Bug 169810 – AX: Expose a new AXSubrole for explicit ARIA "group" role

• Component: Accessibility
• Status: RESOLVED FIXED

[James Craig]

Currently WebKit exposes no difference between explicit groups (ARIA role="group" and ARIA subroles like doc-biblioentry), generic blocks (<div>, etc), generic inlines that happen to be styled as blocks (<span style=“display:block”), and pseudo-elements (::before).

[James Craig]

<rdar://problem/31039693>

[James Craig]

Created attachment 304768 [details]
patch

[James Craig]

Created attachment 304894 [details]
patch

[James Craig]

Created attachment 304895 [details]
patch

[James Craig]

Created attachment 304917 [details]
patch

Updating patch to TOT to potentially fix GTK apply failure.

[James Craig]

Comment on attachment 304917 [details]
patch

Additional info: The fix allows the WebKit Accessibility Inspector to return more accurate role descriptions to distinguish ARIA groups from generic block-level elements in the render tree. (See updated tests.) The fix also provides more differentiation between these groups to each platform accessibility API, providing a way for assistive technologies to convey the differences to the user if desired.

[chris fleizach]

Comment on attachment 304917 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=304917&action=review

> Source/WebKit/win/ChangeLog:9
> +        Account for Windows platform role mapping.

did you modify GTK files to map this new role

> LayoutTests/accessibility/list-detection2-expected.txt:-20
> -PASS: ul w/ background image (NOT A LIST) -> group. 

these objects are no longer groups? that seems unexpected

[James Craig]

Comment on attachment 304917 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=304917&action=review

>> Source/WebKit/win/ChangeLog:9
>> +        Account for Windows platform role mapping.
> 
> did you modify GTK files to map this new role

Looking into that.

>> LayoutTests/accessibility/list-detection2-expected.txt:-20
>> -PASS: ul w/ background image (NOT A LIST) -> group. 
> 
> these objects are no longer groups? that seems unexpected

This is expected. Those elements are still mapped to "AXGroup" (with no subrole) but no longer return "group" as the computed ARIA role string. Layout lists match neither role="list" nor role="group", so web inspector users can now see these have "No exact ARIA role match" rather than the misleading "group" role that previously appeared.

I think a future version of ARIA may include a "generic" or some equivalent role for non-semantic container elements like <div> and other elements heuristically determined to be mere layout items. If/when it does, <div>s and layout lists, would vend that role.

[James Craig]

Created attachment 304945 [details]
patch with Atk mapping

[WebKit Commit Bot]

Comment on attachment 304945 [details]
patch with Atk mapping

Rejecting attachment 304945 [details] from commit-queue.

Failed to run "['/Volumes/Data/EWS/WebKit/Tools/Scripts/webkit-patch', '--status-host=webkit-queues.webkit.org', '--bot-id=webkit-cq-02', 'land-attachment', '--force-clean', '--non-interactive', '--parent-command=commit-queue', 304945, '--port=mac']" exit_code: 2 cwd: /Volumes/Data/EWS/WebKit

Last 500 characters of output:
LayoutTests
:040000 040000 5e136b7a9332a5b00b05c6663fb7515d390b1967 7d6fb45c28996bcf8ba5c4bff9fa0a21c5e2ede9 M	Source
Current branch master is up to date.
ERROR: Not all changes have been committed into SVN, however the committed
ones (if any) seem to be successfully integrated into the working tree.
Please see the above messages for details.


Failed to run "['git', 'svn', 'dcommit', '--rmdir']" exit_code: 1 cwd: /Volumes/Data/EWS/WebKit
Updating OpenSource
Current branch master is up to date.

Full output: http://webkit-queues.webkit.org/results/3388201

[Joanmarie Diggs]

Is the plan to re-CQ+ this?

[James Craig]

(In reply to Joanmarie Diggs (irc: joanie) from comment #11)
> Is the plan to re-CQ+ this?

Yeah, I filed a WebKit Tools bug in the meantime to diagnose the ambiguous failure, but I'll update to TOT and re-CQ soon.

[James Craig]

Others with the same failure after the site migration.

https://bugs.webkit.org/buglist.cgi?chfieldfrom=3w&chfieldto=Now&list_id=2621140&longdesc=Not%20all%20changes%20have%20been%20committed%20into%20SVN&longdesc_type=allwordssubstr&query_format=advanced

[James Craig]

Comment on attachment 304945 [details]
patch with Atk mapping

First attempting another CQ+ since the failure appears to be transient.

[WebKit Commit Bot]

Comment on attachment 304945 [details]
patch with Atk mapping

Clearing flags on attachment: 304945

Committed r214623: <http://trac.webkit.org/changeset/214623>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.