Closed Bug 696009 Opened 12 years ago Closed 12 years ago

[Linux 32bit PGO] Intermittent test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas"

Categories

(Core :: Graphics: CanvasWebGL, defect)

Product:
Core
Component:
Graphics: CanvasWebGL
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED

People

(Reporter: mak, Unassigned)

References

Details

(Keywords: intermittent-failure) 

https://tbpl.mozilla.org/php/getParsedLog.php?id=6945760&tree=Mozilla-Inbound

Rev3 Fedora 12 mozilla-inbound pgo test mochitests-1/5 on 2011-10-20 00:54:16 PDT for push 67673422f7d2

WebGL mochitest: starting page conformance/glsl/reserved/_webgl_variable.vert.html
ResizeOffscreenFBO failed with AA, retrying without...
ResizeOffscreenFBO failed with AA, retrying without...
45996 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas" (URL: conformance/glsl/reserved/_webgl_variable.vert.html)
45997 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "context does not exist" (URL: conformance/glsl/reserved/_webgl_variable.vert.html)
WebGL test error: test page failure: conformance/glsl/reserved/_webgl_variable.vert.html
WebGL mochitest: starting page conformance/glsl/reserved/webgl_field.vert.html
ResizeOffscreenFBO failed with AA, retrying without...
ResizeOffscreenFBO failed with AA, retrying without...
45998 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas" (URL: conformance/glsl/reserved/webgl_field.vert.html)
45999 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "context does not exist" (URL: conformance/glsl/reserved/webgl_field.vert.html)
WebGL test error: test page failure: conformance/glsl/reserved/webgl_field.vert.html
WebGL mochitest: starting page conformance/glsl/reserved/webgl_function.vert.html
ResizeOffscreenFBO failed with AA, retrying without...
ResizeOffscreenFBO failed with AA, retrying without...
46000 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas" (URL: conformance/glsl/reserved/webgl_function.vert.html)
46001 ERROR TEST-UNEXPECTED-FAIL | /tests/content/canvas/test/webgl/test_webgl_conformance_test_suite.html | Test failed, "context does not exist" (URL: conformance/glsl/reserved/webgl_function.vert.html)
WebGL test error: test page failure: conformance/glsl/reserved/webgl_function.vert.html
The timing suggests a regression from bug 615976; this could easily be the same as bug 696093 which just got fixed, in which case we won't see it anymore.
(In reply to Benoit Jacob [:bjacob] from comment #7)
> The timing suggests a regression from bug 615976; this could easily be the
> same as bug 696093 which just got fixed, in which case we won't see it
> anymore.

Nope, comment 8 is against this push, so it's not fixed.
Summary: Intermittent test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas" → [Linux 32bit PGO] Intermittent test/webgl/test_webgl_conformance_test_suite.html | Test failed, "Unable to fetch WebGL rendering context for Canvas"
A theory and a patch are being considered on bug 695947. Try push:

https://bugzilla.mozilla.org/show_bug.cgi?id=696093
(In reply to Benoit Jacob [:bjacob] from comment #14)
> That didn't have PGO enabled. new try:
> https://tbpl.mozilla.org/?tree=Try&rev=789cf49fdc63

All green after 12 retriggers! Will land tomorrow.
http://hg.mozilla.org/mozilla-central/rev/a0784cecf9bd

Keeping this bug open until the next m-c -> m-i merge
Backed out in https://hg.mozilla.org/mozilla-central/rev/66ab3bbae4c5 to investigate the unlikely, but none the less apparently true, way that the followup push turned bug 674244, a failure in a browser-chrome test, permaorange on Linux32 debug. How a change to a test that runs in mochitest-1 could affect a test that runs in a totally different suite is... interesting.
Relanded (with the two folded together) in https://hg.mozilla.org/mozilla-central/rev/1fa31fa85082 - on your push, the original test run and five retriggers all hit bug 674244, and the original test run on the next push had hit it as well, making you look pretty bad. Then after I backed you out, I triggered some more on both your push and the next (and the previous, for that matter), and every single one of those was green, more like it was a timebomb or something like that.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
Whiteboard: [orange]
You need to log in before you can comment on or make changes to this bug.