Yes, we turned it back on yesterday because all known issues have been resolved.

This flake here is different from the actual crashers as it crashes intentionally in the verifier, meaning that we miss a write barrier. Since it has not happened before I assume the last few CLs changed the structure in a way that we now miss a WB.

Taking a look.

 Issue 671791  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7737c8e7d1f6cbf4fdae1fd319709b311254d0e3

commit 7737c8e7d1f6cbf4fdae1fd319709b311254d0e3
Author: mlippautz <mlippautz@chromium.org>
Date: Wed Dec 07 11:55:10 2016

[wrapper-tracing] WebGLRenderingContextBase: Fix wrapper tracing

BUG= 671942 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2560473005
Cr-Commit-Position: refs/heads/master@{#436919}

[modify] https://crrev.com/7737c8e7d1f6cbf4fdae1fd319709b311254d0e3/third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp

As far as I can tell there are no more flakes for WebglConformance_conformance_context_context_lost_restored on the mac bots. 

Please assign any related issues directly to me or reopen this one.

It looks like misc_uninitialized_test_2 on Win7 Release (NVidia) is still failing with the same assertion:

WebglConformance_conformance2_misc_uninitialized_test_2 (gpu_tests.webgl_conformance_integration_test.WebGLConformanceIntegrationTest) ... [3456:2820:1207/095255.973:FATAL:scriptwrappablevisitorverifier.h(54)] Check failed: false. 

https://build.chromium.org/p/chromium.gpu.fyi/builders/Win7%20x64%20Release%20%28NVIDIA%29/builds/8307

mlippautz@, could you take a look?

Is this the same bug as  issue 671791 ? Ken I suspect that might not be the case.

Just had another flake of uninitialized_test_2

https://build.chromium.org/p/tryserver.chromium.angle/builders/linux_angle_rel_ng/builds/2934

Ufz, 

The stacktraces all indicate failures in the verifier: (1) WebGLContextGroup and (2) WebGLFrameBuffer

(1) is for sure related to the re-architecturing done last week
(2) is new to me and definitely didn't happen the first time it was in the tree for a couple days. So might also be related to the recent changes.

I will have a look tomorrow.

Sorry for the inconvenience but I am trying to not revert this time. Every time I postpone the feature I have to play catch-up with changes in some other component. If you feel this cannot wait revert yourself at https://codereview.chromium.org/2540743003/. 

I un-duplicated  Issue 671791  from this one. Sorry, I didn't realize that this was a bug in the wrapper tracing for WebGLContext and the other one is for WebGLFramebuffer.

Looking a bit into the WebGLFramebuffer's management of its attachments, but might need to hand that off to mlippautz@ if I can't figure it out.

Re-closing this as fixed.