[Webkit-unassigned] [Bug 83227] [chromium] wrong justification for arabic/persian page in cr-win

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=83227





--- Comment #12 from Kenichi Ishibashi <bashi at chromium.org>  2012-04-05 01:44:28 PST ---
(From update of attachment 135716)
View in context: https://bugs.webkit.org/attachment.cgi?id=135716&action=review

>>> Source/WebCore/platform/graphics/chromium/UniscribeHelper.cpp:207
>>> +            int characterIndex = m_runs[run].iCharPos + i;
>> 
>> run -> runIndex?
> 
> I think it should be 'run' as m_runs and run are logical, while runIndex and shaping is visual.

Ok. Let me make sure if I understood right. Suppose there are two rtl runs like:

ab|cde

I think values of related variables are:
m_runs[0].iCharPos = 0, m_shapes[0].charLength() = 2
m_runs[1].iCharPos = 2, m_shapes[1].charLength() = 3
m_screenOrder = [1,0]

In this case, the first loop will access the first three characters of m_input. Is this intentional? or is my understanding wrong?
I glanced UniscribeHelper::draw() and UniscribeHelper::xToCharacter() and could see m_runs and m_shapes are accessed via the same index parameter.

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