Closed Bug 599253 Opened 15 years ago Closed 15 years ago

TEST-UNEXPECTED-FAIL | browser_586068-cascaded_restore.js | Test timed out

Categories

(Firefox :: Session Restore, defect)

Product:
Firefox
Component:
Session Restore
Platform:
x86
Windows Server 2003
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
VERIFIED FIXED

People

(Reporter: zpao, Assigned: dwitte)

References

Details

(Keywords: intermittent-failure)

Attachments

(2 files)

extend timeout
1.24 KB, patch
Dolske
: review+
Dolske
: approval2.0+
Details | Diff | Splinter Review
close windows
1.40 KB, patch
zpao
: review+
Details | Diff | Splinter Review
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1285304453.1285307511.21032.gz#err0 TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Test timed out TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Found a tab after previous test timed out: about:blank TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Found a tab after previous test timed out: about:blank TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Found a tab after previous test timed out: about:blank TEST-UNEXPECTED-FAIL | chrome://mochikit/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Found a browser window after previous test timed out Might just need to extend the timeout. The test does a lot of window opening/closing. The tests do all finish & pass, it just looks like it got stuck finishing. If the test stays orange, I'll disable on Windows.
Whiteboard: [orange]
WINNT 5.2 mozilla-central debug test mochitest-other on 2010/09/24 11:37:35 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1285353455.1285356509.18131.gz
For my reference, this is failing after starting test_setBrowserStateInterrupted http://tinderbox.mozilla.org/showlog.cgi?log=MozillaTry/1285637882.1285640916.14103.gz&fulltext=1 makes it a bit more clear since it has another test and adds an info() to separate the tests.
I'll be extending the timeout in bug 602555. Hope that's all it is. Otherwise I'll look into this more closely.
I extended the timeout and refactored the CSR code in bug 602555, landed http://hg.mozilla.org/mozilla-central/rev/bfa9a991f78e - hopefully one, the other, or both make this go away.
This random orange and bug 603536 occur regularly on 50%+ win debug oth runs on try server. Maybe it's time we disable this test for windows until we find a fix?
Attached patch extend timeout — — Splinter Review
Assignee: nobody → paul
Attachment #484457 - Flags: review?(dolske)
Attachment #484457 - Flags: review?(dolske)
Attachment #484457 - Flags: review+
Attachment #484457 - Flags: approval2.0+
Pushed http://hg.mozilla.org/mozilla-central/rev/f0b57753440c I'll leave this open until I'm convinced that actually fixed it (and so other branches can find this until they merge m-c)
2 observations: 1. the rate of reports has gone down significantly. 2. The reports coming in are all timing out in the same test, right around the same spot. TEST-INFO | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | running test_reload TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 1 - browser loaded correct url TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 1 - correct tab was restored TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 2 - browser loaded correct url TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 2 - correct tab was restored TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 3 - browser loaded correct url TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 3 - correct tab was restored TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 4 - browser loaded correct url TEST-PASS | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | test_reload: load 4 - correct tab was restored TEST-INFO | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Longer timeout required, waiting longer... Remaining timeouts: 3 TEST-INFO | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Longer timeout required, waiting longer... Remaining timeouts: 2 TEST-INFO | chrome://mochitests/content/browser/browser/components/sessionstore/test/browser/browser_586068-cascaded_restore.js | Longer timeout required, waiting longer... Remaining timeouts: 1
Depends on: 611313
Attached patch close windows — — Splinter Review
I've tried testing this locally, but it's pretty hard to get a read on whether it fixes the problem, because (when running in a loop for a long time) I keep randomly failing in other browser-chrome tests. :( So, I say we try this on m-c and cross our fingers. Try's chewing on it now.
Attachment #490596 - Flags: review?(paul)
Comment on attachment 490596 [details] [diff] [review] close windows I know this left us a little bit unsatisfied, but it definitely seemed to help while we were debugging this. For posterity's sake: the issue we were having is that setBrowserState was sometimes closing the wrong window and so the window which mochitest opens and actually runs the test from was closed, even with a waitForFocus(stuff, window). So while this fixes it, there's still the actual root problem, which is that Services.wm.getMostRecentWindow doesn't match up with what is thought to be focused. I would rather we fix that than work around it. That said I'd take this to fix orange so long as you file something about getMostRecentWindow lying sometimes (there's already 528706 but that's not the same and we protect against that in SS)
Attachment #490596 - Flags: review?(paul) → review+
Comment on attachment 490596 [details] [diff] [review] close windows http://hg.mozilla.org/mozilla-central/rev/9013327d4a59 We'll see what happens; I'll file said followup if it sticks.
Filed bug 618470 for the followup.
Assignee: paul → dwitte
Status: NEW → RESOLVED
Closed: 15 years ago
Resolution: --- → FIXED
Status: RESOLVED → VERIFIED
Whiteboard: [orange]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: