Adding new nnmail-split-fancy construct

Adding new nnmail-split-fancy construct

[Tonny Madsen]

Hi'

I occasionally need to split messages, by searching them directly for
a regular expression. In most cases, though, it is simply to single
out all messages, where the the subject line starting with a specific
word.

This cannot be done with the current nnmail-split-fancy variable and
function. To deal with this I have added a catch-all split expression,
(/ REGEXP SPLIT), that will use SPLIT if REGEXP matches the message. I
would prefer a solution where one could specify that the header line
must *start* with a specific regexp, but this was easier to implement.

Note that the current implementation will convert the split expression
("subject" "patch" SPLIT) to the regexp
""^\\(\\(subject\\):.*\\)\\<\\(patch\\)\\>". This will match any
subject, which includes the word 'patch'. I want to word to be the
first word of the subject.

/tonny

Patch:

--- nnmail.el~	Sun Jul 20 22:34:48 1997
+++ nnmail.el	Wed Oct 22 10:18:19 1997
@@ -354,6 +354,9 @@
 
 \(& SPLIT...): Process each SPLIT expression.
 
+\(/ REGEXP SPLIT): If the message matches REGEXP, store the messages
+  as specified by SPLIT.
+
 \(: FUNCTION optional args): Call FUNCTION with the optional args, in
   the buffer containing the message headers.  The return value FUNCTION
   should be a split, which is then recursively processed.
@@ -1210,6 +1213,13 @@
    ;; Builtin : operation.
    ((eq (car split) ':)
     (nnmail-split-it (eval (cdr split))))
+
+   ;; Builtin / operation.
+   ((eq (car split) '/)
+    (goto-char (point-max))
+    ;; Same problems as specified for the next condition.
+    (when (re-search-backward (nth 1 split) nil t)
+      (nnmail-split-it (nth 2 split))))
 
    ;; Check the cache for the regexp for this split.
    ;; FIX FIX FIX could avoid calling assq twice here

Re: Adding new nnmail-split-fancy construct

[Lars Magne Ingebrigtsen]

Tonny Madsen <tma@nettest.dk> writes:

> I occasionally need to split messages, by searching them directly for
> a regular expression. In most cases, though, it is simply to single
> out all messages, where the the subject line starting with a specific
> word.
> 
> This cannot be done with the current nnmail-split-fancy variable and
> function. To deal with this I have added a catch-all split expression,
> (/ REGEXP SPLIT), that will use SPLIT if REGEXP matches the message.

I think you should use the `:' thing instead -- it allows you to
specify an arbitrary function to be run.

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

Re: Adding new nnmail-split-fancy construct

[Tonny Madsen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Tonny Madsen <tma@nettest.dk> writes:
> 
> > I occasionally need to split messages, by searching them directly for
> > a regular expression. In most cases, though, it is simply to single
> > out all messages, where the the subject line starting with a specific
> > word.
> > 
> > This cannot be done with the current nnmail-split-fancy variable and
> > function. To deal with this I have added a catch-all split expression,
> > (/ REGEXP SPLIT), that will use SPLIT if REGEXP matches the message.
> 
> I think you should use the `:' thing instead -- it allows you to
> specify an arbitrary function to be run.

True, but I would have prefered to keep all the different search
patterns in nnmail-split-fancy and not to make a number of
aux. functions...

Anyway, I find it currious, that I should be the first to request
this...

/tonny