RESOLVED FIXED 62761
Layout Test panScroll-click-hyperlink.html flaky on Chrome/Windows, fails on Chrome/Linux
https://bugs.webkit.org/show_bug.cgi?id=62761
Summary Layout Test panScroll-click-hyperlink.html flaky on Chrome/Windows, fails on ...
Stephen White
Reported 2011-06-15 15:32:01 PDT
Layout test fast/events/panScroll-click-hyperlink.html flaky on Chrome/Windows, fails on Chrome/Linux: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fevents%2FpanScroll-click-hyperlink.html&group=%40ToT%20-%20chromium.org&showExpectations=true Seems to have been that way since landing.
Attachments
Test case with added debug statement (12.13 KB, text/html)
2011-06-16 14:16 PDT, Daniel Bates
no flags
Layout Test - Chromium Linux expected failure results (1.48 KB, patch)
2011-06-16 14:19 PDT, Daniel Bates
no flags
Daniel Bates
Comment 1 2011-06-15 15:37:38 PDT
Does Chromium Linux enable PAN_SCROLLING? If not, then this test is expected to fail on that platform and we should land expected failure results. I'll look into this issue for Chromium Linux and Windows shortly.
Daniel Bates
Comment 2 2011-06-16 14:16:13 PDT
Created attachment 97494 [details] Test case with added debug statement I'm unsure why the test is flacky on Windows. It could be the the use of setTimeout() and/or the length of the time we wait (500ms) is insufficient. I'm curious of the value of testContainer.scrollTop when it fails on Windows. Attached is a modified version of the test case that adds the line: shouldBe(testContainer.scrollTop, testContainer.scrollHeight - testContainer.clientHeight); The test also inlines the js-test-pre.js and js-test-style.css so that the test is standalone. I don't have a Chromium Windows build at the moment. I would appreciate if someone could run this modified test with new-webkit-tests and report the result. Clearly, we're more interested in the test failing.
Daniel Bates
Comment 3 2011-06-16 14:19:06 PDT
Created attachment 97495 [details] Layout Test - Chromium Linux expected failure results From looking at WebKit/chromium/features.gypi and JavaScriptCore/wtf/Platform.h, Chromium Linux doesn't build with PAN_SCROLLING enabled. Hence, this test is expected to fail.
Mihai Parparita
Comment 4 2011-07-02 20:42:28 PDT
FWIW, http://trac.webkit.org/changeset/90235/ added failing expectations for the Windows port, so the test "passes" now for Chromium. Bug 63809 is open about that Windows port.
Ryosuke Niwa
Comment 5 2011-09-14 21:26:40 PDT
Comment on attachment 97495 [details] Layout Test - Chromium Linux expected failure results Remove r? flag since the rebaseline is landed.
Note You need to log in before you can comment on or make changes to this bug.