Status Update
Comments
uc...@google.com <uc...@google.com> #2
Please share below details
1. logcat logs
2. Run emulator from command line with $ANDROID_SDK_ROOT/emulator/emulator -avd <NameOfAVD> -verbose -show-kernel and send the resulting logs
3. Emulator version
4. adb shell top
1. logcat logs
2. Run emulator from command line with $ANDROID_SDK_ROOT/emulator/emulator -avd <NameOfAVD> -verbose -show-kernel and send the resulting logs
3. Emulator version
4. adb shell top
do...@gmail.com <do...@gmail.com> #3
The first idea.log (22KB) is from a cold start of Android Studio, using the AVD manager to start the emulator, and repeating the bug as described above. The Play Store version of Homer was installed on the device already. There was no logcat because I didn't install from Android Studio and thus the USB connection did not get established. Installing would install MY version instead of the Play Store version on that emulation if I allowed it to remove the Play Store version. (If there's a way to get the USB connection started without installing an app to debug, I don't know what it is.)
The second idea.log (170KB) is on the non-Google version of the emulator (from cold start), and includes setting the display timeout to 15 seconds (from the default "none'). The bug repeats nicely. (The version of Homer is the same source as the Play Store version, recompiled so I could run it with debugging and establish the USB connection.) The Logcat is logcat.1.log.
Terminal.1.log is the log from starting the emulator from the command line (as shown). Bug is present.
Emulator version (pasted from Emulator Help screen) is 29.0.11-5598178.
top being a self-refreshing command, I'm not sure exactly what you wanted, so here's a chunk of the output as top.1.log
If you need more stuff I can uniquely provide, I'm happy to do so. However, please be very specific about what you want and how to get it.
The second idea.log (170KB) is on the non-Google version of the emulator (from cold start), and includes setting the display timeout to 15 seconds (from the default "none'). The bug repeats nicely. (The version of Homer is the same source as the Play Store version, recompiled so I could run it with debugging and establish the USB connection.) The Logcat is logcat.1.log.
Terminal.1.log is the log from starting the emulator from the command line (as shown). Bug is present.
Emulator version (pasted from Emulator Help screen) is 29.0.11-5598178.
top being a self-refreshing command, I'm not sure exactly what you wanted, so here's a chunk of the output as top.1.log
If you need more stuff I can uniquely provide, I'm happy to do so. However, please be very specific about what you want and how to get it.
lf...@google.com <lf...@google.com> #4
Hi, thanks for the report and the awesomely detailed repro instructions! We'll look into it.
Description
AI-183.6156.11.34.5522156, JRE 1.8.0_152-release-1343-b01x64 JetBrains s.r.o, OS Windows 10(amd64) v10.0 , screens 1920x1080
Android Gradle Plugin: 3.4.1
Gradle: 5.4
NDK: from local.properties: (not specified); latest from SDK: (not found);
LLDB: pinned revision 3.1 not found; latest from SDK: (package not found);
CMake: from local.properties: (not specified); latest from SDK: (not found); from PATH: (not found);
IMPORTANT: Please read
The symptom is that the image on the screen stops changing, but from the point of view of the application itself everything is fine: clicks work as expected (as if the screen changed), logging indicates perfectly normal behavior, but the display doesn't reflect the state of the app. This only occurs after the (emulated) display times out, twice.
The bug shows up both in my version of the app (currently actively being developed) and in an older version that's currently available in the Play Store. The owner of the Play Store version can repro the bug in an emulator (which I doubt is on the same HW I'm using), but has not been able to repro it on Android HW when he tried. I've asked others who have Pie installed on HW, and they're unable to repro it either. The owner of the older version also has no reports of the bug "in the wild".
I put extensive debugging printout in my version (for test) and from that point of view everything is working exactly as it should - it reports changes in state that you'd expect from your actions, even though the display does not change.
Both my version and the original are on GitHub. I can send you a patch file of the debugging printouts (that would apply to mine) if that would help (I doubt it would, but it's there). I have also asked about it on Stack Overflow with no suggestions even on how to debug it. See
Here's a repro procedure:
Create a Nexus 5X API 28 emulator with Google (so Google store is there). It also is known to repeat on Nexus 4 API 28 without Google if you install it some other way. Both are x86 emulations. This is on Windows OS build 17134.829. It's on Intel i3-8100 (3.60GHz). (Note that your tools that created the system version info above claim it's AMD: that's wrong (your bug - I'll file a followup bug about that).) Since this could be a display card issue, the display card is NVIDIA GeForce GTX 1060 6GB with current drivers.
Take the minimum installed system on the emulator. No additional stuff or settings changes from the default except as noted below.
Download
Run Homer. Download the sample Audio files that it prompts you to download (takes a few seconds, they're short).
Play an audio file and start and stop playback so you know what to expect. Remember where the "stop" and "start" buttons (on different screens) appear on the device (they'll become sort-of invisible when the bug occurs). (The start and stop screens are actually different Fragments, FWIW). You should be able to hear the audio playing if you have headset or speakers.
Set Android system display timeout to 15 seconds (or wait around until it times out, either way, but that's quicker).
Let the emulated device time out while playing and wake it up with the emulator power button (the audio should continue with the screen dark). Press stop. If the bug occurs it will stop playing but the screen will not change. It may take two (but so far never more than two) timeouts in a row to make this happen.
You'll be able to press where "start" should be and it will start the audio, and stop it by pressing stop. Using the back and "overview" buttons will bring it back to a correct display (until it times out again.) If you do that in stopped (supposedly showing book title) state, the restored image will be that for the stopped state that it should be in. If you let it time out (twice?) in the stopped (book title displayed) state, you'll be able to start the audio, but the display will not change.
On my version the start and stop buttons can be configured to blink the text. The blinking freezes when it hits the bug, although debug printouts confirm that the blinking has been started.
Yeah, I know... I hated bugs like this back in the day. Hopefully you have some tools to be able to figure out what's going on, or at least access to a selection of Pie HW to confirm (or not) that it's really an emulator issue.
Let me know if I can help.