[Fribidi-discuss] Re: BiDi WINE status and fribidi

Behdad Esfahbod behdad at bamdad.org
Fri Aug 23 17:21:05 EST 2002


On Fri, 23 Aug 2002, Shachar Shemesh wrote:

> I would like to dispell some myths that may or may not be surrounding 
> the MS Windows implementation. The myth says that it has a single 
> implementation of anything.
> 
> More answers in body of email.
> 
> >General issue:
> >We use flags (in FriBidiEnv) to tailor the FriBidi algorithm to various
> >situations.  In general, they should be usable in lieu of #ifdef's.
> >However, some people may need a lightweight version of FriBidi, omitting
> >code which supports certain flag settings.
> >Therefore, it may be necessary to accompany each flag with #ifdef around
> >code which implements it.
> >  
> >
> Thing is, these #ifdefs are only relevant when doing UTF-16 processing. 
> Under UTF-32, no surrogates are necessary at all. I don't think 
> compiling fribidi to support UTF-32 and UTF16 SIMULTANIOUSLY is an 
> option, though :-( Please correct me if I'm wrong.

You are right :(.

> >Does this mean that MS-Windows does not follow exactly the Unicode
> >standard when handling the special directional override character codes?
> 
> This doesn't mean a thing. This merely means that, when calling CERTAIN 
> functions, you can specifically say that the third character should be 
> treated as an Hebrew character, regardless of what the Unicode tables say.
> 
> As far as my testings show, Windows DOES support the various Unicode 
> overrides (RLO, RLE, etc.). The problem is that no program will generate 
> them, and so the question is rather mute. For one thing - they are only 
> available if you are using Unicode (and almost all programs are ANSI). 
> Another is that I have seen no interface that allows the Unicode BiDi 
> algorithm to select the base direction. It is always explicetly either 
> RTL or LTR, but never depending on the string printed.

This should be no problem for you, fribidi supports a wider range 
of paragraph directions that windows does, this is not a problem, 
just a feature that you do not use.

> The character override, I believe, is rather a mute point, as I can't 
> see anyone actually using it. It is, however, the deciding factor 
> against using fribidi in Wine. See the thread at 
> http://www.winehq.com/hypermail/wine-devel/2002/08/0201.html, and 
> http://www.winehq.com/hypermail/wine-devel/2002/08/0231.html and the 
> answer at http://www.winehq.com/hypermail/wine-devel/2002/08/0236.html 
> in particular.

We will implement it.

>                 Shachar

-- 
Behdad Esfahbod		2 Shahrivar 1381, 2002 Aug 24 
http://behdad.org/	[Finger for Geek Code]

#define is_persian_leap(y) ((((y)-474)%2820+2820)%2820*31%128<31)





More information about the FriBidi mailing list