[Webkit-unassigned] [Bug 255159] New: text-decoration-skip-ink behaves inconsistently even when specifying "all"

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=255159

            Bug ID: 255159
           Summary: text-decoration-skip-ink behaves inconsistently even
                    when specifying "all"
           Product: WebKit
           Version: Safari 16
          Hardware: All
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: CSS
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: nick at nickheer.com

Created attachment 465811

  --> https://bugs.webkit.org/attachment.cgi?id=465811&action=review

Screenshot showing expected behaviour (in green) and incorrect behaviour (in red) on same page

When using text-decoration-skip-ink:auto or :all, the text-decoration sometimes still clashes with descenders. 

It is an inconsistent bug. Sometimes, I see both the correct and incorrect behaviours within the same page, as in the attached screenshot (website is my own, https://pxlnv.com). In this case, bold text appears to be composited correctly but normal weight text is not.

I assumed this was due to the font-weight or size, but it does not appear to be the case. Playing with this CodePen (https://codepen.io/andyadams/pen/YMKqrN) indicates text-decoration-skip-ink will behave as expected in Safari even down to very small font sizes of about 9px.

Furthermore, upon refresh, even the bold text shown in the attached screenshot will sometimes composite unexpectedly.

I have tried text-decoration-skip-ink:auto and text-decoration-skip-ink:all. Both are inconsistent. I have been able to reproduce this in Safari Version 16.4 (18615.1.26.11.22) and Safari Technology Preview Release 167 (Safari 16.4, WebKit 18616.1.8.2) on MacOS 13.3 (22E252), and in Safari up to iOS 16.5 (20F5028e).

For comparison, text-decoration-skip-link behaves expectedly in Chrome running on the same MacOS system.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230407/a0043f43/attachment.htm>