Re: problems with nnvirtual and marks

[dme@dme.org]

* dme@dme.org [2002-01-03 13:18:40]
> Recently I've been playing with reading most of my (new) mail/news
> through nnvirtual groups.  The behaviour of ticked articles in these
> groups seems erratic.

After spending some more time looking, I came to the conclusion that
some of the (mark related) changes that I initiate in an nnvirtual
group are not getting passed through to component groups.

For example, if a virtual group comprised of various nnml groups has a
ticked article, pressing "d" on the article in the nnvirtual group
doesn't result in the "tick" mark being removed from the corresponding
nnml group's .marks file (after quitting the group, gnus, emacs, etc).

Based on my poor and particularly incomplete knowledge of gnus
internals, I decided that this might be because nnvirtual is missing a
"request-set-mark" function.  So I wrote one:

(deffoo nnvirtual-request-set-mark (group actions &optional server)
  (nnvirtual-possibly-change-server server)
  (dolist (action actions)
      (let ((range (nth 0 action))
	    (what  (nth 1 action))
	    (marks (nth 2 action)))
	(assert (or (eq what 'add) (eq what 'del)) t
		"Unknown request-set-mark action: %s" what)
	(dolist (article (gnus-uncompress-range range))
	  (let* ((amap (nnvirtual-map-article article))
		 (cgroup (car amap)))
	    (dolist (mark marks)
	      (gnus-request-set-mark cgroup (list (list (list (cdr amap)) what marks)))))))))

This improves the specific problem mentioned above - the removal of
"tick" marks now passes through from nnvirtual to nnml and the .marks
file gets updated.

However, when nnvirtual-request-set-mark is called with a large range
(over a couple of thousand articles), which seems to happen when I
enter a group sometimes, the iteration over the expanded range causes
emacs to go away and hide for a while (not sure how long - enough to
be "irritating", which probably means >10 seconds :-).

Is it possible that I'm anywhere near the source of my problems ?

Of course, I don't understand how this explains the random corruption
of seen/unseen marks which I've seen.

dme.