Re: spam handling in No Gnus 0.3

[Alberto L]

>>>>> "Ted" == Ted Zlatanov <tzz@lifelogs.com> wrote:

T>On 19 Mar 2006, alberto.l@nospam.net wrote:
T>> [...]
T>>
T>> It is quite easy to fix the above, one has to move
T>> "spam-mark-junk-as-spam-routine" to before (rather than after)
T>> initializing the "spam-old-articles" in the prepare-summary spam hook:
T>>
T>
T>(defun spam-summary-prepare ()
T>  (spam-mark-junk-as-spam-routine)
T>  (setq spam-old-articles
T>  (list (cons 'ham (spam-list-articles gnus-newsgroup-articles 'ham))
T>        (cons 'spam (spam-list-articles gnus-newsgroup-articles 'spam)))))
T>
T>> I recommend adopting the above modification.
T>
T>The whole point of the order is that old articles are from before
T>unseen articles got marked as spam.  So I don't think this fix will do
T>it.

Well, it fixes the problem for me.  With the No Gnus 0.3 ordering,
this is what happens when a spam group is visited:

- on entry:
  - gnus takes note of spam/ham classification of articles
    (all articles are unseen)
  - gnus marks all unseen articles as spam
- on exiting:
  - gnus checks the article marks
  - w.r.t. what noted on entry, all marks have changed, so
    all articles are entered in the "changed" list
  - gnus-registry notes that the articles were registered as spam
    (BTW, this is consistent with their mark being "spam" now...)
  - since the marks changed, all articles are unregistered as spam
  - since this is a spam group, and spam destination is nil,
    all "spam" marked articles are deleted

By changing the order, on entering Gnus notes that all articles are
marked "spam", on exiting they are again "spam", nothing is entered in
the "changed" list, nothing is unregistered.  If the user marks
articles as "ham", then their mark is changed w.r.t. "spam", so they
are (correctly) unregistered as spam.

My impression is that the modification should do no harm for everybody
else, but I may be wrong.

T>
T>Could you use the CVS version of Gnus, and we can start thinking of a
T>solution against it?
T>

OK I will try that.

[...]

Greetings,
--
Alberto