Fork join
Since AndroidAnnotations 1.0
Fork / Join for the poor Android dev
Let's say you want to split a background operation into two separate operations that run concurrently, and then do something on the UI thread when both are done.假设你想将一个后台操作分成单独的两部分,并且需要同时运行,然后两部分都完成时在UI线程更新。
Here is a simple way to implement this, thanks to @Background and @UiThread.这里有一个简单的方法来实现,感谢 @Background 和@UiThread。
@EActivity public class MyActivity extends Activity { // Instances should only be accessed from the UI Thread to guarantee thread safety static class ResultHolder { ResultA resultA; ResultB resultB; } // Multiple clicks will start multiple distinguished computations @Click(R.id.myButton) void startForkableComputation() { ResultHolder resultHolder = new ResultHolder(); computeResultA(resultHolder); computeResultB(resultHolder); } @Background void computeResultA(ResultHolder resultHolder) { ResultA resultA = new ResultA(); // Do some stuff with resultA joinWork(resultHolder, resultA, null); } @Background void computeResultB(ResultHolder resultHolder) { ResultB resultB = new ResultB(); // Do some stuff with resultB joinWork(resultHolder, null, resultB); } @UiThread void joinWork(ResultHolder resultHolder, ResultA resultA, ResultB resultB) { if (resultA != null) resultHolder.resultA = resultA; if (resultB != null) resultHolder.resultB = resultB; if (resultHolder.resultA == null || resultHolder.resultB == null) { return; } // Show the results on the UI Thread } }
This works because we are joining the work in the Ui Thread, which is always the same thread. We also ensure that we always access ResultHolder from the Ui Thread.合并工作总是在UI线程中进行,我们保证时刻访问 ResultHolder 。
Note by @pyricau: although I'm quite sure everything here is right and safe, I just received Java Concurrency in Practice from Amazon, so I'll read it with this fork / join use case in mind!Note by @pyricau:最然我很确信万事OK,但我刚刚收到来自Amazon的 Java Concurrency in Practice ,所以我会带着 fork / join 这个例子读这本书。

本文介绍了一种使用@Background和@UiThread注解简化Android应用中并发操作的方法。通过将后台任务分解为两个并行执行的部分,并在UI线程上同步它们的结果,确保了线程安全性和操作效率。
575

被折叠的 条评论
为什么被折叠?



