909970 - Crash in testSystemPages (Galaxy Nexus, x86?)

Status: RESOLVED WORKSFORME

(Firefox for Android Graveyard :: General, defect)

Description

[Geoff Brown [:gbrown]]

:margaret noted an unexpected result in robocop testSystemPages at https://bugzilla.mozilla.org/show_bug.cgi?id=907720#c336.

I have noted that random robocop failures (crashes, anr's) in the coming-soon x86 emulator environment seem to happen in testSystemPages, or just after.

When I run testSystemPages locally on a Galaxy Nexus in landscape mode, the test seems to loop almost-continuously: after opening a 5th tab, the test seems to restart. I caught a bit of this in logcat:

08-27 16:05:37.359 I/Robocop ( 1967): 29 INFO TEST-PASS | testSystemPages | Browser toolbar URL stayed the same - about:feedback should equal about:feedback^M
08-27 16:05:37.460 I/Robocop ( 1967): 30 INFO TEST-PASS | testSystemPages | The correct number of tabs are opened - 5 should equal 5^M
08-27 16:05:37.460 D/Robocop ( 1967): waiting for Tab:Added^M
08-27 16:05:37.460 D/Robocop ( 1967): waiting for DOMContentLoaded^M
08-27 16:05:37.617 I/WindowManager(  182): createSurface Window{41a96218 PopupWindow:41a37820 paused=false}: DRAW NOW PENDING^M
08-27 16:05:37.835 D/dalvikvm( 1967): GC_EXPLICIT freed 733K, 23% free 17969K/23175K, paused 2ms+5ms^M
08-27 16:05:37.843 E/StrictMode( 1967): class org.mozilla.gecko.GeckoPreferences; instances=2; limit=1^M
08-27 16:05:37.843 E/StrictMode( 1967): android.os.StrictMode$InstanceCountViolation: class org.mozilla.gecko.GeckoPreferences; instances=2; limit=1^M
08-27 16:05:37.843 E/StrictMode( 1967):         at android.os.StrictMode.setClassInstanceLimit(StrictMode.java:1)^M
08-27 16:06:02.812 D/AndroidRuntime( 2144): ^M
08-27 16:06:02.812 D/AndroidRuntime( 2144): >>>>>> AndroidRuntime START com.android.internal.os.RuntimeInit <<<<<<^M
08-27 16:06:02.812 D/AndroidRuntime( 2144): CheckJNI is ON^M
08-27 16:06:02.976 D/AndroidRuntime( 2144): Calling main entry com.android.commands.am.Am^M
08-27 16:06:02.984 I/ActivityManager(  182): Force stopping package org.mozilla.fennec_mozdev uid=10038^M
08-27 16:06:02.984 W/ActivityManager(  182): Force removing ActivityRecord{41a9da70 org.mozilla.fennec_mozdev/.App}: app died, no saved state^M
08-27 16:06:02.992 I/Process (  182): Sending signal. PID: 1967 SIG: 9^M
08-27 16:06:02.992 W/ActivityManager(  182): Crash of app org.mozilla.fennec_mozdev running instrumentation ComponentInfo{org.mozilla.roboexample.test/org.mozilla.fennec_mozdev.FennecInstrumentationTestRunner}^M
08-27 16:06:02.992 I/ActivityManager(  182): Force stopping package org.mozilla.fennec_mozdev uid=10038^M
08-27 16:06:03.000 I/WindowManager(  182): createSurface Window{41a3e8c0 com.android.launcher/com.android.launcher2.Launcher paused=false}: DRAW NOW PENDING^M
08-27 16:06:03.007 D/dalvikvm( 2154): Not late-enabling CheckJNI (already on)^M
08-27 16:06:03.007 I/ActivityManager(  182): Start proc org.mozilla.fennec_mozdev for added application org.mozilla.fennec_mozdev: pid=2154 uid=10038 gids={3003, 1015, 1006}^M
08-27 16:06:03.062 I/WindowManager(  182): WIN DEATH: Window{41bc8658 org.mozilla.fennec_mozdev/org.mozilla.fennec_mozdev.App paused=false}^

I do not see evidence of failures on tbpl, but it feels like there's something not-quite-right about this test.

Comment 1

[Geoff Brown [:gbrown]]

I can no longer reproduce this.