WorkManager不適用于自定義配置

我正在嘗試使用 運行同步作業，WorkManager但似乎傳遞的配置被部分忽略。

下面是我的實作：

應用程式：

class App : Application(), HasAndroidInjector, Configuration.Provider {
@Inject
    lateinit var workerConfiguration: Configuration
override fun getWorkManagerConfiguration(): Configuration {
        Timber.tag(javaClass.name).d("RRR config ${workerConfiguration.executor.run { Thread.currentThread().name }}")
        return workerConfiguration
    }
}

模塊注入：如您所見，我設定了 1 個 SingleThreaded Executor 與一些特定ThreadFactory和自定義WorkerFactory

@Singleton
@Provides
    fun provideWorkManagerConfiguration(
        globalWorkerFactory: GlobalWorkerFactory
    ): Configuration {
        val threadFactory = ThreadFactoryBuilder().setNameFormat("sync-worker-thread-%d").build()
        val executor = Executors.newSingleThreadExecutor(threadFactory)
        executor.submit { Log.d(javaClass.name,"RRR executor thread = ${Thread.currentThread().name}") }
        return Configuration.Builder()
            .setWorkerFactory(globalWorkerFactory)
            .setExecutor(executor)
            .build()
    }

GlobalWorkingFactory：用于向worker注入一些引數

@Singleton
class GlobalWorkerFactory @Inject constructor(
    api: API
) : DelegatingWorkerFactory() {
    init {
        Log.d(javaClass.name, "RRR adding factory ${Thread.currentThread().name}")
        addFactory(WorkerFactory(api))
    }
}

工人工廠：

CustomWorkerFactory(
    private val api: API,
) : WorkerFactory() {

    override fun createWorker(
        appContext: Context,
        workerClassName: String,
        workerParameters: WorkerParameters
    ): CoroutineWorker? {
        return when (workerClassName) {
            SyncWorker::class.java.name -> {
                Log.d(javaClass.name, "RRR gettting sync worker on ${Thread.currentThread().name}")
                SyncWorker(appContext, workerParameters, api)
            }
            else ->
                null
        }
    }
}

同步作業執行緒執行一些長時間運行的呼叫：

class SyncWorker constructor(
    appContext: Context,
    workerParams: WorkerParameters,
    private val api: API):
    CoroutineWorker(appContext, workerParams) {

override suspend fun doWork(): Result {

Timber.d("RRR START sync file ${uploadingFile.file.name} on thread ${Thread.currentThread().name}")

syncfile()

Timber.d("RRR returning result for ${uploadingFile.file.name}")

}

最后是我對作業進行排隊的方式：

fun addFileToWorker(context: Context, uploadingFile: UploadingFile) {
            val constraints = Constraints.Builder()
                .setRequiredNetworkType(NetworkType.UNMETERED)
                .build()

            val syncWorkerRequest = OneTimeWorkRequestBuilder<SyncWorker>()
                .setInputData(
                    workDataOf(
                        "uploadingFile" to uploadingFileAdapter.toJson(
                            uploadingFile
                        )
                    )
                )
                .addTag(SYNC_WORK_TAG)
                .setConstraints(constraints)
                .setBackoffCriteria(BackoffPolicy.EXPONENTIAL, 10, TimeUnit.SECONDS)
                .build()

            val operation = WorkManager.getInstance(context)
                .enqueueUniqueWork(
                    Sunique_ID_for_each_job
                    ExistingWorkPolicy.KEEP,
                    syncWorkerRequest)
                .result
            operation.addListener({ Timber.d("RRR added file ${uploadingFile.file.name} ${Thread.currentThread().name}")}, { it.run() })
        }

以下是我如何獲取它們的日志，它表明檔案正在與分配的執行緒不同的執行緒上處理：

core.worker.GlobalWorkerFactory: RRR adding factory main
di.module.ApplicationModule: RRR executor thread = sync-worker-thread-0
Activity: RRR calling process sessions
App: RRR config main
Utils$Companion: RRR added file 0.aes pool-14-thread-2
Utils$Companion: RRR added file 1635858822 pool-14-thread-3
CustomWorkerFactory: RRR gettting sync worker on pool-14-thread-1
CustomWorkerFactory: RRR gettting sync worker on pool-14-thread-2
SyncWorker: RRR START sync file 0.aes on thread DefaultDispatcher-worker-1
SyncWorker: RRR START sync file 1635858822 on thread DefaultDispatcher-worker-4
SyncWorker: RRR START sync file 0.aes on thread DefaultDispatcher-worker-3

由于我已經明確定義了 1 個執行緒并且我成功地將我的執行緒的配置傳遞給 GlobalWorkerFactory，它不應該尊重第一個日志中看到的執行緒嗎？即：sync-worker-thread-0

I saw other posts suggesting to just APPEND the work to one worker, but I thought that's not how it's meant to be used. And even then, it's not the thread I specified that is used for the work with or without the APPEND.

uj5u.com熱心網友回復：

來自關于工人的 Android 檔案：Doc

注意 CoroutineWorker.doWork() 是一個掛起函式。與 Worker 不同的是，此代碼不會在您的配置中指定的 Executor 上運行。相反，它默認為 Dispatchers.Default。你可以通過提供你自己的 CoroutineContext 來定制它。在上面的示例中，您可能希望在 Dispatchers.IO 上執行此作業，如下所示：

class CoroutineDownloadWorker(
    context: Context,
    params: WorkerParameters
) : CoroutineWorker(context, params) {

    override suspend fun doWork(): Result {
        withContext(Dispatchers.IO) {
            val data = downloadSynchronously("https://www.google.com")
            saveData(data)
            return Result.success()
        }
    }
}

就我而言，我創建了一個newSingleThreadContext并將其傳遞給我的工人。

標籤：android multithreading android-workmanager

上一篇：函式代替靜態實用方法

下一篇：使用beautifulsoup刮<h2>