The period is not missing from the end of the sentence, as the test suggests. --- /tmp/layout-test-results/fast/events/keydown-1-expected.txt 2011-04-27 00:56:22.484711997 -0700 +++ /tmp/layout-test-results/fast/events/keydown-1-actual.txt 2011-04-27 00:56:22.484711997 -0700 @@ -3,4 +3,4 @@ keydown in subframe: [object KeyboardEvent] keydown in subframe: [object KeyboardEvent] -This sentence should not end with a period +This sentence should not end with a period.
FWIW, the test was converted to text only in <trac.webkit.org/r84897>. This looks more like an actual bug in the port at a first glance though.
(In reply to comment #1) > FWIW, the test was converted to text only in <trac.webkit.org/r84897>. This looks more like an actual bug in the port at a first glance though. Yes, even before the text conversion this test was incredibly flaky across different machines. The text conversion provides good impetus to finally squash this bug.
I believe so. The pixel result may falsely consider it's passing because one dot doesn't contribute much in the whole image. This test also fails on Qt but it was skipped. I suspect it's a key binding issue. Probably we should rewrite the test to use common key codes.
Created attachment 91970 [details] fix patch
Comment on attachment 91970 [details] fix patch Thanks!
The commit-queue encountered the following flaky tests while processing attachment 91970 [details]: inspector/debugger/debugger-scripts.html bug 59921 (authors: pfeldman@chromium.org and podivilov@chromium.org) The commit-queue is continuing to process your patch.
Comment on attachment 91970 [details] fix patch Clearing flags on attachment: 91970 Committed r85528: <http://trac.webkit.org/changeset/85528>
All reviewed patches have been landed. Closing bug.