Fixed
Status Update
Comments
il...@google.com <il...@google.com> #2
Can you please provide some more information? I assume this stacktrace is from an exception, but I don't know what type of exception it is or which version of Android is involved.
Can you also give details on how you got this to happen and how often it happens? Do you have a sample project that causes this to occur?
Can you also give details on how you got this to happen and how often it happens? Do you have a sample project that causes this to occur?
t....@gmail.com <t....@gmail.com> #3
Sorry for not making it too clear.
The Exception it's throwing is an IllegalArgumentException in android.net.ConnectivityManager in unregisterNetworkCallback (Invalid NetworkCallback)
This is happening in our production app (after we introduced workmanager) and it's been happening for some time now.
It's happening only in Nvidia Shield Tablet and Nvidia Shield Tablet K1. We have around 200 crashes per day.
The Exception it's throwing is an IllegalArgumentException in android.net.ConnectivityManager in unregisterNetworkCallback (Invalid NetworkCallback)
This is happening in our production app (after we introduced workmanager) and it's been happening for some time now.
It's happening only in Nvidia Shield Tablet and Nvidia Shield Tablet K1. We have around 200 crashes per day.
il...@google.com <il...@google.com> #4
Which Android API version is this?
t....@gmail.com <t....@gmail.com> #5
It's happening 100% on API 24
il...@google.com <il...@google.com> #6
Do you mind attaching a logcat and a sample of your code?
t....@gmail.com <t....@gmail.com> #7
I'm unable to reproduce this crash in my dev device. It's happening in out prod build. I've attached the crash log here - Not really sure how helpful that can be
https://gist.github.com/supreethprakash/01bc952ab80eba2d9f2478070cba627f
Our code is not in play here, it could be happening when some scheduled job is about to run
Our code is not in play here, it could be happening when some scheduled job is about to run
si...@gmail.com <si...@gmail.com> #8
Project: platform/frameworks/support
Branch: androidx-master-dev
commit 56776a6d2eddef14d8275828f665df1177d52ac6
Author: Sumir Kataria <sumir@google.com>
Date: Thu Nov 15 14:53:31 2018
Speculative fix for IAE on NVIDIA Shield tablets.
We can't reproduce this on other devices, but Shield Tablets on
API 24 can return an IllegalArgumentException for some reason.
Since this isn't something that is fatal to WorkManager, we
will catch the exception and move on safely.
Bug: 119484416
Test: Ran existing tests.
Change-Id: If42c1c9257ebf7d36ddab11a3f24401ead731636
M work/workmanager/src/main/java/androidx/work/impl/constraints/trackers/NetworkStateTracker.java
https://android-review.googlesource.com/828360
https://goto.google.com/android-sha1/56776a6d2eddef14d8275828f665df1177d52ac6
Branch: androidx-master-dev
commit 56776a6d2eddef14d8275828f665df1177d52ac6
Author: Sumir Kataria <sumir@google.com>
Date: Thu Nov 15 14:53:31 2018
Speculative fix for IAE on NVIDIA Shield tablets.
We can't reproduce this on other devices, but Shield Tablets on
API 24 can return an IllegalArgumentException for some reason.
Since this isn't something that is fatal to WorkManager, we
will catch the exception and move on safely.
Bug: 119484416
Test: Ran existing tests.
Change-Id: If42c1c9257ebf7d36ddab11a3f24401ead731636
M work/workmanager/src/main/java/androidx/work/impl/constraints/trackers/NetworkStateTracker.java
il...@google.com <il...@google.com> #9
This has been fixed internally and will be avilable in Fragment 1.3.0-alpha08.
Note: this fix relies on using the
t....@gmail.com <t....@gmail.com> #10
Excellent. Thanks so much!
sm...@gmail.com <sm...@gmail.com> #11
I'm experiencing a similiar issue with version 1.3.1. I wrote some instrumented tests using androidx.fragment:fragment-testing:1.2.5 which are supposed to simulate Fragment being minimized (home button) and brought back again. I perform that by calling FragmentScenario.moveToState() to go RESUMED -> CREATED -> RESUMED. With 1.2.5 everything was fine, but after upgrade to 1.3.1 these tests started failing due to onCreateView() being called twice. I've described my problem with greater detail on SO https://stackoverflow.com/questions/66629817/testing-androidx-fragment-lifecycle-by-stopping-and-resuming-with-fragmentscenar .
This issue is already fixed in 1.3.1, so this probably is a problem in my configuration. Or maybe it's a regression ? It'd be helpful if someone experienced in this topic would take a look. I could provide sample project reproducing the issue on github if needed.
This issue is already fixed in 1.3.1, so this probably is a problem in my configuration. Or maybe it's a regression ? It'd be helpful if someone experienced in this topic would take a look. I could provide sample project reproducing the issue on github if needed.
il...@google.com <il...@google.com> #12
Re #11 - going from RESUMED
-> CREATED
will destroy your view (the CREATED
state is the state immediately after onCreate()
- i.e., before onCreateView()
), moving back to up RESUMED
will recreate the View. It sounds like Fragment 1.3.1 is working as intended.
Description
Version used:
androidx.navigation:navigation-fragment-ktx:2.2.0-rc01
androidx.navigation:navigation-ui-ktx:2.2.0-rc01
Devices/Android versions reproduced on:
Pixel 3
Android 10
When switching tabs (using BottomNavigation / setupWithNavController()), when the destination Fragment contains a call to postponeEnterTransition() in onCreate() or onCreateView(), the onCreateView() lifecycle method gets called twice. This behavior is visible to the user, via a sort of 'flicker' (depending on the layout), as the fragment's view is recreated.
I suspect this may be due to some code in
`FragmentManager.addAddedFragments(@NonNull ArraySet<Fragment> added)`
It appears the destination Fragment (which invoked postponeEnterTransition()) has moveToState() called on it twice (vie addAddedFragments()). My guess, is that the destination fragment hasn't completed its first moveToState() before the second moveToState() is called, so the secondMoveToState() is attempting to move from a stale state.
I'm not able to provide a sample project for you at this stage.