Gradle任务依赖排序问题：指定子项目任务执行顺序无效

I get it, trying to get those subproject tasks to run in a specific sequence can be frustrating! The issue with your current setup is that when you write task(":module1:task") { ... } in the root project, you’re actually creating a brand new task with that path instead of referencing the existing task in the module1 subproject. That’s why your mustRunAfter rules aren’t taking effect—they’re applied to a task that never actually runs.

Here’s how to fix this properly:

Step 1: Reference Existing Subproject Tasks Correctly #

Instead of creating new tasks, you need to retrieve the existing tasks from each subproject using Gradle’s task lookup methods.

Step 2: Define the Execution Order #

Use mustRunAfter on the actual subproject tasks to set up your desired sequence. Adjust the order below to match your specific needs.

Step 3: Ensure bigTask Depends on All Tasks #

Make sure your bigTask explicitly depends on all these subproject tasks so Gradle knows to run them before executing bigTask.

Groovy DSL (root build.gradle) #

// Get references to existing subproject tasks
def module1Task = subprojects.findByName("module1")?.tasks?.named("task")
def module2Task = subprojects.findByName("module2")?.tasks?.named("task")
def module3Task = subprojects.findByName("module3")?.tasks?.named("task")
def module4Task = subprojects.findByName("module4")?.tasks?.named("task")
def module5Task = subprojects.findByName("module5")?.tasks?.named("task")

// Set your desired execution order (example sequence here)
module2Task?.mustRunAfter(module3Task)
module1Task?.mustRunAfter(module2Task)
module4Task?.mustRunAfter(module1Task)
module5Task?.mustRunAfter(module4Task)

// Configure your bigTask
tasks.named("bigTask") {
    // Add all subproject tasks as dependencies
    dependsOn(module1Task, module2Task, module3Task, module4Task, module5Task)
    // Optional: Reinforce sequential execution by linking outputs
    inputs.files(module1Task?.outputs, module2Task?.outputs, module3Task?.outputs, module4Task?.outputs, module5Task?.outputs)
}

Kotlin DSL (root build.gradle.kts) #

// Get references to existing subproject tasks
val module1Task = subprojects.find { it.name == "module1" }?.tasks?.named("task")
val module2Task = subprojects.find { it.name == "module2" }?.tasks?.named("task")
val module3Task = subprojects.find { it.name == "module3" }?.tasks?.named("task")
val module4Task = subprojects.find { it.name == "module4" }?.tasks?.named("task")
val module5Task = subprojects.find { it.name == "module5" }?.tasks?.named("task")

// Set your desired execution order (example sequence here)
module2Task?.mustRunAfter(module3Task)
module1Task?.mustRunAfter(module2Task)
module4Task?.mustRunAfter(module1Task)
module5Task?.mustRunAfter(module4Task)

// Configure your bigTask
tasks.named<JacocoReport>("bigTask") {
    // Add all subproject tasks as dependencies
    dependsOn(module1Task, module2Task, module3Task, module4Task, module5Task)
    // Optional: Reinforce sequential execution by linking outputs
    inputs.files(module1Task?.outputs, module2Task?.outputs, module3Task?.outputs, module4Task?.outputs, module5Task?.outputs)
}

Why This Works #

• By using subprojects.findByName(...).tasks.named(...), you’re accessing the actual tasks that exist in each subproject, not creating new ones.
• mustRunAfter establishes a strict execution order: each task will only run after the preceding task in your sequence has completed successfully.
• Adding all tasks as dependencies to bigTask ensures Gradle includes them in the execution graph when you run bigTask.

If you want to guarantee no parallel execution whatsoever (even if Gradle would normally parallelize independent tasks), you can add maxParallelForks = 1 to each subproject task, but the mustRunAfter rules alone should enforce sequential execution as long as dependencies are correctly configured.

内容的提问来源于stack exchange，提问作者jgm