gnus-cite-parse terribly slow on underscores

[Christoph Rohland <cr@sap.com>]

Hi Folks,

I am running Oort Gnus v0.05
XEmacs 21.4 (patch 4) "Artificial Intelligence" [Lucid] (i386-suse-linux, Mule) of Mon Sep 24 2001 on lhospital


Since upgrading to ognus I had a lot of problems with some mailing
lists and several other mails. 

Displaying these mails did take a really long time or locked up
totally. After a while I did get the pattern: They all had signatures
with a lot of underscores like e.g. yahoo adds.

I then did debug-on-quit and realised that gnus-cite-parse did lockup
in (looking-at gnus-supercite-regexp) or taking a long time in
(re-search-forward prefix-regexp (1- end) t).

Since I realised the underscore in the messages I examined
message-cite-prefix-regexp and gnus-supercite-regexp and noticed the
following common pattern: \(\w\|[-_.]\)+ and modified it to
\(\w\|[-.]\)+ and AHA! now it works.

I do not understand all the interactions, but at least in my xemacs _
is covered by \w and apparently this messes up looking-at and
re-search-forward. Perhaps it is a bug in my xemacs since its info
says:

,----
|    The XEmacs regular expression syntax most closely resembles that of
| `ed', or `grep', the GNU versions of which all utilize the GNU `regex'
| library.  XEmacs' version of `regex' has recently been extended with
| some Perl-like capabilities, described in the next section.
`----

Any insight?

Greetings
		Christoph