Around 21 o'clock on Nov 4, "Maarten L. Hekkelman" wrote:

> I'm writing a unicode text editor that does its own characterset
> conversions when necessary. Now I'm running into the problem that there are
> only a few fonts available for unicode. Why is it that e.g. the truetype
> nfonts are not available in the iso10646 encoding? Take cyberbit which has
> quite a few encodings in one font, why isn't it possible to access this
> font with the unicode encoding?

You should be able to use Xft with client-side fonts today, and with
server-side fonts in the near future.  Xft provides a Unicode interface for
text output, and will (soon, I hope) support conversion for server-side 
fonts.

Providing server-side fonts in 10646 encoding is very expensive as the 
server must rasterize the entire font and deliver all of the metrics as 
soon as the font is opened; Xft will hide this behind an API that converts 
your Unicode chars into whatever encoding the font uses automatically.

> Please don't tell me to use utf8 locales and fontsets, I tried it but fontsets
> are slow and inflexible and utf8 locales seem to be a Linux only feature.

Yup; having an internal API dependent on locale seems like a bad idea to 
many of us.  Codeset conversion belongs at the application interface to 
persistant data; internal APIs should use Unicode.

[EMAIL PROTECTED]        XFree86 Core Team              SuSE, Inc.


_______________________________________________
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n

Reply via email to