Bug 129492 – [ATK] Expose missing functionalities of AtkTableCell to AT.

• Component: Accessibility
• Status: RESOLVED FIXED

[Krzysztof Czech]

Implements, get_column_span, get_row_span, get_position, get_table.

[Radar WebKit Bug Importer]

<rdar://problem/16196432>

[Krzysztof Czech]

Created attachment 225646 [details]
patch

[Krzysztof Czech]

Created attachment 225647 [details]
patch

[Krzysztof Czech]

Created attachment 225648 [details]
patch

[Krzysztof Czech]

Created attachment 225649 [details]
patch

[Mario Sanchez Prada]

Comment on attachment 225649 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=225649&action=review

Thanks for doing this. The patch looks good to me in general, but still needs some changes IMHO (see below). Also, table cell is a pretty new thing to ATK so you probably want to use ATK_CHECK_VERSION in some places too.

> Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:49
> +    accessibilityObjectGetOrReturn(cell, axObject, nullptr);
>  
> -    AccessibilityObject* axObject = core(cell);
> -    if (!axObject || !axObject->isTableCell())
> +    if (!axObject->isTableCell())

I'd rather keep the old idiom in place. This new macro is confusing IMHO, specially because of containing an out parameter, and thus decreases readability

> Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:103
> +gboolean webkitAccessibleTableCellGetPosition(AtkTableCell* cell, gint *row, gint* column)

misplaced *

> Source/WebCore/accessibility/atk/WebKitAccessibleUtil.h:70
> +    } while (0)

Because of the particular nature of this macro, you can not place the whole block inside G_STMT_START/END guards which, besides being probably not good enough for places where the compiler expects the macro to expand in only one line, looks a bit awkward, IMO

This should go away if you go back to the old idiom, though

> Tools/WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:213
> +    gint row = -1, column= -1, rowSpan = -1, columnSpan = -1;

Please declare each variable in a separate line and use int instead of gint.

[chris fleizach]

Comment on attachment 225649 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=225649&action=review

> Source/WebCore/ChangeLog:8
> +        Implemented  missing API of AtkTableCell.

too many spaces after Implemented

> Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:134
> +    AtkObject* table = atk_object_get_parent(axObject->wrapper());

should you verify that this is actually the table element before returning?

> Source/WebCore/accessibility/atk/WebKitAccessibleUtil.h:63
> +    AccessibilityObject* axObject = nullptr; \

It's a little non-intuitive that a local variable will be declared inside the macro, that you reference outside the macro.

Maybe it would be better to write a function that does this same check

AccessibilityObject* axObject = verifyAccessibilityObjectIsValid(...);
if (!axObject)
return;

[Krzysztof Czech]

Created attachment 225773 [details]
patch

[Krzysztof Czech]

(In reply to comment #6)
> (From update of attachment 225649 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=225649&action=review
> 
> Thanks for doing this. The patch looks good to me in general, but still needs some changes IMHO (see below). Also, table cell is a pretty new thing to ATK so you probably want to use ATK_CHECK_VERSION in some places too.
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:49
> > +    accessibilityObjectGetOrReturn(cell, axObject, nullptr);
> >  
> > -    AccessibilityObject* axObject = core(cell);
> > -    if (!axObject || !axObject->isTableCell())
> > +    if (!axObject->isTableCell())
> 
> I'd rather keep the old idiom in place. This new macro is confusing IMHO, specially because of containing an out parameter, and thus decreases readability

Done.

> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:103
> > +gboolean webkitAccessibleTableCellGetPosition(AtkTableCell* cell, gint *row, gint* column)
> 
> misplaced *
Done
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleUtil.h:70
> > +    } while (0)
> 
> Because of the particular nature of this macro, you can not place the whole block inside G_STMT_START/END guards which, besides being probably not good enough for places where the compiler expects the macro to expand in only one line, looks a bit awkward, IMO
> 
> This should go away if you go back to the old idiom, though
> 
> > Tools/WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:213
> > +    gint row = -1, column= -1, rowSpan = -1, columnSpan = -1;
> 
> Please declare each variable in a separate line and use int instead of gint.
Done

[Krzysztof Czech]

(In reply to comment #7)
> (From update of attachment 225649 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=225649&action=review
> 
> > Source/WebCore/ChangeLog:8
> > +        Implemented  missing API of AtkTableCell.
Done.
> 
> too many spaces after Implemented
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:134
> > +    AtkObject* table = atk_object_get_parent(axObject->wrapper());
> 
> should you verify that this is actually the table element before returning?
Yes exactly, you are right. Thanks.
> 
> > Source/WebCore/accessibility/atk/WebKitAccessibleUtil.h:63
> > +    AccessibilityObject* axObject = nullptr; \
> 
> It's a little non-intuitive that a local variable will be declared inside the macro, that you reference outside the macro.
> 
> Maybe it would be better to write a function that does this same check
Yes, this sounds as good idea. We already have similar function that returns AccessibilityObject for respective AtkObject and it's called webkitAccessibleGetAccessibilityObject. I think it can be slightly extended for adding some additional checks. I think it should be modified in a separate patch because it touches many files that are not related to this changes.
> 
> AccessibilityObject* axObject = verifyAccessibilityObjectIsValid(...);
> if (!axObject)
> return;
Thanks

[Mario Sanchez Prada]

Comment on attachment 225773 [details]
patch

Looks good to me now. Thanks!

[WebKit Commit Bot]

Comment on attachment 225773 [details]
patch

Clearing flags on attachment: 225773

Committed r165107: <http://trac.webkit.org/changeset/165107>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.