By defining your receiver with android:process=":remote"
you basically run your receiver in a different process (= VM). For typical use-cases, you don't need to run this in a different process, and whatever you want to do can probably run just fine locally (in your APK process).
The drawback of using android:process=":remote"
is that you need additional resources for it to run (in this case a separate process). When doing so, you're basically dealing with 2 VMs, and some patterns like singletons, static fields can no longer be shared between your app and your remote service.
The benefit of using android:process=":remote"
is that for some use-cases, it might be handy to start a service that will keep on running (in its own process) after you've shutdown your application, or if you want remote clients to be able to bind to your service. Your broadcast receiver will not block your applications main thread when running in a separate process upon calling the onReceive()
method (however, there are other ways of implementing this).
I've found that most of the time, for most common use-cases, you can get away without using android:process=":remote"
.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…