Behdad Esfahbod wrote:
>On Sat, 15 Jun 2002, Shachar Shemesh wrote:
>
>
>
>>Not according to the character types it announces, and not according to
>>some people's (Mati from IBM, for one) experience. I have positive
>>confirmation that the character types they are publishing are 2.0
>>complient (even though the standard clearly suggests that 1.0 is the
>>standard they are implementing).
>>
>>Whether we want to allow an algorythm more correct than the one Windows
>>is using is an open question at this stage. Like I said before, I am
>>leaning torwards implementing that as an option.
>>
>> Shachar
>>
>>
>
>Ok, but what components do you need? We have enough bidi
>implementations, and I can back-port fribidi for your needs too.
>Also from other projects...
>
>
>
What I was thinking of doing was to implement either the 1.0 or the 2.0
algorythm inline (unless you point me to an already existing
implementation), and if fribidi is available at compile time, allow a
run time option that will use it instead. This way, the default mode is
compatible, but someone who hates the Windows behaviour can get an
upgrade mode. I am not currently aware of any 1.0 or 2.0 implementation
though.
Using a customized UTF-16 fribidi is an interesting idea, but I don't
think it is worth the hassle. Why did you do it UTF-32 to begin with,
btw? None of the bidi languages need two UTF-16 chars.
Shachar
=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]
