Hello, I am sorry for misunderstanding the levels. I have lost a lot of valuable and not recoverable work because of this, and the only level it matched, as I understood, was critical. Anyway, I have found out a way to reproduce the bug. First you must notice that by default I have activated the fly spell mode when loading a LaTeX file with
(add-hook 'LaTeX-mode-hook 'flyspell-mode) and the default dictionary is "castellano8", I do not know if with other languages there is a similar problem. To reproduce the error you must load the attached file and then switch to LaTeX-mode with (M-x LaTeX-mode). Luis Llana. -- http://antares.sip.ucm.es/~luis TEL: +34 91 394 4527 FAX: +34 91 394 4654
\bdfn Let $M=(S,I,O,\delta,s_0)$ be a \PFSM. We write the {\it
generalized} transition $s
\tranap{(i_1/o_1,\ldots,i_n/o_n)}{\overline{p}} s'$ if there exist
$s_1,\ldots,s_{n-1}\in S$ and $\overline{p}_1,\ldots,
\overline{p}_n\in\simbP $ such that
$$s\tranp{i_1/o_1}{\overline{p}_1}s_1\tranp{i_2/o_2}{\overline{p}_2}s_2\;
\cdots \; s_{n-1}\tranp{i_n/o_n}{\overline{p}_n}s'$$
%
where $\overline{p}=\prod \overline{p}_i$.
We say that $\rho=(i_1/o_1,\ldots,i_n/o_n)$ is a {\it
non-probabilistic trace}, or simply a {\it trace}, of $M$ if there
exist $s'\in S$ and $\overline{p}\in\simbP $ such that
$s_0\tranap{\rho}{\overline{p}} s'$.
Let $\rho=(i_1/o_1,\ldots,i_n/o_n)$ and $\overline{p}\in \simbP$. We say that
$\overline{\rho}=(\rho,\overline{p})$ is a {\it probabilistic trace} of $M$ if
there exists $s'\in S$
%,\overline{p}\in \simbP$
such that $s_0\tranap{\rho}{\overline{p}} s'$.
We denote by $\Traces{M}$ and $\probTraces{M}$ the sets of
non-probabilistic and probabilistic traces of $M$, respectively.
\edfn
pgpsXRUeTddvD.pgp
Description: PGP signature
