Bug 120373 – AX: aria-relevant does not expose AXARIARelevant

• Status: RESOLVED
• Resolution: FIXED
• Component: Accessibility

[Jon Gunderson]

In ARIA testcase 509 aria-relavent is not being exposed, the aria-live property should cancel the aria-presentation property
https://www.w3.org/WAI/PF/testharness/testcases/edit?testsuite_id=1&testcase_id=509 

This is an important feature for ARIA 1.0 specification to go to recommendation

[Radar WebKit Bug Importer]

<rdar://problem/14848518>

[James Craig]

Public test case: Previous link is to the private test harness.
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/presentation-role/509.html

[James Craig]

From the dupe:

Here are the test cases that fail:(radio, checkbox)
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/754.html
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/755.html
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/759.html
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/760.html

Here are some of the test cases that pass (text, file, button):
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/753.html
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/756.html
https://dvcs.w3.org/hg/pfwg/raw-file/default/ARIA/1.0/tests/test-files/name-computation-input/757.html

[James Craig]

Ignore that last comment. It was intended for bug 116570.

[chris fleizach]

Created attachment 210622 [details]
patch

[Darin Adler]

Comment on attachment 210622 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=210622&action=review

> Source/WebCore/accessibility/AccessibilityObject.cpp:1562
> +        || hasAttribute(aria_atomicAttr)
> +        || hasAttribute(aria_busyAttr)
> +        || hasAttribute(aria_controlsAttr)
> +        || hasAttribute(aria_describedbyAttr)
> +        || hasAttribute(aria_disabledAttr)
> +        || hasAttribute(aria_haspopupAttr)
> +        || hasAttribute(aria_invalidAttr)
> +        || hasAttribute(aria_labelAttr)
> +        || hasAttribute(aria_labelledbyAttr)
> +        || hasAttribute(aria_relevantAttr);

Eventually this is going to get really slow. Every one of these is a separate hash table lookup.

[chris fleizach]

(In reply to comment #6)
> (From update of attachment 210622 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=210622&action=review
> 
> > Source/WebCore/accessibility/AccessibilityObject.cpp:1562
> > +        || hasAttribute(aria_atomicAttr)
> > +        || hasAttribute(aria_busyAttr)
> > +        || hasAttribute(aria_controlsAttr)
> > +        || hasAttribute(aria_describedbyAttr)
> > +        || hasAttribute(aria_disabledAttr)
> > +        || hasAttribute(aria_haspopupAttr)
> > +        || hasAttribute(aria_invalidAttr)
> > +        || hasAttribute(aria_labelAttr)
> > +        || hasAttribute(aria_labelledbyAttr)
> > +        || hasAttribute(aria_relevantAttr);
> 
> Eventually this is going to get really slow. Every one of these is a separate hash table lookup.

Do you think we should cache this result? Or do you know of another way to check for the presence of a list of attributes?

Thanks

[WebKit Commit Bot]

Comment on attachment 210622 [details]
patch

Clearing flags on attachment: 210622

Committed r155191: <http://trac.webkit.org/changeset/155191>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.