[Webkit-unassigned] [Bug 262751] New: AX: WebKit adds additional space to inline elements when computing the label of an ancestor

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=262751

            Bug ID: 262751
           Summary: AX: WebKit adds additional space to inline elements
                    when computing the label of an ancestor
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: All
                OS: All
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: Accessibility
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: jcraig at apple.com
                CC: andresg_22 at apple.com,
                    webkit-bug-importer at group.apple.com

# Background:

As part of https://github.com/web-platform-tests/wpt/pull/42354 I noticed WebKit behaved differently than Gecko and Chromium. See last three failures https://wpt.fyi/results/accname/name/comp_name_from_content.html?diff&filter=ADC&run_id=6197887781830656&run_id=6282535614545920 Search for "button name from content for each child (no space, inline)" if you're unsure which tests...

WebKit is technically the only one to implement that portion of the AccName spec as written today, so I raised this issuse for the ARIA WG. https://github.com/w3c/accname/issues/205

WG Resolution was that the spec is wrong, and will be shifting to match the behavior in Gecko and Chromium. I generally agree with the decision and think WebKit should also shift in anticipation of this spec change, and to match the other implementations. 

The PR hasn't landed yet, but when it does, those tests will be the last three in: https://github.com/web-platform-tests/wpt/blob/master/accname/name/comp_name_from_content.html


# Change Needed:

When a role supporting "name from content" has `display:inline` descendants with no inter-element whitespace, the AccName algorithm should not add the additional whitespace joiner.

<h1>simple w/ for each child (w/o spaces and display:inline)</h1>
<button data-expectedlabel="onetwothree"><span>one</span><span>two</span><span>three</span></button><br>
<h3 data-expectedlabel="onetwothree"><span>one</span><span>two</span><span>three</span></h3>
<a href="#" data-expectedlabel="onetwothree"><span>one</span><span>two</span><span>three</span></a><br>
<br>

But make sure we don't regress the tests that immediately follow that one, because those elements don't use an inline display.

<h1>simple w/ for each child (w/o spaces and display:block)</h1>
<button data-expectedlabel="one two three" style="display:block;"><span>one</span><span>two</span><span>three</span></button><br>
<h3 data-expectedlabel="one two three" style="display:block;"><span>one</span><span>two</span><span>three</span></h3>
<a href="#" data-expectedlabel="one two three" style="display:block;"><span>one</span><span>two</span><span>three</span></a><br>
<br>

<h1>simple w/ for each child (w/o spaces and display:inline block)</h1>
<button data-expectedlabel="one two three" style="display:inline-block;"><span>one</span><span>two</span><span>three</span></button><br>
<h3 data-expectedlabel="one two three" style="display:inline-block;"><span>one</span><span>two</span><span>three</span></h3>
<a href="#" data-expectedlabel="one two three" style="display:inline-block;"><span>one</span><span>two</span><span>three</span></a><br>
<br>

-- 
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/20231006/5f693517/attachment-0001.htm>