fast/forms/search-event-delay.html fails sometimes. For example the latest fail on the Qt bot: http://build.webkit.org/results/Qt%20Linux%20Release/r77916%20%2827877%29/fast/forms/search-event-delay-pretty-diff.html --- /home/webkitbuildbot/slaves/release32bit/buildslave/qt-linux-release/build/layout-test-results/fast/forms/search-event-delay-expected.txt 2011-02-08 01:30:54.550056855 -0800 +++ /home/webkitbuildbot/slaves/release32bit/buildslave/qt-linux-release/build/layout-test-results/fast/forms/search-event-delay-actual.txt 2011-02-08 01:30:54.550056855 -0800 @@ -6,4 +6,4 @@ The two rows below should match. 0.5 0.4 0.3 0.2 0.2 0 -0.5 0.4 0.3 0.2 0.2 0 +0.5 0.4 0.3 0.2 0.4 0
This test introduced in http://trac.webkit.org/changeset/19336 . I cc-ed Darin as author and Maciej as reviewer of this changeset. Have you got any idea what cause this rare flakey fail and how can we fix it?
My guess is the flakiness is caused by the fact that it’s a test that involves real time.
If the bot is heavily loaded, then the test will fail. That number means that it took 0.2 seconds. To make this test more reliable we could come up with a way to measure time that is independent of wall time and make WebCore's logic use that time system, at least when running tests.
Marked as flaky in <http://trac.webkit.org/r156732>. The test frequently fails on debug bots. It also failed on my local run of release build (with --child-processes=1, so CPU was not heavily loaded). My local result: The two rows below should match. 0.5 0.4 0.3 0.2 0.2 0 -0.5 0.4 0.3 0.2 0.2 0 +0.6 0.5 0.4 0.3 0.3 0.1
Test is gone - https://github.com/WebKit/WebKit/commit/99ff08340f12db8d867a0ed04f176802b7e0a355