Grails also supports a simple mechanism for developers to contribute their own codecs that will be recognized at runtime.
Codec Classes
A Grails codec class is one that may contain an encode closure, a decode closure or both. When a Grails application starts up the Grails framework dynamically loads codecs from the
grails-app/utils/
directory.
The framework looks under
grails-app/utils/
for class names that end with the convention
Codec
. For example one of the standard codecs that ships with Grails is
HTMLCodec
.
If a codec contains an
encode
closure Grails will create a dynamic
encode
method and add that method to the
Object
class with a name representing the codec that defined the encode closure. For example, the
HTMLCodec
class defines an
encode
closure, so Grails attaches it with the name
encodeAsHTML
.
The
HTMLCodec
and
URLCodec
classes also define a
decode
closure, so Grails attaches those with the names
decodeHTML
and
decodeURL
respectively. Dynamic codec methods may be invoked from anywhere in a Grails application. For example, consider a case where a report contains a property called 'description' which may contain special characters that must be escaped to be presented in an HTML document. One way to deal with that in a GSP is to encode the description property using the dynamic encode method as shown below:
${report.description.encodeAsHTML()}
Decoding is performed using
value.decodeHTML()
syntax.
Encoder and Decoder interfaces for staticly compiled code
A preferred way to use codecs is to use the codecLookup bean to get hold of
Encoder
and
Decoder
instances .
package org.codehaus.groovy.grails.support.encoding;public interface CodecLookup {
public Encoder lookupEncoder(String codecName);
public Decoder lookupDecoder(String codecName);
}
example of using
CodecLookup
and
Encoder
interface
import org.codehaus.groovy.grails.support.encoding.CodecLookupclass CustomTagLib {
CodecLookup codecLookup def myTag = { Map attrs, body ->
out << codecLookup.lookupEncoder('HTML').encode(attrs.something)
}
}
Standard Codecs
HTMLCodecThis codec performs HTML escaping and unescaping, so that values can be rendered safely in an HTML page without creating any HTML tags or damaging the page layout. For example, given a value "Don't you know that 2 > 1?" you wouldn't be able to show this safely within an HTML page because the > will look like it closes a tag, which is especially bad if you render this data within an attribute, such as the value attribute of an input field.
Example of usage:
<input name="comment.message" value="${comment.message.encodeAsHTML()}"/>
Note that the HTML encoding does not re-encode apostrophe/single quote so you must use double quotes on attribute values to avoid text with apostrophes affecting your page.
HTMLCodec defaults to HTML4 style escaping (legacy HTMLCodec implementation in Grails versions before 2.3.0) which escapes non-ascii characters.
You can use plain XML escaping instead of HTML4 escaping by setting this config property in Config.groovy:
grails.views.gsp.htmlcodec = 'xml'
XMLCodecThis codec performs XML escaping and unescaping. It escapes & , < , > , " , ' , \ , @ , ` , non breaking space (\u00a0), line separator (\u2028) and paragraph separator (\u2029).
HTMLJSCodecThis codec performs HTML and JS encoding. It is used for preventing some DOM-XSS vulnerabilities. See
OWASP - DOM based XSS Prevention Cheat Sheet for guidelines of preventing DOM based XSS attacks.
URLCodecURL encoding is required when creating URLs in links or form actions, or any time data is used to create a URL. It prevents illegal characters from getting into the URL and changing its meaning, for example "Apple & Blackberry" is not going to work well as a parameter in a GET request as the ampersand will break parameter parsing.
Example of usage:
<a href="/mycontroller/find?searchKey=${lastSearch.encodeAsURL()}">
Repeat last search
</a>
Base64CodecPerforms Base64 encode/decode functions. Example of usage:
Your registration code is: ${user.registrationCode.encodeAsBase64()}
JavaScriptCodecEscapes Strings so they can be used as valid JavaScript strings. For example:
Element.update('${elementId}',
'${render(template: "/common/message").encodeAsJavaScript()}')
HexCodecEncodes byte arrays or lists of integers to lowercase hexadecimal strings, and can decode hexadecimal strings into byte arrays. For example:
Selected colour: #${[255,127,255].encodeAsHex()}
MD5CodecUses the MD5 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a lowercase hexadecimal string. Example of usage:
Your API Key: ${user.uniqueID.encodeAsMD5()}
MD5BytesCodecUses the MD5 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a byte array. Example of usage:
byte[] passwordHash = params.password.encodeAsMD5Bytes()
SHA1CodecUses the SHA1 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a lowercase hexadecimal string. Example of usage:
Your API Key: ${user.uniqueID.encodeAsSHA1()}
SHA1BytesCodecUses the SHA1 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a byte array. Example of usage:
byte[] passwordHash = params.password.encodeAsSHA1Bytes()
SHA256CodecUses the SHA256 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a lowercase hexadecimal string. Example of usage:
Your API Key: ${user.uniqueID.encodeAsSHA256()}
SHA256BytesCodecUses the SHA256 algorithm to digest byte arrays or lists of integers, or the bytes of a string (in default system encoding), as a byte array. Example of usage:
byte[] passwordHash = params.password.encodeAsSHA256Bytes()
Custom Codecs
Applications may define their own codecs and Grails will load them along with the standard codecs. A custom codec class must be defined in the
grails-app/utils/
directory and the class name must end with
Codec
. The codec may contain a
static
encode
closure, a
static
decode
closure or both. The closure must accept a single argument which will be the object that the dynamic method was invoked on. For Example:
class PigLatinCodec {
static encode = { str ->
// convert the string to pig latin and return the result
}
}
With the above codec in place an application could do something like this:
${lastName.encodeAsPigLatin()}
16.4 Authentication
Grails has no default mechanism for authentication as it is possible to implement authentication in many different ways. It is however, easy to implement a simple authentication mechanism using either
interceptors or
filters. This is sufficient for simple use cases but it's highly preferable to use an established security framework, for example by using the
Spring Security or the
Shiro plugin.
Filters let you apply authentication across all controllers or across a URI space. For example you can create a new set of filters in a class called
grails-app/conf/SecurityFilters.groovy
by running:
grails create-filters security
and implement your interception logic there:
class SecurityFilters {
def filters = {
loginCheck(controller: '*', action: '*') {
before = {
if (!session.user && actionName != "login") {
redirect(controller: "user", action: "login")
return false
}
}
}
}
}
Here the
loginCheck
filter intercepts execution
before all actions except
login
are executed, and if there is no user in the session then redirect to the
login
action.
The
login
action itself is simple too:
def login() {
if (request.get) {
return // render the login view
} def u = User.findByLogin(params.login)
if (u) {
if (u.password == params.password) {
session.user = u
redirect(action: "home")
}
else {
render(view: "login", model: [message: "Password incorrect"])
}
}
else {
render(view: "login", model: [message: "User not found"])
}
}
16.5 SSL in Development
You can configure the embedded Tomcat server to use SSL in development mode. Typically with Tomcat you would edit the
server.xml
file to define these properties. However in Grails you need to hook into Tomcat by configuring
grails-app/scripts/_Events.groovy
. In the example below I am using environment variables to pass in the values of the keystore and truststore. Notice also that we must add the keystore/truststore properties twice. Using
System.setProperty("javax.net.ssl.keyStore", "$keystorepath")
will define the keystore path for Grails. You will also need to also define the keystore/truststore paths for the Tomcat connector. Here is an example in
grails-app/scripts/_Events.groovy
:
import org.apache.catalina.connector.*
import grails.util.Environment eventConfigureTomcat = { org.apache.catalina.startup.Tomcat tomcat ->
if (Environment.getCurrent() == Environment.DEVELOPMENT) {
String keystorepass = System.getenv("KEY_STORE_PASS")
String keystorepath = System.getenv("KEY_STORE_PATH")
String truststorepass = System.getenv("TRUST_STORE_PASS")
String truststorepath = System.getenv("TRUST_STORE_PATH") System.setProperty("javax.net.debug", "ssl") //use this to confirm grails adds proper keystore/truststore settings
System.setProperty("javax.net.ssl.keyStoreType", "jks")
System.setProperty("javax.net.ssl.keyStore", "$keystorepath")
System.setProperty("javax.net.ssl.keyStorePassword", "$keystorepass")
System.setProperty("javax.net.ssl.trustStoreType", "jks")
System.setProperty("javax.net.ssl.trustStore", "$truststorepath")
System.setProperty("javax.net.ssl.trustStorePassword", "$truststorepass") def connector = new org.apache.catalina.connector.Connector()
connector.port = 8443
connector.protocol = 8443
connector.scheme = "https"
connector.setProperty("maxThreads", "150")
connector.setProperty("SSLEnabled", "true")
connector.setProperty("secure", "true")
connector.setProperty("clientAuth", "true")
connector.setProperty("sslProtocol", "TLS")
connector.setProperty("keystoreType", "JKS")
connector.setProperty("keystoreFile", "$keystorepath")
connector.setProperty("keystorePass", "$keystorepass")
connector.setProperty("truststoreType", "JKS")
connector.setProperty("truststoreFile", "$truststorepath")
connector.setProperty("truststorePass", "$truststorepass") tomcat.service.addConnector connector
println "SSL configuration complete"
}
}
16.6 Security Plugins
If you need more advanced functionality beyond simple authentication such as authorization, roles etc. then you should consider using one of the available security plugins.
16.6.1 Spring Security
The Spring Security plugins are built on the
Spring Security project which provides a flexible, extensible framework for building all sorts of authentication and authorization schemes. The plugins are modular so you can install just the functionality that you need for your application. The Spring Security plugins are the official security plugins for Grails and are actively maintained and supported.
There is a
Core plugin which supports form-based authentication, encrypted/salted passwords, HTTP Basic authentication, etc. and secondary dependent plugins provide alternate functionality such as
OpenID authentication,
ACL support,
single sign-on with Jasig CAS,
LDAP authentication,
Kerberos authentication, and a plugin providing
user interface extensions and security workflows.
See the
Core plugin page for basic information and the
user guide for detailed information.
16.6.2 Shiro
Shiro is a Java POJO-oriented security framework that provides a default domain model that models realms, users, roles and permissions. With Shiro you extend a controller base class called
JsecAuthBase
in each controller you want secured and then provide an
accessControl
block to setup the roles. An example below:
class ExampleController extends JsecAuthBase {
static accessControl = {
// All actions require the 'Observer' role.
role(name: 'Observer') // The 'edit' action requires the 'Administrator' role.
role(name: 'Administrator', action: 'edit') // Alternatively, several actions can be specified.
role(name: 'Administrator', only: [ 'create', 'edit', 'save', 'update' ])
}
…
}
For more information on the Shiro plugin refer to the
documentation.
17 Plugins
Grails is first and foremost a web application framework, but it is also a platform. By exposing a number of extension points that let you extend anything from the command line interface to the runtime configuration engine, Grails can be customised to suit almost any needs. To hook into this platform, all you need to do is create a plugin.
Extending the platform may sound complicated, but plugins can range from trivially simple to incredibly powerful. If you know how to build a Grails application, you'll know how to create a plugin for
sharing a data model or some static resources.
17.1 Creating and Installing Plugins
Creating Plugins
Creating a Grails plugin is a simple matter of running the command:
grails create-plugin [PLUGIN NAME]
This will create a plugin project for the name you specify. For example running
grails create-plugin example
would create a new plugin project called
example
.
Make sure the plugin name does not contain more than one capital in a row, or it won't work. Camel case is fine, though.
The structure of a Grails plugin is very nearly the same as a Grails application project's except that in the root of the plugin directory you will find a plugin Groovy file called the "plugin descriptor".
The only plugins included in a new plugin project are Tomcat and Release. Hibernate is not included by default.
Being a regular Grails project has a number of benefits in that you can immediately test your plugin by running:
Plugin projects don't provide an index.gsp by default since most plugins don't need it. So, if you try to view the plugin running in a browser right after creating it, you will receive a page not found error. You can easily create a grails-app/views/index.gsp
for your plugin if you'd like.
The plugin descriptor name ends with the convention
GrailsPlugin
and is found in the root of the plugin project. For example:
class ExampleGrailsPlugin {
def version = "0.1" …
}
All plugins must have this class in the root of their directory structure. The plugin class defines the version of the plugin and other metadata, and optionally various hooks into plugin extension points (covered shortly).
You can also provide additional information about your plugin using several special properties:
title
- short one-sentence description of your plugin
version
- The version of your plugin. Valid values include example "0.1", "0.2-SNAPSHOT", "1.1.4" etc.
grailsVersion
- The version of version range of Grails that the plugin supports. eg. "1.2 > *" (indicating 1.2 or higher)
author
- plugin author's name
authorEmail
- plugin author's contact e-mail
description
- full multi-line description of plugin's features
documentation
- URL of the plugin's documentation
Here is an example from the
Quartz Grails plugin:
class QuartzGrailsPlugin {
def version = "0.1"
def grailsVersion = "1.1 > *"
def author = "Sergey Nebolsin"
def authorEmail = "nebolsin@gmail.com"
def title = "Quartz Plugin"
def description = '''\
The Quartz plugin allows your Grails application to schedule jobs\
to be executed using a specified interval or cron expression. The\
underlying system uses the Quartz Enterprise Job Scheduler configured\
via Spring, but is made simpler by the coding by convention paradigm.\
'''
def documentation = "http://grails.org/plugin/quartz" …
}
Installing Local Plugins
To make your plugin available for use in a Grails application run the
maven-install
command:
This will install the plugin into your local Maven cache. Then to use the plugin within an application declare a dependency on the plugin in your
grails-app/conf/BuildConfig.groovy
file:
Notes on excluded Artefacts
Although the
create-plugin command creates certain files for you so that the plugin can be run as a Grails application, not all of these files are included when packaging a plugin. The following is a list of artefacts created, but not included by
package-plugin:
grails-app/conf/BootStrap.groovy
grails-app/conf/BuildConfig.groovy
(although it is used to generate dependencies.groovy
)
grails-app/conf/Config.groovy
grails-app/conf/DataSource.groovy
(and any other *DataSource.groovy
)
grails-app/conf/UrlMappings.groovy
grails-app/conf/spring/resources.groovy
- Everything within
/web-app/WEB-INF
- Everything within
/web-app/plugins/**
- Everything within
/test/**
- SCM management files within
**/.svn/**
and **/CVS/**
If you need artefacts within
WEB-INF
it is recommended you use the
_Install.groovy
script (covered later), which is executed when a plugin is installed, to provide such artefacts. In addition, although
UrlMappings.groovy
is excluded you are allowed to include a
UrlMappings
definition with a different name, such as
MyPluginUrlMappings.groovy
.
Customizing the plugin contents
You can specify what to exclude in addition to the default excludes by adding elements to the
pluginExcludes
descriptor property (described below). In addition, there are two ways to configure the contents of the plugin ZIP or JAR file.
One is to create an event handler for the
CreatePluginArchiveStart
event, which is fired after all of the plugin files have been copied to the staging directory. By adding an event handler you can add, modify, or delete files as needed. Add the handler to
_Events.groovy
in the
scripts
directory, for example
eventCreatePluginArchiveStart = { stagingDir ->
// update staging directory contents here
}
You can customize the location of the staging directory with the
grails.project.plugin.staging.dir
attribute in
BuildConfig.groovy
or as as system property.
Note that there is also a
CreatePluginArchiveEnd
event which is fired after the ZIP or JAR is packaged.
You can also do this work in a Closure in
BuildConfig.groovy
with the property
grails.plugin.resources
which is analogous to the
grails.war.resources
property, e.g.
grails.plugin.resources = { stagingDir ->
// update staging directory contents here
}
Specifying Plugin Locations
An application can load plugins from anywhere on the file system, even if they have not been installed. Specify the location of the (unpacked) plugin in the application's
grails-app/conf/BuildConfig.groovy
file:
// Useful to test plugins you are developing.
grails.plugin.location.shiro =
"/home/dilbert/dev/plugins/grails-shiro"// Useful for modular applications where all plugins and
// applications are in the same directory.
grails.plugin.location.'grails-ui' = "../grails-grails-ui"
This is particularly useful in two cases:
- You are developing a plugin and want to test it in a real application without packaging and installing it first.
- You have split an application into a set of plugins and an application, all in the same "super-project" directory.
The Artifactory repository for Grails now includes all the dependencies for published plugins. So, if you are using inline plugins that have dependencies, it is necessary to do a secondary resolve because these dependencies might not be in the repository. Therefore, you should set legacyResolve
to true
in your BuildConfig.groovy
if you are using inline plugins with dependencies.
17.2 Plugin Repositories
Distributing Plugins in the Grails Central Plugin Repository
The preferred way to distribute plugin is to publish to the official Grails Central Plugin Repository. This will make your plugin visible to the
list-plugins command:
which lists all plugins that are in the central repository. Your plugin will also be available to the
plugin-info command:
grails plugin-info [plugin-name]
which prints extra information about it, such as its description, who wrote, etc.
If you have created a Grails plugin and want it to be hosted in the central repository, you'll find instructions for getting an account on this wiki page.
When you have access to the Grails Plugin repository, install the
Release Plugin by declaring it as a 'build' scoped dependency in
grails-app/conf/BuildConfig.groovy
file:
grails.project.dependency.resolution = {
…
plugins {
build ':release:3.0.0'
}
}
And execute the
publish-plugin
command to release your plugin:
This will automatically publish the plugin to the central repository. If the command is successful, it will immediately be available on the plugin portal at http://grails.org/plugin/<pluginName>. You can find out more about the Release plugin and its other features in
its user guide.
Configuring Additional Repositories
The process for configuring repositories in Grails differs between versions. For version of Grails 1.2 and earlier please refer to the
Grails 1.2 documentation on the subject. The following sections cover Grails 1.3 and above.
Grails 1.3 and above use Ivy under the hood to resolve plugin dependencies. The mechanism for defining additional plugin repositories is largely the same as
defining repositories for JAR dependencies. For example you can define a remote Maven repository that contains Grails plugins using the following syntax in
grails-app/conf/BuildConfig.groovy
:
repositories {
mavenRepo "http://repository.codehaus.org" // ...or with a name
mavenRepo name: "myRepo",
root: "http://myserver:8081/artifactory/plugins-snapshots-local"
}
You can also define a SVN-based Grails repository (such as the one hosted at
http://plugins.grails.org) using the
grailsRepo
method:
repositories {
grailsRepo "http://myserver/mygrailsrepo" // ...or with a name
grailsRepo "http://myserver/svn/grails-plugins", "mySvnRepo"
}
There is a shortcut to setup the Grails central repository:
repositories {
grailsCentral()
}
The order in which plugins are resolved is based on the ordering of the repositories. So in this case the Grails central repository will be searched last:
repositories {
grailsRepo "http://myserver/mygrailsrepo"
grailsCentral()
}
Publishing to Maven Compatible Repositories
In general it is recommended for Grails 1.3 and above to use standard Maven-style repositories to self host plugins. The benefits of doing so include the ability for existing tooling and repository managers to interpret the structure of a Maven repository.
You use the Release plugin to publish a plugin to a Maven repository. Please refer to the section of the
Maven deployment user guide on the subject.
17.3 Understanding a Plugin's Structure
As as mentioned previously, a plugin is basically a regular Grails application with a plugin descriptor. However when installed, the structure of a plugin differs slightly. For example, take a look at this plugin directory structure:
+ grails-app
+ controllers
+ domain
+ taglib
etc.
+ lib
+ src
+ java
+ groovy
+ web-app
+ js
+ css
When a plugin is installed the contents of the
grails-app
directory will go into a directory such as
plugins/example-1.0/grails-app
. They
will not be copied into the main source tree. A plugin never interferes with a project's primary source tree.
Dealing with static resources is slightly different. When developing a plugin, just like an application, all static resources go in the
web-app
directory. You can then link to static resources just like in an application. This example links to a JavaScript source:
<g:resource dir="js" file="mycode.js" />
When you run the plugin in development mode the link to the resource will resolve to something like
/js/mycode.js
. However, when the plugin is installed into an application the path will automatically change to something like
/plugin/example-0.1/js/mycode.js
and Grails will deal with making sure the resources are in the right place.
There is a special
pluginContextPath
variable that can be used whilst both developing the plugin and when in the plugin is installed into the application to find out what the correct path to the plugin is.
At runtime the
pluginContextPath
variable will either evaluate to an empty string or
/plugins/example
depending on whether the plugin is running standalone or has been installed in an application
Java and Groovy code that the plugin provides within the lib and
src/java
and
src/groovy
directories will be compiled into the main project's
web-app/WEB-INF/classes
directory so that they are made available at runtime.
17.4 Providing Basic Artefacts
Adding a new Script
A plugin can add a new script simply by providing the relevant Gant script in its scripts directory:
+ MyPlugin.groovy
+ scripts <-- additional scripts here
+ grails-app
+ controllers
+ services
+ etc.
+ lib
Adding a new grails-app artifact (Controller, Tag Library, Service, etc.)
A plugin can add new artifacts by creating the relevant file within the
grails-app
tree. Note that the plugin is loaded from where it is installed and not copied into the main application tree.
+ ExamplePlugin.groovy
+ scripts
+ grails-app
+ controllers <-- additional controllers here
+ services <-- additional services here
+ etc. <-- additional XXX here
+ lib
Providing Views, Templates and View resolution
When a plugin provides a controller it may also provide default views to be rendered. This is an excellent way to modularize your application through plugins. Grails' view resolution mechanism will first look for the view in the application it is installed into and if that fails will attempt to look for the view within the plugin. This means that you can override views provided by a plugin by creating corresponding GSPs in the application's
grails-app/views
directory.
For example, consider a controller called
BookController
that's provided by an 'amazon' plugin. If the action being executed is
list
, Grails will first look for a view called
grails-app/views/book/list.gsp
then if that fails it will look for the same view relative to the plugin.
However if the view uses templates that are also provided by the plugin then the following syntax may be necessary:
<g:render template="fooTemplate" plugin="amazon"/>
Note the usage of the
plugin
attribute, which contains the name of the plugin where the template resides. If this is not specified then Grails will look for the template relative to the application.
Excluded Artefacts
By default Grails excludes the following files during the packaging process:
grails-app/conf/BootStrap.groovy
grails-app/conf/BuildConfig.groovy
(although it is used to generate dependencies.groovy
)
grails-app/conf/Config.groovy
grails-app/conf/DataSource.groovy
(and any other *DataSource.groovy
)
grails-app/conf/UrlMappings.groovy
grails-app/conf/spring/resources.groovy
- Everything within
/web-app/WEB-INF
- Everything within
/web-app/plugins/**
- Everything within
/test/**
- SCM management files within
**/.svn/**
and **/CVS/**
If your plugin requires files under the
web-app/WEB-INF
directory it is recommended that you modify the plugin's
scripts/_Install.groovy
Gant script to install these artefacts into the target project's directory tree.
In addition, the default
UrlMappings.groovy
file is excluded to avoid naming conflicts, however you are free to add a UrlMappings definition under a different name which
will be included. For example a file called
grails-app/conf/BlogUrlMappings.groovy
is fine.
The list of excludes is extensible with the
pluginExcludes
property:
// resources that are excluded from plugin packaging
def pluginExcludes = [
"grails-app/views/error.gsp"
]
This is useful for example to include demo or test resources in the plugin repository, but not include them in the final distribution.
17.5 Evaluating Conventions
Before looking at providing runtime configuration based on conventions you first need to understand how to evaluate those conventions from a plugin. Every plugin has an implicit
application
variable which is an instance of the
GrailsApplication interface.
The
GrailsApplication
interface provides methods to evaluate the conventions within the project and internally stores references to all artifact classes within your application.
Artifacts implement the
GrailsClass interface, which represents a Grails resource such as a controller or a tag library. For example to get all
GrailsClass
instances you can do:
for (grailsClass in application.allClasses) {
println grailsClass.name
}
GrailsApplication
has a few "magic" properties to narrow the type of artefact you are interested in. For example to access controllers you can use:
for (controllerClass in application.controllerClasses) {
println controllerClass.name
}
The dynamic method conventions are as follows:
*Classes
- Retrieves all the classes for a particular artefact name. For example application.controllerClasses
.
get*Class
- Retrieves a named class for a particular artefact. For example application.getControllerClass("PersonController")
is*Class
- Returns true
if the given class is of the given artefact type. For example application.isControllerClass(PersonController)
The
GrailsClass
interface has a number of useful methods that let you further evaluate and work with the conventions. These include:
getPropertyValue
- Gets the initial value of the given property on the class
hasProperty
- Returns true
if the class has the specified property
newInstance
- Creates a new instance of this class.
getName
- Returns the logical name of the class in the application without the trailing convention part if applicable
getShortName
- Returns the short name of the class without package prefix
getFullName
- Returns the full name of the class in the application with the trailing convention part and with the package name
getPropertyName
- Returns the name of the class as a property name
getLogicalPropertyName
- Returns the logical property name of the class in the application without the trailing convention part if applicable
getNaturalName
- Returns the name of the property in natural terms (e.g. 'lastName' becomes 'Last Name')
getPackageName
- Returns the package name
For a full reference refer to the
javadoc API.
17.6 Hooking into Build Events
Post-Install Configuration and Participating in Upgrades
Grails plugins can do post-install configuration. This is achieved using a specially named script under the
scripts
directory of the plugin -
_Install.groovy
.
_Install.groovy
is executed after the plugin has been installed.
This scripts is a
Gant script, so you can use the full power of Gant. An addition to the standard Gant variables there is also a
pluginBasedir
variable which points at the plugin installation basedir.
As an example this
_Install.groovy
script will create a new directory type under the
grails-app
directory and install a configuration template:
ant.mkdir(dir: "${basedir}/grails-app/jobs")ant.copy(file: "${pluginBasedir}/src/samples/SamplePluginConfig.groovy",
todir: "${basedir}/grails-app/conf")
The
pluginBasedir
variable is not available in custom scripts, but you can use
fooPluginDir
, where
foo
is the name of your plugin.
Scripting events
It is also possible to hook into command line scripting events. These are events triggered during execution of Grails target and plugin scripts.
For example, you can hook into status update output (i.e. "Tests passed", "Server running") and the creation of files or artefacts.
A plugin just has to provide an
_Events.groovy
script to listen to the required events. Refer the documentation on
Hooking into Events for further information.
17.7 Hooking into Runtime Configuration
Grails provides a number of hooks to leverage the different parts of the system and perform runtime configuration by convention.
Hooking into the Grails Spring configuration
First, you can hook in Grails runtime configuration by providing a property called
doWithSpring
which is assigned a block of code. For example the following snippet is from one of the core Grails plugins that provides
i18n support:
import org.springframework.web.servlet.i18n.CookieLocaleResolver
import org.springframework.web.servlet.i18n.LocaleChangeInterceptor
import org.springframework.context.support.ReloadableResourceBundleMessageSourceclass I18nGrailsPlugin { def version = "0.1" def doWithSpring = {
messageSource(ReloadableResourceBundleMessageSource) {
basename = "WEB-INF/grails-app/i18n/messages"
}
localeChangeInterceptor(LocaleChangeInterceptor) {
paramName = "lang"
}
localeResolver(CookieLocaleResolver)
}
}
This plugin configures the Grails
messageSource
bean and a couple of other beans to manage Locale resolution and switching. It using the
Spring Bean Builder syntax to do so.
Participating in web.xml Generation
Grails generates the
WEB-INF/web.xml
file at load time, and although plugins cannot change this file directly, they can participate in the generation of the file. A plugin can provide a
doWithWebDescriptor
property that is assigned a block of code that gets passed the
web.xml
as an
XmlSlurper
GPathResult
.
Add servlet
and servlet-mapping
Consider this example from the
ControllersPlugin
:
def doWithWebDescriptor = { webXml -> def mappingElement = webXml.'servlet-mapping' def lastMapping = mappingElement[mappingElement.size() - 1]
lastMapping + {
'servlet-mapping' {
'servlet-name'("grails")
'url-pattern'("*.dispatch")
}
}
}
Here the plugin gets a reference to the last
<servlet-mapping>
element and appends Grails' servlet after it using XmlSlurper's ability to programmatically modify XML using closures and blocks.
Add filter
and filter-mapping
Adding a filter with its mapping works a little differently. The location of the
<filter>
element doesn't matter since order is not important, so it's simplest to insert your custom filter definition immediately after the last
<context-param>
element. Order
is important for mappings, but the usual approach is to add it immediately after the last
<filter>
element like so:
def doWithWebDescriptor = { webXml -> def contextParam = webXml.'context-param' contextParam[contextParam.size() - 1] + {
'filter' {
'filter-name'('springSecurityFilterChain')
'filter-class'(DelegatingFilterProxy.name)
}
} def filter = webXml.'filter'
filter[filter.size() - 1] + {
'filter-mapping'{
'filter-name'('springSecurityFilterChain')
'url-pattern'('/*')
}
}
}
In some cases you need to ensure that your filter comes after one of the standard Grails filters, such as the Spring character encoding filter or the SiteMesh filter. Fortunately you can insert filter mappings immediately after the standard ones (more accurately, any that are in the template web.xml file) like so:
def doWithWebDescriptor = { webXml ->
... // Insert the Spring Security filter after the Spring
// character encoding filter.
def filter = webXml.'filter-mapping'.find {
it.'filter-name'.text() == "charEncodingFilter"
} filter + {
'filter-mapping'{
'filter-name'('springSecurityFilterChain')
'url-pattern'('/*')
}
}
}
Doing Post Initialisation Configuration
Sometimes it is useful to be able do some runtime configuration after the Spring
ApplicationContext has been built. In this case you can define a
doWithApplicationContext
closure property.
class SimplePlugin { def name = "simple"
def version = "1.1" def doWithApplicationContext = { appCtx ->
def sessionFactory = appCtx.sessionFactory
// do something here with session factory
}
}
17.8 Adding Dynamic Methods at Runtime
The Basics
Grails plugins let you register dynamic methods with any Grails-managed or other class at runtime. This work is done in a
doWithDynamicMethods
closure.
For Grails-managed classes like controllers, tag libraries and so forth you can add methods, constructors etc. using the
ExpandoMetaClass mechanism by accessing each controller's
MetaClass:
class ExamplePlugin {
def doWithDynamicMethods = { applicationContext ->
for (controllerClass in application.controllerClasses) {
controllerClass.metaClass.myNewMethod = {-> println "hello world" }
}
}
}
In this case we use the implicit application object to get a reference to all of the controller classes' MetaClass instances and add a new method called
myNewMethod
to each controller. If you know beforehand the class you wish the add a method to you can simply reference its
metaClass
property.
For example we can add a new method
swapCase
to
java.lang.String
:
class ExamplePlugin { def doWithDynamicMethods = { applicationContext ->
String.metaClass.swapCase = {->
def sb = new StringBuilder()
delegate.each {
sb << (Character.isUpperCase(it as char) ?
Character.toLowerCase(it as char) :
Character.toUpperCase(it as char))
}
sb.toString()
} assert "UpAndDown" == "uPaNDdOWN".swapCase()
}
}
Interacting with the ApplicationContext
The
doWithDynamicMethods
closure gets passed the Spring
ApplicationContext
instance. This is useful as it lets you interact with objects within it. For example if you were implementing a method to interact with Hibernate you could use the
SessionFactory
instance in combination with a
HibernateTemplate
:
import org.springframework.orm.hibernate3.HibernateTemplateclass ExampleHibernatePlugin { def doWithDynamicMethods = { applicationContext -> for (domainClass in application.domainClasses) { domainClass.metaClass.static.load = { Long id->
def sf = applicationContext.sessionFactory
def template = new HibernateTemplate(sf)
template.load(delegate, id)
}
}
}
}
Also because of the autowiring and dependency injection capability of the Spring container you can implement more powerful dynamic constructors that use the application context to wire dependencies into your object at runtime:
class MyConstructorPlugin { def doWithDynamicMethods = { applicationContext ->
for (domainClass in application.domainClasses) {
domainClass.metaClass.constructor = {->
return applicationContext.getBean(domainClass.name)
}
}
}
}
Here we actually replace the default constructor with one that looks up prototyped Spring beans instead!
17.9 Participating in Auto Reload Events
Monitoring Resources for Changes
Often it is valuable to monitor resources for changes and perform some action when they occur. This is how Grails implements advanced reloading of application state at runtime. For example, consider this simplified snippet from the Grails
ServicesPlugin
:
class ServicesGrailsPlugin {
…
def watchedResources = "file:./grails-app/services/*Service.groovy" …
def onChange = { event ->
if (event.source) {
def serviceClass = application.addServiceClass(event.source)
def serviceName = "${serviceClass.propertyName}"
def beans = beans {
"$serviceName"(serviceClass.getClazz()) { bean ->
bean.autowire = true
}
}
if (event.ctx) {
event.ctx.registerBeanDefinition(
serviceName,
beans.getBeanDefinition(serviceName))
}
}
}
}
First it defines
watchedResources
as either a String or a List of strings that contain either the references or patterns of the resources to watch. If the watched resources specify a Groovy file, when it is changed it will automatically be reloaded and passed into the
onChange
closure in the
event
object.
The
event
object defines a number of useful properties:
event.source
- The source of the event, either the reloaded Class
or a Spring Resource
event.ctx
- The Spring ApplicationContext
instance
event.plugin
- The plugin object that manages the resource (usually this
)
event.application
- The GrailsApplication
instance
event.manager
- The GrailsPluginManager
instance
These objects are available to help you apply the appropriate changes based on what changed. In the "Services" example above, a new service bean is re-registered with the
ApplicationContext
when one of the service classes changes.
Influencing Other Plugins
In addition to reacting to changes, sometimes a plugin needs to "influence" another.
Take for example the Services and Controllers plugins. When a service is reloaded, unless you reload the controllers too, problems will occur when you try to auto-wire the reloaded service into an older controller Class.
To get around this, you can specify which plugins another plugin "influences". This means that when one plugin detects a change, it will reload itself and then reload its influenced plugins. For example consider this snippet from the
ServicesGrailsPlugin
:
def influences = ['controllers']
Observing other plugins
If there is a particular plugin that you would like to observe for changes but not necessary watch the resources that it monitors you can use the "observe" property:
def observe = ["controllers"]
In this case when a controller is changed you will also receive the event chained from the controllers plugin.
It is also possible for a plugin to observe all loaded plugins by using a wildcard:
The Logging plugin does exactly this so that it can add the
log
property back to
any artefact that changes while the application is running.
17.10 Understanding Plugin Load Order
Controlling Plugin Dependencies
Plugins often depend on the presence of other plugins and can adapt depending on the presence of others. This is implemented with two properties. The first is called
dependsOn
. For example, take a look at this snippet from the Hibernate plugin:
class HibernateGrailsPlugin { def version = "1.0" def dependsOn = [dataSource: "1.0",
domainClass: "1.0",
i18n: "1.0",
core: "1.0"]
}
The Hibernate plugin is dependent on the presence of four plugins: the
dataSource
,
domainClass
,
i18n
and
core
plugins.
The dependencies will be loaded before the Hibernate plugin and if all dependencies do not load, then the plugin will not load.
The
dependsOn
property also supports a mini expression language for specifying version ranges. A few examples of the syntax can be seen below:
def dependsOn = [foo: "* > 1.0"]
def dependsOn = [foo: "1.0 > 1.1"]
def dependsOn = [foo: "1.0 > *"]
When the wildcard * character is used it denotes "any" version. The expression syntax also excludes any suffixes such as -BETA, -ALPHA etc. so for example the expression "1.0 > 1.1" would match any of the following versions:
- 1.1
- 1.0
- 1.0.1
- 1.0.3-SNAPSHOT
- 1.1-BETA2
Controlling Load Order
Using
dependsOn
establishes a "hard" dependency in that if the dependency is not resolved, the plugin will give up and won't load. It is possible though to have a weaker dependency using the
loadAfter
and
loadBefore
properties:
def loadAfter = ['controllers']
Here the plugin will be loaded after the
controllers
plugin if it exists, otherwise it will just be loaded. The plugin can then adapt to the presence of the other plugin, for example the Hibernate plugin has this code in its
doWithSpring
closure:
if (manager?.hasGrailsPlugin("controllers")) {
openSessionInViewInterceptor(OpenSessionInViewInterceptor) {
flushMode = HibernateAccessor.FLUSH_MANUAL
sessionFactory = sessionFactory
}
grailsUrlHandlerMapping.interceptors << openSessionInViewInterceptor
}
Here the Hibernate plugin will only register an
OpenSessionInViewInterceptor
if the
controllers
plugin has been loaded. The
manager
variable is an instance of the
GrailsPluginManager interface and it provides methods to interact with other plugins.
You can also use the
loadBefore
property to specify one or more plugins that your plugin should load before:
def loadBefore = ['rabbitmq']
Scopes and Environments
It's not only plugin load order that you can control. You can also specify which environments your plugin should be loaded in and which scopes (stages of a build). Simply declare one or both of these properties in your plugin descriptor:
def environments = ['development', 'test', 'myCustomEnv']
def scopes = [excludes:'war']
In this example, the plugin will only load in the 'development' and 'test' environments. Nor will it be packaged into the WAR file, because it's excluded from the 'war' phase. This allows
development-only
plugins to not be packaged for production use.
The full list of available scopes are defined by the enum
BuildScope, but here's a summary:
test
- when running tests
functional-test
- when running functional tests
run
- for run-app and run-war
war
- when packaging the application as a WAR file
all
- plugin applies to all scopes (default)
Both properties can be one of:
- a string - a sole inclusion
- a list - a list of environments or scopes to include
- a map - for full control, with 'includes' and/or 'excludes' keys that can have string or list values
For example,
def environments = "test"
will only include the plugin in the test environment, whereas
def environments = ["development", "test"]
will include it in both the development
and test environments. Finally,
def environments = [includes: ["development", "test"]]
will do the same thing.
17.11 The Artefact API
You should by now understand that Grails has the concept of artefacts: special types of classes that it knows about and can treat differently from normal Groovy and Java classes, for example by enhancing them with extra properties and methods. Examples of artefacts include domain classes and controllers. What you may not be aware of is that Grails allows application and plugin developers access to the underlying infrastructure for artefacts, which means you can find out what artefacts are available and even enhance them yourself. You can even provide your own custom artefact types.
17.11.1 Asking About Available Artefacts
As a plugin developer, it can be important for you to find out about what domain classes, controllers, or other types of artefact are available in an application. For example, the
Searchable plugin needs to know what domain classes exist so it can check them for any
searchable
properties and index the appropriate ones. So how does it do it? The answer lies with the
grailsApplication
object, and instance of
GrailsApplication that's available automatically in controllers and GSPs and can be
injected everywhere else.
The
grailsApplication
object has several important properties and methods for querying artefacts. Probably the most common is the one that gives you all the classes of a particular artefact type:
for (cls in grailsApplication.<artefactType>Classes) {
…
}
In this case,
artefactType
is the property name form of the artefact type. With core Grails you have:
- domain
- controller
- tagLib
- service
- codec
- bootstrap
- urlMappings
So for example, if you want to iterate over all the domain classes, you use:
for (cls in grailsApplication.domainClasses) {
…
}
and for URL mappings:
for (cls in grailsApplication.urlMappingsClasses) {
…
}
You need to be aware that the objects returned by these properties are not instances of
Class. Instead, they are instances of
GrailsClass that has some particularly useful properties and methods, including one for the underlying
Class
:
shortName
- the class name of the artefact without the package (equivalent of Class.simpleName
).
logicalPropertyName
- the artefact name in property form without the 'type' suffix. So MyGreatController
becomes 'myGreat'.
isAbstract()
- a boolean indicating whether the artefact class is abstract or not.
getPropertyValue(name)
- returns the value of the given property, whether it's a static or an instance one. This works best if the property is initialised on declaration, e.g. static transactional = true
.
The artefact API also allows you to fetch classes by name and check whether a class is an artefact:
- get<type>Class(String name)
- is<type>Class(Class clazz)
The first method will retrieve the
GrailsClass
instance for the given name, e.g. 'MyGreatController'. The second will check whether a class is a particular type of artefact. For example, you can use
grailsApplication.isControllerClass(org.example.MyGreatController)
to check whether
MyGreatController
is in fact a controller.
17.11.2 Adding Your Own Artefact Types
Plugins can easily provide their own artefacts so that they can easily find out what implementations are available and take part in reloading. All you need to do is create an
ArtefactHandler
implementation and register it in your main plugin class:
class MyGrailsPlugin {
def artefacts = [ org.somewhere.MyArtefactHandler ]
…
}
The
artefacts
list can contain either handler classes (as above) or instances of handlers.
So, what does an artefact handler look like? Well, put simply it is an implementation of the
ArtefactHandler interface. To make life a bit easier, there is a skeleton implementation that can readily be extended:
ArtefactHandlerAdapter.
In addition to the handler itself, every new artefact needs a corresponding wrapper class that implements
GrailsClass. Again, skeleton implementations are available such as
AbstractInjectableGrailsClass, which is particularly useful as it turns your artefact into a Spring bean that is auto-wired, just like controllers and services.
The best way to understand how both the handler and wrapper classes work is to look at the Quartz plugin:
Another example is the
Shiro plugin which adds a realm artefact.
17.12 Binary Plugins
Regular Grails plugins are packaged as zip files containing the full source of the plugin. This has some advantages in terms of being an open distribution system (anyone can see the source), in addition to avoiding problems with the source compatibility level used for compilation.
As of Grails 2.0 you can pre-compile Grails plugins into regular JAR files known as "binary plugins". This has several advantages (and some disadvantages as discussed in the advantages of source plugins above) including:
- Binary plugins can be published as standard JAR files to a Maven repository
- Binary plugins can be declared like any other JAR dependency
- Commercial plugins are more viable since the source isn't published
- IDEs have a better understanding since binary plugins are regular JAR files containing classes
Packaging
To package a plugin in binary form you can use the package-plugin command and the
--binary
flag:
grails package-plugin --binary
Supported artefacts include:
- Grails artifact classes such as controllers, domain classes and so on
- I18n Message bundles
- GSP Views, layouts and templates
You can also specify the packaging in the plugin descriptor:
in which case the packaging will default to binary.
Using Binary Plugins
The packaging process creates a JAR file in the
target
directory of the plugin, for example
target/foo-plugin-0.1.jar
. There are two ways to incorporate a binary plugin into an application.
One is simply placing the plugin JAR file in your application's
lib
directory. The other is to publish the plugin JAR to a compatible Maven repository and declare it as a dependency in
grails-app/conf/BuildConfig.groovy
:
dependencies {
compile "mycompany:myplugin:0.1"
}
Since binary plugins are packaged as JAR files, they are declared as dependencies in the dependencies
block, not in the plugins
block as you may be naturally inclined to do. The plugins
block is used for declaring traditional source plugins packaged as zip files
18 Grails and Spring
This section is for advanced users and those who are interested in how Grails integrates with and builds on the
Spring Framework. It is also useful for
plugin developers considering doing runtime configuration Grails.
18.1 The Underpinnings of Grails
Grails is actually a
Spring MVC application in disguise. Spring MVC is the Spring framework's built-in MVC web application framework. Although Spring MVC suffers from some of the same difficulties as frameworks like Struts in terms of its ease of use, it is superbly designed and architected and was, for Grails, the perfect framework to build another framework on top of.
Grails leverages Spring MVC in the following areas:
- Basic controller logic - Grails subclasses Spring's DispatcherServlet and uses it to delegate to Grails controllers
- Data Binding and Validation - Grails' validation and data binding capabilities are built on those provided by Spring
- Runtime configuration - Grails' entire runtime convention based system is wired together by a Spring ApplicationContext
- Transactions - Grails uses Spring's transaction management in GORM
In other words Grails has Spring embedded running all the way through it.
The Grails ApplicationContext
Spring developers are often keen to understand how the Grails
ApplicationContext
instance is constructed. The basics of it are as follows.
- Grails constructs a parent
ApplicationContext
from the web-app/WEB-INF/applicationContext.xml
file. This ApplicationContext
configures the GrailsApplication instance and the GrailsPluginManager.
- Using this
ApplicationContext
as a parent Grails' analyses the conventions with the GrailsApplication
instance and constructs a child ApplicationContext
that is used as the root ApplicationContext
of the web application
Configured Spring Beans
Most of Grails' configuration happens at runtime. Each
plugin may configure Spring beans that are registered in the
ApplicationContext
. For a reference as to which beans are configured, refer to the reference guide which describes each of the Grails plugins and which beans they configure.
18.2 Configuring Additional Beans
Using the Spring Bean DSL
You can easily register new (or override existing) beans by configuring them in
grails-app/conf/spring/resources.groovy
which uses the Grails
Spring DSL. Beans are defined inside a
beans
property (a Closure):
beans = {
// beans here
}
As a simple example you can configure a bean with the following syntax:
import my.company.MyBeanImplbeans = {
myBean(MyBeanImpl) {
someProperty = 42
otherProperty = "blue"
}
}
Once configured, the bean can be auto-wired into Grails artifacts and other classes that support dependency injection (for example
BootStrap.groovy
and integration tests) by declaring a public field whose name is your bean's name (in this case
myBean
):
class ExampleController { def myBean
…
}
Using the DSL has the advantage that you can mix bean declarations and logic, for example based on the
environment:
import grails.util.Environment
import my.company.mock.MockImpl
import my.company.MyBeanImplbeans = {
switch(Environment.current) {
case Environment.PRODUCTION:
myBean(MyBeanImpl) {
someProperty = 42
otherProperty = "blue"
}
break case Environment.DEVELOPMENT:
myBean(MockImpl) {
someProperty = 42
otherProperty = "blue"
}
break
}
}
The
GrailsApplication
object can be accessed with the
application
variable and can be used to access the Grails configuration (amongst other things):
import grails.util.Environment
import my.company.mock.MockImpl
import my.company.MyBeanImplbeans = {
if (application.config.my.company.mockService) {
myBean(MockImpl) {
someProperty = 42
otherProperty = "blue"
}
} else {
myBean(MyBeanImpl) {
someProperty = 42
otherProperty = "blue"
}
}
}
If you define a bean in resources.groovy
with the same name as one previously registered by Grails or an installed plugin, your bean will replace the previous registration. This is a convenient way to customize behavior without resorting to editing plugin code or other approaches that would affect maintainability.
Using XML
Beans can also be configured using a
grails-app/conf/spring/resources.xml
. In earlier versions of Grails this file was automatically generated for you by the
run-app
script, but the DSL in
resources.groovy
is the preferred approach now so it isn't automatically generated now. But it is still supported - you just need to create it yourself.
This file is typical Spring XML file and the Spring documentation has an
excellent reference on how to configure Spring beans.
The
myBean
bean that we configured using the DSL would be configured with this syntax in the XML file:
<bean id="myBean" class="my.company.MyBeanImpl">
<property name="someProperty" value="42" />
<property name="otherProperty" value="blue" />
</bean>
Like the other bean it can be auto-wired into any class that supports dependency injection:
class ExampleController { def myBean
}
Referencing Existing Beans
Beans declared in
resources.groovy
or
resources.xml
can reference other beans by convention. For example if you had a
BookService
class its Spring bean name would be
bookService
, so your bean would reference it like this in the DSL:
beans = {
myBean(MyBeanImpl) {
someProperty = 42
otherProperty = "blue"
bookService = ref("bookService")
}
}
or like this in XML:
<bean id="myBean" class="my.company.MyBeanImpl">
<property name="someProperty" value="42" />
<property name="otherProperty" value="blue" />
<property name="bookService" ref="bookService" />
</bean>
The bean needs a public setter for the bean reference (and also the two simple properties), which in Groovy would be defined like this:
package my.companyclass MyBeanImpl {
Integer someProperty
String otherProperty
BookService bookService // or just "def bookService"
}
or in Java like this:
package my.company;class MyBeanImpl { private BookService bookService;
private Integer someProperty;
private String otherProperty; public void setBookService(BookService theBookService) {
this.bookService = theBookService;
} public void setSomeProperty(Integer someProperty) {
this.someProperty = someProperty;
} public void setOtherProperty(String otherProperty) {
this.otherProperty = otherProperty;
}
}
Using
ref
(in XML or the DSL) is very powerful since it configures a runtime reference, so the referenced bean doesn't have to exist yet. As long as it's in place when the final application context configuration occurs, everything will be resolved correctly.
For a full reference of the available beans see the plugin reference in the reference guide.
18.3 Runtime Spring with the Beans DSL
This Bean builder in Grails aims to provide a simplified way of wiring together dependencies that uses Spring at its core.
In addition, Spring's regular way of configuration (via XML and annotations) is static and difficult to modify and configure at runtime, other than programmatic XML creation which is both error prone and verbose. Grails'
BeanBuilder changes all that by making it possible to programmatically wire together components at runtime, allowing you to adapt the logic based on system properties or environment variables.
This enables the code to adapt to its environment and avoids unnecessary duplication of code (having different Spring configs for test, development and production environments)
The BeanBuilder class
Grails provides a
grails.spring.BeanBuilder class that uses dynamic Groovy to construct bean definitions. The basics are as follows:
import org.apache.commons.dbcp.BasicDataSource
import org.codehaus.groovy.grails.orm.hibernate.ConfigurableLocalSessionFactoryBean
import org.springframework.context.ApplicationContext
import grails.spring.BeanBuilderdef bb = new BeanBuilder()bb.beans { dataSource(BasicDataSource) {
driverClassName = "org.h2.Driver"
url = "jdbc:h2:mem:grailsDB"
username = "sa"
password = ""
} sessionFactory(ConfigurableLocalSessionFactoryBean) {
dataSource = ref('dataSource')
hibernateProperties = ["hibernate.hbm2ddl.auto": "create-drop",
"hibernate.show_sql": "true"]
}
}ApplicationContext appContext = bb.createApplicationContext()
Within plugins and the grails-app/conf/spring/resources.groovy file you don't need to create a new instance of BeanBuilder
. Instead the DSL is implicitly available inside the doWithSpring
and beans
blocks respectively.
This example shows how you would configure Hibernate with a data source with the
BeanBuilder
class.
Each method call (in this case
dataSource
and
sessionFactory
calls) maps to the name of the bean in Spring. The first argument to the method is the bean's class, whilst the last argument is a block. Within the body of the block you can set properties on the bean using standard Groovy syntax.
Bean references are resolved automatically using the name of the bean. This can be seen in the example above with the way the
sessionFactory
bean resolves the
dataSource
reference.
Certain special properties related to bean management can also be set by the builder, as seen in the following code:
sessionFactory(ConfigurableLocalSessionFactoryBean) { bean ->
// Autowiring behaviour. The other option is 'byType'. [autowire]
bean.autowire = 'byName'
// Sets the initialisation method to 'init'. [init-method]
bean.initMethod = 'init'
// Sets the destruction method to 'destroy'. [destroy-method]
bean.destroyMethod = 'destroy'
// Sets the scope of the bean. [scope]
bean.scope = 'request'
dataSource = ref('dataSource')
hibernateProperties = ["hibernate.hbm2ddl.auto": "create-drop",
"hibernate.show_sql": "true"]
}
The strings in square brackets are the names of the equivalent bean attributes in Spring's XML definition.
Using BeanBuilder with Spring MVC
Include the
grails-spring-<version>.jar
file in your classpath to use BeanBuilder in a regular Spring MVC application. Then add the following
<context-param>
values to your
/WEB-INF/web.xml
file:
<context-param>
<param-name>contextConfigLocation</param-name>
<param-value>/WEB-INF/applicationContext.groovy</param-value>
</context-param><context-param>
<param-name>contextClass</param-name>
<param-value>
org.codehaus.groovy.grails.commons.spring.GrailsWebApplicationContext
</param-value>
</context-param>
Then create a
/WEB-INF/applicationContext.groovy
file that does the rest:
import org.apache.commons.dbcp.BasicDataSourcebeans {
dataSource(BasicDataSource) {
driverClassName = "org.h2.Driver"
url = "jdbc:h2:mem:grailsDB"
username = "sa"
password = ""
}
}
Loading Bean Definitions from the File System
You can use the
BeanBuilder
class to load external Groovy scripts that define beans using the same path matching syntax defined here. For example:
def bb = new BeanBuilder()
bb.loadBeans("classpath:*SpringBeans.groovy")def applicationContext = bb.createApplicationContext()
Here the
BeanBuilder
loads all Groovy files on the classpath ending with
SpringBeans.groovy
and parses them into bean definitions. An example script can be seen below:
import org.apache.commons.dbcp.BasicDataSource
import org.codehaus.groovy.grails.orm.hibernate.ConfigurableLocalSessionFactoryBeanbeans { dataSource(BasicDataSource) {
driverClassName = "org.h2.Driver"
url = "jdbc:h2:mem:grailsDB"
username = "sa"
password = ""
} sessionFactory(ConfigurableLocalSessionFactoryBean) {
dataSource = dataSource
hibernateProperties = ["hibernate.hbm2ddl.auto": "create-drop",
"hibernate.show_sql": "true"]
}
}
Adding Variables to the Binding (Context)
If you're loading beans from a script you can set the binding to use by creating a Groovy
Binding
:
def binding = new Binding()
binding.maxSize = 10000
binding.productGroup = 'finance'def bb = new BeanBuilder()
bb.binding = binding
bb.loadBeans("classpath:*SpringBeans.groovy")def ctx = bb.createApplicationContext()
Then you can access the
maxSize
and
productGroup
properties in your DSL files.
18.4 The BeanBuilder DSL Explained
Using Constructor Arguments
Constructor arguments can be defined using parameters to each bean-defining method. Put them after the first argument (the Class):
bb.beans {
exampleBean(MyExampleBean, "firstArgument", 2) {
someProperty = [1, 2, 3]
}
}
This configuration corresponds to a
MyExampleBean
with a constructor that looks like this:
MyExampleBean(String foo, int bar) {
…
}
Configuring the BeanDefinition (Using factory methods)
The first argument to the closure is a reference to the bean configuration instance, which you can use to configure factory methods and invoke any method on the
AbstractBeanDefinition class:
bb.beans {
exampleBean(MyExampleBean) { bean ->
bean.factoryMethod = "getInstance"
bean.singleton = false
someProperty = [1, 2, 3]
}
}
As an alternative you can also use the return value of the bean defining method to configure the bean:
bb.beans {
def example = exampleBean(MyExampleBean) {
someProperty = [1, 2, 3]
}
example.factoryMethod = "getInstance"
}
Using Factory beans
Spring defines the concept of factory beans and often a bean is created not directly from a new instance of a Class, but from one of these factories. In this case the bean has no Class argument and instead you must pass the name of the factory bean to the bean defining method:
bb.beans { myFactory(ExampleFactoryBean) {
someProperty = [1, 2, 3]
} myBean(myFactory) {
name = "blah"
}
}
Another common approach is provide the name of the factory method to call on the factory bean. This can be done using Groovy's named parameter syntax:
bb.beans { myFactory(ExampleFactoryBean) {
someProperty = [1, 2, 3]
} myBean(myFactory: "getInstance") {
name = "blah"
}
}
Here the
getInstance
method on the
ExampleFactoryBean
bean will be called to create the
myBean
bean.
Creating Bean References at Runtime
Sometimes you don't know the name of the bean to be created until runtime. In this case you can use a string interpolation to invoke a bean defining method dynamically:
def beanName = "example"
bb.beans {
"${beanName}Bean"(MyExampleBean) {
someProperty = [1, 2, 3]
}
}
In this case the
beanName
variable defined earlier is used when invoking a bean defining method. The example has a hard-coded value but would work just as well with a name that is generated programmatically based on configuration, system properties, etc.
Furthermore, because sometimes bean names are not known until runtime you may need to reference them by name when wiring together other beans, in this case using the
ref
method:
def beanName = "example"
bb.beans { "${beanName}Bean"(MyExampleBean) {
someProperty = [1, 2, 3]
} anotherBean(AnotherBean) {
example = ref("${beanName}Bean")
}
}
Here the example property of
AnotherBean
is set using a runtime reference to the
exampleBean
. The
ref
method can also be used to refer to beans from a parent
ApplicationContext
that is provided in the constructor of the
BeanBuilder
:
ApplicationContext parent = ...//
def bb = new BeanBuilder(parent)
bb.beans {
anotherBean(AnotherBean) {
example = ref("${beanName}Bean", true)
}
}
Here the second parameter
true
specifies that the reference will look for the bean in the parent context.
Using Anonymous (Inner) Beans
You can use anonymous inner beans by setting a property of the bean to a block that takes an argument that is the bean type:
bb.beans { marge(Person) {
name = "Marge"
husband = { Person p ->
name = "Homer"
age = 45
props = [overweight: true, height: "1.8m"]
}
children = [ref('bart'), ref('lisa')]
} bart(Person) {
name = "Bart"
age = 11
} lisa(Person) {
name = "Lisa"
age = 9
}
}
In the above example we set the
marge
bean's husband property to a block that creates an inner bean reference. Alternatively if you have a factory bean you can omit the type and just use the specified bean definition instead to setup the factory:
bb.beans { personFactory(PersonFactory) marge(Person) {
name = "Marge"
husband = { bean ->
bean.factoryBean = "personFactory"
bean.factoryMethod = "newInstance"
name = "Homer"
age = 45
props = [overweight: true, height: "1.8m"]
}
children = [ref('bart'), ref('lisa')]
}
}
Abstract Beans and Parent Bean Definitions
To create an abstract bean definition define a bean without a
Class
parameter:
class HolyGrailQuest {
def start() { println "lets begin" }
}
class KnightOfTheRoundTable { String name
String leader
HolyGrailQuest quest KnightOfTheRoundTable(String name) {
this.name = name
} def embarkOnQuest() {
quest.start()
}
}
import grails.spring.BeanBuilderdef bb = new BeanBuilder()
bb.beans {
abstractBean {
leader = "Lancelot"
}
…
}
Here we define an abstract bean that has a
leader
property with the value of
"Lancelot"
. To use the abstract bean set it as the parent of the child bean:
bb.beans {
…
quest(HolyGrailQuest) knights(KnightOfTheRoundTable, "Camelot") { bean ->
bean.parent = abstractBean
quest = ref('quest')
}
}
When using a parent bean you must set the parent property of the bean before setting any other properties on the bean!
If you want an abstract bean that has a
Class
specified you can do it this way:
import grails.spring.BeanBuilderdef bb = new BeanBuilder()
bb.beans { abstractBean(KnightOfTheRoundTable) { bean ->
bean.'abstract' = true
leader = "Lancelot"
} quest(HolyGrailQuest) knights("Camelot") { bean ->
bean.parent = abstractBean
quest = quest
}
}
In this example we create an abstract bean of type
KnightOfTheRoundTable
and use the bean argument to set it to abstract. Later we define a knights bean that has no
Class
defined, but inherits the
Class
from the parent bean.
Using Spring Namespaces
Since Spring 2.0, users of Spring have had easier access to key features via XML namespaces. You can use a Spring namespace in BeanBuilder by declaring it with this syntax:
xmlns context:"http://www.springframework.org/schema/context"
and then invoking a method that matches the names of the Spring namespace tag and its associated attributes:
context.'component-scan'('base-package': "my.company.domain")
You can do some useful things with Spring namespaces, such as looking up a JNDI resource:
xmlns jee:"http://www.springframework.org/schema/jee"jee.'jndi-lookup'(id: "dataSource", 'jndi-name': "java:comp/env/myDataSource")
This example will create a Spring bean with the identifier
dataSource
by performing a JNDI lookup on the given JNDI name. With Spring namespaces you also get full access to all of the powerful AOP support in Spring from BeanBuilder. For example given these two classes:
class Person { int age
String name void birthday() {
++age;
}
}
class BirthdayCardSender { List peopleSentCards = [] void onBirthday(Person person) {
peopleSentCards << person
}
}
You can define an aspect that uses a pointcut to detect whenever the
birthday()
method is called:
xmlns aop:"http://www.springframework.org/schema/aop"fred(Person) {
name = "Fred"
age = 45
}birthdayCardSenderAspect(BirthdayCardSender)aop {
config("proxy-target-class": true) {
aspect(id: "sendBirthdayCard", ref: "birthdayCardSenderAspect") {
after method: "onBirthday",
pointcut: "execution(void ..Person.birthday()) and this(person)"
}
}
}
18.5 Property Placeholder Configuration
Grails supports the notion of property placeholder configuration through an extended version of Spring's
PropertyPlaceholderConfigurer, which is typically useful in combination with
externalized configuration.
Settings defined in either
ConfigSlurper scripts or Java properties files can be used as placeholder values for Spring configuration in
grails-app/conf/spring/resources.xml
and
grails-app/conf/spring/resources.groovy
. For example given the following entries in
grails-app/conf/Config.groovy
(or an externalized config):
database.driver="com.mysql.jdbc.Driver"
database.dbname="mysql:mydb"
You can then specify placeholders in
resources.xml
as follows using the familiar ${..} syntax:
<bean id="dataSource"
class="org.springframework.jdbc.datasource.DriverManagerDataSource">
<property name="driverClassName">
<value>${database.driver}</value>
</property>
<property name="url">
<value>jdbc:${database.dbname}</value>
</property>
</bean>
To specify placeholders in
resources.groovy
you need to use single quotes:
dataSource(org.springframework.jdbc.datasource.DriverManagerDataSource) {
driverClassName = '${database.driver}'
url = 'jdbc:${database.dbname}'
}
This sets the property value to a literal string which is later resolved against the config by Spring's PropertyPlaceholderConfigurer.
A better option for
resources.groovy
is to access properties through the
grailsApplication
variable:
dataSource(org.springframework.jdbc.datasource.DriverManagerDataSource) {
driverClassName = grailsApplication.config.database.driver
url = "jdbc:${grailsApplication.config.database.dbname}"
}
Using this approach will keep the types as defined in your config.
18.6 Property Override Configuration
Grails supports setting of bean properties via
configuration. This is often useful when used in combination with
externalized configuration.
You define a
beans
block with the names of beans and their values:
beans {
bookService {
webServiceURL = "http://www.amazon.com"
}
}
The general format is:
[bean name].[property name] = [value]
The same configuration in a Java properties file would be:
beans.bookService.webServiceURL=http://www.amazon.com
19 Grails and Hibernate
If
GORM (Grails Object Relational Mapping) is not flexible enough for your liking you can alternatively map your domain classes using Hibernate, either with XML mapping files or JPA annotations. You will be able to map Grails domain classes onto a wider range of legacy systems and have more flexibility in the creation of your database schema. Best of all, you will still be able to call all of the dynamic persistent and query methods provided by GORM!
19.1 Using Hibernate XML Mapping Files
Mapping your domain classes with XML is pretty straightforward. Simply create a
hibernate.cfg.xml
file in your project's
grails-app/conf/hibernate
directory, either manually or with the
create-hibernate-cfg-xml command, that contains the following:
<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE hibernate-configuration PUBLIC
"-//Hibernate/Hibernate Configuration DTD 3.0//EN"
"http://hibernate.sourceforge.net/hibernate-configuration-3.0.dtd">
<hibernate-configuration>
<session-factory>
<mapping resource="org.example.Book.hbm.xml"/>
…
</session-factory>
</hibernate-configuration>
The individual mapping files, like 'org.example.Book.hbm.xml' in the above example, also go into the
grails-app/conf/hibernate
directory. To find out how to map domain classes with XML, check out the
Hibernate manual.
If the default location of the
hibernate.cfg.xml
file doesn't suit you, you can change it by specifying an alternative location in
grails-app/conf/DataSource.groovy
:
hibernate {
config.location = "file:/path/to/my/hibernate.cfg.xml"
}
or even a list of locations:
hibernate {
config.location = ["file:/path/to/one/hibernate.cfg.xml",
"file:/path/to/two/hibernate.cfg.xml"]
}
Grails also lets you write your domain model in Java or reuse an existing one that already has Hibernate mapping files. Simply place the mapping files into
grails-app/conf/hibernate
and either put the Java files in
src/java
or the classes in the project's
lib
directory if the domain model is packaged as a JAR. You still need the
hibernate.cfg.xml
though!
19.2 Mapping with Hibernate Annotations
To map a domain class with annotations, create a new class in
src/java
and use the annotations defined as part of the EJB 3.0 spec (for more info on this see the
Hibernate Annotations Docs):
package com.books;import javax.persistence.Entity;
import javax.persistence.GeneratedValue;
import javax.persistence.Id;@Entity
public class Book {
private Long id;
private String title;
private String description;
private Date date; @Id
@GeneratedValue
public Long getId() {
return id;
} public void setId(Long id) {
this.id = id;
} public String getTitle() {
return title;
} public void setTitle(String title) {
this.title = title;
} public String getDescription() {
return description;
} public void setDescription(String description) {
this.description = description;
}
}
Then register the class with the Hibernate
sessionFactory
by adding relevant entries to the
grails-app/conf/hibernate/hibernate.cfg.xml
file as follows:
<!DOCTYPE hibernate-configuration SYSTEM
"http://hibernate.sourceforge.net/hibernate-configuration-3.0.dtd">
<hibernate-configuration>
<session-factory>
<mapping package="com.books" />
<mapping class="com.books.Book" />
</session-factory>
</hibernate-configuration>
See the previous section for more information on the
hibernate.cfg.xml
file.
When Grails loads it will register the necessary dynamic methods with the class. To see what else you can do with a Hibernate domain class see the section on
Scaffolding.
19.3 Adding Constraints
You can still use GORM validation even if you use a Java domain model. Grails lets you define constraints through separate scripts in the
src/java
directory. The script must be in a directory that matches the package of the corresponding domain class and its name must have a
Constraints suffix. For example, if you had a domain class
org.example.Book
, then you would create the script
src/java/org/example/BookConstraints.groovy
.
Add a standard GORM
constraints
block to the script:
constraints = {
title blank: false
author blank: false
}
Once this is in place you can validate instances of your domain class!
20 Scaffolding
Scaffolding lets you generate some basic CRUD interfaces for a domain class, including:
- The necessary views
- Controller actions for create/read/update/delete (CRUD) operations
As of Grails 2.3, the scaffolding feature has been moved to a plugin. By default this is configured for installation in new applications, but if you are upgrading from a previous version of Grails you will need to add the following configuration to your
BuildConfig.groovy
file:
plugins {
…
compile ":scaffolding:2.0.0"
…
}
Version 1.0.0 of the plugin provides the same scaffolding seen in Grails 2.2.x and below. Version 2.0.x of the scaffolding plugin includes different scaffolding templates that are aligned with the new REST APIs introduced in Grails 2.3 and above.
Dynamic Scaffolding
The simplest way to get started with scaffolding is to enable it with the
scaffold
property. Set the
scaffold
property in the controller to
true
for the
Book
domain class:
class BookController {
static scaffold = true
}
This works because the
BookController
follows the same naming convention as the
Book
domain class. To scaffold a specific domain class we could reference the class directly in the scaffold property:
class SomeController {
static scaffold = Author
}
With this configured, when you start your application the actions and views will be auto-generated at runtime. The following actions are dynamically implemented by default by the runtime scaffolding mechanism:
- index
- show
- edit
- delete
- create
- save
- update
A CRUD interface will also be generated. To access this open
http://localhost:8080/app/book
in a browser.
If you prefer to keep your domain model in Java and
mapped with Hibernate you can still use scaffolding, simply import the domain class and set its name as the
scaffold
argument.
You can add new actions to a scaffolded controller, for example:
class BookController { static scaffold = Book def changeAuthor() {
def b = Book.get(params.id)
b.author = Author.get(params["author.id"])
b.save() // redirect to a scaffolded action
redirect(action:show)
}
}
You can also override the scaffolded actions:
class BookController { static scaffold = Book // overrides scaffolded action to return both authors and books
def index() {
[bookInstanceList: Book.list(),
bookInstanceTotal: Book.count(),
authorInstanceList: Author.list()]
} def show() {
def book = Book.get(params.id)
log.error(book)
[bookInstance : book]
}
}
All of this is what is known as "dynamic scaffolding" where the CRUD interface is generated dynamically at runtime.
By default, the size of text areas in scaffolded views is defined in the CSS, so adding 'rows' and 'cols' attributes will have no effect.Also, the standard scaffold views expect model variables of the form <propertyName>InstanceList
for collections and <propertyName>Instance
for single instances. It's tempting to use properties like 'books' and 'book', but those won't work.
Customizing the Generated Views
The views adapt to
Validation constraints. For example you can change the order that fields appear in the views simply by re-ordering the constraints in the builder:
def constraints = {
title()
releaseDate()
}
You can also get the generator to generate lists instead of text inputs if you use the
inList
constraint:
def constraints = {
title()
category(inList: ["Fiction", "Non-fiction", "Biography"])
releaseDate()
}
Or if you use the
range
constraint on a number:
def constraints = {
age(range:18..65)
}
Restricting the size with a constraint also effects how many characters can be entered in the generated view:
def constraints = {
name(size:0..30)
}
Static Scaffolding
Grails also supports "static" scaffolding.
The above scaffolding features are useful but in real world situations it's likely that you will want to customize the logic and views. Grails lets you generate a controller and the views used to create the above interface from the command line. To generate a controller type:
grails generate-controller Book
or to generate the views:
grails generate-views Book
or to generate everything:
If you have a domain class in a package or are generating from a
Hibernate mapped class remember to include the fully qualified package name:
grails generate-all com.bookstore.Book
Customizing the Scaffolding templates
The templates used by Grails to generate the controller and views can be customized by installing the templates with the
install-templates command.
21 Deployment
Grails applications can be deployed in a number of ways, each of which has its pros and cons.
"grails run-app"
You should be very familiar with this approach by now, since it is the most common method of running an application during the development phase. An embedded Tomcat server is launched that loads the web application from the development sources, thus allowing it to pick up an changes to application files.
This approach is not recommended at all for production deployment because the performance is poor. Checking for and loading changes places a sizable overhead on the server. Having said that,
grails prod run-app
removes the per-request overhead and lets you fine tune how frequently the regular check takes place.
Setting the system property "disable.auto.recompile" to
true
disables this regular check completely, while the property "recompile.frequency" controls the frequency. This latter property should be set to the number of seconds you want between each check. The default is currently 3.
"grails run-war"
This is very similar to the previous option, but Tomcat runs against the packaged WAR file rather than the development sources. Hot-reloading is disabled, so you get good performance without the hassle of having to deploy the WAR file elsewhere.
WAR file
When it comes down to it, current java infrastructures almost mandate that web applications are deployed as WAR files, so this is by far the most common approach to Grails application deployment in production. Creating a WAR file is as simple as executing the
war command:
There are also many ways in which you can customise the WAR file that is created. For example, you can specify a path (either absolute or relative) to the command that instructs it where to place the file and what name to give it:
grails war /opt/java/tomcat-5.5.24/foobar.war
Alternatively, you can add a line to
grails-app/conf/BuildConfig.groovy
that changes the default location and filename:
grails.project.war.file = "foobar-prod.war"
Any command line argument that you provide overrides this setting.
It is also possible to control what libraries are included in the WAR file, for example to avoid conflicts with libraries in a shared directory. The default behavior is to include in the WAR file all libraries required by Grails, plus any libraries contained in plugin "lib" directories, plus any libraries contained in the application's "lib" directory. As an alternative to the default behavior you can explicitly specify the complete list of libraries to include in the WAR file by setting the property
grails.war.dependencies
in BuildConfig.groovy to either lists of Ant include patterns or closures containing AntBuilder syntax. Closures are invoked from within an Ant "copy" step, so only elements like "fileset" can be included, whereas each item in a pattern list is included. Any closure or pattern assigned to the latter property will be included in addition to
grails.war.dependencies
.
Be careful with these properties: if any of the libraries Grails depends on are missing, the application will almost certainly fail. Here is an example that includes a small subset of the standard Grails dependencies:
def deps = [
"hibernate3.jar",
"groovy-all-*.jar",
"standard-${servletVersion}.jar",
"jstl-${servletVersion}.jar",
"oscache-*.jar",
"commons-logging-*.jar",
"sitemesh-*.jar",
"spring-*.jar",
"log4j-*.jar",
"ognl-*.jar",
"commons-*.jar",
"xstream-1.2.1.jar",
"xpp3_min-1.1.3.4.O.jar" ]grails.war.dependencies = {
fileset(dir: "libs") {
for (pattern in deps) {
include(name: pattern)
}
}
}
This example only exists to demonstrate the syntax for the properties. If you attempt to use it as is in your own application, the application will probably not work. You can find a list of dependencies required by Grails in the "dependencies.txt" file in the root directory of the unpacked distribution. You can also find a list of the default dependencies included in WAR generation in the "War.groovy" script - see the
DEFAULT_DEPS
and
DEFAULT_J5_DEPS
variables.
The remaining two configuration options available to you are
grails.war.copyToWebApp
and
grails.war.resources
. The first of these lets you customise what files are included in the WAR file from the "web-app" directory. The second lets you do any extra processing you want before the WAR file is finally created.
// This closure is passed the command line arguments used to start the
// war process.
grails.war.copyToWebApp = { args ->
fileset(dir:"web-app") {
include(name: "js/**")
include(name: "css/**")
include(name: "WEB-INF/**")
}
}// This closure is passed the location of the staging directory that
// is zipped up to make the WAR file, and the command line arguments.
// Here we override the standard web.xml with our own.
grails.war.resources = { stagingDir, args ->
copy(file: "grails-app/conf/custom-web.xml",
tofile: "${stagingDir}/WEB-INF/web.xml")
}
Application servers
Ideally you should be able to simply drop a WAR file created by Grails into any application server and it should work straight away. However, things are rarely ever this simple. The
Grails website contains a list of application servers that Grails has been tested with, along with any additional steps required to get a Grails WAR file working.
22 Contributing to Grails
Grails is an open source project with an active community and we rely heavily on that community to help make Grails better. As such, there are various ways in which people can contribute to Grails. One of these is by
writing useful plugins and making them publicly available. In this chapter, we'll look at some of the other options.
22.1 Report Issues in JIRA
Grails uses
JIRA to track issues in the core framework, its documentation, its website, and many of the public plugins. If you've found a bug or wish to see a particular feature added, this is the place to start. You'll need to create a (free) JIRA account in order to either submit an issue or comment on an existing one.
When submitting issues, please provide as much information as possible and in the case of bugs, make sure you explain which versions of Grails and various plugins you are using. Also, an issue is much more likely to be dealt with if you attach a reproducible sample application (which can be packaged up using the
grails bug-report
command).
Reviewing issues
There are quite a few old issues in JIRA, some of which may no longer be valid. The core team can't track down these alone, so a very simple contribution that you can make is to verify one or two issues occasionally.
Which issues need verification? A shared
JIRA filter will display all issues that haven't been resolved and haven't been reviewed by someone else in the last 6 months. Just pick one or two of them and check whether they are still relevant.
Once you've verified an issue, simply edit it and set the "Last Reviewed" field to today. If you think the issue can be closed, then also check the "Flagged" field and add a short comment explaining why. Once those changes are saved, the issue will disappear from the results of the above filter. If you've flagged it, the core team will review and close if it really is no longer relevant.
One last thing: you can easily set the above filter as a favourite on
this JIRA screen so that it appears in the "Issues" drop down. Just click on the star next to a filter to make it a favourite.
22.2 Build From Source and Run Tests
If you're interested in contributing fixes and features to the core framework, you will have to learn how to get hold of the project's source, build it and test it with your own applications. Before you start, make sure you have:
- A JDK (1.6 or above)
- A git client
Once you have all the pre-requisite packages installed, the next step is to download the Grails source code, which is hosted at
GitHub in several repositories owned by the
"grails" GitHub user. This is a simple case of cloning the repository you're interested in. For example, to get the core framework run:
git clone http://github.com/grails/grails-core.git
This will create a "grails-core" directory in your current working directory containing all the project source files. The next step is to get a Grails installation from the source.
Creating a Grails installation
If you look at the project structure, you'll see that it doesn't look much like a standard
GRAILS_HOME
installation. But, it's very simple to turn it into one. Just run this from the root directory of the project:
This will fetch all the standard dependencies required by Grails and then build a
GRAILS_HOME
installation. Note that this target skips the extensive collection of Grails test classes, which can take some time to complete.
Once the above command has finished, simply set the
GRAILS_HOME
environment variable to the checkout directory and add the "bin" directory to your path. When you next type run the
grails
command, you'll be using the version you just built.
Running the test suite
All you have to do to run the full suite of tests is:
These will take a while (15-30 mins), so consider running individual tests using the command line. For example, to run the test spec
BinaryPluginSpec
simply execute the following command:
./gradlew :grails-core:test --tests *.BinaryPluginSpec
Note that you need to specify the sub-project that the test case resides in, because the top-level "test" target won't work....
Developing in IntelliJ IDEA
You need to run the following gradle task:
Then open the project file which is generated in IDEA. Simple!
Developing in STS / Eclipse
You need to run the following gradle task:
./gradlew cleanEclipse eclipse
Before importing projects to STS do the following action:
- Edit grails-scripts/.classpath and remove the line "<classpathentry kind="src" path="../scripts"/>".
Use "Import->General->Existing Projects into Workspace" to import all projects to STS. There will be a few build errors. To fix them do the following:
- Add the springloaded-core JAR file in $GRAILS_HOME/lib/org.springsource.springloaded/springloaded-core/jars to grails-core's classpath.
- Remove "src/test/groovy" from grails-plugin-testing's source path GRECLIPSE-1067
- Add the jsp-api JAR file in $GRAILS_HOME/lib/javax.servlet.jsp/jsp-api/jars to the classpath of grails-web
- Fix the source path of grails-scripts. Add linked source folder linking to "../scripts". If you get build errors in grails-scripts, do "../gradlew cleanEclipse eclipse" in that directory and edit the .classpath file again (remove the line "<classpathentry kind="src" path="../scripts"/>"). Remove possible empty "scripts" directory under grails-scripts if you are not able to add the linked folder.
- Do a clean build for the whole workspace.
- To use Eclipse GIT scm team provider: Select all projects (except "Servers") in the navigation and right click -> Team -> Share project (not "Share projects"). Choose "Git". Then check "Use or create repository in parent folder of project" and click "Finish".
- Get the recommended code style settings from the mailing list thread (final style not decided yet, currently profile.xml). Import the code style xml file to STS in Window->Preferences->Java->Code Style->Formatter->Import . Grails code uses spaces instead of tabs for indenting.
Debugging Grails or a Grails application
To enable debugging, run:
grails --debug-fork run-app
By default Grails forks a JVM to run the application in. The
--debug-fork
argument causes the debugger to be associated with the forked JVM. In order to instead attach the debugger to the build system which is going to fork the JVM use the
-debug
option:
22.3 Submit Patches to Grails Core
If you want to submit patches to the project, you simply need to fork the repository on GitHub rather than clone it directly. Then you will commit your changes to your fork and send a pull request for a core team member to review.
Forking and Pull Requests
One of the benefits of
GitHub is the way that you can easily contribute to a project by
forking the repository and
sending pull requests with your changes.
What follows are some guidelines to help ensure that your pull requests are speedily dealt with and provide the information we need. They will also make your life easier!
Create a local branch for your changes
Your life will be greatly simplified if you create a local branch to make your changes on. For example, as soon as you fork a repository and clone the fork locally, execute
This will create a new local branch called "mine" based off the "master" branch. Of course, you can name the branch whatever you like - you don't have to use "mine".
Create JIRAs for non-trivial changes
For any non-trivial changes, raise a JIRA issue if one doesn't already exist. That helps us keep track of what changes go into each new version of Grails.
Include JIRA issue ID in commit messages
This may not seem particularly important, but having a JIRA issue ID in a commit message means that we can find out at a later date why a change was made. Include the ID in any and all commits that relate to that issue. If a commit isn't related to an issue, then there's no need to include an issue ID.
Make sure your fork is up to date
Since the core developers must merge your commits into the main repository, it makes life much easier if your fork on GitHub is up to date before you send a pull request.
Let's say you have the main repository set up as a remote called "upstream" and you want to submit a pull request. Also, all your changes are currently on the local "mine" branch but not on "master". The first step involves pulling any changes from the main repository that have been added since you last fetched and merged:
git checkout master
git pull upstream
This should complete without any problems or conflicts. Next, rebase your local branch against the now up-to-date master:
git checkout mine
git rebase master
What this does is rearrange the commits such that all of your changes come after the most recent one in master. Think adding some cards to the top of a deck rather than shuffling them into the pack.
You'll now be able to do a clean merge from your local branch to master:
git checkout master
git merge mine
Finally, you must push your changes to your remote repository on GitHub, otherwise the core developers won't be able to pick them up:
You're now ready to send the pull request from the GitHub user interface.
Say what your pull request is for
A pull request can contain any number of commits and it may be related to any number of issues. In the pull request message, please specify the IDs of all issues that the request relates to. Also give a brief description of the work you have done, such as: "I refactored the data binder and added support for custom number editors (GRAILS-xxxx)".
22.4 Submit Patches to Grails Documentation
Contributing to the documentation is simpler for the core framework because there is a public fork of the
http://github.com/grails/grails-doc project that anyone can request commit access to. So, if you want to submit patches to the documentation, simply request commit access to the following repository
http://github.com/pledbrook/grails-doc by sending a GitHub message to 'pledbrook' and then commit your patches just as you would to any other GitHub repository.
Building the Guide
To build the documentation, simply type:
Be warned: this command can take a while to complete and you should probably increase your Gradle memory settings by giving the
GRADLE_OPTS
environment variable a value like
export GRADLE_OPTS="-Xmx512m -XX:MaxPermSize=384m"
Fortunately, you can reduce the overall build time with a couple of useful options. The first allows you to specify the location of the Grails source to use:
./gradlew -Dgrails.home=/home/user/projects/grails-core docs
The Grails source is required because the guide links to its API documentation and the build needs to ensure it's generated. If you don't specify a
grails.home
property, then the build will fetch the Grails source - a download of 10s of megabytes. It must then compile the Grails source which can take a while too.
Additionally you can create a local.properties file with this variable set:
grails.home=/home/user/projects/grails-core
or
grails.home=../grails-core
The other useful option allows you to disable the generation of the API documentation, since you only need to do it once:
./gradlew -Ddisable.groovydocs=true docs
Again, this can save a significant amount of time and memory.
The main English user guide is generated in the
build/docs
directory, with the
guide
sub-directory containing the user guide part and the
ref
folder containing the reference material. To view the user guide, simply open
build/docs/index.html
.
Publishing
The publishing system for the user guide is the same as
the one for Grails projects. You write your chapters and sections in the gdoc wiki format which is then converted to HTML for the final guide. Each chapter is a top-level gdoc file in the
src/<lang>/guide
directory. Sections and sub-sections then go into directories with the same name as the chapter gdoc but without the suffix.
The structure of the user guide is defined in the
src/<lang>/guide/toc.yml
file, which is a YAML file. This file also defines the (language-specific) section titles. If you add or remove a gdoc file, you must update the TOC as well!
The
src/<lang>/ref
directory contains the source for the reference sidebar. Each directory is the name of a category, which also appears in the docs. Hence the directories need different names for the different languages. Inside the directories go the gdoc files, whose names match the names of the methods, commands, properties or whatever that the files describe.
Translations
This project can host multiple translations of the user guide, with
src/en
being the main one. To add another one, simply create a new language directory under
src
and copy into it all the files under
src/en
. The build will take care of the rest.
Once you have a copy of the original guide, you can use the
{hidden}
macro to wrap the English text that you have replaced, rather than remove it. This makes it easier to compare changes to the English guide against your translation. For example:
{hidden}
When you create a Grails application with the [create-app|commandLine] command,
Grails doesn't automatically create an Ant build.xml
file but you can generate
one with the [integrate-with|commandLine] command:
{hidden}Quando crias uma aplicação Grails com o comando [create-app|commandLine], Grails
não cria automaticamente um ficheiro de construção Ant build.xml
mas podes gerar
um com o comando [integrate-with|commandLine]:
Because the English text remains in your gdoc files,
diff
will show differences on the English lines. You can then use the output of
diff
to see which bits of your translation need updating. On top of that, the
{hidden}
macro ensures that the text inside it is not displayed in the browser, although you can display it by adding this URL as a bookmark:
javascript:toggleHidden();
(requires you to build the user guide with Grails 2.0 M2 or later).
Even better, you can use the
left_to_do.groovy
script in the root of the project to see what still needs translating. You run it like so:
This will then print out a recursive diff of the given translation against the reference English user guide. Anything in
{hidden}
blocks that hasn't changed since being translated will
not appear in the diff output. In other words, all you will see is content that hasn't been translated yet and content that has changed since it was translated. Note that
{code}
blocks are ignored, so you
don't need to include them inside
{hidden}
macros.
To provide translations for the headers, such as the user guide title and subtitle, just add language specific entries in the 'resources/doc.properties' file like so:
es.title=El Grails Framework
es.subtitle=...
For each language translation, properties beginning
<lang>
. will override the standard ones. In the above example, the user guide title will be El Grails Framework for the Spanish translation. Also, translators can be credited by adding a '<lang>.translators' property:
fr.translators=Stéphane Maldini
This should be a comma-separated list of names (or the native language equivalent) and it will be displayed as a "Translated by" header in the user guide itself.
You can build specific translations very easily using the
publishGuide_*
and
publishPdf_*
tasks. For example, to build both the French HTML and PDF user guides, simply execute
Each translation is generated in its own directory, so for example the French guide will end up in
build/docs/fr
. You can then view the translated guide by opening
build/docs/<lang>/index.html
.
All translations are created as part of the
Hudson CI build for the grails-doc project, so you can easily see what the current state is without having to build the docs yourself.