Re: Re: Arabic support

Hello Eli,

Many thanks for your quick and detailled feedback. Any hints on how we 
could go about to figure out which of the factors below is the culprit?
(e.g. check which libraries we linked against,...)

Regards,    Martin.

On 2010/09/02 18:31, Eli Zaretskii wrote:
>> From: 大嶋 俊祐<[email protected]>
>> Date: Thu, 2 Sep 2010 16:45:02 +0900
>> Cc: [email protected], [email protected], [email protected]
>>
>>   This appended picture shows a test of bidi-display-reordering. The upper case is the case that bidi-display-reordering is nil. The lower case is non-nil. The problem is that the characters are ordered left to right even they should been right to left when bidi-display-reordering is non-nil. This test is in Emacs Ver. 24.0.50.1 (i686-pc-cygwin) on Microsoft Windows XP Professional Version 2002 Service Pack 3.
>
> Thanks.  This is a Cygwin build, which is supposed to use (a port of)
> libotf.  It isn't supposed to use Uniscribe directly.
>
> So this particular problem could be the result of one or more of the
> following factors:
>
>   . Some problem in the build process (I think Handa-san mentioned some
>     special actions for a proper build, like link against specific
>     libraries).
>
>   . Some bug in the ported libotf, which screws up bidirectional
>     display when the text is already reordered.
>
>   . Some bug in Emacs related to display bidirectional text.
>
> The last one seems extremely unlikely, since AFAIU it works for
> Handa-san on GNU/Linux and for me on MS-Windows, with Hebrew text.
>
> To summarize, I don't think this problem is the same one as seen in
> the native Windows build.
>
>

-- 
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:[email protected]