Problem with gnus-group-mark-group

Problem with gnus-group-mark-group

[Ulf Stegemann]

XEmacs 21.4 (patch 17) "Jumbo Shrimp" [Lucid] (i686-pc-linux, Mule)
No Gnus v0.3

I recently switched form Gnus 5.10.6 to No Gnus 0.3 and encountered a problem
when hitting M-g in a summary buffer (gnus-summary-rescan-group) or the
group buffer (gnus-group-get-new-news-this-group).

Each M-g produces an additional group entry in the group buffer leaving the
existing one(s) borked ("no group on current line" when trying to enter).

I was able to identify gnus-group-mark-group as responsible for the faulty
behavior. The function changed significantly since 5.10.6 and replacing No
Gnus' gnus-group-mark-group with the v5.10.6 version apparently fixed the
problem.

Unfortunately, my lack of time (and sufficient lisp knowledge, probably)
let's me hesitate to further investigate this issue. Maybe there's someone
out there who knows what the reason for the described behavior is?


Ulf

Re: Problem with gnus-group-mark-group

[Reiner Steib]

On Tue, Mar 08 2005, Ulf Stegemann wrote:

> XEmacs 21.4 (patch 17) "Jumbo Shrimp" [Lucid] (i686-pc-linux, Mule)
> No Gnus v0.3
>
> I recently switched form Gnus 5.10.6 to No Gnus 0.3 and encountered a problem
> when hitting M-g in a summary buffer (gnus-summary-rescan-group) or the
> group buffer (gnus-group-get-new-news-this-group).
>
> Each M-g produces an additional group entry in the group buffer leaving the
> existing one(s) borked ("no group on current line" when trying to enter).
>
> I was able to identify gnus-group-mark-group as responsible for the faulty
> behavior. The function changed significantly since 5.10.6 and replacing No
> Gnus' gnus-group-mark-group with the v5.10.6 version apparently fixed the
> problem.

The only relevant change in `gnus-group-mark-group' is this one
(AFAICS):

--8<---------------cut here---------------start------------->8---
--- gnus-group.el	5 Jan 2004 16:30:15 -0000	7.4
+++ gnus-group.el	8 Jan 2004 16:33:41 -0000	7.5
@@ -1689,15 +1689,14 @@
 	;; Go to the mark position.
 	(beginning-of-line)
 	(forward-char (or (cdr (assq 'process gnus-group-mark-positions)) 2))
-	(subst-char-in-region
-	 (point) (1+ (point)) (char-after)
-	 (if unmark
-	     (progn
-	       (setq gnus-group-marked (delete group gnus-group-marked))
-	       ? )
+	(delete-char 1)
+	(if unmark
+	    (progn
+	      (setq gnus-group-marked (delete group gnus-group-marked))
+	      (insert-char ? 1 t))
 	   (setq gnus-group-marked
 		 (cons group (delete group gnus-group-marked)))
-	   gnus-process-mark)))
+	   (insert-char gnus-process-mark 1 t)))
       (unless no-advance
 	(gnus-group-next-group 1))
       (decf n))
--8<---------------cut here---------------end--------------->8---

Maybe there is some incompatibility with XEmacs in this change?  Ideas
anyone?

Could you try to edebug the function `gnus-group-mark-group'?  Go to
the function definition and do `M-x edebug-defun RET'.  When the
function is called you can step through it with SPC and hopefully see
where the problem is.  Cancel the debug mode by re-evaluation the
function with `M-x eval-defun RET'.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/