Pablo Saratxaga <[EMAIL PROTECTED]> writes:
> Kaixo!
>
> On Thu, Jul 05, 2001 at 05:33:59PM -0400, Owen Taylor wrote:
>
> > Syriac is not supported by Pango, but (given fonts) it probably
> > would be a few hours to add such support.
> >
> > Basically, all that is needed is to take the Arabic support and
> > change it for the Syriac joining classes, and for the encoding
> > of the font.
>
> Well, not exactly.
> Arabic support is easy, because all presentation forms have code points
> in unicode; that is fonts that provide those glyphs use those codepoints
> it is standardized and life is easy.
>
> Syriac, while similar in concept to Arabic script, only has unicode code
> points for letters, not for presentation forms.
> Which means there is the same font problem as with indic languages: how
> are the presentation forms stored in a font?
The presentation forms are only in Unicode for compatibility reasons,
and while they make for a convenient "one font" hack, and provide
a bit of a standard they don't actually make rendering any easier.
There are basically three approaches you can take to font
encoding and Pango:
- Used a fixed encoding of the presentation forms. As well as the
Unicode presentation form encoding of the Arabic script Pango
supports 3 other encodings (mulearabic, LangBox, naqshfont, as you
know :-)
- Use BDF properties to encode information about the
character => glyph properties, as Robert Brady did
for Indic. This is probably overkill for Syriac.
- If you have OpenType fonts, use OpenType tables to encode
the character => glyph mapping and use the Xft backend.
It really depends a lot on what's available for fonts.
Regards,
Owen
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/linux-utf8/
