1124996 - dom/canvas/test/webgl-conformance/_wrappers/test_conformance__extensions__webgl-compressed-texture-s3tc.html fails on OS X 10.10

Status: RESOLVED FIXED

(Core :: Graphics: CanvasWebGL, defect)

Description

[Phil Ringnalda (:philor)]

Attachment: log fragment

No idea what it all means, but here's the section of the log with tons of things like

19:55:31     INFO -  137 INFO TEST-UNEXPECTED-FAIL | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__extensions__webgl-compressed-texture-s3tc.html | at (4, 0) expected: 255,0,0,255 was 255,255,0,255 - expected PASS
19:55:31     INFO -  FAIL: at (4, 0) expected: 255,0,0,255 was 255,255,0,255
19:55:31     INFO -  STACK TRACE:
19:55:31     INFO -  compareRect@http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:593:1
19:55:31     INFO -  testDXTTexture@http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:521:13
19:55:31     INFO -  testDXTTextures@http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:257:9
19:55:31     INFO -  testDXT1_RGB@http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:197:5
19:55:31     INFO -  runTestExtension@http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:176:5
19:55:31     INFO -  @http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/extensions/webgl-compressed-texture-s3tc.html:108:9

Comment 1

[Bas Schouten (:bas.schouten)]

There's several of these, I'm going to assume you will have a look at these or redirect these to someone else Jeff.

Comment 2

[Jeff Muizelaar [:jrmuizel]]

Does this fail consistently?

Comment 3

[Phil Ringnalda (:philor)]

Yep, they were all 100% consistent, that's why I got away with fails-if() in bug 1124996.

Comment 4

[Phil Ringnalda (:philor)]

Err, this is that, I mean in bug 1125003.

Comment 5

[u480271]

But I'm using s3tc textures on OSX 10.10. *puzzled*

Comment 6

[u480271]

I just tried the tests locally on MacBook Pro (Retina, 15", Early 2013) and all the tests pass both with the version in our tree and the latest version at khronos.org.

Can we get the specs of the machine that this is failing on?

Comment 7

[Phil Ringnalda (:philor)]

The specs are "one of the slaves that is currently running 10.8 tests, reimaged with 10.10." If you have very specific questions, :kmoir would be a good relenger to ask, but they are the same things we've been using for years (strictly speaking, one is a former 10.8 test slave and two are former build slaves, but since all three of them fail in exactly the same way, that difference doesn't matter).

Comment 8

[u480271]

I that case, I suspect this might be the same issue as 10.9 (Bug 1009642)

Comment 9

[u480271]

https://tbpl.mozilla.org/?tree=Try&rev=60be1f10ad59

Comment 10

[Phil Ringnalda (:philor)]

Because we like to be difficult, you won't get 10.10 (or for that matter 10.8) tests on that. You only get 10.10 with -u mochitest-gl[10.10].

Comment 11

[u480271]

Attachment: Block S3TC on OSX 10.10 & HD3000

Extend blocking of S3TC textures on Yosemite when HD3000 is present.

(This matches the testing infra H/W)

Comment 12

[u480271]

try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=5550e5216d6c

Comment 13

[u480271]

Now with more yosemite special sauce: https://treeherder.mozilla.org/#/jobs?repo=try&revision=7dbe81bf6374

Comment 14

[u480271]

Unexpected pass! Well I've disabled the feature, so of course.

Comment 15

[u480271]

https://hg.mozilla.org/integration/mozilla-inbound/rev/71966b47b3f9

Comment 16

[u480271]

Attachment: Remove expected fail on OSX 10.10

Comment 17

[u480271]

Once "Remove expected fail on OSX 10.10" patch is reviewed, it can be landed to remove the "expected 1 fail" messages (https://treeherder.mozilla.org/logviewer.html#?job_id=7644713&repo=mozilla-inbound)

Comment 18

[u480271]

https://hg.mozilla.org/integration/mozilla-inbound/rev/eb29765267c7

Comment 19

[Carsten Book [:Tomcat]]

sorry had to back this out since this caused still orange results like https://treeherder.mozilla.org/logviewer.html#?job_id=7648481&repo=mozilla-inbound

Comment 20

[Kim Moir [:kmoir] ET]

Not sure if you still want me to review the patch since you landed it and then it was backed out, let me know

Comment 21

[u480271]

Ok. I'll talk it over with :jgilbert.

Comment 22

[u480271]

Attachment: Remove expected fail on OSX 10.10

Updated errata and regenerated wrappers.

Comment 23

[u480271]

https://tbpl.mozilla.org/?tree=Try&rev=ed2750eaef23

Comment 24

[u480271]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9fed7e2884fe
https://hg.mozilla.org/integration/mozilla-inbound/rev/6aaf86f2c5d8

Comment 25

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/9fed7e2884fe
https://hg.mozilla.org/mozilla-central/rev/6aaf86f2c5d8