1283780 - Intermittent webgl-color-test.html?16bpp&readback&premult&_____ == http://10.0.2.2:8854/tests/dom/canvas/test/reftest/wrapper.html?colors-no-alpha.png | application crashed [@ jemalloc_crash]

Status: RESOLVED WORKSFORME

(Core :: Graphics: CanvasWebGL, defect, P5)

Description

[Carsten Book [:Tomcat]]

https://treeherder.mozilla.org/logviewer.html#?job_id=155762&repo=autoland#L2712

 00:58:30 INFO - REFTEST PROCESS-CRASH | http://10.0.2.2:8854/tests/dom/canvas/test/reftest/webgl-color-test.html?16bpp&readback&premult&_____ == http://10.0.2.2:8854/tests/dom/canvas/test/reftest/wrapper.html?colors-no-alpha.png | application crashed [@ jemalloc_crash]

Comment 1

[Intermittent Failures Robot]

21 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 13
* fx-team: 5
* autoland: 3

Platform breakdown:
* android-4-3-armv7-api15: 21

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1283780&startday=2016-07-04&endday=2016-07-10&tree=all

Comment 2

[Intermittent Failures Robot]

23 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 9
* fx-team: 7
* autoland: 6
* mozilla-central: 1

Platform breakdown:
* android-4-3-armv7-api15: 23

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1283780&startday=2016-07-11&endday=2016-07-17&tree=all

Comment 3

[Intermittent Failures Robot]

23 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 10
* autoland: 9
* mozilla-central: 2
* fx-team: 2

Platform breakdown:
* android-4-3-armv7-api15: 23

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1283780&startday=2016-07-18&endday=2016-07-24&tree=all

Comment 4

[BMO Automation]

Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.

Comment 5

[Kelsey Gilbert [:jgilbert]]

Seems to have stopped.

Comment 6

[Phil Ringnalda (:philor)]

Stopped like "I looked at all the failures in the hidden Android webgl tests, and this particular one no longer happens," or like "oh, I forgot these bugs were all Android and Android webgl tests have been hidden since early September"?

Comment 7

[Phil Ringnalda (:philor)]

Tough to tell: current mozilla-release looks like I remember them looking, swathes of jemalloc crashes and the like, mozilla-beta introduced an odd permafailure in gl5 while retaining close to the same volume of jemalloc crashes, aurora and trunk still have the gl5 failure, added a lot of assertions and crashes in mozilla::MediaShutdownManager::BlockShutdown, and might have dropped the percentage of jemalloc crashes, not sure.

Comment 8

[Kelsey Gilbert [:jgilbert]]

Well I'm not seeing anything at https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1283780&endday=2016-11-28&startday=2016-09-28&tree=all .

Please let me know how else I should quantify.

When we un-hide android webgl, we'll probably see this reopened.

Comment 9

[Ryan VanderMeulen [:RyanVM]]

Jeff, did you actually look at Treeherder before claiming this is gone? OrangeFactor is only going to tell you information on failures that were actually classified by someone, which isn't going to be happening if the test suite is hidden by default unless someone on the Graphics team has been monitoring those results and classifying failures as they occur.

Still looks pretty terrible AFAICT, though as philor already pointed out, it's pretty hard to make any reasonable conclusions about the current state of this specific failure when Android M(gl) is permafailing for a multitude of other reasons at the moment.
https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&filter-searchStr=android%20webgl&exclusion_profile=false

Comment 10

[Kelsey Gilbert [:jgilbert]]

Ok.