Also on:
https://uberchromegw.corp.google.com/i/chromeos/builders/cyan-chrome-pfq/builds/200

 Issue 650441  has been merged into this issue.

This failure happens consistently on amd64-generic-telemetry starting with build 9817 [1].

1: https://build.chromium.org/p/chromiumos.chromium/builders/amd64-generic-telemetry/builds/9817

This seems to be bringing down x86-generic-tot-asan-informational[1] now.

I've found a way to repro locally, though. Here[2] are logs from a local run. It looks like Chrome is hitting a CHECK, but we are not leaving behind any logs that Chrome crashed.

[22427:22427:0929/154640:FATAL:synthetic_gesture_target_base.cc(61)] Check failed: web_touch.touches[i].state != WebTouchPoint::StatePressed || PointIsWithinContents(web_touch.touches[i].position.x, web_touch.touches[i].position.y). Touch coordinates are not within content bounds on TouchStart.
#0 0x7f7e65966356 base::debug::StackTrace::StackTrace()
#1 0x7f7e6597e44e logging::LogMessage::~LogMessage()
#2 0x7f7e649b534d content::SyntheticGestureTargetBase::DispatchInputEventToPlatform()
#3 0x7f7e649b6726 content::SyntheticSmoothMoveGesture::ForwardTouchInputEvents()
#4 0x7f7e649b6d20 content::SyntheticSmoothMoveGesture::ForwardInputEvents()
#5 0x7f7e649b38da content::SyntheticGestureController::Flush()
#6 0x7f7e64a18e8a content::RenderWidgetHostImpl::FlushInput()
#7 0x7f7e64a2685a content::RenderWidgetHostViewAura::OnBeginFrame()
#8 0x7f7e6657761b cc::DelayBasedBeginFrameSource::OnTimerTick()
#9 0x7f7e659f6f21 base::debug::TaskAnnotator::RunTask()
#10 0x7f7e65986077 base::MessageLoop::RunTask()
#11 0x7f7e6598646e base::MessageLoop::DeferOrRunPendingTask()
#12 0x7f7e659876e0 base::MessageLoop::DoDelayedWork()
#13 0x7f7e65988c02 base::MessagePumpLibevent::Run()
#14 0x7f7e659a81fa base::RunLoop::Run()
#15 0x7f7e6559a883 ChromeBrowserMainParts::MainMessageLoopRun()
#16 0x7f7e64788d50 content::BrowserMainLoop::RunMainMessageLoopParts()
#17 0x7f7e6478ba95 content::BrowserMainRunnerImpl::Run()
#18 0x7f7e64784a79 content::BrowserMain()
#19 0x7f7e6551d9ed content::ContentMainRunnerImpl::Run()
#20 0x7f7e6551b969 content::ContentMain()
#21 0x7f7e638e09f6 ChromeMain
#22 0x7f7e613cbfb6 __libc_start_main
#23 0x7f7e638e0844 <unknown>


For reproducing, I deployed an origin/master Chrome build using the simple chrome workflow to a link and then ran this command from inside the chrome simple chrome chroot:

$ ./third_party/catapult/telemetry/bin/run_tests --browser=cros-chrome --remote=$(chromeos_ip_get) ScrollActionTest.testScrollAction

An example run is at [2].

1: https://build.chromium.org/p/chromiumos.chromium/builders/x86-generic-tot-asan-informational
2: https://paste.googleplex.com/5640570500284416?raw

https://codereview.chromium.org/2372873003 seems very suspicious - assigning to dtapuska@.

Actually, I don't think this is caused by that patch. When I revert it locally I get this stack trace:

[30977:30977:0929/161326:FATAL:synthetic_gesture_target_base.cc(61)] Check failed: web_touch.touches[i].state != WebTouchPoint::StatePressed || PointIsWithinContents(web_touch.touches[i].position.x, web_touch.touches[i].position.y). Touch coordinates are not within content bounds on TouchStart.
#0 0x7fbce76c02b6 base::debug::StackTrace::StackTrace()
#1 0x7fbce76d83ae logging::LogMessage::~LogMessage()
#2 0x7fbce670f34d content::SyntheticGestureTargetBase::DispatchInputEventToPlatform()
#3 0x7fbce6710726 content::SyntheticSmoothMoveGesture::ForwardTouchInputEvents()
#4 0x7fbce6710d20 content::SyntheticSmoothMoveGesture::ForwardInputEvents()
#5 0x7fbce670d8da content::SyntheticGestureController::Flush()
#6 0x7fbce6772e8a content::RenderWidgetHostImpl::FlushInput()
#7 0x7fbce772ab2a base::Timer::RunScheduledTask()
#8 0x7fbce7750e81 base::debug::TaskAnnotator::RunTask()
#9 0x7fbce76dffd7 base::MessageLoop::RunTask()
#10 0x7fbce76e03ce base::MessageLoop::DeferOrRunPendingTask()
#11 0x7fbce76e1640 base::MessageLoop::DoDelayedWork()
#12 0x7fbce76e2b62 base::MessagePumpLibevent::Run()
#13 0x7fbce770215a base::RunLoop::Run()
#14 0x7fbce72f47e3 ChromeBrowserMainParts::MainMessageLoopRun()
#15 0x7fbce64e2d50 content::BrowserMainLoop::RunMainMessageLoopParts()
#16 0x7fbce64e5a95 content::BrowserMainRunnerImpl::Run()
#17 0x7fbce64dea79 content::BrowserMain()
#18 0x7fbce727794d content::ContentMainRunnerImpl::Run()
#19 0x7fbce72758c9 content::ContentMain()
#20 0x7fbce563a9f6 ChromeMain
#21 0x7fbce3125fb6 __libc_start_main
#22 0x7fbce563a844 <unknown>

+cc to owners.

aelias@,  dtapuska@, tdresser@: Any ideas what could be causing this?

So is this a dupe of https://bugs.chromium.org/p/chromium/issues/detail?id=648665? That bug has been around for a while so seems unlikely?

I suspect this is a different failure, but they may be related.

Do we know when it started to fail? Producing a range of changes might be helpful. I agree I wouldn't expect my change to cause it. And it is clear from the stack trace in #7 that you can reproduce it without it.

As the other issue is from the 20th. I wonder if something in aura views changed because the check that is happening is comparing the touch event coords to the bounds of the content.

If you can add some LOG(ERROR) in here and get some context of the rect and the point; it might be helpful...

https://cs.chromium.org/chromium/src/content/browser/renderer_host/input/synthetic_gesture_target_base.cc?q=DispatchInputEventToPlatform&sq=package:chromium&l=129

Starting today, it seems that this test works as of the latest master. Marking as WontFix. crbug.com/652420 contains info for the other crashes similar to this one.