Re: Bug? pgnus 0.95

[Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann)]

Lee Willis <lee@gbdirect.co.uk> writes:

> I tootle around and read my mail, then say for example I want to compose
> a new mail. I'm in the Group buffer with point over some indeterminate
> group (At this point I don't care and shouldn't have to care which!) and
> hit 'm' to compose a new message.

FWIW, I have the following instead of the gcc-self mechanism.  It
pretty much does the same thing but provides a default.  Maybe you
would like to avoid your problem altogether and use something like
this?

(defun kai-gnus-outgoing-message-group ()
  (cond ((and gnus-newsgroup-name
              (stringp gnus-newsgroup-name)
              (not (string= "" gnus-newsgroup-name))
              (string-match "^\\(nnml\\|nnimap\\):" gnus-newsgroup-name))
         gnus-newsgroup-name)
        ((and gnus-newsgroup-name
              (stringp gnus-newsgroup-name)
              (string-match "^nntodo:" gnus-newsgroup-name))
         nil)
        (t "nnimap:INBOX")))

(setq gnus-outgoing-message-group 'kai-gnus-outgoing-message-group)

(defun message-toggle-gcc ()
  (interactive)
  (save-excursion
    (save-restriction
      (let ((pmin (progn (beginning-of-buffer) (point)))
            (pmax (progn
                    (beginning-of-buffer)
                    (re-search-forward (concat "^" mail-header-separator
                                               "$"))
                    (beginning-of-line)
                    (point))))
        (beginning-of-buffer)
        (narrow-to-region pmin pmax)
        (if (re-search-forward "^gcc: .*" nil t)
            (progn
              (beginning-of-line)
              (kill-line 1))
          (goto-char pmax)
          (insert (concat "Gcc: "
                          (kai-gnus-outgoing-message-group)
                          "\n")))))))

The last function is useful in the message buffer.

kai
-- 
I like BOTH kinds of music.