824276 - Perma-orange on "Android 4.0 opt" mozilla-central, after 12/22 merge from inbound, with "java.lang.RuntimeException: Your TabHost must have a TabWidget whose id attribute is 'android.R.id.tabs'"

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect)

Description

[Daniel Holbert [:dholbert]]

Today's merge from inbound to central was fully orange/red for one platform, "Android 4.0 opt": https://tbpl.mozilla.org/?rev=ea373e534245
with this error:
{
12-22 06:46:39.242 I/GeckoViewsFactory( 2175): Creating custom Gecko view: RemoteTabs
12-22 06:46:39.242 D/AndroidRuntime( 2175): Shutting down VM
12-22 06:46:39.265 E/GeckoAppShell( 2175): >>> REPORTING UNCAUGHT EXCEPTION FROM THREAD 1 ("main")
12-22 06:46:39.265 E/GeckoAppShell( 2175): java.lang.RuntimeException: Unable to start activity ComponentInfo{org.mozilla.fennec/org.mozilla.fennec.App}: java.lang.RuntimeException: Your TabHost must have a TabWidget whose id attribute is 'android.R.id.tabs'
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1956)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread.handleLaunchActivity(ActivityThread.java:1981)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread.access$600(ActivityThread.java:123)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread$H.handleMessage(ActivityThread.java:1147)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.os.Handler.dispatchMessage(Handler.java:99)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.os.Looper.loop(Looper.java:137)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread.main(ActivityThread.java:4424)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at java.lang.reflect.Method.invokeNative(Native Method)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at java.lang.reflect.Method.invoke(Method.java:511)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at com.android.internal.os.ZygoteInit$MethodAndArgsCaller.run(ZygoteInit.java:784)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at com.android.internal.os.ZygoteInit.main(ZygoteInit.java:551)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at dalvik.system.NativeStart.main(Native Method)
12-22 06:46:39.265 E/GeckoAppShell( 2175): Caused by: java.lang.RuntimeException: Your TabHost must have a TabWidget whose id attribute is 'android.R.id.tabs'
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.widget.TabHost.setup(TabHost.java:124)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at org.mozilla.gecko.TabsPanel.initialize(TabsPanel.java:122)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at org.mozilla.gecko.TabsPanel.onFinishInflate(TabsPanel.java:116)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.view.LayoutInflater.rInflate(LayoutInflater.java:747)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.view.LayoutInflater.rInflate(LayoutInflater.java:742)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.view.LayoutInflater.inflate(LayoutInflater.java:489)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.view.LayoutInflater.inflate(LayoutInflater.java:396)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.view.LayoutInflater.inflate(LayoutInflater.java:352)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at com.android.internal.policy.impl.PhoneWindow.setContentView(PhoneWindow.java:251)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.Activity.setContentView(Activity.java:1835)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at org.mozilla.gecko.GeckoApp.onCreate(GeckoApp.java:1597)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at org.mozilla.gecko.BrowserApp.onCreate(BrowserApp.java:208)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.Activity.performCreate(Activity.java:4465)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.Instrumentation.callActivityOnCreate(Instrumentation.java:1049)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	at android.app.ActivityThread.performLaunchActivity(ActivityThread.java:1920)
12-22 06:46:39.265 E/GeckoAppShell( 2175): 	... 11 more
12-22 06:46:39.265 F/libc    ( 2175): Fatal signal 11 (SIGSEGV) at 0x00000000 (code=1)
}
https://tbpl.mozilla.org/php/getParsedLog.php?id=18203410&full=1&branch=mozilla-central

Facts:
* That platform was green on the previous mozilla-central push, so it's almost certainly something in that merge that regressed it.
* Inbound apparently doesn't run tests on this platform ("Android 4.0 opt"), so we can't track down when this regressed on inbound. :-/

gcp suggests that bug 817749 looks involved.  Not sure whether we need a backout or just a clobber; tried a clobber for now, but this may need a backout later.

Comment 1

[Daniel Holbert [:dholbert]]

This is holding the tree closed, FWIW.

Comment 2

[Jonathan Kew [:jfkthame]]

(In reply to Daniel Holbert [:dholbert] from comment #0)
> gcp suggests that bug 817749 looks involved.  Not sure whether we need a
> backout or just a clobber; tried a clobber for now, but this may need a
> backout later.

If that doesn't help, I'd also be a bit suspicious of bug 817706.

Comment 3

[Daniel Holbert [:dholbert]]

Good point. If the clobber doesn't fix this, I suggest we just back out both of those.

Comment 4

[Daniel Holbert [:dholbert]]

Hooray -- there are some green test-runs on the post-clobber cycle[1], so it looks like the clobber fixed this.

Resolving as FIXED by the clobber.

[1] https://tbpl.mozilla.org/?rev=84320dffec6e

Comment 5

[Daniel Holbert [:dholbert]]

(NOTE: I don't actually know offhand which android build these devices run, so I clobbered all of the m-c builders that had "android" in the name, which is currently the top 9 rows at https://secure.pub.build.mozilla.org/clobberer/?branch=mozilla-central .  Apparently, one of those clobbers did the trick.)

Comment 6

[Ed Morley [:emorley]]

(In reply to Daniel Holbert [:dholbert] from comment #5)
> (NOTE: I don't actually know offhand which android build these devices run

The Android 4.0 Fennec runs are done on the Panda boards (compared to Android 2.2 on Tegras) and have been switched on for m-c only in the last ~week (see bug 725544). I believe the intention is to switch them on for other trees very soon, once some more of the automation is smoothed out.