You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Android Hybrid - Search - App crashes when switching between expenses of the same report. #51584
👋 Friendly reminder that deploy blockers are time-sensitive ⏱ issues! Check out the open `StagingDeployCash` deploy checklist to see the list of PRs included in this release, then work quickly to do one of the following:
Identify the pull request that introduced this issue and revert it.
@lanitochka17 for debugging this issue, are you able to reproduce the bug on your device, connect it to a computer and run adb bugreport from a terminal and share the zip file here with us?
(The crash from the logs is an ANR, and in that log file we have no info to what happened. This bugreport zip file will help us to understand what caused the crash)
Looking at the ANRs from your device we can see that the main/UI and JS thread are in a deadlock, presumably due to some sync calls caused by reanimated:
JS Thread
Main Thread
Another thing i can see from your report is that your device was pretty low on memory when the ANR happens, only 200mb left. The code i added might introduced issues with high memory, but then, that code never gets executed in the bug's flow (it only gets executed when opening the search page):
Okay, i can reproduce this bug when building a release app from main (where the accused PR is reverted on), so i don't really think its caused by the PR. Let me test the adhoc build from the revert PR though
(happy to help investigate whats causing the bug though!)
@hannojg Hmm, shoot. I tried reproducing the bug for nearly 10 minutes on the AdHoc build and couldn't get it, but I was able to reproduce it within a minute on staging. I'm so sorry if I made a mistake reverting your PR.
Feel free to revert the revert again 😅 and I'd be happy to help you merge it. If this bug is really hard to reproduce, it makes sense that we could only find it now.
Do you have any recommendations on which PRs specifically would be causing this regression?
Ah hmm, that being said, @NikkiWines mentioned that the revert solved a deploy blocker that she was assigned to #51527, so we may still want to investigate that. It also doesn't really seem like your code is invoked there, so it may be unrelated and just flaky like this one.
Yes, i will make a new PR that addresses the other blockers soon!
For the crash: i can reproduce the bug very easily 100% of the time, so it should be easy for me to investigate. I can test a few likely PRs by reverting them to identify the faulty PR tmrw!
(I think you guys have a set of git commands that you use to revert a PR right, if so, could you share that one please?)
Note: i am still able to reproduce this bug on version 9.0.53-0, which is over one week old. Going further back …
// Edit: Also reproducible on 9.0.50-0, two weeks old
// Edit: also reproducible on 9.0.45-0, one month old
at this point i am not sure if it makes sense to go further back (i have the feeling i would find that its since we turned on the new arch). Maybe instead we should try to just fix the bug?
Let us know if you want us to investigate whats going on (feel free to assign to me, either i will work on it or someone else from our margelo team can take a look!)
Activity
melvin-bot commentedon Oct 28, 2024
Triggered auto assignment to @jasperhuangg (
DeployBlockerCash), see https://stackoverflowteams.com/c/expensify/questions/9980/ for more details.melvin-bot commentedon Oct 28, 2024
💬 A slack conversation has been started in #expensify-open-source
github-actions commentedon Oct 28, 2024
👋 Friendly reminder that deploy blockers are time-sensitive ⏱ issues! Check out the open `StagingDeployCash` deploy checklist to see the list of PRs included in this release, then work quickly to do one of the following:
lanitochka17 commentedon Oct 28, 2024
Production:
Screenrecorder-2024-10-28-16-48-40-587.mp4
lanitochka17 commentedon Oct 28, 2024
We think that this bug might be related to #wave-collect - Release 1
jasperhuangg commentedon Oct 28, 2024
At the dentist, will tackle this when I'm back
jasperhuangg commentedon Oct 28, 2024
Building a revert that might be related
hannojg commentedon Oct 28, 2024
@lanitochka17 for debugging this issue, are you able to reproduce the bug on your device, connect it to a computer and run
adb bugreportfrom a terminal and share the zip file here with us?(The crash from the logs is an ANR, and in that log file we have no info to what happened. This bugreport zip file will help us to understand what caused the crash)
jasperhuangg commentedon Oct 28, 2024
@hannojg I'm testing right now, I'll try that out
hannojg commentedon Oct 28, 2024
Thanks! (Btw, the pr hasn't touched any UI related stuff and nothing in regards to reports, so i am not sure how likely it is related - but lets see!)
14 remaining items
hannojg commentedon Oct 29, 2024
This is so weird. The code I have added is only used in
SearchRouter. The code i added isn't even invoked in the flow that this bug describes.hannojg commentedon Oct 29, 2024
Looking at the ANRs from your device we can see that the main/UI and JS thread are in a deadlock, presumably due to some sync calls caused by reanimated:
Another thing i can see from your report is that your device was pretty low on memory when the ANR happens, only 200mb left. The code i added might introduced issues with high memory, but then, that code never gets executed in the bug's flow (it only gets executed when opening the search page):
hannojg commentedon Oct 29, 2024
Aha, interesting, i can reproduce on release, but not on debug
hannojg commentedon Oct 29, 2024
Okay, i can reproduce this bug when building a release app from
main(where the accused PR is reverted on), so i don't really think its caused by the PR. Let me test the adhoc build from the revert PR though(happy to help investigate whats causing the bug though!)
hannojg commentedon Oct 29, 2024
I can reproduce this with the adhoc build as well:
hannojg commentedon Oct 29, 2024
I think the issue is caused by our native stack PR/bridgeless PR/or some general bug in reanimated that we've only discovered now for some reason
jasperhuangg commentedon Oct 29, 2024
@hannojg Hmm, shoot. I tried reproducing the bug for nearly 10 minutes on the AdHoc build and couldn't get it, but I was able to reproduce it within a minute on staging. I'm so sorry if I made a mistake reverting your PR.
Feel free to revert the revert again 😅 and I'd be happy to help you merge it. If this bug is really hard to reproduce, it makes sense that we could only find it now.
Do you have any recommendations on which PRs specifically would be causing this regression?
jasperhuangg commentedon Oct 29, 2024
Ah hmm, that being said, @NikkiWines mentioned that the revert solved a deploy blocker that she was assigned to #51527, so we may still want to investigate that. It also doesn't really seem like your code is invoked there, so it may be unrelated and just flaky like this one.
hannojg commentedon Oct 29, 2024
Yes, i will make a new PR that addresses the other blockers soon!
For the crash: i can reproduce the bug very easily 100% of the time, so it should be easy for me to investigate. I can test a few likely PRs by reverting them to identify the faulty PR tmrw!
(I think you guys have a set of git commands that you use to revert a PR right, if so, could you share that one please?)
hannojg commentedon Oct 30, 2024
Note: i am still able to reproduce this bug on version 9.0.53-0, which is over one week old. Going further back …
// Edit: Also reproducible on 9.0.50-0, two weeks old
// Edit: also reproducible on 9.0.45-0, one month old
at this point i am not sure if it makes sense to go further back (i have the feeling i would find that its since we turned on the new arch). Maybe instead we should try to just fix the bug?
melvin-bot commentedon Oct 30, 2024
If you are the assigned CME please investigate whether the linked PR caused a regression and leave a comment with the results.
If a regression has occurred and you are the assigned CM follow the instructions here.
If this regression could have been avoided please consider also proposing a recommendation to the PR checklist so that we can avoid it in the future.
Beamanator commentedon Oct 31, 2024
I'm marking this NAB since @hannojg confirmed this can be reproduced on older builds, cool? 👍
Thanks so much for your investigation, both of y'all! 🙏
hannojg commentedon Nov 4, 2024
Let us know if you want us to investigate whats going on (feel free to assign to me, either i will work on it or someone else from our margelo team can take a look!)