SurfaceSync + Viz Android TouchInputBrowserTest.TouchHandlerConsume Failure |
||||||||
Issue descriptionOS: Android Test suite: content_browsertests Test: TouchInputBrowserTest.TouchHandlerConsume Args: --enable-surface-synchronization Seeing a consistent failure of this test, with flaking failures of TouchHandlerNoConsume. Example failure: https://ci.chromium.org/p/chromium/builders/luci.chromium.try/android-marshmallow-arm64-rel/23013 Stack Trace: ./../content/browser/renderer_host/input/touch_input_browsertest.cc:191: Failure Expected equality of these values: INPUT_EVENT_ACK_STATE_CONSUMED Which is: 1 filter->WaitForAck() Which is: 4 Stack trace: Searching for native crashes in: /b/swarming/w/itGdaHYa/tmpviMKkb Unknown Android release, consider passing --packed-lib. Reading Android symbols from: /b/swarming/w/ir Searching for Chrome symbols from within: /b/swarming/w/ir/out/Release/lib.unstripped:/b/swarming/w/ir/out/Release Stack Trace: RELADDR FUNCTION FILE:LINE 000000000215b4db content::TouchInputBrowserTest_TouchHandlerConsume_Test::RunTestOnMainThread() ??:0:0 0000000003e3d053 content::BrowserTestBase::ProxyRunTestOnMainThreadLoop() ??:0:0 0000000001fb94e3 void base::internal::Invoker<base::internal::BindState<void (content::GenerateMHTMLAndExitRendererMessageFilter::*)(), base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >, void ()>::RunImpl<void (content::GenerateMHTMLAndExitRendererMessageFilter::*)(), std::__ndk1::tuple<base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >, 0ul>(void (content::GenerateMHTMLAndExitRendererMessageFilter::*&&)(), std::__ndk1::tuple<base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >&&, std::__ndk1::integer_sequence<unsigned long, 0ul>) ??:0:0 0000000003ea3307 content::ShellBrowserMainParts::PreMainMessageLoopRun() ??:0:0 0000000002b4a07b content::BrowserMainLoop::PreMainMessageLoopRun() ??:0:0 0000000001fb94e3 void base::internal::Invoker<base::internal::BindState<void (content::GenerateMHTMLAndExitRendererMessageFilter::*)(), base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >, void ()>::RunImpl<void (content::GenerateMHTMLAndExitRendererMessageFilter::*)(), std::__ndk1::tuple<base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >, 0ul>(void (content::GenerateMHTMLAndExitRendererMessageFilter::*&&)(), std::__ndk1::tuple<base::internal::UnretainedWrapper<content::GenerateMHTMLAndExitRendererMessageFilter> >&&, std::__ndk1::integer_sequence<unsigned long, 0ul>) ??:0:0 0000000002e9b50b content::StartupTaskRunner::RunAllTasksNow() ??:0:0 0000000002b492e3 content::BrowserMainLoop::CreateStartupTasks() ??:0:0 0000000002b4bd6b content::BrowserMainRunnerImpl::Initialize(content::MainFunctionParams const&, std::__ndk1::unique_ptr<content::BrowserProcessSubThread, std::__ndk1::default_delete<content::BrowserProcessSubThread> >) ??:0:0 0000000002b47977 content::BrowserMain(content::MainFunctionParams const&, std::__ndk1::unique_ptr<content::BrowserProcessSubThread, std::__ndk1::default_delete<content::BrowserProcessSubThread> >) ??:0:0 0000000003e3cde7 content::BrowserTestBase::SetUp() ??:0:0 0000000003e3232b content::ContentBrowserTest::SetUp() ??:0:0 000000000281f943 testing::Test::Run() ??:0:0 [ RUN ] TouchInputBrowserTest.TouchHandlerConsume
,
Jun 22 2018
,
Jun 22 2018
,
Jun 29 2018
I think this test is flaky regardless. Hmm...
,
Jul 9
,
Jul 23
Bump, this is the remaining Viz Android testing failure.
,
Jul 23
Grabbing to look at myself
,
Jul 23
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/59d19bd1a9da98e7b1b4d598936538ae65b6fb64 commit 59d19bd1a9da98e7b1b4d598936538ae65b6fb64 Author: Jonathan Ross <jonross@chromium.org> Date: Mon Jul 23 22:53:05 2018 Speculatively re-enable TouchInputBrowserTest.TouchHandlerConsume So TouchInputBrowserTest.TouchHandlerConsume was failing somewhat consistently. However I cannot reproduce the error locally. Nor on trybot runs. I want to re-enable it on the FYI bot Mojo Android to see if it still reproduces, or has happened to be fixed. TBR=fsamuel@chromium.org TEST=TouchInputBrowserTest.TouchHandlerConsume Bug: 855532 Change-Id: Ib6b6bfe013f69b692b8975d70077fed92401299a Reviewed-on: https://chromium-review.googlesource.com/1147070 Reviewed-by: Jonathan Ross <jonross@chromium.org> Cr-Commit-Position: refs/heads/master@{#577313} [modify] https://crrev.com/59d19bd1a9da98e7b1b4d598936538ae65b6fb64/testing/buildbot/filters/viz.android.content_browsertests.filter
,
Jul 24
Issue 514456 has been merged into this issue.
,
Jul 27
,
Jul 27
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/9db0a14b711415d66efdcec8eb88971a9b2efdb0 commit 9db0a14b711415d66efdcec8eb88971a9b2efdb0 Author: jonross <jonross@chromium.org> Date: Fri Jul 27 21:50:52 2018 Update TouchInputBrowserTest Frame Synchronizing The TouchInputBrowserTests were flaking across OSes. This was due to how it was waiting for frames. It was waiting for a visual properties synchronization to have been started. However, while this leads to a frame being generated, it does not block until that frame is produced. This change updates the frame syncing. Now it will use RenderFrameSubmisisonObserver to wait for a frame from the URL load to be generated. For non-Android it will then set a size, which triggers a visual properties sync. Then it will wait for a frame to be generated. On Android, this size is set in the initial page load, so there is no synchronizing here. Furthermore for Viz Hit Testing, a wait for hit test data is added. TEST=TouchInputBrowserTest Bug: 346363 , 514456 , 855532 Change-Id: Ifa9179af9234d98c5be5575876c061763edb9f4e Reviewed-on: https://chromium-review.googlesource.com/1147322 Reviewed-by: Dirk Pranke <dpranke@chromium.org> Reviewed-by: Sadrul Chowdhury <sadrul@chromium.org> Commit-Queue: Jonathan Ross <jonross@chromium.org> Cr-Commit-Position: refs/heads/master@{#578819} [modify] https://crrev.com/9db0a14b711415d66efdcec8eb88971a9b2efdb0/content/browser/renderer_host/input/touch_input_browsertest.cc [modify] https://crrev.com/9db0a14b711415d66efdcec8eb88971a9b2efdb0/testing/buildbot/chromium.fyi.json [modify] https://crrev.com/9db0a14b711415d66efdcec8eb88971a9b2efdb0/testing/buildbot/filters/BUILD.gn [delete] https://crrev.com/74963d4150ed161e053de56206e41d345e7c8161/testing/buildbot/filters/surface_sync.content_browsertests.filter [delete] https://crrev.com/74963d4150ed161e053de56206e41d345e7c8161/testing/buildbot/filters/viz.android.content_browsertests.filter [modify] https://crrev.com/9db0a14b711415d66efdcec8eb88971a9b2efdb0/testing/buildbot/test_suites.pyl
,
Jul 27
|
||||||||
►
Sign in to add a comment |
||||||||
Comment 1 by jonr...@chromium.org
, Jun 22 2018Appears to also occur in the Viz config. Example from TouchHanlderNoComsume: ../../content/browser/renderer_host/input/touch_input_browsertest.cc:168: Failure Expected equality of these values: INPUT_EVENT_ACK_STATE_NOT_CONSUMED Which is: 2 filter->WaitForAck() Which is: 4