Fix for LayoutTests/accessibility/mac/search-text/search-text.html in IsolatedTree mode.

[Andres Gonzalez]

Fix for LayoutTests/accessibility/mac/search-text/search-text.html in IsolatedTree mode.

[Andres Gonzalez]

Created attachment 392111 [details]
Patch

[chris fleizach]

Comment on attachment 392111 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=392111&action=review

> Source/WebCore/accessibility/AccessibilityObject.h:759
> +    AccessibilityObject* webAreaObject() const override { return nullptr; }

the type here is AccessibilityObject, but in the interface it's AXCoreObject

> Source/WebCore/accessibility/AccessibilityScrollView.h:44
> +    AccessibilityObject* webAreaObject() const override;

interface class defines this as AXCoreObject

> Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp:229
> +            && item.m_wrapper)

this line can go on the last one. it looks like it's short enough

[Andres Gonzalez]

Created attachment 392134 [details]
Patch

[Andres Gonzalez]

(In reply to chris fleizach from comment #2)
> Comment on attachment 392111 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=392111&action=review
> 
> > Source/WebCore/accessibility/AccessibilityObject.h:759
> > +    AccessibilityObject* webAreaObject() const override { return nullptr; }
> 
> the type here is AccessibilityObject, but in the interface it's AXCoreObject

It is a long chain to change this since callers are assigning the return value to AccessibilityObject. It works because the override is returning a subclass. There are other methods whose overrides also return a subclass. I propose we leave it like this for now, and make a complete pass through to clean them up all at once in a separate patch.
> 
> > Source/WebCore/accessibility/AccessibilityScrollView.h:44
> > +    AccessibilityObject* webAreaObject() const override;
> 
> interface class defines this as AXCoreObject

Same comment as above.
> 
> > Source/WebCore/accessibility/isolatedtree/AXIsolatedTree.cpp:229
> > +            && item.m_wrapper)
> 
> this line can go on the last one. it looks like it's short enough

Done.

[Andres Gonzalez]

Created attachment 392180 [details]
Patch

[Andres Gonzalez]

(In reply to Andres Gonzalez from comment #5)
> Created attachment 392180 [details]
> Patch

Fix for GTK build.

[WebKit Commit Bot]

Comment on attachment 392180 [details]
Patch

Clearing flags on attachment: 392180

Committed r257739: <https://trac.webkit.org/changeset/257739>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Radar WebKit Bug Importer]

<rdar://problem/59969174>

[Ryan Haddad]

This change appears to have introduced layout test crashes, as indicated by EWS: https://ews-build.webkit.org/#/builders/31/builds/4027

See: https://bugs.webkit.org/show_bug.cgi?id=208488