1321574 - PROCESS-CRASH | testFlingCorrectness | application crashed [@ mozilla::jni::NativeStub<mozilla::java::NativePanZoomController::SetIsLongpressEnabled_t, nsWindow::NPZCSupport, mozilla::jni::Args<bool> >::Wrap<&nsWindow::NPZCSupport::SetIsLongpressEnabled>]

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: Toolbar, defect, P3)

Description

[Bob Clary [:bc] (inactive)]

Testing Pixels on the unit tests found the *debug only* Robocop crash [@ mozilla::jni::NativeStub<mozilla::java::NativePanZoomController::SetIsLongpressEnabled_t, nsWindow::NPZCSupport, mozilla::jni::Args<bool> >::Wrap<&nsWindow::NPZCSupport::SetIsLongpressEnabled>]

testFlingCorrectness, testPanCorrectness reproduce this crash reliably. testAxisLocking and testStumblerSetting reproduce this about 50% of the time.

Not sure if this is Pixel/Android 7.1 only or if it reproduces on other versions as well.

https://treeherder.allizom.org/#/jobs?repo=mozilla-central&filter-searchStr=autophone&tochange=28eb27fd7d141a8f9f2a6dbfae892ad78ecb40a5&fromchange=05328d3102efd4d5fc0696489734d7771d24459f&selectedJob=5481535

Example crash:
https://autophone-dev.s3.amazonaws.com/v1/task/JLMaH73zTmGVIvbyIDVfBg/runs/0/artifacts/public/build/robocoptest-robocoptests-settings.ini-2-pixel-05-5722fb3f-573c-4a8e-9c1f-3055d2145774.log

Comment 1

[Geoff Brown [:gbrown] (pto Apr 11-18)]

11-29 22:30:08.878 F/MOZ_Assert(20428): Assertion failure: AndroidBridge::IsJavaUiThread(), at /home/worker/workspace/build/src/widget/android/nsWindow.cpp:561

https://dxr.mozilla.org/mozilla-central/rev/cd4cdcc9ad6c45dad8b8d8c0d40e459db2bca8a1/widget/android/nsWindow.cpp#561

Comment 2

[Kartikaya Gupta (email:[e-mail])]

I believe this function is only used from testing infra anyway, we can probably just remove the assert. Although it's surprising that the test is running on a different thread on this device/version than in current automation platforms.

Comment 3

[Randall Barker [:rbarker]]

Attachment: Bug 1321574 - Remove ASSERT checking if nsWindow::NPZCSupport::SetIsLongpressEnabled is called on Java UI thread. The APZCTreeManager ensures the call is dispatched to correct thread

Review commit: https://reviewboard.mozilla.org/r/151848/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/151848/

Comment 4

[Eugen Sawin [:esawin]]

Comment on attachment 8880509 [details]
Bug 1321574 - Remove ASSERT checking if nsWindow::NPZCSupport::SetIsLongpressEnabled is called on Java UI thread. The APZCTreeManager ensures the call is dispatched to correct thread

https://reviewboard.mozilla.org/r/151848/#review156886

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/4e2b30513c88
Remove ASSERT checking if nsWindow::NPZCSupport::SetIsLongpressEnabled is called on Java UI thread. The APZCTreeManager ensures the call is dispatched to correct thread r=esawin

Comment 6

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/4e2b30513c88