Bug 62718 – [GTK] Consider rows being ignored when adding children to tables

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

[Mario Sanchez Prada]

After bug 62235 was fixed, the accessibility/aria-tables.html test started to fail in GTK because the applied patch assumes that accessibility is not being ignored for rows, which is not the case for GTK at the moment.

Hence we need to fix this for GTK in order to be able to make the test pass again, skipping it in the meanwhile.

Note: See comment 12 in bug 62335 for more details (https://bugs.webkit.org/show_bug.cgi?id=62335#c12)

[Mario Sanchez Prada]

Created attachment 97278 [details]
Patch proposal + unskipped layout test

Attaching patch according to proposal explained in https://bugs.webkit.org/show_bug.cgi?id=62335#c12

[Mario Sanchez Prada]

Created attachment 98358 [details]
Patch proposal + unskipped layout test

Attaching patch using a different approach to hide rows in GTK, as discussed by mail with Chris

[chris fleizach]

Comment on attachment 98358 [details]
Patch proposal + unskipped layout test

View in context: https://bugs.webkit.org/attachment.cgi?id=98358&action=review

r=me with suggested fixes

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:309
> +

looks like you could put this after the !coreParent check, then you wouldn't need to check coreParent in this if statement

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:319
> +    unsigned tableChildrenCount = tableChildren.size();

i think this unsigned should be a size_t, along with cellsCount (size vector.size() returns size_t i believe)

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:323
> +    for (unsigned i = 0; i < tableChildrenCount; ++i) {

size_t for the i as well

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:352
> +

size_t change

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:358
> +            if (static_cast<unsigned>(index) < cellsCount + rowChildrenCount)

you could probably static_cast<index> one time before the loop, so it doesn't have to be done everytime within the loop

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:403
> +    AccessibilityObject* grandParent = parent->parentObjectUnignored();

you should probably verify parent != 0 before getting gradparent

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:410
> +

size_t

> Source/WebCore/accessibility/gtk/AccessibilityObjectWrapperAtk.cpp:418
> +

size_t

[Mario Sanchez Prada]

(In reply to comment #3)
> (From update of attachment 98358 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=98358&action=review
> 
> r=me with suggested fixes
>
Committed after making those changes:

http://trac.webkit.org/changeset/89660

Thanks!