Bug 177520 – Deferred image size change makes image-load-on-delay.html flaky.

• Component: Accessibility
• Status: RESOLVED FIXED

[zalan]

Related to bug 177446

[Radar WebKit Bug Importer]

<rdar://problem/34674299>

[zalan]

Created attachment 322036 [details]
Patch

[chris fleizach]

Comment on attachment 322036 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=322036&action=review

> Source/WebCore/accessibility/AXObjectCache.cpp:2803
> +        m_deferredRecomputeIsIgnoredList.add(element);

is this a set or array? do we need to worry about duplicates being added?

> Source/WebCore/accessibility/AXObjectCache.cpp:2805
> +    }

we could save some code if we did

if (rendererNeedsDeferredUpdate(*renderer)) {
       m_deferredRecomputeIsIgnoredList.add(element);
else
   recomputeIsIgnored(renderer);

> Source/WebCore/rendering/RenderImage.cpp:269
> +    if (AXObjectCache* cache = document().existingAXObjectCache())

are you sure you want to lose the accessibilityEnabled() check?

[zalan]

(In reply to chris fleizach from comment #3)
> Comment on attachment 322036 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=322036&action=review
> 
> > Source/WebCore/accessibility/AXObjectCache.cpp:2803
> > +        m_deferredRecomputeIsIgnoredList.add(element);
> 
> is this a set or array? do we need to worry about duplicates being added?
It's a set. Dupes are fine.

> 
> > Source/WebCore/accessibility/AXObjectCache.cpp:2805
> > +    }
> 
> we could save some code if we did
> 
> if (rendererNeedsDeferredUpdate(*renderer)) {
>        m_deferredRecomputeIsIgnoredList.add(element);
> else
>    recomputeIsIgnored(renderer);
>
I don't mind doing that but in WebCore we prefer early returns (unless it has changed).


> > Source/WebCore/rendering/RenderImage.cpp:269
> > +    if (AXObjectCache* cache = document().existingAXObjectCache())
> 
> are you sure you want to lose the accessibilityEnabled() check?
Not sure how much that helps and if you look at all the call sites, they don't really check for that (not even the hot ones).

[zalan]

Created attachment 322039 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 322039 [details]
Patch

Clearing flags on attachment: 322039

Committed r222594: <http://trac.webkit.org/changeset/222594>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.