api / org.gradle.testkit.runner / GradleRunner

GradleRunner

abstract class GradleRunner

Executes a Gradle build, allowing inspection of the outcome.

A Gradle runner can be used to functionally test build logic, by executing a contrived build. Assertions can then be made on the outcome of the build, such as the state of files created by the build, or what tasks were actually executed during the build.

A runner can be created via the #create() method.

Typically, the test code using the runner will programmatically create a build (e.g. by writing Gradle build files to a temporary space) to execute. The build to execute is effectively specified by the #withProjectDir(File)} method. It is a requirement that a project directory be set.

The #withArguments(String...) method allows the build arguments to be specified, just as they would be on the command line.

The #build() method can be used to invoke the build when it is expected to succeed, while the #buildAndFail() method can be used when the build is expected to fail.

GradleRunner instances are not thread safe and cannot be used concurrently. However, multiple instances are able to be used concurrently.

Please see the Gradle TestKit User Guide chapter for more information.

Since
2.6

Constructors

<init>

GradleRunner()

Executes a Gradle build, allowing inspection of the outcome.

A Gradle runner can be used to functionally test build logic, by executing a contrived build. Assertions can then be made on the outcome of the build, such as the state of files created by the build, or what tasks were actually executed during the build.

A runner can be created via the #create() method.

Typically, the test code using the runner will programmatically create a build (e.g. by writing Gradle build files to a temporary space) to execute. The build to execute is effectively specified by the #withProjectDir(File)} method. It is a requirement that a project directory be set.

The #withArguments(String...) method allows the build arguments to be specified, just as they would be on the command line.

The #build() method can be used to invoke the build when it is expected to succeed, while the #buildAndFail() method can be used when the build is expected to fail.

GradleRunner instances are not thread safe and cannot be used concurrently. However, multiple instances are able to be used concurrently.

Please see the Gradle TestKit User Guide chapter for more information.

Functions

build

abstract fun build(): BuildResult

Executes a build, expecting it to complete without failure.

buildAndFail

abstract fun buildAndFail(): BuildResult

Executes a build, expecting it to complete with failure.

create

open static fun create(): GradleRunner

Creates a new Gradle runner.

The runner requires a Gradle distribution (and therefore a specific version of Gradle) in order to execute builds. This method will find a Gradle distribution, based on the filesystem location of this class. That is, it is expected that this class is loaded from a Gradle distribution.

When using the runner as part of tests being executed by Gradle (i.e. a build using the gradleTestKit() dependency), this means that the same distribution of Gradle that is executing the tests will be used by runner returned by this method.

When using the runner as part of tests being executed by an IDE, this means that the same distribution of Gradle that was used when importing the project will be used.

forwardOutput

abstract fun forwardOutput(): GradleRunner

Forwards the output of executed builds to the System.out stream.

The output of the build is always available via BuildResult#getOutput(). This method can be used to additionally forward the output to System.out of the process using the runner.

This method does not separate the standard output and error output. The two streams will be merged as they typically are when using Gradle from a command line interface. If you require separation of the streams, you can use #forwardStdOutput(Writer) and #forwardStdError(Writer) directly.

Calling this method will negate the effect of previously calling #forwardStdOutput(Writer) and/or #forwardStdError(Writer).

forwardStdError

abstract fun forwardStdError(writer: Writer): GradleRunner

Configures the runner to forward standard error output from builds to the given writer.

The output of the build is always available via BuildResult#getOutput(). This method can be used to additionally capture the error output.

Calling this method will negate the effect of previously calling #forwardOutput().

The given writer will not be closed by the runner.

forwardStdOutput

abstract fun forwardStdOutput(writer: Writer): GradleRunner

Configures the runner to forward standard output from builds to the given writer.

The output of the build is always available via BuildResult#getOutput(). This method can be used to additionally capture the output.

Calling this method will negate the effect of previously calling #forwardOutput().

The given writer will not be closed by the runner.

When executing builds with Gradle versions earlier than 2.9 in debug mode, any output produced by the build that was written directly to System.out or System.err will not be represented in BuildResult#getOutput(). This is due to a defect that was fixed in Gradle 2.9.

getArguments

abstract fun getArguments(): MutableList<String>

The build arguments.

Effectively, the command line arguments to Gradle. This includes all tasks, flags, properties etc.

The returned list is immutable.

getPluginClasspath

abstract fun getPluginClasspath(): MutableList<out File>

The injected plugin classpath for the build.

The returned list is immutable. Returns an empty list if no classpath was provided with #withPluginClasspath(Iterable).

getProjectDir

abstract fun getProjectDir(): File

The directory that the build will be executed in.

This is analogous to the current directory when executing Gradle from the command line.

isDebug

abstract fun isDebug(): Boolean

Indicates whether the build should be executed “in process” so that it is debuggable.

If debug support is not enabled, the build will be executed in an entirely separate process. This means that any debugger that is attached to the test execution process will not be attached to the build process. When debug support is enabled, the build is executed in the same process that is using the Gradle Runner, allowing the build to be debugged.

Debug support is off (i.e. false) by default. It can be enabled by setting the system property org.gradle.testkit.debug to true for the test process, or by using the #withDebug(boolean) method.

withArguments

abstract fun withArguments(arguments: MutableList<String>): GradleRunner
abstract fun withArguments(vararg arguments: String): GradleRunner

Sets the build arguments.

withDebug

abstract fun withDebug(flag: Boolean): GradleRunner

Sets whether debugging support is enabled.

withGradleDistribution

abstract fun withGradleDistribution(distribution: URI): GradleRunner

Configures the runner to execute the build using the distribution of Gradle specified.

The given URI must point to a valid Gradle distribution ZIP file. This method is typically used as an alternative to #withGradleVersion(String), where it is preferable to obtain the Gradle runtime from “local” servers.

Unless previously downloaded, this method will cause the Gradle runtime at the given URI to be downloaded. The download will be cached beneath the Gradle User Home directory, the location of which is determined by the following in order of precedence:

  1. The system property "gradle.user.home"
  2. The environment variable "GRADLE_USER_HOME"

If neither are present, "~/.gradle" will be used, where "~" is the value advertised by the JVM's "user.dir" system property. The system property and environment variable are read in the process using the runner, not the build process.

withGradleInstallation

abstract fun withGradleInstallation(installation: File): GradleRunner

Configures the runner to execute the build using the installation of Gradle specified.

The given file must be a directory containing a valid Gradle installation.

Alternatively, you may use #withGradleVersion(String) to use an automatically installed Gradle version.

withGradleVersion

abstract fun withGradleVersion(versionNumber: String): GradleRunner

Configures the runner to execute the build with the version of Gradle specified.

Unless previously downloaded, this method will cause the Gradle runtime for the version specified to be downloaded over the Internet from Gradle's distribution servers. The download will be cached beneath the Gradle User Home directory, the location of which is determined by the following in order of precedence:

  1. The system property "gradle.user.home"
  2. The environment variable "GRADLE_USER_HOME"

If neither are present, "~/.gradle" will be used, where "~" is the value advertised by the JVM's "user.dir" system property. The system property and environment variable are read in the process using the runner, not the build process.

Alternatively, you may use #withGradleInstallation(File) to use an installation already on the filesystem.

To use a non standard Gradle runtime, or to obtain the runtime from an alternative location, use #withGradleDistribution(URI).

withPluginClasspath

abstract fun withPluginClasspath(): GradleRunner

Sets the plugin classpath based on the Gradle plugin development plugin conventions.

The 'java-gradle-plugin' generates a file describing the plugin under test and makes it available to the test runtime. This method configures the runner to use this file. Please consult the Gradle documentation of this plugin for more information.

This method looks for a file named plugin-under-test-metadata.properties on the runtime classpath, and uses the implementation-classpath as the classpath, which is expected to a File#pathSeparatorChar joined string. If the plugin metadata file cannot be resolved an InvalidPluginMetadataException is thrown.

Plugins from classpath are able to be resolved using the plugins { } syntax in the build under test. Please consult the TestKit Gradle User Guide chapter for more information and usage examples.

Calling this method will replace any previous classpath specified via #withPluginClasspath(Iterable) and vice versa.

Note: this method will cause an InvalidRunnerConfigurationException to be emitted when the build is executed, if the version of Gradle executing the build (i.e. not the version of the runner) is earlier than Gradle 2.8 as those versions do not support this feature. Please consult the TestKit Gradle User Guide chapter alternative strategies that can be used for older Gradle versions.

abstract fun withPluginClasspath(classpath: MutableIterable<File>): GradleRunner

Sets the injected plugin classpath for the build.

Plugins from the given classpath are able to be resolved using the plugins { } syntax in the build under test. Please consult the TestKit Gradle User Guide chapter for more information and usage examples.

Note: this method will cause an InvalidRunnerConfigurationException to be emitted when the build is executed, if the version of Gradle executing the build (i.e. not the version of the runner) is earlier than Gradle 2.8 as those versions do not support this feature. Please consult the TestKit Gradle User Guide chapter alternative strategies that can be used for older Gradle versions.

withProjectDir

abstract fun withProjectDir(projectDir: File): GradleRunner

Sets the directory that the Gradle will be executed in.

This is typically set to the root project of the build under test.

A project directory must be set. This method must be called before #build() or #buildAndFail().

All builds executed with the runner effectively do not search parent directories for a settings.gradle file. This suppresses Gradle's default behaviour of searching upwards through the file system in order to find the root of the current project tree. This default behaviour is often utilised when focusing on a particular build within a multi-project build. This behaviour is suppressed due to test builds being executed potentially being created within a “real build” (e.g. under the /build directory of the plugin's project directory).

withTestKitDir

abstract fun withTestKitDir(testKitDir: File): GradleRunner

Sets the directory to use for TestKit's working storage needs.

This directory is used internally to store various files required by the runner. If no explicit Gradle user home is specified via the build arguments (i.e. the -g «dir» option}), this directory will also be used for the Gradle user home for the test build.

If no value has been specified when the build is initiated, a directory unique to the current operating system user will be created and used within the JVM's temporary directory as advertised by the java.io.tmpdir system property. This directory is not deleted by the runner after the test build.

You may wish to specify a location that is within your project and regularly cleaned, such as the project's build directory.

The actual contents of this directory are an internal implementation detail and may change at any time.