Bug 116133 – Printing in 1Password app is broken with screen fonts disabled

• Component: Printing
• Status: RESOLVED FIXED

[Beth Dakin]

When we disabled screen fonts with http://trac.webkit.org/changeset/129593 we regressed printing in the 1Password app.

<rdar://problem/13162981>

[Beth Dakin]

Created attachment 201788 [details]
Patch

[Anders Carlsson]

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.

[Build Bot]

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

[Build Bot]

Comment on attachment 201788 [details]
Patch

Attachment 201788 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/466824

[Beth Dakin]

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

[Ryosuke Niwa]

It seems like this patch broke two tests:
http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r150101%20(9981)/results.html

[Beth Dakin]

(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.

[Ryosuke Niwa]

(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

[Ryosuke Niwa]

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

[Beth Dakin]

(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!

[Ryosuke Niwa]

(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.