Bug 233907
| Summary: | AX: ARIA "form" role should vend AXSubrole: AXLandmarkForm. | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | James Craig <jcraig> |
| Component: | Accessibility | Assignee: | Nobody <webkit-unassigned> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | andresg_22, scottaohara, webkit-bug-importer |
| Priority: | P2 | Keywords: | InRadar |
| Version: | WebKit Nightly Build | ||
| Hardware: | All | ||
| OS: | All | ||
James Craig
AX: ARIA "form" role should vend AXSubrole: AXLandmarkForm. AFAIK, AXRole should still be AXGroup.
Seems like this landmark may have been overlooked when the other landmarks were added. There may be additional Radars needed on the platforms to support AT rotor navigation, for example.
| Attachments | ||
|---|---|---|
| Add attachment proposed patch, testcase, etc. |
Radar WebKit Bug Importer
<rdar://problem/86135614>
James Craig
From: https://github.com/w3c/core-aam/issues/101
James Craig
Update... [AXSubrole: AXLandmarkForm] for role=form but only if the form is labeled, e.g. with aria-label.
Note: There is some ambiguity on whether the landmark mapping should apply to the native <form> element if it has a label like <form title="foo">. Seeking comment from ARIA contributors.
James Craig
one of the related ARIA discussions: https://github.com/w3c/core-aam/issues/100
James Craig
Joanie made the related ATK change here: https://bugs.webkit.org/attachment.cgi?id=323323&action=prettypatch
James Craig
I think it's probably fine to change the subrole for <form title="foo"> too. It's already exposed as a labeled AXGroup, so it'd be highly unlikely to cause an interop regression. The new role description and landmark navigation functionality would be an innocuous improvement.
Unlabeled <div role=form> and <form> should remain unmapped.
Scott
"Unlabeled <div role=form> and <form> should remain unmapped."
I agree with this statement, and per our conversation in the ARIA wg today:
<article role=form> - would fall back to "article"
and
<div role="form article">...</div> would also fall back to article due to the lack of an explicit name.
Scott
sorry, misread some of the above. but also since the naming of a form element with native HTML is only possible with the title attribute, then yes - form title=foo should become a form landmark
James Craig
Thanks Scott.
Rahim Abdi
Pull request: https://github.com/WebKit/WebKit/pull/44316
EWS
Committed 294748@main (a8728728ea0d): <https://commits.webkit.org/294748@main>
Reviewed commits have been landed. Closing PR #44316 and removing active labels.