ricea: Any idea what's going on? The test is marked as slow but it doesn't seem to be timing out. There seems to be a crash in the logs:

https://test-results.appspot.com/data/layout_results/Linux_Tests__dbg__1_/74914/layout-test-results/results.html

Disabling at https://chromium-review.googlesource.com/1268518 while ricea@ investigates.

ricea@: If you are not the appropriate owner, please add the Sheriff-Chromium label back to the bug.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1c97b8752976687e34b96483474fbf2cfc97a02e

commit 1c97b8752976687e34b96483474fbf2cfc97a02e
Author: Joshua Pawlicki <waffles@chromium.org>
Date: Mon Oct 08 17:29:01 2018

Disable flaky release-handle-crash test.

TBR=olka@chromium.org

Bug:  893015 
Change-Id: Ida8106cf8d29659b3a66ed7d85d007302acfc5f6
Reviewed-on: https://chromium-review.googlesource.com/c/1268518
Commit-Queue: Joshua Pawlicki <waffles@chromium.org>
Reviewed-by: Joshua Pawlicki <waffles@chromium.org>
Cr-Commit-Position: refs/heads/master@{#597596}
[modify] https://crrev.com/1c97b8752976687e34b96483474fbf2cfc97a02e/third_party/WebKit/LayoutTests/TestExpectations

Fails about 1 time in 200 locally on my machine. I got a crash with this stack trace:

STDERR: [1:1:1010/164420.625390:FATAL:script_promise_resolver.h(60)] Check failed: state_ == kDetached || !is_promise_called_ || !GetScriptState()->ContextIsValid() || !GetExecutionContext() || GetExecutionContext()->IsContextDestroyed(). 
STDERR: #0 0x7fae9ed06e7c base::debug::StackTrace::StackTrace()
STDERR: #1 0x7fae9ec30e9b logging::LogMessage::~LogMessage()
STDERR: #2 0x7fae97b43b30 blink::ScriptPromiseResolver::~ScriptPromiseResolver()
STDERR: #3 0x7fae9690b1ab blink::NormalPage::Sweep()
STDERR: #4 0x7fae96906369 blink::BaseArena::CompleteSweep()
STDERR: #5 0x7fae969170da blink::ThreadState::EagerSweep()
STDERR: #6 0x7fae96916ad6 blink::ThreadState::AtomicPauseEpilogue()
STDERR: #7 0x7fae9691ab06 blink::ThreadState::AtomicPauseSweepAndCompact()
STDERR: #8 0x7fae9691a5c2 blink::ThreadState::RunAtomicPause()
STDERR: #9 0x7fae96911c7b blink::ThreadState::CollectGarbage()
STDERR: #10 0x7fae9691681c blink::ThreadState::IncrementalMarkingFinalize()
STDERR: #11 0x7fae969161e7 blink::ThreadState::RunScheduledGC()
STDERR: #12 0x7fae969198de blink::ThreadState::SafePoint()
STDERR: #13 0x7fae9ec955b5 base::sequence_manager::internal::SequenceManagerImpl::NotifyDidProcessTask()
STDERR: #14 0x7fae9ec95320 base::sequence_manager::internal::SequenceManagerImpl::DidRunTask()
STDERR: #15 0x7fae9eca3f5f base::sequence_manager::internal::ThreadControllerImpl::DoWork()
STDERR: #16 0x7fae9eca6088 _ZN4base8internal7InvokerINS0_9BindStateIMNS_16sequence_manager8internal20ThreadControllerImplEFvNS5_8WorkTypeEEJNS_7WeakPtrIS5_EES6_EEEFvvEE3RunEPNS0_13BindStateBaseE
STDERR: #17 0x7fae9ec11a7d base::debug::TaskAnnotator::RunTask()
STDERR: #18 0x7fae9ec3eaee base::MessageLoop::RunTask()
STDERR: #19 0x7fae9ec3efb2 base::MessageLoop::DoWork()
STDERR: #20 0x7fae9ec42916 base::MessagePumpDefault::Run()
STDERR: #21 0x7fae9ec3e5c1 base::MessageLoop::Run()
STDERR: #22 0x7fae9ec72556 base::RunLoop::Run()
STDERR: #23 0x7fae9e344332 content::RendererMain()
STDERR: #24 0x7fae9e414f87 content::RunZygote()
STDERR: #25 0x7fae9e416197 content::ContentMainRunnerImpl::Run()
STDERR: #26 0x7fae9a74cb09 service_manager::Main()
STDERR: #27 0x7fae9e414541 content::ContentMain()
STDERR: #28 0x0000002ea14b main
STDERR: #29 0x7fae93c6f2b1 __libc_start_main
STDERR: #30 0x0000002ea02a _start

Appears to be unrelated to outofblink-cors, as the normal version of the test http/tests/fetch/chromium/release-handle-crash.html also crashes 2% of the time.

There are no crashes or failures when DCHECK is disabled, so lowering priority.

The test wasn't flaky when it first landed in 035ce793d12. I will bisect it when I have time.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/607b9a3aa9c6a03cacd8a6fdabcbce3de6b6a440

commit 607b9a3aa9c6a03cacd8a6fdabcbce3de6b6a440
Author: Adam Rice <ricea@chromium.org>
Date: Thu Oct 11 05:52:00 2018

Mark release-handle-crash.html as crashy

Layout test http/tests/fetch/chromium/release-handle-crash.html crashes around
2% of the time in DCHECK-enabled builds.

Add a crash expectation to TestExpectations.

BUG= 893015 

Change-Id: I7b3b208ab76027eb52e80facfb1c7d1a6b41023b
Reviewed-on: https://chromium-review.googlesource.com/c/1273175
Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
Commit-Queue: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#598683}
[modify] https://crrev.com/607b9a3aa9c6a03cacd8a6fdabcbce3de6b6a440/third_party/WebKit/LayoutTests/TestExpectations

I have discovered I can make the crash reliable by causing a garbage collection at the end of the test.

I am running a bisect to find the CL that introduced the crash.

3cbb8c9aeac96ef999816bf2ce2ffd826a08cfe4 is the first bad commit
commit 3cbb8c9aeac96ef999816bf2ce2ffd826a08cfe4
Author: Dong-hee Na <donghee.na92@gmail.com>
Date:   Wed Sep 19 15:15:39 2018 +0000

    Fetch: Throw TypeError for response from disturbed/locked streams
    
    If the Response body object is a readable stream then check whether
    the object is disturbed or locked.
    If the condition is true then throw TypeError.
    
    refernce: https://fetch.spec.whatwg.org/#concept-bodyinit-extract
    
    Bug:  878281 
    Change-Id: I0593d048e7bd146863efdc9f1d355bc70857cdc0
    Reviewed-on: https://chromium-review.googlesource.com/1226718
    Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
    Commit-Queue: Jinho Bang <jinho.bang@samsung.com>
    Cr-Commit-Position: refs/heads/master@{#592402}

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/467f2d999ad0f3c63a856dbd516f0e1f334304e8

commit 467f2d999ad0f3c63a856dbd516f0e1f334304e8
Author: Adam Rice <ricea@chromium.org>
Date: Mon Oct 15 08:31:47 2018

Resolve ScriptPromiseResolver when Body methods fail

When blink::BodyBuffer::StartLoading failed , the ScriptPromiseResolver was left
unresolved, resulting in a DCHECK when it was destroyed. Resolve it explicitly
in this case, to avoid the DCHECK and retain the invariant.

BUG= 893015 

Change-Id: Ib437986c6dc0d8b40e8ed0040f8796387d94feb6
Reviewed-on: https://chromium-review.googlesource.com/c/1278613
Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
Commit-Queue: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#599589}
[modify] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/SlowTests
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/resources/response-method-gc-crash.js
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/response-array-buffer-gc-crash.html
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/response-blob-gc-crash.html
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/response-form-data-gc-crash.html
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/response-json-gc-crash.html
[add] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/WebKit/LayoutTests/http/tests/fetch/chromium/response-text-gc-crash.html
[modify] https://crrev.com/467f2d999ad0f3c63a856dbd516f0e1f334304e8/third_party/blink/renderer/core/fetch/body.cc

 Issue 895246  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/57f2f7c7f2d9224eb6d00c32f0e8b9ff16be5579

commit 57f2f7c7f2d9224eb6d00c32f0e8b9ff16be5579
Author: Adam Rice <ricea@chromium.org>
Date: Tue Oct 16 04:28:49 2018

Remove crash expectations for release-handle-crash.html

http/tests/fetch/chromium/release-handle-crash.html no longer crashes
since https://chromium-review.googlesource.com/c/1278613 landed. Remove
the flaky expectations.

BUG= 893015 

Change-Id: I79738c8d44e001a3fad05fdbe11d687d895a3dae
Reviewed-on: https://chromium-review.googlesource.com/c/1280074
Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
Commit-Queue: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/master@{#599848}
[modify] https://crrev.com/57f2f7c7f2d9224eb6d00c32f0e8b9ff16be5579/third_party/WebKit/LayoutTests/TestExpectations