In bug 104945 we saw flakiness in one of my mouse cursor tests as a result of a typo in one of the two copies of a URL. It sucks that accidentally missing or mistyping one of the URLs in the preload list leads to random flakiness instead of a deterministic failure. I need to come up with some way to fix this - ideally by having only a single list of URLs that need to be preloaded.
Created attachment 180520 [details] Patch
Comment on attachment 180520 [details] Patch I think this approach is a better trade-off between simplicity and brittleness. Eric, what do you think?
Comment on attachment 180520 [details] Patch Attachment 180520 [details] did not pass mac-ews (mac): Output: http://queues.webkit.org/results/15454435 New failing tests: fast/events/mouse-cursor-multiframecur.html
Created attachment 180546 [details] Fix mac failure
Ping - eric, do you want to review this?
Comment on attachment 180546 [details] Fix mac failure Very cool.
Comment on attachment 180546 [details] Fix mac failure Rejecting attachment 180546 [details] from commit-queue. Failed to run "[u'/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', u'--status-host=queues.webkit.org', ..." exit_code: 1 cwd: /mnt/git/webkit-commit-queue /mnt/git/webkit-commit-queue/LayoutTests/ChangeLog neither lists a valid reviewer nor contains the string "Unreviewed" or "Rubber stamp" (case insensitive). Full output: http://queues.webkit.org/results/15700236
Created attachment 181368 [details] Patch for landing
Comment on attachment 181368 [details] Patch for landing Clearing flags on attachment: 181368 Committed r138857: <http://trac.webkit.org/changeset/138857>
All reviewed patches have been landed. Closing bug.