[NTG-context] XeTeX and built-in glyph generation

Philipp Reichmuth reichmuth at web.de
Thu Nov 2 11:04:26 CET 2006


I have a question regarding ConTeXt, XeTeX and the use of Unicode fonts 
with diacritic characters which I use quite a lot.  (Note also that I'm 
a novice user.)  Consider the following extreme example (using Gentium 

\definetypeface[lucida][rm][Xserif][Lucida Bright]

Ṯḥíş ıš ā pìėçẹ ȯf ṭēxt.

If you typeset this with 'texexec --xtx', you get very different results 
with Gentium vis-a-vis Lucida because of the different glyph repertoires 
of both fonts.  This is a bit unsatisfactory because ConTeXt has all the 
logic already in place to construct most of these these from base 
characters plus accents where needed.  Of course I can go through my 
Unicode text manually and change all the missing characters to something 
like \={e}, but that is not a good solution either; having to sift 
through your entire document because your publisher demands a different 
font makes you feel more than a bit stupid :)

Is there a way to tell ConTeXt what Unicode characters are available in 
a font and which ones it has to build using TeX's accent functions?  I 
guess in the above example it would be difficult because no regime gets 
loaded at all.  What would be ideal in my case would be a configurable 
UTF-8 regime where I could specify the following on a per-typeface basis:

- character[ range]s which get passed directly to the underlying engine 
(useful mainly with XeTeX and possibly with Aleph/Omega)
- character[ range]s which get constructed from accents
- character[ range]s which get used from a different typescript; for 
example if I have passages or quotations in a script which isn't 
available in my typescript, it would be nice to say that "all 
IPA/Greek/Old Church Slavonic" should come from a substitution 
typescript where these scripts are available and look good.

Is such a thing possible?


More information about the ntg-context mailing list