Closed Bug 464946 Opened 15 years ago Closed 13 years ago

mochitest-chrome: intermittent "test_bug428288.html | Test timed out"

Categories

(Core :: DOM: Navigation, defect)

Product:
Core
Component:
DOM: Navigation
Platform:
x86
All
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla2.0b7

People

(Reporter: myk, Assigned: philor)

References

(
URL
)

Details

(Keywords: intermittent-failure) 

moz2-win32-slave08 turned orange this morning on:

TEST-UNEXPECTED-FAIL | chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html | Test timed out.
The tinderbox turned green the next cycle.
Uh...  I don't see how that test can time out unless the browser crashes or the test harness is completely broken.  It doesn't even waitForExplicitFinish(): just runs some code and that's it.
Unless the load of the document in general timed out (say because disk load was through the roof or something).
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1241562873.1241569105.19472.gz
OS X 10.4 comm-central unit test on 2009/05/05 15:34:33
Blocks: 438871
OS: Windows XP → All
Summary: 358 ERROR TEST-UNEXPECTED-FAIL because test_bug428288.html timed out → mochitest-chrome: intermittent "test_bug428288.html | Test timed out"
Whiteboard: [orange]
(In reply to comment #4)
> http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1241562873.1241569105.19472.gz
> OS X 10.4 comm-central unit test on 2009/05/05 15:34:33

Full test log is
{
...
*** 424 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html...
*** 425 INFO TEST-PASS | chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html | Crashes if bug 428288 is present
NEXT ERROR *** 426 ERROR TEST-UNEXPECTED-FAIL | chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html | Test timed out.
*** 428 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug449778.xul...
...
}
(In reply to comment #5)
> *** 428 INFO Running
> chrome://mochikit/content/chrome/docshell/test/chrome/test_bug449778.xul...
> ...

Fwiw, test_bug449778.xul did not report anything on this run:
{
*** 428 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug449778.xul...
*** 430 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug449780.xul...
*** 431 INFO TEST-PASS | chrome://mochikit/content/chrome/docshell/test/chrome/test_bug449780.xul | Expected hide then show on first loaded page
}
whereas it usually does work as expected.
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1241668403.1241672909.23912.gz
Mac OS X mozilla-central unit test on 2009/05/06
This just happened on Mac:

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1242261742.1242264722.17345.gz
OS X 10.5.2 mozilla-central unit test on 2009/05/13 17:42:22
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258081555.1258086853.21997.gz#err0
WINNT 5.2 mozilla-central debug test everythingelse on 2009/11/12 19:05:55
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1260603687.1260609452.28990.gz#err0
WINNT 5.2 mozilla-central debug test everythingelse on 2009/12/11 23:41:27  
s: win32-slave28
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1260739320.1260746416.20171.gz
WINNT 5.2 mozilla-central debug test everythingelse on 2009/12/13 13:22:00
s: win32-slave35
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1270593722.1270594118.29175.gz
OS X 10.5.2 mozilla-central opt test mochitest-other on 2010/04/06 15:42:02
s: bm-xserve18

I think this is the same bug -- the process was killed (by the harness due to a timeout?) during this same test:
{
2035 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html...
process killed by signal 30
program finished with exit code -1
}
Likewise, I think:

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1270593722.1270594118.29175.gz
OS X 10.5.2 mozilla-central opt test mochitest-other on 2010/04/06 15:42:02
s: bm-xserve18

2035 INFO Running chrome://mochikit/content/chrome/docshell/test/chrome/test_bug428288.html...
process killed by signal 30
program finished with exit code -1
elapsedTime=42.251634
TinderboxPrint: mochitest-chrome<br/><em class="testfail">T-FAIL</em>
Unknown Error: command finished with exit code: -1
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1274807690.1274808986.10409.gz
Linux mozilla-central opt test mochitest-other on 2010/05/25 10:14:50
s: moz2-linux-slave08

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1274807455.1274809092.10647.gz
Rev3 Fedora 12x64 mozilla-central debug test mochitest-other on 2010/05/25 10:10:55
s: talos-r3-fed64-014
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1274808166.1274809876.13788.gz
WINNT 5.2 mozilla-central debug test mochitest-other on 2010/05/25 10:22:46
s: win32-slave27
We apparently went 10 weeks without seeing this between comment 32 and comment 33, and now it's back with a vengeance. If it's that much more reproducible now will there be enough evidence to fix it?
<philor>: bz: does http://mxr.mozilla.org/mozilla-central/source/docshell/test/chrome/test_bug428288.html?force=1 actually need to be loading mozilla.com? I don't understand why it's a test for the bug, so I don't know whether we could maybe fix bug 464946 by having it load something more local
<bz>: it doesn't matter what it loads
<bz>: pretty sure about:blank would work there

Tryserver seems to agree that about:blank works (though the frequency isn't quite high enough to hit it on every push), and agrees that about:blank + the patch from bug 428288 reverted still properly crashes, so

http://hg.mozilla.org/mozilla-central/rev/192c38466579
I think that's enough runs on mozilla-central and on tracemonkey to say that it's still happening quite a bit without my patch, and not happening with it. Worst comes to worst, we can always reopen.
Assignee: nobody → philringnalda
Status: NEW → RESOLVED
Closed: 13 years ago
Flags: in-testsuite+
Resolution: --- → FIXED
Target Milestone: --- → mozilla2.0b7
Whiteboard: [orange]
You need to log in before you can comment on or make changes to this bug.