Closed Bug 885324 Opened 11 years ago Closed 10 years ago

Intermittent TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | uncaught exception - TypeError: SpecialPowers.wrap(...).location is null at http://mochi.test:8888/tests/docshell/test/test_bug529119-2.html:64

Categories

(Core :: DOM: Navigation, defect)

Product:
Core
Component:
DOM: Navigation
Platform:
ARM
Gonk (Firefox OS)
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla34
Tracking Flags:
Tracking Status
firefox32 --- wontfix
firefox33 --- fixed
firefox34 --- fixed
firefox-esr24 --- wontfix
firefox-esr31 --- fixed
b2g-v1.4 --- fixed
b2g-v2.0 --- fixed
b2g-v2.1 --- fixed

People

(Reporter: cbook, Assigned: martijn.martijn)

References

(
URL
)

Details

(Keywords: intermittent-failure)

Attachments

(1 file)

885324.diff
1.60 KB, patch
mayhemer
: review+
Details | Diff | Splinter Review
https://tbpl.mozilla.org/php/getParsedLog.php?id=24377812&tree=Mozilla-Inbound b2g_ics_armv7a_gecko_emulator_vm mozilla-inbound opt test mochitest-3 on 2013-06-20 02:42:45 PDT for push 457bf850bf04 slave: tst-linux64-ec2-397 03:23:54 INFO - 35885 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | uncaught exception - TypeError: SpecialPowers.wrap(...).location is null at http://mochi.test:8888/tests/docshell/test/test_bug529119-2.html:64 03:23:54 INFO - 35886 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | [SimpleTest.finish()] this test already called finish! 03:23:54 INFO - 35887 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | called finish() multiple times 03:2 03:23:36 INFO - 35877 INFO TEST-START | /tests/docshell/test/test_bug529119-2.html 03:23:36 INFO - creating 1! 03:23:36 INFO - [TabChild] SHOW (w,h)= (0, 0) 03:23:36 INFO - ###################################### forms.js loaded 03:23:36 INFO - ############################### browserElementPanning.js loaded 03:23:36 INFO - ######################## BrowserElementChildPreload.js loaded 03:23:36 INFO - 35878 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Waiting for error page succeeded 03:23:54 INFO - 35879 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Waiting for error page succeeded 03:23:54 INFO - 35880 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Waiting for original page succeeded 03:23:54 INFO - 35881 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Is on an error page 03:23:54 INFO - 35882 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Is not on the previous page 03:23:54 INFO - 35883 INFO TEST-END | /tests/docshell/test/test_bug529119-2.html | finished in 7909ms 03:23:54 INFO - 35884 INFO TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Waiting for original page succeeded 03:23:54 INFO - 35885 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | uncaught exception - TypeError: SpecialPowers.wrap(...).location is null at http://mochi.test:8888/tests/docshell/test/test_bug529119-2.html:64 03:23:54 INFO - 35886 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | [SimpleTest.finish()] this test already called finish! 03:23:54 INFO - 35887 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/test_bug529119-2.html | called finish() multiple times 03:23:54
Closing inactive keywords:intermittent-failure bugs where the TBPLbot has previously commented and the test isn't marked as disabled; filter on orange-cleanup-201401.
Status: NEW → RESOLVED
Closed: 11 years ago
Resolution: --- → WORKSFORME
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Martijn, I'm somehow guessing that this recent spike is related to the patches of yours that have landed lately.
I assume you mean bug 945781, bug 943152 or bug 941459, right? I don't see how they could be related to this intermittent failure, though. What I see in the log is that this line is seen 2 times: TEST-PASS | /tests/docshell/test/test_bug529119-2.html | Waiting for error page succeeded In a normal test run, this would only happen 1 time. In docshell/test/test_bug529119-2.html, I see this: <body onload="window.opener.windowLoaded();"> When I change that line to this: <body onload="window.opener.windowLoaded();window.opener.windowLoaded();"> I can see a result similar as to the errors that are shown in this bug. I don't know how this would happen with this test, that test_bug529119-2.html would fire the load event 2 times, but I can certainly a potential for this test to go wrong here. For instance, when I set the browser.sessionhistory.max_total_viewers pref to 0 (to turn of bfcache), this test runs perpetually. So I think some guard to make sure that windowLoaded() would run only once, would be good here.
Attached patch 885324.diff — — Splinter Review
I think this might fix the intermittent failure. It also makes this work: ./mach mochitest-plain docshell/test/test_bug529119-2.html --setpref browser.sessionhistory.max_total_viewers=1 Which currently would cause it to run endlessly with that test. I wonder if we should run all mochitests with bfcache turned off. It seems to me that all tests should handle that (except the ones that are explicitly tied to bfcache, perhaps).
Assignee: nobody → martijn.martijn
Attachment #8459559 - Flags: review?(honzab.moz)
Comment on attachment 8459559 [details] [diff] [review] 885324.diff Review of attachment 8459559 [details] [diff] [review]: ----------------------------------------------------------------- How would that happen the onload gets called twice? I think something else here is broken. Is there some parallelism involved?
(In reply to Martijn Wargers [:mwargers] (QA) from comment #43) > It also makes this work: > ./mach mochitest-plain docshell/test/test_bug529119-2.html --setpref > browser.sessionhistory.max_total_viewers=1 > Which currently would cause it to run endlessly with that test. Actually, it should be browser.sessionhistory.max_total_viewers=0 where the test would run endlessly. (In reply to Honza Bambas (:mayhemer) from comment #51) > How would that happen the onload gets called twice? I think something else > here is broken. Is there some parallelism involved? The test is expecting the workingURL page to be loaded in the bfcache. Here it goes back to the workingURL page and expects it not to cause a load event to fire again: http://mxr.mozilla.org/mozilla-central/source/docshell/test/test_bug529119-2.html?force=1#55 If for some reason bfcache would fail, then the load event would fire again. Or am I seeing it wrong?
Flags: needinfo?(bambas)
Please don't use bambas@acepoint.cz anymore, where the hell did you get that email anyway?
(In reply to Martijn Wargers [:mwargers] (QA) from comment #52) > The test is expecting the workingURL page to be loaded in the bfcache. > Here it goes back to the workingURL page and expects it not to cause a load > event to fire again: > http://mxr.mozilla.org/mozilla-central/source/docshell/test/test_bug529119-2. > html?force=1#55 > > If for some reason bfcache would fail, then the load event would fire again. > Or am I seeing it wrong? Aha! Now I got it (completely forgot how the test worked). Then yes, if there is a bfcache race, this is the fix. Going to r+, thanks.
Attachment #8459559 - Flags: review?(honzab.moz) → review+
(In reply to Honza Bambas (:mayhemer) from comment #53) > Please don't use bambas@acepoint.cz anymore, where the hell did you get that > email anyway? Sorry, this turns up in Bugzilla autocomplete. I guess you didn't mark the account with that mail address as inactive/disabled? If you would do that, it won't show up anymore in Bugzilla autocomplete.
Flags: needinfo?(bambas) → needinfo?(martijn.martijn)
Keywords: checkin-needed
Flags: needinfo?(martijn.martijn)
(In reply to Martijn Wargers [:mwargers] (QA) from comment #43) > I wonder if we should run all mochitests with bfcache turned off. It seems > to me that all tests should handle that (except the ones that are explicitly > tied to bfcache, perhaps). I filed bug 1057499 for this.
https://hg.mozilla.org/mozilla-central/rev/8eeec1587d58
Status: REOPENED → RESOLVED
Closed: 11 years ago → 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla34
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: