What's wrong with this regexp?

What's wrong with this regexp?

[Frank Schmitt]

Is it possible that regexp's in nnmail-split-methods are broken?
I've got the following in my .gnus:

(setq nnmail-split-methods
           '(("NT-Emacs" "^\(CC:\|To:\).*help-emacs-windows@gnu.org.*")
	     ("NT-Emacs" "^\(CC:\|To:\).*ntemacs-users@cs.washington.edu.*")
	     ("XEmacs-NT" "^\(To:\|CC:\).*xemacs-???nt@xemacs.org.*")
	     ("Hamster-src" "^\(CC:\|To:\).*hamster-sourcen@\(yahoogroups\.\(de\|com\)\|egroups\.com\).*")
	     ("Replies" "^\(CC:\|To:\).*usenet@Frank-Schmitt.net.*")
	     ("Uni" "^\(CC:\|To:\).*fschmitt@uni-koblenz.de.*")
             ("mail.general" "")))

But everything goes in mail.general. I have absolutely no idea why.
Some examples of mails which go to mail.general but should definitely go
to a different group:

To: hamster-sourcen@egroups.com
Message-ID: <9jemnj.3vuh011.1@server.heimweb.de>
From: "Andreas Heim" <liste@heimweb.de>
Date: Sun, 22 Jul 2001 14:07:47 +0200
Subject: Re: [Fix]Hamster Events hamwindows.hsm
Xref: FRANK-FMB8GC9RR mail.general:1437
Lines: 17

To: xemacs-nt@xemacs.org
Subject: ldap, win32
From: Fabrice Popineau <Fabrice.Popineau@supelec.fr>
Date: Fri, 07 Sep 2001 16:17:14 +0200
Message-ID: <vgiv2ihh.fsf@supelec.fr>
Lines: 25
Xref: FRANK-FMB8GC9RR XEmacs-NT:360

Message-Id: <200110261117.HAA04621@glass.lplizard.com>
From: Chris Green <chrisg@lplizard.com>
To: help-emacs-windows@gnu.org
Subject: [h-e-w] emacs 21 under win2k
Date: Fri, 26 Oct 2001 07:17:38 -0400
Lines: 5
Xref: FRANK-FMB8GC9RR NT-Emacs:743

(I've snipped irrelevant headers)

-- 
One Ring to rule them all, One Ring to find them,
One Ring to bring them all and in the darkness bind them
In the Land of Mordor where the Shadows lie.
19. Dezember 2001

Re: What's wrong with this regexp?

[Norbert Koch]

Frank Schmitt <usereplyto@Frank-Schmitt.net> writes:

> Is it possible that regexp's in nnmail-split-methods are broken?
> I've got the following in my .gnus:
>
> (setq nnmail-split-methods
>            '(("NT-Emacs" "^\(CC:\|To:\).*help-emacs-windows@gnu.org.*")
> 	     ("NT-Emacs" "^\(CC:\|To:\).*ntemacs-users@cs.washington.edu.*")
> 	     ("XEmacs-NT" "^\(To:\|CC:\).*xemacs-???nt@xemacs.org.*")
> 	     ("Hamster-src" "^\(CC:\|To:\).*hamster-sourcen@\(yahoogroups\.\(de\|com\)\|egroups\.com\).*")
> 	     ("Replies" "^\(CC:\|To:\).*usenet@Frank-Schmitt.net.*")
> 	     ("Uni" "^\(CC:\|To:\).*fschmitt@uni-koblenz.de.*")
>              ("mail.general" "")))

Don't you have to use 

^\\(Cc\\To\\):.*...

etc in the regexp?

norbert.

Re: What's wrong with this regexp?

[Frank Schmitt]

Norbert Koch <nk@LF.net> writes:

>>(setq nnmail-split-methods
>>           '(("NT-Emacs" "^\(CC:\|To:\).*help-emacs-windows@gnu.org.*")
>>	     ("NT-Emacs" "^\(CC:\|To:\).*ntemacs-users@cs.washington.edu.*")
>>	     ("XEmacs-NT" "^\(To:\|CC:\).*xemacs-???nt@xemacs.org.*")
>>	     ("Hamster-src" "^\(CC:\|To:\).*hamster-sourcen@\(yahoogroups\.\(de\|com\)\|egroups\.com\).*")
>>	     ("Replies" "^\(CC:\|To:\).*usenet@Frank-Schmitt.net.*")
>>	     ("Uni" "^\(CC:\|To:\).*fschmitt@uni-koblenz.de.*")
>>             ("mail.general" "")))
>
>Don't you have to use 
>
>^\\(Cc\\To\\):.*...
>
>etc in the regexp?

Don't we use the regexp scheme documented under 
(x)emacs->Searching and Replacement->Regexp search?
If we don't, where is the regexp engine of Gnus documented?

-- 
One Ring to rule them all, One Ring to find them,
One Ring to bring them all and in the darkness bind them
In the Land of Mordor where the Shadows lie.
19. Dezember 2001

Re: What's wrong with this regexp?

[Karl Kleinpaste]

Frank Schmitt <usereplyto@Frank-Schmitt.net> writes:
> Don't we use the regexp scheme documented under 
> (x)emacs->Searching and Replacement->Regexp search?

Yes, but it's a matter of the lisp reader's context: You have to
double most occurrences of `\' in order to get them into a lisp
expression.  What you type in the minibuffer as "\(a\|b\)" must become
"\\(a\\|b\\)" when used in a string expression in elisp.

BTW, nnmail-split-methods usage is case-insensitive unless you
explicitly turn off case-fold-search in nnmail-split-hook.

Re: What's wrong with this regexp?

[Frank Schmitt]

Karl Kleinpaste <karl@charcoal.com> writes:

>Frank Schmitt <usereplyto@Frank-Schmitt.net> writes:
>>Don't we use the regexp scheme documented under 
>>(x)emacs->Searching and Replacement->Regexp search?
>
>Yes, but it's a matter of the lisp reader's context: You have to
>double most occurrences of `\' in order to get them into a lisp
>expression.  What you type in the minibuffer as "\(a\|b\)" must become
>"\\(a\\|b\\)" when used in a string expression in elisp.

Is this documented somewhere and if it is, where?

-- 
One Ring to rule them all, One Ring to find them,
One Ring to bring them all and in the darkness bind them
In the Land of Mordor where the Shadows lie.
19. Dezember 2001

Re: What's wrong with this regexp?

[Norbert Koch]

Frank Schmitt <usereplyto@Frank-Schmitt.net> writes:

>>Yes, but it's a matter of the lisp reader's context: You have to
>>double most occurrences of `\' in order to get them into a lisp
>>expression.  What you type in the minibuffer as "\(a\|b\)" must become
>>"\\(a\\|b\\)" when used in a string expression in elisp.
>
> Is this documented somewhere and if it is, where?

Sorry, I should have been more explicit in my first message.  You find
something about this in the lispref manual:

(lispref)Regular Expressions

norbert.

Re: What's wrong with this regexp?

[Martin Monsorno]

Karl Kleinpaste <karl@charcoal.com> writes:

> Frank Schmitt <usereplyto@Frank-Schmitt.net> writes:
> > Don't we use the regexp scheme documented under 
>> (x)emacs->Searching and Replacement->Regexp search?
>
> Yes, but it's a matter of the lisp reader's context: You have to
> double most occurrences of `\' in order to get them into a lisp
> expression.  What you type in the minibuffer as "\(a\|b\)" must become
> "\\(a\\|b\\)" when used in a string expression in elisp.
>

I had the same problem a few days ago, and finally found the solution
by searching through some of the "other peoples .gnus-files" on the
net. I suppose, just one little example with "\\(To\\|cc\\)" in the
info file about mail-splitting would help very much.

> BTW, nnmail-split-methods usage is case-insensitive unless you
>explicitly turn off case-fold-search in nnmail-split-hook.
>

This would also be nice to be in that info page.

-- 
Martin Monsorno
mailto:monsorno@gmx.de

Re: What's wrong with this regexp?

[Lars Magne Ingebrigtsen]

Martin Monsorno <monsorno@gmx.de> writes:

>> Yes, but it's a matter of the lisp reader's context: You have to
>> double most occurrences of `\' in order to get them into a lisp
>> expression.  What you type in the minibuffer as "\(a\|b\)" must become
>> "\\(a\\|b\\)" when used in a string expression in elisp.
>
> I had the same problem a few days ago, and finally found the solution
> by searching through some of the "other peoples .gnus-files" on the
> net. I suppose, just one little example with "\\(To\\|cc\\)" in the
> info file about mail-splitting would help very much.

Well, the examples in that section of the manual does use \\...

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen