[pdftex] ucs.sty & others

G. D. Brettschneider pdftex at GDBrettschneider.de
Sun Jul 13 19:04:32 CEST 2003

Jean-Christophe Dubacq wrote:

>I would like to know if there is some enhanced method to get pdftex to
>use a (ttf) unicode font in all its charsets (without defining a specific
>encoding for each language). I can do this with latex (some problems
>arise, tough, but none that I cannot fix) by using utf-8 encoding.
>However, I can only have pdftex use the pk files generated by ttf2pk,
>not the ttf directly (...)
TeX3 can only handle character sets with 256 characters, so you must 
create multiple TeX font metrics (tfm) and virtual fonts (vf) out of one 
single unicode font, one for each row. You can do this using ttf2afm, 
afm2tfm and vptovf, as described in the pdftex manual. However, you have 
to create the unicode row encoding files by yourself. Inspect the 
unicode data file (www.unicode org) and the Adobe Glyph List (AGL, 
http://partners.adobe.com) for the information required. Open the ttf 
file with a text editor to find out which glyph names are supported. 
Last you have to write a few TeX macros to deal with several font files 
using the same font-family name.

In this way I can embed e.g. Arabic letters from arial.ttf without 
problems, but I could't use ArialUnicodeMS or Code2000 which do not 
contain glyph names. These were the error messages I got (on both 
Windows and Linux):
- "No names available in 'post' table, printing by index forced" 
(ttf2afm/ ArialUnicodeMS)
- "Unknown version of OS/2 table (3)" (pdfeTeX 1.0b / Code2000)

I'm working on "some enhanced method", please be patient...



More information about the pdftex mailing list