/*
* Copyright 2015 the original author or authors.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.gradle.testkit.runner;
import org.gradle.api.Incubating;
import org.gradle.testkit.runner.internal.DefaultGradleRunner;
import java.io.File;
import java.io.Writer;
import java.net.URI;
import java.util.List;
/**
* Executes a Gradle build, allowing inspection of the outcome.
* <p>
* 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.
* <p>
* A runner can be created via the {@link #create()} method.
* <p>
* 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 {@link #withProjectDir(File)}} method.
* It is a requirement that a project directory be set.
* <p>
* The {@link #withArguments(String...)} method allows the build arguments to be specified,
* just as they would be on the command line.
* <p>
* The {@link #build()} method can be used to invoke the build when it is expected to succeed,
* while the {@link #buildAndFail()} method can be used when the build is expected to fail.
* <p>
* GradleRunner instances are not thread safe and cannot be used concurrently.
* However, multiple instances are able to be used concurrently.
* <p>
* Please see <a href="https://docs.gradle.org/current/userguide/test_kit.html" target="_top">the Gradle TestKit User Guide chapter</a> for more information.
*
* @since 2.6
*/
@Incubating
public abstract class GradleRunner {
/**
* Creates a new Gradle runner.
* <p>
* 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.
* <p>
* When using the runner as part of tests <i>being executed by Gradle</i> (i.e. a build using the {@code gradleTestKit()} dependency),
* this means that the same distribution of Gradle that is executing the tests will be used by runner returned by this method.
* <p>
* When using the runner as part of tests <i>being executed by an IDE</i>,
* this means that the same distribution of Gradle that was used when importing the project will be used.
*
* @return a new Gradle runner
*/
public static GradleRunner create() {
return new DefaultGradleRunner();
}
/**
* Configures the runner to execute the build with the version of Gradle specified.
* <p>
* 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:
* <ol>
* <li>The system property {@code "gradle.user.home"}</li>
* <li>The environment variable {@code "GRADLE_USER_HOME"}</li>
* </ol>
* <p>
* If neither are present, {@code "~/.gradle"} will be used, where {@code "~"} is the value advertised by the JVM's {@code "user.dir"} system property.
* The system property and environment variable are read in the process using the runner, not the build process.
* <p>
* Alternatively, you may use {@link #withGradleInstallation(File)} to use an installation already on the filesystem.
* <p>
* To use a non standard Gradle runtime, or to obtain the runtime from an alternative location, use {@link #withGradleDistribution(URI)}.
*
* @param versionNumber the version number (e.g. "2.9")
* @return this
* @since 2.9
* @see #withGradleInstallation(File)
* @see #withGradleDistribution(URI)
*/
public abstract GradleRunner withGradleVersion(String versionNumber);
/**
* Configures the runner to execute the build using the installation of Gradle specified.
* <p>
* The given file must be a directory containing a valid Gradle installation.
* <p>
* Alternatively, you may use {@link #withGradleVersion(String)} to use an automatically installed Gradle version.
*
* @param installation a valid Gradle installation
* @return this
* @since 2.9
* @see #withGradleVersion(String)
* @see #withGradleDistribution(URI)
*/
public abstract GradleRunner withGradleInstallation(File installation);
/**
* Configures the runner to execute the build using the distribution of Gradle specified.
* <p>
* The given URI must point to a valid Gradle distribution ZIP file.
* This method is typically used as an alternative to {@link #withGradleVersion(String)},
* where it is preferable to obtain the Gradle runtime from “local” servers.
* <p>
* 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:
* <ol>
* <li>The system property {@code "gradle.user.home"}</li>
* <li>The environment variable {@code "GRADLE_USER_HOME"}</li>
* </ol>
* <p>
* If neither are present, {@code "~/.gradle"} will be used, where {@code "~"} is the value advertised by the JVM's {@code "user.dir"} system property.
* The system property and environment variable are read in the process using the runner, not the build process.
*
* @param distribution a URI pointing at a valid Gradle distribution zip file
* @return this
* @since 2.9
* @see #withGradleVersion(String)
* @see #withGradleInstallation(File)
*/
public abstract GradleRunner withGradleDistribution(URI distribution);
/**
* Sets the directory to use for TestKit's working storage needs.
* <p>
* 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 {@code -g «dir»} option}),
* this directory will also be used for the Gradle user home for the test build.
* <p>
* 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 {@code java.io.tmpdir} system property.
* This directory is not deleted by the runner after the test build.
* <p>
* You may wish to specify a location that is within your project and regularly cleaned, such as the project's build directory.
* <p>
* The actual contents of this directory are an internal implementation detail and may change at any time.
*
* @param testKitDir the TestKit directory
* @return {@code this}
* @since 2.7
*/
public abstract GradleRunner withTestKitDir(File testKitDir);
/**
* The directory that the build will be executed in.
* <p>
* This is analogous to the current directory when executing Gradle from the command line.
*
* @return the directory to execute the build in
*/
public abstract File getProjectDir();
/**
* Sets the directory that the Gradle will be executed in.
* <p>
* This is typically set to the root project of the build under test.
* <p>
* A project directory must be set.
* This method must be called before {@link #build()} or {@link #buildAndFail()}.
* <p>
* All builds executed with the runner effectively implicitly add the {@code --no-search-upward} argument.
* 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 {@code /build} directory of the plugin's project directory).
*
* @param projectDir the project directory
* @return {@code this}
* @see #getProjectDir()
*/
public abstract GradleRunner withProjectDir(File projectDir);
/**
* The build arguments.
* <p>
* Effectively, the command line arguments to Gradle.
* This includes all tasks, flags, properties etc.
* <p>
* The returned list is immutable.
*
* @return the build arguments
*/
public abstract List<String> getArguments();
/**
* Sets the build arguments.
*
* @param arguments the build arguments
* @return this
* @see #getArguments()
*/
public abstract GradleRunner withArguments(List<String> arguments);
/**
* Sets the build arguments.
*
* @param arguments the build arguments
* @return this
* @see #getArguments()
*/
public abstract GradleRunner withArguments(String... arguments);
/**
* The injected plugin classpath for the build.
* <p>
* The returned list is immutable.
* Returns an empty list if no classpath was provided with {@link #withPluginClasspath(Iterable)}.
*
* @return the classpath of plugins to make available to the build under test
* @since 2.8
*/
public abstract List<? extends File> getPluginClasspath();
/**
* Sets the plugin classpath based on the Gradle plugin development plugin conventions.
* <p>
* 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.
* <p>
* This method looks for a file named {@code plugin-under-test-metadata.properties} on the runtime classpath,
* and uses the {@code implementation-classpath} as the classpath, which is expected to a {@link File#pathSeparatorChar} joined string.
* If the plugin metadata file cannot be resolved an {@link InvalidPluginMetadataException} is thrown.
* <p>
* Plugins from classpath are able to be resolved using the <code>plugins { }</code> syntax in the build under test.
* Please consult the TestKit Gradle User Guide chapter for more information and usage examples.
* <p>
* Calling this method will replace any previous classpath specified via {@link #withPluginClasspath(Iterable)} and vice versa.
* <p>
* <b>Note:</b> this method will cause an {@link 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.
*
* @return this
* @see #withPluginClasspath(Iterable)
* @see #getPluginClasspath()
* @since 2.13
*/
public abstract GradleRunner withPluginClasspath() throws InvalidPluginMetadataException;
/**
* Sets the injected plugin classpath for the build.
* <p>
* Plugins from the given classpath are able to be resolved using the <code>plugins { }</code> syntax in the build under test.
* Please consult the TestKit Gradle User Guide chapter for more information and usage examples.
* <p>
* <b>Note:</b> this method will cause an {@link 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.
*
* @param classpath the classpath of plugins to make available to the build under test
* @return this
* @see #getPluginClasspath()
* @since 2.8
*/
public abstract GradleRunner withPluginClasspath(Iterable<? extends File> classpath);
/**
* Indicates whether the build should be executed “in process” so that it is debuggable.
* <p>
* 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.
* <p>
* Debug support is off (i.e. {@code false}) by default.
* It can be enabled by setting the system property {@code org.gradle.testkit.debug} to {@code true} for the test process,
* or by using the {@link #withDebug(boolean)} method.
*
* @return whether the build should be executed in the same process
* @since 2.9
*/
public abstract boolean isDebug();
/**
* Sets whether debugging support is enabled.
*
* @see #isDebug()
* @param flag the debug flag
* @return this
* @since 2.9
*/
public abstract GradleRunner withDebug(boolean flag);
/**
* Configures the runner to forward standard output from builds to the given writer.
* <p>
* The output of the build is always available via {@link BuildResult#getOutput()}.
* This method can be used to additionally capture the output.
* <p>
* Calling this method will negate the effect of previously calling {@link #forwardOutput()}.
* <p>
* The given writer will not be closed by the runner.
* <p>
* When executing builds with Gradle versions earlier than 2.9 <b>in debug mode</b>, any output produced by the build that
* was written directly to {@code System.out} or {@code System.err} will not be represented in {@link BuildResult#getOutput()}.
* This is due to a defect that was fixed in Gradle 2.9.
*
* @param writer the writer that build standard output should be forwarded to
* @return this
* @since 2.9
* @see #forwardOutput()
* @see #forwardStdError(Writer)
*/
public abstract GradleRunner forwardStdOutput(Writer writer);
/**
* Configures the runner to forward standard error output from builds to the given writer.
* <p>
* The output of the build is always available via {@link BuildResult#getOutput()}.
* This method can be used to additionally capture the error output.
* <p>
* Calling this method will negate the effect of previously calling {@link #forwardOutput()}.
* <p>
* The given writer will not be closed by the runner.
*
* @param writer the writer that build standard error output should be forwarded to
* @return this
* @since 2.9
* @see #forwardOutput()
* @see #forwardStdOutput(Writer)
*/
public abstract GradleRunner forwardStdError(Writer writer);
/**
* Forwards the output of executed builds to the {@link System#out System.out} stream.
* <p>
* The output of the build is always available via {@link BuildResult#getOutput()}.
* This method can be used to additionally forward the output to {@code System.out} of the process using the runner.
* <p>
* 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 {@link #forwardStdOutput(Writer)} and {@link #forwardStdError(Writer)} directly.
* <p>
* Calling this method will negate the effect of previously calling {@link #forwardStdOutput(Writer)} and/or {@link #forwardStdError(Writer)}.
*
* @return this
* @since 2.9
* @see #forwardStdOutput(Writer)
* @see #forwardStdError(Writer)
*/
public abstract GradleRunner forwardOutput();
/**
* Executes a build, expecting it to complete without failure.
*
* @throws InvalidRunnerConfigurationException if the configuration of this runner is invalid (e.g. project directory not set)
* @throws UnexpectedBuildFailure if the build does not succeed
* @return the build result
*/
public abstract BuildResult build() throws InvalidRunnerConfigurationException, UnexpectedBuildFailure;
/**
* Executes a build, expecting it to complete with failure.
*
* @throws InvalidRunnerConfigurationException if the configuration of this runner is invalid (e.g. project directory not set)
* @throws UnexpectedBuildSuccess if the build succeeds
* @return the build result
*/
public abstract BuildResult buildAndFail() throws InvalidRunnerConfigurationException, UnexpectedBuildSuccess;
}