Bug 231316 – Regression (r281291): [ iOS 15 ] fast/text/accessibility-bold-system-font-2.html is image failing

• Status: NEW
• Component: Text

fast/text/accessibility-bold-system-font-2.html

Is constantly image failing on iOS 15 Debug and flaky (very flaky) image failing on iOS 15 Release.

History: https://results.webkit.org/?suite=layout-tests&test=fast%2Ftext%2Faccessibility-bold-system-font-2.html

According to the history, it started failing after updating to iOS 15.

Result page: https://build.webkit.org/results/Apple-iOS-15-Simulator-Debug-WK2-Tests/r283612%20(72)/results.html

Created attachment 440398 [details]
image diff

[Radar WebKit Bug Importer]

<rdar://problem/83944188>

Marked test expectations https://trac.webkit.org/changeset/283642/webkit

I was able to reproduce the image failure locally on iOS 15 sim using run-webkit-tests --force --debug --ios-simulator fast/text/accessibility-bold-system-font-2.html

As I was bisecting, the test failed on r281291 and newer but passed on r281290 and older.

[Myles C. Maxfield]

Created attachment 457510 [details]
Patch

[Darin Adler]

Comment on attachment 457510 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=457510&action=review

> Source/WebCore/platform/graphics/cocoa/FontCacheCoreText.cpp:392
> -        auto& after = keyframes[i];
> +        auto& after = keyframes[i + 1];

Great fix, but not mentioned in the change log. Do we have test coverage?

> Source/WebCore/platform/graphics/cocoa/FontCacheCoreText.cpp:417
> -        auto& after = keyframes[i];
> +        auto& after = keyframes[i + 1];

Ditto.

[Myles C. Maxfield]

Comment on attachment 457510 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=457510&action=review

>> Source/WebCore/platform/graphics/cocoa/FontCacheCoreText.cpp:392
>> +        auto& after = keyframes[i + 1];
> 
> Great fix, but not mentioned in the change log. Do we have test coverage?

This patch is not marked for review 🤔

We do; I’m handling it in https://bugs.webkit.org/show_bug.cgi?id=239374. That bug has a big chain of blocking bugs because I’m trying to figure out a performance regression, and I’m lumping in the fix in with that effort.

[Brent Fulgham]

This actually seems to be tracked by:
rdar://88249241.