editing/selection/directionality-after-undo-replace.html has been failing since it was checked.
Thanks for filling this bug. The issue here seems to be that there are so many leapForwards that the test is exceeding the double-click timeout on GTK+. This change causes the test to start passing: --- a/LayoutTests/editing/selection/directionality-after-undo-replace.html +++ b/LayoutTests/editing/selection/directionality-after-undo-replace.html @@ -23,12 +23,11 @@ if (window.layoutTestController) { var test = document.getElementById('target'); eventSender.mouseMoveTo(target.offsetLeft + target.offsetWidth / 2, target.offsetTop + target.offsetHeight / 2); eventSender.mouseDown(); - eventSender.leapForward(200); eventSender.mouseUp(); eventSender.leapForward(200); + eventSender.mouseDown(); - eventSender.leapForward(200); eventSender.mouseUp(); document.execCommand('InsertText', false, 'a');
Created attachment 99543 [details] Patch
Comment on attachment 99543 [details] Patch Clearing flags on attachment: 99543 Committed r90309: <http://trac.webkit.org/changeset/90309>
All reviewed patches have been landed. Closing bug.
Now the test is hitting an assertion :( http://build.webkit.org/results/GTK%20Linux%2064-bit%20Debug/r90310%20(23936)/editing/selection/directionality-after-undo-replace-stderr.txt
This might be related to https://bugs.webkit.org/show_bug.cgi?id=63440.
It's certainly related to https://bugs.webkit.org/show_bug.cgi?id=63611.
That is awesome!!, We can get to the root cause and have more inspiration to fix this. Fixing this might fix many textiterator read crashes.
This seems okay now on GTK+.