When we disabled screen fonts with http://trac.webkit.org/changeset/129593 we regressed printing in the 1Password app. <rdar://problem/13162981>
Created attachment 201788 [details] Patch
Comment on attachment 201788 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=201788&action=review > Source/WebCore/platform/mac/WebCoreSystemInterface.mm:213 > +#if __MAC_OS_X_VERSION_MIN_REQUIRED >= 1090 > +bool (*wkExecutableWasLinkedOnOrBeforeMountainLion)(void); > +#endif Doesn't look like this is called from WebCore, you can remove it.
Comment on attachment 201788 [details] Patch Attachment 201788 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.appspot.com/results/466821
Comment on attachment 201788 [details] Patch Attachment 201788 [details] did not pass mac-ews (mac): Output: http://webkit-queues.appspot.com/results/466824
Thank you! I removed the WebCore and WK2 code, and I believe I fixed the build problems as well: http://trac.webkit.org/changeset/150101
It seems like this patch broke two tests: http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r150101%20(9981)/results.html
(In reply to comment #6) > It seems like this patch broke two tests: > http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r150101%20(9981)/results.html I rolled my change out of my tree locally and was still able to reproduce these two test failures, so I am not sure that this change is the one responsible.
(In reply to comment #7) > (In reply to comment #6) > > It seems like this patch broke two tests: > > http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r150101%20(9981)/results.html > > I rolled my change out of my tree locally and was still able to reproduce these two test failures, so I am not sure that this change is the one responsible. There is exactly one changeset in the blame list: Pass: http://build.webkit.org/builders/Apple%20MountainLion%20Release%20WK1%20%28Tests%29/builds/9980 Fail: http://build.webkit.org/builders/Apple%20MountainLion%20Release%20WK1%20%28Tests%29/builds/9981
It also seems quite unlikely that 7 builders happen to observe the same flakiness: http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=platform%2Fmac%2Ffast%2Ftext%2Fthai-combining-mark-positioning.html
(In reply to comment #9) > It also seems quite unlikely that 7 builders happen to observe the same flakiness: > http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=platform%2Fmac%2Ffast%2Ftext%2Fthai-combining-mark-positioning.html Ugh. It is hard to deny that evidence! Unfortunately I haven't made any more progress on this though. I reverted my whole tree to the revision before my change and did a clean build; I was still able to reproduce the failure. On top of that, the failure is subtle, but you can clearly see it if you line up the Safari windows right. I am able to reproduce the failure with Nightly builds from before my change even though the bots didn't start seeing it until last night. I wish I had an explanation!
(In reply to comment #10) > > Unfortunately I haven't made any more progress on this though. I reverted my whole tree to the revision before my change and did a clean build; I was still able to reproduce the failure. On top of that, the failure is subtle, but you can clearly see it if you line up the Safari windows right. I am able to reproduce the failure with Nightly builds from before my change even though the bots didn't start seeing it until last night. I wish I had an explanation! Maybe builders got the OS update at the same time? It seems like we should just rebaseline the tests unless new results are somehow worse than the old ones.