528654 - Intermittent timeout waiting for onload to fire in js1_8_1/trace/trace-test.js

Status: RESOLVED FIXED

(Core :: JavaScript Engine, defect)

Description

[Phil Ringnalda (:philor)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258130171.1258138406.31047.gz#err3
Linux mozilla-central debug test everythingelse on 2009/11/13 08:36:11
s: moz2-linux-slave05

TEST-PASS | trace-test.js | testInterpreterReentry
getInterface
e
x
testBug474769_b
{many many lines elided}
atob
btoa
frameElement
showModalDialog
postMessage
localStorage
 FAILED! uncaught exception: [Exception... "Operation is not supported"  code: "9" nsresult: "0x80530009 (NS_ERROR_DOM_NOT_SUPPORTED_ERR)"  location: "file:///builds/moz2_slave/mozilla-central-linux-debug-unittest-everythingelse/build/jsreftest/tests/js1_8_1/trace/trace-test.js Line: 4273"]
JavaScript error: , line 0: uncaught exception: [Exception... "Operation is not supported"  code: "9" nsresult: "0x80530009 (NS_ERROR_DOM_NOT_SUPPORTED_ERR)"  location: "file:///builds/moz2_slave/mozilla-central-linux-debug-unittest-everythingelse/build/jsreftest/tests/js1_8_1/trace/trace-test.js Line: 4273"]
REFTEST TEST-UNEXPECTED-FAIL | file:///builds/moz2_slave/mozilla-central-linux-debug-unittest-everythingelse/build/jsreftest/tests/jsreftest.html?test=js1_8_1/trace/trace-test.js | timed out waiting for onload to fire

Comment 1

[Bob Clary [:bc] (inactive)]

the timeout occurred right after Josh's checkin http://hg.mozilla.org/mozilla-central/rev/88b83c564d3c which he backed out in http://hg.mozilla.org/mozilla-central/rev/0e01045dad63

I don't think this has occurred since, but I'll leave open for tracking for now. js1_8_1/trace/trace-test.js is too horking big though and should be split up.

Comment 2

[Phil Ringnalda (:philor)]

While it did occur then, that was

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258109629.1258117678.13319.gz
Linux mozilla-central debug test everythingelse on 2009/11/13 02:53:49
s: moz2-linux-slave11

and comment 0 is from a build six hours later. Plus, if a Linux build was affected by that changeset that touched nsLocalFileOSX.mm and a chunk of nsAppRunner.cpp inside an |#elif defined(XP_MACOSX)|, we've got way bigger problems than a jsreftest timeout :)

Comment 3

[Phil Ringnalda (:philor)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258300816.1258308475.11657.gz
Linux mozilla-central debug test everythingelse on 2009/11/15 08:00:16
s: moz2-linux-slave05

(Yeah yeah, that's another "run while something busted was in the tree" failure.)

Comment 4

[Dave Townsend [:mossop]]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258383970.1258392001.8887.gz&fulltext=1

Comment 5

[Timothy Nikkel (:tnikkel)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1258413623.1258425944.6884.gz

Comment 6

[Bob Clary [:bc] (inactive)]

I don't want to change js1_8_1/trace/trace-test.js since there are several open bugs on the existing js1_8_1/trace/trace-test.js which may depend on its large size. I think the appropriate solution will be to disable trace-test.js in the jsreftests and to create two additional tests by splitting the original.

Comment 7

[Bob Clary [:bc] (inactive)]

http://hg.mozilla.org/mozilla-central/rev/36022c7ab623
http://hg.mozilla.org/tracemonkey/rev/2ca2a236da89

I wrapped the exceptions in try catch blocks and removed the fails marker. Hopefully that will resolve the issue.

marking fixed for now.