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

[bugzilla-daemon at webkit.org]

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





--- Comment #13 from Krzysztof Czech <k.czech at samsung.com>  2014-01-23 08:07:46 PST ---
> > > > Source/WebCore/accessibility/AccessibilityNodeObject.h:185
> > > > +    virtual void elementsFromAttribute(Vector<Element*>& elements, const QualifiedName&) const override;
> > > 
> > > I think we normally write OVERRIDE (uppercase letters). Also, why do you need to make this method virtual if the implementation in the parent is going to be an empty one?
> > 
> > Regarding override instead of OVERRIDE, this is after r162139
> 
> Forget about this. It's 'override' from now on, not OVERRIDE (see webkit-dev)
> 
> > I always thought that AccessibilityObject class is kind of an abstract class that provides an interface for AccessibilityNodeObject and for AccessibilityRenderObject so that virtual callings may happen. This is a reason this method has an empty body.
> > 
> > AccessibilityRenderObject derives the whole interface from AO and ANO and it will have this method out of the box.
> > 
> > Other thing is that ANO is something between AO and ARO and having body of this method there might look more natural.
> > > 
> > > I see the point of being able to call it from a AccessibilityObject, but I'm not sure that alone justifies doing this, unless it made sense to move the whole method (implementation too) up to AccessibilityObject
> > 
> > I realized this might be a good idea of moving its body directly to AccessibilityObject.
> > 
> Ok. I buy it :)
> 

I moved this method to AccessibilityObject and I think it does not have to be virtual either. It provides a common interface for both ANO and ARO and I doubt it could ever be customized by those classes.

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