David Kastrup <[EMAIL PROTECTED]> writes:

> "Harry Schmidt" <[EMAIL PROTECTED]> writes:
>
>> In the following minimal example only the preview for the first $A$ is
>> rendered:
>>
>> \documentclass{revtex4}
>> \begin{document}
>> $A$
>> \cite{a:b_c,x.2003}
>> $A$
>> \begin{thebibliography}{2}
>> \bibitem{a:b_c} asdf
>> \bibitem{x.2003} defg
>> \end{thebibliography}
>> \end{document}
>>
>>   I think this has to do with the document class.  Changing revtex4
>> to article will remove the problem.
>
> Not really.  Please complain to the author of natbib.sty.  He uses
> \edef (not even [EMAIL PROTECTED]) on \mbox.  It is more or less pure
> luck that this does not break with the default \mbox command.  If he
> were to use [EMAIL PROTECTED] instead, this would be no problem with
> preview.sty.
>
> I might figure out a working fix involving eTeX (where available),
> but this should really be fixed in natbib.sty.  Using \edef instead
> of [EMAIL PROTECTED] on LaTeX macros is plainly a bug and very wrong.

I checked in an eTeX-based workaround into preview.dtx.  You might
want to apply the patch (and fix the checksum afterwards) if things
are urgent for you.  It will only work on eTeX-based `latex' commands.

--- preview.dtx	15 Aug 2006 12:47:32 +0200	1.121
+++ preview.dtx	25 Aug 2006 12:29:11 +0200	
@@ -527,11 +527,14 @@
 %   Now [EMAIL PROTECTED] needs its helper macro.  In order to avoid
 %   recursive definitions, we advise only macros that are not yet
 %   advised.  Or, more exactly, we throw away the old advice and only
-%   take the new one.
+%   take the new one.  We use e\TeX's \cmd{\protected} where available
+%   for some extra robustness.
 %    \begin{macrocode}
 [EMAIL PROTECTED]@delay{%
   \ifx#1\relax \let#1#2\fi
-  [EMAIL PROTECTED]@}}}
+  [EMAIL PROTECTED]
+  [EMAIL PROTECTED] \else \protected\fi
+  [EMAIL PROTECTED]
 %    \end{macrocode}
 %\end{macro}
 %
-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum
_______________________________________________
bug-auctex mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-auctex

Reply via email to