I am using the native-artifacts plugin which defines a number of tasks that extract the Nar dependencies of dependencies that are built elsewhere (and stored in Nexus/Maven). I need to ensure these tasks are called before the build of a binary, otherwise the headers that these Nars include are not found.
My question is, how do I define a system/plugin-defined task as a dependency of one of my tasks?
I'd like something like:
binaries.all {binary ->
dependencies {
// this next line is now the same as the plugin-defined task I want to have called before
// before the build takes place
compile "extractNarDeps${binary.name.capitalize()}"
}
}
Sadly, this doesn't build. How can I achieve this please? I have a component called unitTests that is a C++ component and is used to create unitTestsExecutable. I want extractNarDepsxxx called before compileUnitTests is called.
First of all, do you have any list named binaries? If yes, then you can try something like below in your build.gradle file as dependencies will be executed before your defined task:
dependencies {
binaries.each {binary ->
// this next line is now the same as the plugin-defined task I want to have called before
// before the build takes place
compile "extractNarDeps${binary.name.capitalize()}"
}
}
// Your defined task here
The way I did it was to specify my tasks as dependencies of the plugin's relevant task. The missing syntax link was in using "tasks" as a means of referencing the tasks in the project.
The answer isn't mine, it was very well explained here: tasks._applied_plugin_task_name_here.dependsOn(myTask)
Related
I have a Gradle-based library that is imported as a dependency into consuming applications. In other words, an application that consumes my library will have a build.gradle file with a list of dependencies that includes both my library as well as any other dependencies they wish to import.
From within my library's build.gradle file, I need to write a Gradle task that can access the full set of dependencies declared by the consuming application. In theory, this should be pretty straightforward, but hours of searching has not yielded a working solution yet.
The closest I've come is to follow this example and define an additional task in the library's build.gradle file that runs after the library is built:
build {
doLast {
project.getConfigurations().getByName('runtime')
.resolvedConfiguration
.firstLevelModuleDependencies
.each { println(it.name) }
}
}
I keep getting an error message that the 'runtime' configuration (passed into getByName and referenced in the Gradle forum post I linked) cannot be found. I have tried other common Gradle configurations that I can think of, but I never get any dependencies back from this code.
So: what is the best way to access the full set of dependencies declared by a consuming application from within the build file of one of those dependencies?
Okay, I mostly figured it out. The code snippet is essentially correct, but the configuration I should have been accessing was 'compileClasspath' or 'runtimeClasspath', not 'runtime'. This page helped me understand the configuration I was looking for.
The final build task in the library looks roughly like this:
build {
doLast {
// ...
def deps = project.getConfigurations().getByName('compileClasspath')
.resolvedConfiguration
.firstLevelModuleDependencies
.each {
// it.name will give you the dependency in the standard Gradle format (e.g."org.springframework.boot:spring-boot:1.5.22.RELEASE")
}
}
}
As can be seen from the top of documentation of class org.gradle.api.tasks.testing.Test, test tasks can be configured using the following piece of code:
test {
// configuration here. For example:
useJUnitPlatform()
}
From usage of method useJUnitPlatform we can assume that method test is called with a Closure which has an instance of aforementioned class Test as delegate.
In Gradle, there are other similar methods which take a Closure. For example, afterEvaluate. Documentation of method afterEvaluate is readily available in documentation of class Project. This is also mentioned in the user guide:
This example uses method Project.afterEvaluate() to add a closure which is executed after the project is evaluated.
Where is the documentation of method test? I could not find it. Or maybe this isn't a method in a class, but inserted via reflection into class Project at runtime or some similar DSL magic?
test in this context is not a method per se, but rather a task named test. To figure out what exactly is going on here requires diving into the internals of Gradle itself which is not part of any public documentation because well, it's not part of the public API.
The only way to figure exactly out what is going on is to debug Gradle during its execution. The easiest way to do that is to generate a plugin project via gradle init. Write a simple Gradle build file such as (build.gradle; I am assuming you are using the Groovy DSL):
plugins {
id("java")
}
test {
useJUnitPlatform()
}
Then write a basic functional test and start debugging. I was curious myself what is going and did just that.
In the following screenshot, you can see the stack trace in the bottom left corner. As you can see, there is a lot of methods called.
There is a mixture of Groovy specific methods and Gradle specific methods. Digging further in, you will come to:
You can see here bottom right that the list of objects is:
Project (root project)
Extra properties
Extensions
Tasks
This aligns with what I mentioned earlier: Gradle will go out of its way to match to what is being asked for. This is also explained in the "A Groovy Build Script Primer" in official documentation (starting from "If the name is unqualified [...]").
Eventually, you will land in some of the public API methods:
getByName is part of NamedDomainObjectContainer which is documented. However, what actually implements that interface is not as you can see from the Javadoc here. The implementation, from debugging, is DefaultTaskContainer.
The rest I will leave to you as an exercise. Hopefully this gives you an idea as to what is going on behind the scenes.
Indeed, test { ... } in this case is not calling a method with name test. This block is a feature of the Gradle API called "Groovy dynamic task configuration block". Per Gradle documentation version 6.1:
// Configure task using Groovy dynamic task configuration block
myCopy {
from 'resources'
into 'target'
}
myCopy.include('**/*.txt', '**/*.xml', '**/*.properties')
This works for any task. Task access is just a shortcut for the tasks.named() (Kotlin) or tasks.getByName() (Groovy) method. It is important to note that blocks used here are for configuring the task and are not evaluated when the task executes.
As such, per this shortcut convention, test { ... } block is used for configuring a task registered in the project – task with name test in this case.
Although nowadays I'd recommend using Gradle's Configuration Avoidance API to configure a task lazily instead of eagerly:
project.tasks.named('test', Test).configure {
it.useJunitPlatform()
}
See getByName replacement in the table "Existing vs New API overview".
My project uses several Gradle source sets for its production code base instead of just main:
domain
dal
rest
test
dbUnitTest
This has proven very useful for limiting dependencies and enforcing separation of concern.
It comes with one downside however: we cannot access classes or methods with visibility internal from within test classes. The reason for this is that the Kotlin compiler places every source set in its own "module":
$ find . -name '*.kotlin_module'
./classes/kotlin/domain/META-INF/contact-management_domain.kotlin_module
./classes/kotlin/dal/META-INF/contact-management_dal.kotlin_module
./classes/kotlin/rest/META-INF/contact-management_dal.kotlin_module
./classes/kotlin/test/META-INF/contact-management.kotlin_module
./classes/kotlin/dbUnitTest/META-INF/contact-management_dbUnitTest.kotlin_module
I would like all sourceset to use the same module name "contact-management", as the main sourceset would by default.
I tried to override the name with the compiler option -module-name:
tasks.withType<KotlinCompile> {
kotlinOptions {
// place all source sets in the same kotlin module, making objects with 'internal' visibility available to every source sets of this project
freeCompilerArgs += listOf("-module-name \"contact-management\")
}
}
Upon running gradlew build, I get
> Task :contact-management:compileDomainKotlin FAILED
e: Invalid argument: -module-name "contact-management"
The reason being that -module-name "contact-management_domain" is set before by the Gradle code invoking the Kotlin compiler as well, but apparently this option is only accepted once.
In a Gradle build, how can I control what is being considered "one module" by the Kotlin compiler?
A related question where the test source set is to be split has no satisfactory answers so far.
You can do that using kotlin compilations. (As far as I understand, a compilation is simply a block of files that are compiled together. A good explanation can be found here)
When you create a sourceset in gradle, the kotlin plugin creates a compilation under the hood (with the same name as the sourceset).
What you can do now with compilations is create associations. If a compilation A is associated with another compilation B, source code in A gets access to internal code units of B.
So in your case, if the test sourceset should get access to the dal sourceset you can simply associate the test compilation with the dal compilation:
kotlin.target.compilations.getByName("test").associateWith(kotlin.target.compilations.getByName("dal"))
PS: It also works the other way around. If you create compilations explicitly, the corresponding sourcesets are created under the hood. So for custom sourcesets you can create compilations and associate them:
val domainCompilation = kotlin.target.compilations.create("domain")
val dalCompilation = kotlin.target.compilations.create("dal") {
associateWith(domainCompilation)
}
In above example, the sourceset domain will have access to internal code units of the sourceset dal.
I have a build.gradle file cobbled together from examples online:
apply plugin: "java"
sourceSets {
java {
srcDirs = ['src']
}
}
repositories {
flatDir {
name "fileRepo"
dirs "repo"
}
}
uploadArchives {
repositories {
add project.repositories.fileRepo
}
}
When I run gradle tasks --all, I can see that "uploadArchives" is a task. How can I tell what is a task by looking at the build.gradle file? If "repositories" and "sourceSets" aren't considered tasks, what are they?
You simply can't.
But, the pure knowledge whether a closure configures a task or something else, won't give you anything. To understand a build script, you will need to understand the basic concept of Gradle and the used plugins, either built-in or third-party.
Each build.gradle script is executed against a Project instance. Everything you can access from the build script belongs to one of the following scopes:
The Project object itself. This scope includes any property getters and setters declared by the Project implementation class. For example, getRootProject() is accessible as the rootProject property. The properties of this scope are readable or writable depending on the presence of the corresponding getter or setter method.
The extra properties of the project. Each project maintains a map of extra properties, which can contain any arbitrary name -> value pair. Once defined, the properties of this scope are readable and writable. See extra properties for more details.
The extensions added to the project by the plugins. Each extension is available as a read-only property with the same name as the extension.
The convention properties added to the project by the plugins. A plugin can add properties and methods to a project through the project's Convention object. The properties of this scope may be readable or writable, depending on the convention objects.
The tasks of the project. A task is accessible by using its name as a property name. The properties of this scope are read-only. For example, a task called compile is accessible as the compile property.
The extra properties and convention properties are inherited from the project's parent, recursively up to the root project. The properties of this scope are read-only.
For your specific example, uploadArchives is a task, repositories belongs to the original Project object (it is available in each build script) and sourceSets is an extension of the java plugin.
Please note, that many plugins do not require or plan direct task configuration. They provide a DSL extension for configuration and then generate the tasks based on this configuration.
I’ve written a plugin (which currently just lives in buildSrc) that creates several tasks whose names are based on values provided by the user. How can I make it so that they execute whenever the build script that applies the plugin is run? It doesn't need to run at any specific point in the execution phase.
To start off with, you work around a basic Gradle concept. A Gradle task is not designed to run on every Gradle invocation. If you really need code to run on each Gradle invocation, execute it directly during configuration phase instead of wrapping it inside a task.
However, there are two causes for a task to run on a Gradle build:
direct selection (via command line or settings.startParameter.taskNames modification)
via one or more task dependencies (dependsOn / finalizedBy)
Of course you can use one of these methods to circumvent Gradle and execute your task on each build (#mkobit used the second method), but since your plugin would basically break basic Gradle principles, your solution may fail at some future time or for a more complex project (since plugins are supposed to be reusable).
As a summary, I would recommend to bundle all your generated tasks in one task with a constant name, so that your user can easily run the task on each Gradle invocation by putting a single line in his settings.gradle file:
startParameter.taskNames.add '<bundleTask>'
One way you could accomplish this is to use the all method on the TaskCollection to add a dependsOn/finalizedBy relationship to all (or some) tasks in the project.
Example to create a single myTask with every task in allproject depending on it:
class MyPlugin implements Plugin<Project> {
void apply(final Project project) {
final myTask = project.tasks.create('myTask')
project.allprojects.each { proj ->
proj.tasks.all {
// Make sure to not add a circular dependency
if (it != myTask) {
it.dependsOn(myTask)
}
}
}
}
}