Apparently a TI regression: I saw a couple of these on Jaegermonkey, and this instance is from the merge to mozilla-central. http://tbpl.allizom.org/php/getParsedLog.php?id=6186291 Rev3 WINNT 5.1 mozilla-central opt test mochitests-1/5 on 2011-08-30 00:46:13 PDT WebGL mochitest: starting page conformance/oes-texture-float.html 43704 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "gl.getExtension("OES_texture_float").myProperty should be 2 (of type number). Was undefined (of type undefined)." (URL: conformance/oes-texture-float.html) WebGL test error: test page failure: conformance/oes-texture-float.html
http://tbpl.allizom.org/php/getParsedLog.php?id=6183059 and http://tbpl.allizom.org/php/getParsedLog.php?id=6183260 for a couple of recent JM ones, suspiciously on the same build, though the third time I retriggered on that build it passed.
Oh, this. This is an actual bug in our code. Somehow it hadn't been causing trouble until now. Test patch coming as I don't have time to fix the actual bug now and it doesn't matter hugely.
...and it does make sense that it would be triggered by a JS engine change. See the discussion in bug 630672
Depends on: 630672
This test is valid, but flaky as it relies on GC happening at a particular time. A recent JS engine change making it actually happen caused it to actually catch a bug we've always had in the OES_texture_float impl, see the discussion on bug 630672, and I don't have time to fix that now.
Attachment #556965 - Flags: review?(jmuizelaar)
Comment on attachment 556965 [details] [diff] [review] disable the uniqueObjectTest Personally, I'd prefer if you included the explanation of the problem inline.
Attachment #556965 - Flags: review?(jmuizelaar) → review+
http://hg.mozilla.org/mozilla-central/rev/7d9ed22b7b49 leaving open, this has just disabled part of the test afaict
Target Milestone: --- → mozilla9
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
This intermittent orange is really fixed by this patch. But we're still ignoring this test on linux as I first noticed it there and did improper analysis. This patch un-ignores this test.
Attachment #557657 - Flags: review?(jmuizelaar)
Attachment #557657 - Flags: review?(jmuizelaar) → review+
You need to log in before you can comment on or make changes to this bug.