Detected 3 new flakes for test/step "mash_browser_tests (with patch)". To see the actual flakes, please visit https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyKgsSBUZsYWtlIh9tYXNoX2Jyb3dzZXJfdGVzdHMgKHdpdGggcGF0Y2gpDA. This message was posted automatically by the chromium-try-flakes app.

Detected 9 new flakes for test/step "mash_browser_tests (with patch)". To see the actual flakes, please visit https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyKgsSBUZsYWtlIh9tYXNoX2Jyb3dzZXJfdGVzdHMgKHdpdGggcGF0Y2gpDA. This message was posted automatically by the chromium-try-flakes app.

Detected 14 new flakes for test/step "mash_browser_tests (with patch)". To see the actual flakes, please visit https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyKgsSBUZsYWtlIh9tYXNoX2Jyb3dzZXJfdGVzdHMgKHdpdGggcGF0Y2gpDA. This message was posted automatically by the chromium-try-flakes app.

jamescook@, mash_browser_tests is failing, and I found you touched the test recently.  Could you take a look and triage appropriately?

The test log looks below:

[0927/111138:FATAL:compositor.cc(486)] Check failed: false. 
#0 0x00000312915e base::debug::StackTrace::StackTrace()
#1 0x00000313efba logging::LogMessage::~LogMessage()
#2 0x000004939703 ui::Compositor::DidFailToInitializeCompositorFrameSink()
#3 0x000007502f97 cc::LayerTreeHostInProcess::DidFailToInitializeCompositorFrameSink()
#4 0x00000750accb cc::SingleThreadProxy::SetCompositorFrameSink()
#5 0x000007502b90 cc::LayerTreeHostInProcess::SetCompositorFrameSink()
#6 0x000004938c7a ui::Compositor::SetCompositorFrameSink()
#7 0x000005a0f690 views::SurfaceContextFactory::CreateCompositorFrameSink()
#8 0x0000049396a1 ui::Compositor::RequestNewCompositorFrameSink()
#9 0x00000750aaf2 cc::SingleThreadProxy::RequestNewCompositorFrameSink()
#10 0x0000031b6814 base::debug::TaskAnnotator::RunTask()
#11 0x000003146ee5 base::MessageLoop::RunTask()
#12 0x0000031471d8 base::MessageLoop::DeferOrRunPendingTask()
#13 0x00000314751b base::MessageLoop::DoWork()
#14 0x0000031489ec base::MessagePumpDefault::Run()
#15 0x000003146a2d base::MessageLoop::RunHandler()
#16 0x000003167eb0 base::RunLoop::Run()
#17 0x00000311c66c (anonymous namespace)::StartChildApp()
#18 0x0000016514b6 _ZN4base8internal7InvokerINS0_9BindStateIPFvN4mojo16InterfaceRequestIN3ash5mojom16SystemTrayClientEEEEJEEES9_E3RunEPNS0_13BindStateBaseEOS8_
#19 0x000002a226b0 shell::ChildProcessMainWithCallback()
#20 0x00000311c3b8 RunMashBrowserTests()
#21 0x00000311c283 main
#22 0x7f0a196787ed __libc_start_main
#23 0x0000006209bd <unknown>

Failing at
https://cs.chromium.org/chromium/src/ui/compositor/compositor.cc?q=ui::Compositor::DidFailToInitializeCompositorFrameSink&sq=package:chromium&l=482

Sadrul, can you find a graphics-y person to own this?

I've also seen it locally, sometimes.

This sounds like  issue 651239 . I will investigate.

 Issue 651239  has been merged into this issue.

Removing this issue from the sheriff queue, since it's assigned to an owner.

Detected 12 new flakes for test/step "mash_browser_tests (with patch)". To see the actual flakes, please visit https://chromium-try-flakes.appspot.com/all_flake_occurrences?key=ahVzfmNocm9taXVtLXRyeS1mbGFrZXNyKgsSBUZsYWtlIh9tYXNoX2Jyb3dzZXJfdGVzdHMgKHdpdGggcGF0Y2gpDA. This message was posted automatically by the chromium-try-flakes app.

Fix up for review: https://codereview.chromium.org/2379263002/

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/6616407d2d6f64c18234c08cdda042827b7295c1

commit 6616407d2d6f64c18234c08cdda042827b7295c1
Author: sadrul <sadrul@chromium.org>
Date: Fri Sep 30 14:47:43 2016

services/ui: Fix setting client id for gpu channels.

This fixes a race in setting the client id for the gpu channels for mus
clients. This race can happen if client requests for gpu channels start
coming in to the window server, before the gpu service is initialized
and the ws is notified. This race was causing the gpu channels in mus
clients to get dropped, leading to unexpected crashes.

BUG= 649567 

Review-Url: https://codereview.chromium.org/2379263002
Cr-Commit-Position: refs/heads/master@{#422109}

[modify] https://crrev.com/6616407d2d6f64c18234c08cdda042827b7295c1/services/ui/ws/gpu_service_proxy.cc

This is now fixed.