[PATCH] [Bug 91955] Bidi algorithm: incorrect implementation of rule N2

[Stephan Kulow]

Am Wednesday 03 November 2004 15:53 schrieb Mitz Pettel:
> Stephan,
> 
> Thanks for your reply.
> 
> On Nov 3, 2004, at 3:56 PM, Stephan Kulow wrote:
> 
> > Your new patch doesn't break anything visible, but it got some not so
> > nice side effects in the regression tests in breaking tons of text 
> > runs into.
> 
> Sorry, into what? Can you give a specific example?
As I said: it's not visible. But in the render tree I get diffs like 

             RenderText {text} at (0,0) size 75x16 text inline minMaxKnown
-               text run at (0,0) width 75: "hapl007.html"
+               text run at (0,0) width 45: "hapl007"
+               text run at (45,0) width 30: ".html"

This has to be reviewed manually for corretness.

Greetings, Stephan