IssueTracker

Is this specific to navigation? Or can this be reproduced with another other Composable component?

It only happens inside of a NavHost. So, yes, it seems to be specific to navigation.

This was caused by the fix to b/182284739 which made BackHandler lifecycle aware. The problem here is that the NavController registers its lifecycle and added an observer to the Activity Lifecycle in composition while the BackHandler registers its observer on the Activity Lifecycle in a DisposbleEffect. This means that the NavController will always get Lifecycle callbacks before the BackHandler, so the components listening the the NavController's lifecycle (like the NavBackStackEntry) will always get their Lifecycle callbacks before the BackHandler as well.

This should be addressed by work coming in 2.7 to integrate the BackHandler into the NavHost.

This has been fixed internally and will be available in Navigation 2.7.0-beta01.

The following release(s) address this bug.It is possible this bug has only been partially addressed:

• androidx.navigation:navigation-compose:2.7.0-beta01

I wouldn't say this is fixed. If you add a back-callback in the activity which is lifecycle aware, and then background->foreground the app, the activity back-callback will be put on top of the callback pile. This leads to that any BackHandler in your composable will NOT be triggered.

The best workaround I have for now is to use the non-lifecycle aware function in the activity.

class MainActivity : FragmentActivity() {
    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        
        // open fragment which hoists the composables
        ...

        if (savedInstanceState == null) {
             onBackPressedDispatcher.addCallback(onBackPressedCallback = someBackHandling())  
            // onBackPressedDispatcher.addCallback(owner = this, onBackPressedCallback = someBackHandling)  // not this one
        }
    }

https://developer.android.com/jetpack/androidx/releases/navigation#2.7.0-beta01 :

Bug Fixes : NavHost in Navigation Compose now correctly intercepts system back calls even after the Activity has been STOPPED and RESUMED.

It works fine for me.

Re comment#14 - please file a new issue with a project that reproduces your problem.

Note that you should never, ever be calling navController.popBackStack() in your own BackHandler as Navigation must be the only one handling its own back stack with the system back button so that Predictive Back works correctly in the future.

These changes broke our use case to override back handling in navigation library for closing the menu drawer. Please see more details in the issue I've reported.

I believe this is still an issue even without using NavHost like mentioned in comment #16. I filed a new issue here

The problem still exists.

Tech info: Android Studio 4.1.3 Build #AI-201.8743.12.41.7199119, built on March 11, 2021 Runtime version: 1.8.0_242-release-1644-b3-6915495 x86_64 VM: OpenJDK 64-Bit Server VM by JetBrains s.r.o macOS 10.15.7

Android Emulator version 30.5.3

Can confirm a workaround for this situation when bluetooth devices used - disconnect and connect headphones again when the emulator is already running. For some reasons after reconnecting the sound is fine. Based on that I believe it is something that happens only when AVD is starting.

P.S. Please stop sending +1, since it is not helpful and spams every other person following this issue. Just star the issue at the top of the page.

This is not fixed! And noise created by emulator now is even harder to get rid sometimes even disconnect/reconnect of earphones doesn't help :(

This is still occurring, but given this issue is marked as fixed I'd recommend starring/commenting on an open one e.g. https://issuetracker.google.com/u/0/issues/183139207

This problem just got worse. I have been able to work around this by setting hw.audioInput=no in my emulator config.ini. However, this recently stopped working.

This is enormously disruptive:

1. If I open or restart an emulator while on a Zoom call, I'm stuck with terrible audio for the remainder of the meeting.
2. Changing the mic back to the internal mic from my WH-1000XM3 doesn't change BT back to a2dp. I have to toggle Bluetooth
3. I can't toggle Bluetooth because I'm using a Bluetooth keyboard and mouse so if I turn it off, I have to go to my laptop, open the lid, and turn it back on.

M1 Pro MacBook + Emulator 31.2.8

Hey, I've been using -noaudio flag when launching an emulator and it's working fine.

Here's my example alias command:

alias androidsim="cd ~/Library/Android/sdk/emulator && ./emulator -avd Nexus_5X_API_28 -noaudio"

It's appalling that this is still a problem 8 years later. During a meeting trying to show case a feature using the Android 10 emulator and suddenly my audio gets killed, my microphone muted, and all I can hear is sounds from the emulator. No one in the meeting can hear me too. Checking the emulator settings shows the mic input is already off, but makes no difference. Solution was the same as #56 . It's deplorable that this is a well known bug and they still pretend to be fixed.

Hi particlecore, please try an API33 or API34 AVD. This is the sound card we use there. If you still observe the audio issues, they are caused by the QEMU audio stack. We are currently upgrading our QEMU (from the very old 2.12 to the most recent one, 8 or 9). We don't have resources to fix this in our copy of 2.12, but this could be fixed in the latest version.

Hi mtrakal, feel free to send us patches (for the new QEMU), we are open source: https://android.googlesource.com/platform/external/qemu/+/refs/heads/emu-dev