spam.el: Bugs, compatibility with v5-10 (was: spam manual update for Gnus)

[Reiner Steib <reinersteib+gmane@imap.cc>]

On Wed, Feb 08 2006, Ted Zlatanov wrote:

> The differences are pretty major, and I would actually like to bring
> the CVS version of spam.el and spam-report.el over to Emacs CVS.

You might have to restore compatibility with Emacs 20.7 if the version
in v5-10 has it.  (The versions in Emacs and v5-10 should be equal.)

> There's a lot of bug fixes, but no new features and no bug reports
> with the CVS version of spam.el/spam-report.el for a while now.  

[ I didn't report it yet because I wanted to test if the problem only
occurs with nnimap or also for nnml groups.  But now that you ask... ]
`ham-process-destination' doesn't work correctly for me.  When I find
a ham message in "Spam/Unsure", I mark it with `d'.  On group exit,
the message is moved to `rs-imap-train-ham-group' but not to
`rs-imap-inbox-group'.  IIRC, it worked correctly when using Gnus 5.11
(from Emacs CVS).  `spam-process-destination' works also with multiple
destinations[1] for the same group.

--8<---------------cut here---------------start------------->8---
(defvar rs-imap-method "nnimap+kiz:"
  "Prefered IMAP method.")
(defvar rs-imap-inbox-group (concat rs-imap-method "INBOX")
  "IMAP inbox.")
(defvar rs-imap-spam-group (concat rs-imap-method "Spam/spamblock")
  "Group for incoming spam.")
(defvar rs-imap-train-ham-group (concat rs-imap-method "bayes/ham")
  "Group for training ham.")
(defvar rs-imap-train-spam-group (concat rs-imap-method "bayes/spam")
  "Group for training spam.")

(setq
 gnus-parameters
 `(;; [other stuff]
   (":Spam/Unsure"
    (ham-process-destination ,rs-imap-train-ham-group
			     ,rs-imap-inbox-group)
    (spam-move-spam-nonspam-groups-only nil)
    (spam-process-destination ,rs-imap-train-spam-group
			      "nnml+personal:misc.Spam.2006")
    (spam-contents gnus-group-spam-classification-spam)
    (auto-expire . nil))
   ;; [other stuff]
   ("^gmane\\."
    (spam-contents gnus-group-spam-classification-ham);; tool bar
    (spam-process
     ,(if (boundp 'spam-summary-exit-behavior)
	  ;;(> (gnus-continuum-version) 5.11)
	  '((spam spam-use-gmane))
	'(gnus-group-spam-exit-processor-report-gmane))))
   ;; [other stuff]
   ))
--8<---------------cut here---------------end--------------->8---

In the "^gmane\\." entry, I need a conditional for v5-10 and trunk,
because the backward compatibility code doesn't work correctly,
see [2].

[ Bring the trunk version of spam.el and spam-report.el over to Emacs
  CVS / v5-10. ]
> Is this possible?

Additionally to those two items, my message from September [3].
Unless these problems are resolved, I'd rather not put it into v5-10.

> That being said, Chong's version of the manual looks good after a
> first read.  I will probably suggest minor changes here and there, but
> I am really glad he wrote something better than my original version :)

Great!

Bye, Reiner.

[1]
,----[ *Messages* ]
| Exiting summary buffer and applying spam rules
| Registering 1 specific articles as spam using backend spam-use-move
| Copying to nnimap+kiz:bayes/spam: (657)...
| nnimap: Setting marks in bayes/spam...done
| nnimap: Updating info for nnimap+kiz:bayes/spam...done
| Copying to nnml+personal:misc.Spam.2006: (657)...
| Wrote /home/dept/ste/Mail/personal/misc.Spam.2006/715
| 1 spam messages were registered by backend spam-use-move.
| nnimap: Setting marks in Spam/Unsure...done
| Retrieving newsgroup: nnimap+kiz:bayes/spam...
`----

[2] <news:v9br2n61qd.fsf@marauder.physik.uni-ulm.de> 
    http://thread.gmane.org/v9br2n61qd.fsf%40marauder.physik.uni-ulm.de

[3] <news:v9vf0ugya0.fsf@marauder.physik.uni-ulm.de>
    http://thread.gmane.org/v9vf0ugya0.fsf%40marauder.physik.uni-ulm.de
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/