kbr, this may be related to a state restore bug that you're working on?

Feel free to re-triage back to me if it's unrelated.

iMac (27-inch, Mid 2010) ATI Radeon HD 5750 1024MB 
MacOS 10.11.6
Chrome 71.0.3578.80

Thank you for reporting this, and for the excellent test case! I'm 99% sure this is the same problem as  Issue 911918 ; have confirmed that the fix for the other bug also fixes this, but want to look more closely at the WebGL calls being made by both to make sure there isn't any other context state leak I may have missed.

Thank you for the fast response!

Please try vary frame numbers, turn antialias on/off before you declare this fixed.

What is the easiest way for me to test the fix when you publish it? Is it possible to do something without building whole chromium?

Will do, and will look into expanding the test case that was written for the other bug.

You'll be able to test this fix in Chrome Canary when it's integrated. I highly recommend testing with Canary on a regular basis and filing *any* bugs you find. Feel free to email me the IDs of bugs you file directly and I'll make sure they get categorized appropriately.

Erik and I investigated this thing for 2 weeks, had to make sure its real bug and not something that PixiJS or a plugin does. We've managed to squash several bugs in pixi and plugins before we got to actual cause.

I will follow your advice, next time I will report a suspicion of a bug earlier.

I already have positive experience regarding resolution of graphics bugs in Skia, Blink-SVG.

Thank you again for reporting this bug. I've expanded the associated WebGL conformance test in https://github.com/KhronosGroup/WebGL/pull/2756 and believe it now covers all of the code paths. Interestingly, all four cases failed without the fix from  Issue 911918 , and all pass with the fix.

Please tell me if you see any missing cases. Since this was the same bug as  Issue 911918 , duplicating this into that one.

Looks fine, I hope it gets in canary soon and i'll test it personally :)

It should be in
Canary 73.0.3637.0
Dev 73.0.3640.0
https://chromiumdash.appspot.com/commit/5fe432ffc97808da0d270c97dc08756758a23e4d

Oops, that's for Android. Slightly different for Mac:
Canary 73.0.3637.0
Dev 73.0.3639.1

Awesome! Thanks. I will put your aliases in the credits for the game I'm working on.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4d1557e84992428767937e98808d3bbad915e53f

commit 4d1557e84992428767937e98808d3bbad915e53f
Author: James Darpinian <jdarpinian@chromium.org>
Date: Wed Dec 19 07:17:47 2018

Roll WebGL 5b6cbd7..a2b3563

https://chromium.googlesource.com/external/khronosgroup/webgl.git/+log/5b6cbd7..a2b3563

Bug: 849576,  809237 ,  911918 ,  angleproject:3012 
Bug: 913301,  912842 ,  913033 ,  563816 , 788439
Bug: angleproject:3033
Change-Id: I9b257afd69d26e75d1635633f556249583fdaa52
Cq-Include-Trybots: luci.chromium.try:win_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_angle_rel_ng;luci.chromium.try:win_angle_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/c/1383329
Commit-Queue: James Darpinian <jdarpinian@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#617754}
[modify] https://crrev.com/4d1557e84992428767937e98808d3bbad915e53f/DEPS
[modify] https://crrev.com/4d1557e84992428767937e98808d3bbad915e53f/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
[modify] https://crrev.com/4d1557e84992428767937e98808d3bbad915e53f/content/test/gpu/gpu_tests/webgl_conformance_revision.txt