DCHECK being hit in release Linux build: events_x_utils.cc:274 |
|||||
Issue descriptionUserAgent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.116 Safari/537.36 Steps to reproduce the problem: 1. Build tip-of-tree chrome in release, no-components, DCHECK always on 2. Run and click anywhere in the browser window What is the expected behavior? Should not crash. What went wrong? It crashes. $ out/Release/chrome --user-data-dir=/tmp/kjsdffhkklfasjkjk [506:506:0831/152919:FATAL:events_x_utils.cc(274)] Check failed: false. #0 0x7f0852b6ad7e base::debug::StackTrace::StackTrace() #1 0x7f0852b81f5b logging::LogMessage::~LogMessage() #2 0x7f08534fd919 ui::EventTypeFromXEvent() #3 0x7f0853e6cdf6 views::DesktopWindowTreeHostX11::DispatchEvent() #4 0x7f0853e6d0cd views::DesktopWindowTreeHostX11::DispatchEvent() #5 0x7f08534dd068 ui::PlatformEventSource::DispatchEvent() #6 0x7f08534ff6e6 ui::X11EventSource::ExtractCookieDataDispatchEvent() #7 0x7f08534ff5fd ui::X11EventSource::DispatchXEvents() #8 0x7f08563cc149 ui::(anonymous namespace)::XSourceDispatch() #9 0x7f084f19fce5 g_main_context_dispatch #10 0x7f084f1a0048 <unknown> #11 0x7f084f1a00ec g_main_context_iteration #12 0x7f0852b8b65d base::MessagePumpGlib::Run() #13 0x7f0852b890e1 base::MessageLoop::RunHandler() #14 0x7f0852bab2f0 base::RunLoop::Run() #15 0x7f0852a145ba ChromeBrowserMainParts::MainMessageLoopRun() #16 0x7f0851e35489 content::BrowserMainLoop::RunMainMessageLoopParts() #17 0x7f0851e38b78 content::BrowserMainRunnerImpl::Run() #18 0x7f0851e308be content::BrowserMain() #19 0x7f08527615fe content::RunNamedProcessTypeMain() #20 0x7f08527620b3 content::ContentMainRunnerImpl::Run() #21 0x7f0852760910 content::ContentMain() #22 0x7f0851198b4b ChromeMain #23 0x7f084a0acf45 __libc_start_main #24 0x7f0851198a21 <unknown> [1:1:0100/000000:ERROR:broker_posix.cc(41)] Invalid node channel message Aborted (core dumped) Crashed report ID: How much crashed? Whole browser Is it a problem with a plugin? N/A Did this work before? Yes Recently, not sure (within the last day or two). Chrome version: Channel: n/a OS Version: Flash Version: Shockwave Flash 22.0 r0
,
Aug 31 2016
I should include that I'm running this on a Linux machine with a touch screen attached as a second monitor, in case that's related.
,
Aug 31 2016
It looks like evtype is XI_Enter/XI_Leave. Not sure what would have caused this problem to start showing up but I can replicate without a touchcsreen by using the --touch-devices flag.
,
Aug 31 2016
This was probably caused by https://codereview.chromium.org/2165083002/ That CL selected XI focus and crossing events. You can probably just return ui::ET_UNKNOWN for XI_Enter/XI_Leave/XI_FocusIn/XI_FocusOut + sadrul to confirm
,
Sep 1 2016
,
Sep 5 2016
Is there a workaround for this? AFAIK I haven't changed anything in my setup, but I'm running into this consistently and it makes developing hard...
,
Sep 5 2016
This is only linux workstation, and I don't have any touch devices attached to it.
,
Sep 5 2016
#6 + #7 kylechar@ has created a CL that has not landed yet https://codereview.chromium.org/2298953003/ Please try applying that patch locally
,
Sep 5 2016
Great, thanks!
,
Sep 6 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a6e7a8cb6df5a0ab7ddc7fe38e279e98ae0bd548 commit a6e7a8cb6df5a0ab7ddc7fe38e279e98ae0bd548 Author: kylechar <kylechar@chromium.org> Date: Tue Sep 06 20:41:27 2016 Handle XI_Enter and XI_Leave events. These events are now expected and are hitting a NOTREACHED statement. Add case to handle them but doesn't do anything to avoid this. BUG= 642856 Review-Url: https://codereview.chromium.org/2298953003 Cr-Commit-Position: refs/heads/master@{#416714} [modify] https://crrev.com/a6e7a8cb6df5a0ab7ddc7fe38e279e98ae0bd548/ui/events/x/events_x_utils.cc [modify] https://crrev.com/a6e7a8cb6df5a0ab7ddc7fe38e279e98ae0bd548/ui/views/widget/desktop_aura/desktop_window_tree_host_x11.cc
,
Sep 6 2016
|
|||||
►
Sign in to add a comment |
|||||
Comment 1 by wjmaclean@chromium.org
, Aug 31 2016Owner: kylec...@chromium.org
Status: Assigned (was: Unconfirmed)