Re: Bibtex and Xelatex problems

25/10/2005, 7pm - Bernd wrote:


> I have started to create a bibliography with BibDesk and some nice  
> scripts that help me gather all the information. Now I wanted to  
> actually print out the whole stuff and had to find out that some  
> unicode characters are plain left out. That is especially for  
> characters that use bars and dots under and above letters such as  
> you would want for transliteration of Arabic script.

I don't think you have a problem.
BibTeX seems clever enough to ignore the actual contents of the files  
it processes -- effectively it ignores the encoding of the file for  
our purposes.

I tested this hypothesis with the following file:

%%% unicode-bibtex-test.tex %%%
\begin{filecontents}{test.bib}
@book{test,
   author = {Will Robertson},
   title = {उद्दिष्टः  
समाहितचित्तस्य योगः ।  
कथं व्युत्थितचित्तोऽपि},
   publisher = {Nobody},
   year = {1842}}
\end{filecontents}

\documentclass{article}
\usepackage{fontspec}
\setromanfont{Code2000}
\begin{document}
hello \cite{test}
\bibliographystyle{plain}
\bibliography{test}
\end{document}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

So your problem stems from one of two areas: (a) BibDesk is doing  
something funny with your bibliography database. Check the plain .bib  
file in a text editor to ensure this file looks as you expect. (Also  
look into BibDesk's Unicode->TeX conversion in the "Files" preference.)

Alternatively, you might simply be printing out the bibliography in a  
font that doesn't *contain* all of the glyphs you're trying to  
output. This can be checked by using a font like Code2000, which has  
every glyph under the sun, albeit at the expense of actually looking  
nice in some cases.

Hope this helps,

Will