[Webkit-unassigned] [Bug 121684] [ATK] Expose aria-describedby with ATK_RELATION_DESCRIBED_BY

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Jan 28 08:54:21 PST 2014


https://bugs.webkit.org/show_bug.cgi?id=121684





--- Comment #18 from Krzysztof Czech <k.czech at samsung.com>  2014-01-28 08:51:44 PST ---
> >> LayoutTests/accessibility/aria-describedby-on-input-expected.txt:3
> >> +which the computer will self-destruct.
> > 
> > I still don't quite understand why the line break shows up here. IMHO, the text from the two divs should be concatenated together and joined with a " ", not a line break.
> > 
> > Chris, could you comment on this issue? Maybe it's me misunderstanding how this aria-describedby thing should work, or maybe it's another bug somewhere else
> 
> yea i'm also a little surprised that ariaDescribedByAttribute() would somehow get the newline
> It would indicate that
> static String accessibleNameForNode(Node* node) -> textUnderElement()
>   returns a string in the first element that has a newline
> that surprises me. i don't think it's a show stopper and I don't think this patch is the cause of the problem, but it's something we should at least understand
Thanks this is something that I did not know. I thought that html is not obligated to take into account aria* attributes. I guess additional bug could be created to dig a bit more into this.
> 
> > Source/WebCore/accessibility/AccessibilityObject.cpp:2141
> > +    for (unsigned i = 0; i < size; ++i) {
> 
> we should use 
> for (auto name : idVector)
Sounds good.
> 
> > Source/WebCore/accessibility/AccessibilityObject.cpp:2144
> > +        if (idElement)
> 
> this can be written
> if (Element* idElement = treeScope...)
>    elements.append()
Sounds good, thanks
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleWrapperAtk.cpp:253
> > +            removeAtkRelationByType(relationSet, ATK_RELATION_DESCRIBED_BY);
> 
> There's a ariaDescribedByAttribute() on the AXObject. We should make that public so you can call that directly, instead of calling the attribute directly on the Node object in the platform code

How about to create something like ?: 
supportsAriaDescribedBy() { return !getAttribute(aria_describedbyAttr).isEmpty(); }

I looked into the code and I see ariaDescribedByAttribute calls elementsFromAttribute. This method would be called twice in this context. 
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleWrapperAtk.cpp:258
> > +                if (!axObject)
> 
> this if can be combined with the line above
Sounds good, thanks.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.


More information about the webkit-unassigned mailing list