Re: [Bug-AUCTeX] 11.85; Biblatex vs. RefTex

[Andreas Breitbach]

Am Samstag, den 27.06.2009, 16:33 +0200 schrieb Ralf Angeli:
> Please keep the mailing list copied!
> 
> * Andreas Breitbach (2009-06-27) writes:
> 
> > Am Samstag, den 27.06.2009, 15:42 +0200 schrieb Ralf Angeli:
> >> 
> >> I still cannot reproduce this, not even with Emacs 22.2, so this will be
> >> difficult to debug without having access to the installation.  Anyway,
> >> what's the result if you set `reftex-cite-prompt-optional-args' to nil?
> > I hope to have the time to cross-check this whole stuff with another
> >  PC(running Ubuntu 8.10, as far as I remember). With the "nil"
> >  applied, it works as it should:
> >  \footcite{InternationalSecurity.1Summer.2000}.

So, I'm at my other PC now(both maschines are ~60 km apart). And guess
what: It works. Even with RefTeX 4.31, which was supplied with "Emacs
22.1.1 (i486-pc-linux-gnu, GTK+ Version 2.12.9) of 2008-05-03 on
terranova, modified by Ubuntu". (This maschine is running Ubuntu 8.0ß4
LTS.)
> Does it help if you apply the following patch to reftex-cite.el,
> recompile and reinstall?  Alternatively you can evaluate a patched
> version of `reftex-do-citation' in a running Emacs session by copying
> its definition into the *scratch* buffer and typing `C-M-x' with point
> in it.
> 
> --- reftex-cite.el.~1.53.~    2009-03-01 15:42:06.000000000 +0100
> +++ reftex-cite.el    2009-06-27 16:27:14.000000000 +0200
> @@ -702,8 +702,9 @@
>                         (equal arg '(4))))
>            (let ((start 0) (nth 0) value)
>              (while (setq start (string-match "\\[\\]" string start))
> -              (setq value (read-string (format "Optional argument %d: "
> -                                               (setq nth (1+ nth)))))
> +              (setq value (save-match-data
> +                         (read-string (format "Optional argument %d: "
> +                                              (setq nth (1+ nth))))))
>                (setq string (replace-match (concat "[" value "]") t t string))
>                (setq start (1+ start)))))
>          ;; Should we cleanup empty optional arguments?
I'll try this one when I'm back there on monday, but so long I'd assume it's 
something with packaging(maybe comparing the filelists of both versions yields 
any results)