1216549 - osx 10.10.5 error in /test_conformance__canvas__viewport-unchanged-upon-resize.html

Status: RESOLVED FIXED

(Core :: Graphics: CanvasWebGL, defect)

Description

[Joel Maher ( :jmaher ) (UTC -8)]

we are trying to get osx 10.10.5 machines up and running, this is coming along well, but we have a few failures, one of them is in the opt/debug runs for the 'gl' job:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=f57ad0d4fd2a

In here we have a failure as seen in this debug log:
09:46:25     INFO -  37 INFO TEST-START | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html
09:46:25     INFO -  ++DOMWINDOW == 129 (0x12c5db800) [pid = 1763] [serial = 129] [outer = 0x120a86c00]
09:46:25     INFO -  ++DOCSHELL 0x12c5d0200 == 33 [pid = 1763] [id = 45]
09:46:25     INFO -  ++DOMWINDOW == 130 (0x12fe72c00) [pid = 1763] [serial = 130] [outer = 0x0]
09:46:25     INFO -  ++DOMWINDOW == 131 (0x12eff8800) [pid = 1763] [serial = 131] [outer = 0x12fe72c00]
09:46:25     INFO -  JavaScript warning: chrome://specialpowers/content/specialpowersAPI.js, line 162: Error: WebGL: Exceeded 32 live WebGL contexts, losing the least recently used one.
09:46:25     INFO -  WebGL(0x12cdace00)::ForceLoseContext
09:46:25     INFO -  ++DOCSHELL 0x12cdaf100 == 34 [pid = 1763] [id = 46]
09:46:25     INFO -  ++DOMWINDOW == 132 (0x12b655c00) [pid = 1763] [serial = 132] [outer = 0x0]
09:46:25     INFO -  ++DOMWINDOW == 133 (0x12d40d400) [pid = 1763] [serial = 133] [outer = 0x12b655c00]
09:46:26     INFO -  [1763] WARNING: NS_ENSURE_TRUE(ParseTypeAttribute(type, &version)) failed: file /builds/slave/try-m64-d-00000000000000000000/build/src/dom/base/nsScriptLoader.cpp, line 495
09:46:26     INFO -  [1763] WARNING: NS_ENSURE_TRUE(ParseTypeAttribute(type, &version)) failed: file /builds/slave/try-m64-d-00000000000000000000/build/src/dom/base/nsScriptLoader.cpp, line 495
09:46:26     INFO -  JavaScript warning: http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/resources/webgl-test.js, line 49: Error: WebGL: Exceeded 16 live WebGL contexts for this principal, losing the least recently used one.
09:46:26     INFO -  WebGL(0x12cdaec00)::ForceLoseContext
09:46:26     INFO -  TEST-INFO | started process screencapture
09:46:26     INFO -  TEST-INFO | screencapture: exit 0
09:46:26     INFO -  38 INFO TEST-PASS | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | A valid string reason is expected
09:46:26     INFO -  39 INFO TEST-PASS | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | Reason cannot be empty
09:46:26     INFO -  40 INFO TEST-FAIL | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | The author of the test has indicated that flaky timeouts are expected.  Reason: untriaged
09:46:26     INFO -  41 INFO TEST-UNEXPECTED-FAIL | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | Pixel at (2, 2) should have been (255, 0, 0, 255), was (0, 0, 255, 255)
09:46:26     INFO -      window.webglTestHarness.reportResults@dom/canvas/test/webgl-conformance/mochi-single.html?conformance/canvas/viewport-unchanged-upon-resize.html:215:7
09:46:26     INFO -      reportTestResultsToHarness@dom/canvas/test/webgl-conformance/resources/js-test-pre.js:45:5
09:46:26     INFO -      testFailed@dom/canvas/test/webgl-conformance/resources/js-test-pre.js:87:5
09:46:26     INFO -      @dom/canvas/test/webgl-conformance/conformance/canvas/viewport-unchanged-upon-resize.html:67:1
09:46:26     INFO -  FAIL: Pixel at (2, 2) should have been (255, 0, 0, 255), was (0, 0, 255, 255)
09:46:26     INFO -  STACK TRACE:
09:46:26     INFO -  @http://mochi.test:8888/tests/dom/canvas/test/webgl-conformance/conformance/canvas/viewport-unchanged-upon-resize.html:67:1
09:46:26     INFO -  Not taking screenshot here: see the one that was previously logged
09:46:26     INFO -  42 INFO TEST-UNEXPECTED-FAIL | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | Expected no failures: conformance/canvas/viewport-unchanged-upon-resize.html
09:46:26     INFO -      OnTestComplete@dom/canvas/test/webgl-conformance/mochi-single.html?conformance/canvas/viewport-unchanged-upon-resize.html:246:1
09:46:26     INFO -      window.webglTestHarness.notifyFinished@dom/canvas/test/webgl-conformance/mochi-single.html?conformance/canvas/viewport-unchanged-upon-resize.html:233:5
09:46:26     INFO -      notifyFinishedToHarness@dom/canvas/test/webgl-conformance/resources/js-test-pre.js:51:5
09:46:26     INFO -      finishTest@dom/canvas/test/webgl-conformance/resources/js-test-pre.js:473:5
09:46:26     INFO -      @dom/canvas/test/webgl-conformance/conformance/canvas/viewport-unchanged-upon-resize.html:109:1
09:46:26     INFO -  MEMORY STAT | vsize 3492MB | residentFast 385MB | heapAllocated 117MB
09:46:26     INFO -  43 INFO TEST-OK | dom/canvas/test/webgl-conformance/_wrappers/test_conformance__canvas__viewport-unchanged-upon-resize.html | took 361ms
09:46:26     INFO -  ++DOMWINDOW == 134 (0x12c5db000) [pid = 1763] [serial = 134] [outer = 0x120a86c00]

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

:jgilbert, can you take a look at this and help us figure out if we need to adjust the machine, test, or manifest.

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

jgilbert, this has been a week with no response.  Any guidance and I would be happy to help out here.

Comment 3

[Kelsey Gilbert [:jgilbert]]

(In reply to Joel Maher (:jmaher) from comment #2)
> jgilbert, this has been a week with no response.  Any guidance and I would
> be happy to help out here.

Just mark it failing for the configuration in dom/canvas/test/webgl-conformance/mochitest-errara.ini and rerun generate-wrappers-and-manifests.py.

Comment 4

[Milan Sreckovic [:milan] (needinfo for best results)]

Get some more info when failing: https://treeherder.mozilla.org/perf.html#/comparechooser?newProject=try&newRevision=c251b80585fb

Comment 5

[Milan Sreckovic [:milan] (needinfo for best results)]

I meant this one: https://treeherder.mozilla.org/#/jobs?repo=try&revision=c251b80585fb even though there are some random patches on that try.

Comment 6

[Milan Sreckovic [:milan] (needinfo for best results)]

What happens if you just run this test directly on those test machines?  e.g., file:///..../dom/canvas/test/webgl-conformance/conformance/canvas/viewport-unchanged-upon-resize.html

Comment 7

[Joel Maher ( :jmaher ) (UTC -8)]

if I just run the single test via mochitest, it fails.

if I load the file directly (via file:///.../viewport-unchanged-upon-resize.html), it fails the same way:
FAIL Pixel at (2, 2) should have been (255, 0, 0, 255), was (0, 0, 255, 255)

If there is something I can do to get more information let me know, otherwise I could just skip this test for osx.

Comment 8

[Joel Maher ( :jmaher ) (UTC -8)]

:milan, let me know what you think of my results.

Comment 9

[Milan Sreckovic [:milan] (needinfo for best results)]

(In reply to Joel Maher (:jmaher) from comment #7)
> if I just run the single test via mochitest, it fails.
> 
> if I load the file directly (via
> file:///.../viewport-unchanged-upon-resize.html), it fails the same way:
> FAIL Pixel at (2, 2) should have been (255, 0, 0, 255), was (0, 0, 255, 255)
> 
> If there is something I can do to get more information let me know,
> otherwise I could just skip this test for osx.

Skip it for now, as Jeff suggested in comment 3, but I'm "glad" that it fails with the direct loading, because that's where I don't with 10.9, and we should be able to get to the bottom of it.  Without blocking you, so disable the test.

Comment 10

[Milan Sreckovic [:milan] (needinfo for best results)]

Markus, do you have a 10.10.5 or some flavour of 10.10, and does the direct loading/single mochitest fail for you (see comment 7)?  It fails on 10.10.5 for Joel, passes on 10.9.5 for me.
Joel, is this a physical machine you have, or some kind of a remote access?

Comment 11

[Joel Maher ( :jmaher ) (UTC -8)]

this is a loaner machine which I have that I remote into.  I will mark the test as disabled, we can choose to leave this bug open if we want to investigate it more.

Comment 12

[Joel Maher ( :jmaher ) (UTC -8)]

Attachment: disable test on mac

using the method recommended to just disable this test, I have done so and uploaded a patch.

Comment 13

[Markus Stange [:mstange]]

(In reply to Milan Sreckovic [:milan] from comment #10)
> Markus, do you have a 10.10.5 or some flavour of 10.10,

I only have 10.11 (Beta (15A279b) !?) at the moment, and the test passes for me.

Comment 14

[Kelsey Gilbert [:jgilbert]]

Comment on attachment 8681150 [details] [diff] [review]
disable test on mac

Review of attachment 8681150 [details] [diff] [review]:
-----------------------------------------------------------------

Thanks!

Comment 15

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/0b638967882d

Comment 16

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/0b638967882d

Comment 17

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/0b638967882d

Comment 18

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/releases/mozilla-b2g44_v2_5/rev/15d4c6707a31

Comment 19

[Pulsebot]

https://hg.mozilla.org/releases/mozilla-esr140/rev/bd82bec9bc34