Splitting subjects with non-latin characters

[Niklas Morberg <niklas.morberg@axis.com>]

I'm using NT-Emacs 21.1 with a CVS gnus from the 13th of
November. I'm accessing an Exchange server over IMAP for
reading email.

The problem I'm having is that I can't split emails with
non-latin characters in the subject. I have a bunch of
different nnimap-split-fancy rules, but the one that fails
is:

	("subject" ".*\\s-\\{3,\\}.+" "spam")	

It works perfectly for latin characters, but as soon as I
get an email with e.g. big5 encoding the splitting stops
abruptly and tells me (in the message buffer):

nnmail-split-it: Search failed: ".*\\s-\\{3,\\}.+"

I have tried this with both an emacs I built myself on
cygwin and the pre-built version from FSF (GNU Emacs 21.1.1
(i386-msvc-nt5.0.2195) of 2001-10-22 on buffy) with the same
result.

(I'm pretty sure that the rule itself has nothing to do with
it; this is my first rule that splits on subject which is
why it shows the error.)

The problem probably has to do with my platform, but is there
a way to not have nnmail-split-it stop when it runs into
problems? I'm perfectly fine with the search failing, but
since it stops so abruptly no other rules will be evaluated
and (more importantly) no other emails will be split.

Niklas