475369 - test_seek1.html randomly failed on Windows unittest box: "Video currentTime should be around 2: 0"

Status: RESOLVED FIXED

(Core :: Audio/Video, defect, P2)

Description

[Daniel Holbert [:dholbert]]

This mochitest just randomly failed:
WINNT 5.2 mozilla-central unit test on 2009/01/26 07:53:42
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1232985222.1232990388.24904.gz
*** 27703 ERROR TEST-UNEXPECTED-FAIL | /tests/content/media/video/test/test_seek1.html | Video currentTime should be around 2: 0.032999999821186066

I say "randomly" because the previous cycle was built from the same revision, and it passed this mochitest (though it failed 2 crashtests) as shown in this log:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1232982889.1232987205.16910.gz

Comment 1

[Daniel Holbert [:dholbert]]

Note: this sounds a bit like bug 469595, but that bug's been resolved as fixed, so I'm filing this as a separate bug.

Comment 2

[Daniel Holbert [:dholbert]]

This test also failed on the same machine 4 cycles earlier, and also during a cycle last night:
WINNT 5.2 mozilla-central unit test on 2009/01/26 03:57:45
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1232971065.1232975290.11372.gz
WINNT 5.2 mozilla-central unit test on 2009/01/25 11:53:42
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1232913222.1232917361.11397.gz

Both of these have the message "Video currentTime should be around 2: 0"

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

We need to add some instrumentation so we can collect debug output on failing seek tests to try to narrow down the problem.

Comment 4

[cajbir (:cajbir)]

Attachment: Pass seeking information through to channel listener [Checked in: Comment 8 & 10]

Comment 5

[cajbir (:cajbir)]

Attachment: Re-enable tests

Comment 6

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I'm not sure if this will fix the orange, but we should try it, by landing both patches together when the tree is not under pressure (i.e. not until beta3 is done) and then backing out the test-enabler patch if we still get failures.

Comment 7

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Actually we should land the code change ASAP. Reenabling the tests should wait.

Comment 8

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Pushed code change as http://hg.mozilla.org/mozilla-central/rev/6739e046d870. Leaving this bug open to until we push the test re-enable and see if they're fixed.

Comment 9

[Robert O'Callahan (:roc) (email my personal email if necessary)]

The code change should land in 1.9.1 anyway.

Comment 10

[Robert O'Callahan (:roc) (email my personal email if necessary)]

http://hg.mozilla.org/releases/mozilla-1.9.1/rev/9fdd476acaf4

Comment 11

[Chris Pearce [:cpearce (Not reading bugmail)]]

I've been testing the seek tests on XP and Vista in both debug and opt builds, and I haven't seen them fail, despite testing in a large number of runs. I think we should try to enable them again, some recent change may have fixed them.

Comment 12

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Downgrading to wanted anyway, since there's nothing here to block on.

Comment 13

[Marco Bonardo [:mak] (Away Apr 25 - May 5)]

test_seek4.html failed on 1.9.1 in a similar way. should i file a new bug?
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.5/1242810170.1242820731.20679.gz

Comment 14

[cajbir (:cajbir)]

Is this test still failing? Can we close?

Comment 15

[cajbir (:cajbir)]

No reply for two weeks, and there appears to be no repeat of the failure. If it happens again or you think it's not fixed please reopen with details.