[Webkit-unassigned] [Bug 52952] Focus ring for anchor with inline image is incorrect

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=52952





--- Comment #36 from ChangSeok Oh <kevin.cs.oh at gmail.com>  2011-04-15 21:33:01 PST ---
(From update of attachment 89225)
View in context: https://bugs.webkit.org/attachment.cgi?id=89225&action=review

Thanks for review again. :)

>> Source/WebCore/rendering/InlineFlowBox.cpp:109
>>      }
> 
> Style nit:  no { } on single line blocks.  I'm surprised the style queue didn't flag this.

My mistake. Done.

>> Source/WebCore/rendering/InlineFlowBox.h:293
>> +    bool m_hasTextChildrenOnly : 1;
> 
> I probably would have named this m_hasOnlyTextChildren.
> 
> So this is only children?  Or is this true of all dependents?  Do we keep this up to date when grafting on sub-trees?  I'm not sure what linebox tree manipulations are allowed that we'd need to worry about.

Done.

m_hasTextChildrenOnly just hold a condition whether InlineFlowBox has only text type InlineBox child.
In my understanding, current paintOutline implementation for css:outline has taken care of only text string. Because of InlineFlowBox's logicalTop.
When the top of outline is decided, the logical top of InlineFlowBox is refered.
Although RootInlineBox has proper top position, it could not be bigger than top of text line.

I tested another case for checking m_hasOnlyTextChildren is kept up to date.
Yes, it is. 
I added a new image child to parentAnchor node using by setTimeout JS function.
The outline was expanded successfully like FF and Opera. Why it does is that m_hasOnlyTextChildren is kept up to date in InlineFlowBox::addToLine when a new InlineBox is added to InlineFlowBox.

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