593463 - browser_awesomescreen.js is causing browser-chrome tests to fail

Status: RESOLVED WORKSFORME

(Firefox for Android Graveyard :: General, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

It appears that we have a nice series of javascript errors:
http://tinderbox.mozilla.org/showlog.cgi?log=Mobile/1283541262.1283542680.26707.gz&fulltext=1

If you run this locally (on a n900), you will get to the same output that the tinderbox log shows, then the test will stop and hang.  I found that if I manually click on the browser (not the browser-chrome controlling xul window which is in focus) the tests will continue, but very problematic.

Can we back this test out or get a fix in soon?

Comment 1

[Mark Finkle (:mfinkle) (use needinfo?)]

This JS errors will likely appear for many of the tests. They might even be harmless - we sometimes check for properties that aren't present in e10s.

The test in question seems to be running multiple times. It should only run once, but seems to be running 3 times in the log.

Comment 2

[Matt Brubeck (:mbrubeck)]

When I try to run this test on a desktop GTK debug build, it fails with:

###!!! ASSERTION: Content should not be calling this: 'Not Reached', file /home/mbrubeck/src/mozilla/mozilla-central/chrome/src/nsChromeRegistryContent.cpp, line 248

(The failure is in nsChromeRegistryContent::GetSelectedLocale.)

Comment 3

[Mark Finkle (:mfinkle) (use needinfo?)]

Can you track down why content is trying to get the selected locale? We should only be doing that in chrome. Is the wrong nsIChromeRegistry impl being created?

Comment 4

[Josh Matthews [:jdm]]

Content calling GetSelectedLocale is a well-established problem that nobody's bothered to look into.

Comment 5

[Mark Finkle (:mfinkle) (use needinfo?)]

(In reply to comment #4)
> Content calling GetSelectedLocale is a well-established problem that nobody's
> bothered to look into.

bug 562408

I think that failure is a red herring to this bug. Something else is the problem.

Running this test locally, on Linux GTK, it works fine for me. All test cases pass.

Comment 6

[Joel Maher ( :jmaher ) (UTC -8)]

this seems to be working, can we close wfm?

Comment 7

[Aki Sasaki (not active)]

Is there some other reason N900 browser-chrome tests are only running 9 to 11 tests?

Comment 8

[Joel Maher ( :jmaher ) (UTC -8)]

it appears to be running non mobile tests:
http://tinderbox.mozilla.org/showlog.cgi?log=Mobile/1285694612.1285697454.16561.gz

I have patches that need landing to resolve some of this.  It seems that more and more stuff keeps cluttering up the browser-chrome test repository for mobile.

Comment 9

[Vivien Nicolas (:vingtetun) (:21) - (NOT reading bugmails, needinfo? please)]

I guess we can close this one?