IssueTracker

Please provide the requested information to proceed further. Unfortunately the issue will be closed within 7 days if there is no further update.

We have shared this with our product and engineering team and will update this issue with more information as it becomes available.

Thanks for reporting this issue.

COARSE_LOCATION typically takes location information from the nearby cell tower. If many users are near the same cell tower, each of those users will be given the same position. Using a FINE position will give much more detailed information.

Also, in certain areas, for privacy reasons, a less-exact location will be given, and that less-exact location might be identical for many users. Again, a fine-location configuration will return more precise location data.

We believe with reference to the above comment, your query has been answered, hence closing the bug. Please feel free to re-open the issue in the future if desired.

Thanks, Emre!

Based on #10, it seems AGP's current behavior makes sense.

Regardless of whether we keep or change the default behavior, it makes sense to me to add an API to either (1) enable/disable stripping per variant / component or (2) set the "local strip mode" per variant / component.

I wrote an internal doc about this. PTAL. We'll discuss at one of the upcoming AGP API council meetings.

This was discussed at the AGP API meeting on Feb 1. There was no consensus about the API, but there was consensus that AS should support debugging, whether the android test APK is stripped or not.

With that in mind, I wrote ag/21235820 to package stripped native libraries in test APKs by default, planning to add support for the use cases pointed out in #10.

As it turns out, all of the following cases of debugging android tests are already broken, regardless of whether the test APK native libraries are stripped:

• Issue 268213434 - Error when trying to debug android test of library from AS
• Issue 268083076 - connected test fails if androidTestImplementation dependency on android native library
• Issue 268223043 - Unable to debug androidTest-specific native code

The good news is that we can merge ag/21235820 without causing any new breakages, which is what I propose we do, since the large test APKs are causing problems for some users.

Any updates here? Could we make progress on this? We still have massively large APKs due to not doing any stripping on camke made binaries

A fix on the AGP side is here: ag/21235820.

Emre, now that Issue 268213434 and Issue 268223043 have been fixed, would you be able to update the code linked in #10 to pick up the stripped debug symbols from the androidTest-specific native code?

After the change I did in ag/21325527, we now find the facet that corresponds to the app.androidTest module or the app.main module (as opposed to the app holder module), and propagate that throughout the native debugger in Android Studio.

However, when we get the module dependencies to collect symbols in AndroidNativeAppDebugProcess.java;l=624, we still use the app.main module.

There can be other places where we call facet.getMainModule() which we might need to update.

Before looking into this, it would be nice to create a simple test app that demonstrates the problem.

androidx-main has several projects that hit this, you can just look for cmake usage in build.gradle files.

Before looking into this, it would be nice to create a simple test app that demonstrates the problem.

Would it make sense to go ahead and merge ag/21235820 on the AGP side, and then I expect the repro project from Issue 268223043 would demonstrate the problem after updating to the version of AGP with the fix.

The test app from Issue 268223043 (MyApplication29.zip) is rather trivial, debugging the instrumented test in that app already works (after patching your ag/21235820 + building/using AGP 8.2.0-dev) without any additional modifications on Android Studio native debuggers side.

A more complex test app would have:

• An 'app' module with native code.
• A 'lib' module with native code.
• An app-to-lib androidTestImplementation dependency.

And, test cases:

1. Run instrumented test of 'app' module, verify:
   a. hit native breakpoints in 'app' module
   b. hit native breakpoints in 'lib' module
2. Run instrumented test of 'lib' module, verify:
   a. hit native breakpoints in 'lib' module.

Current Status (studio-main, around Hedgehog canary 3):

• The following tests pass: 1(a), 2(a)
• The following tests fail: 1(b)

I prepared ag/23066587, which also makes 1(b) pass. I need to add some integration tests to that change.

spollom@ Feel free to submit your ag/21235820.

Thanks, Emre!

spollom@ Feel free to submit your ag/21235820.

Just merged it.

I just merged ag/23066587, which should fix the debugger for 1(b) in comment19. Adding integration tests is still WIP, it's done but is flaky due to some unexpected resource leaks.

Is it too risky to cp this to 8.1.0?

ag/21235820 is not too risky, but I assume if we CP it, we'd need to CP ag/23066587 and any other necessary changes in the IDE... not sure if those are too risky.

I'd love to see https://googleplex-android-review.git.corp.google.com/c/platform/tools/base/+/21235820 in 8.1.0

Could someone share in which version of AGP this issue is (planned to be) fixed?

The fix is in 8.2.0 and newer. It's currently in beta, but a stable version will be available soon.