Re: Problem with nnvirtual in 0.72?

Re: Problem with nnvirtual in 0.72?

[St. Suika Fenderson Roberts]

> Unfortunately after I upgraded to 0.72 I found that after reading all of
> the articles in the group and calling gnus-group-get-new-news the unread
> article count would be some large number 

I experienced the same thing, and haven't said anything since my site still
hasn't upgraded from 19.29, and I couldn't be sure if that wasn't the problem.

Since it didn't happen in 0.70 we know aproximately where to look, but I haven't
even tried.  Too much school.

	Suika
--
		wrobert2@mailhost.tcs.tulane.edu
"The dew fell with a particularly sickening thud this morning." 
								-Rhonda Rubin
   <a href="http://studentweb.tulane.edu/~wrobert2/">Suika no homepage</a>

Re: Problem with nnvirtual in 0.72?

[Jason L Tibbitts III]

>>>>> "SSFR" == St Suika Fenderson Roberts <wrobert2@mailhost.tcs.tulane.edu> writes:

SSFR> I experienced the same thing, and haven't said anything since my site
SSFR> still hasn't upgraded from 19.29, and I couldn't be sure if that
SSFR> wasn't the problem.

Well, I'm at 19.34, so I don't think that's it.  Hopefully I'll have some
time to delve into this later.

 - J<

Re: Problem with nnvirtual in 0.72?

[Han Pilmeyer]

Hi,

I noticed this too yesterday immediately after upgrading to 0.72. I've
seen it on at least 2 nnvirtual groups with both XEmacs and Emacs.

It's very fast though ;-)

Cheers.
-- 
Han Pilmeyer, han@uto.dec.com, Phone: +31 30 2832671, FAX: +31 30 2832686
Digital Equipment B.V., Europalaan 44, 3526 KS Utrecht, The Netherlands
|         Answers are the easy part, questions raise the doubt   -JB          |
All opinions are those of the author, not of Digital Equipment!

Re: Problem with nnvirtual in 0.72?

[St. Suika Fenderson Roberts]

I don't think this problem is in nnvirtual, for the simple reason that
nnvirtual worked in 0.70 and didn't in 0.72, and no changes were made to
_it_ in that span.

It just seems to be loosing marks.  I haven't run catchup on any of my
groups where I've noticed this, but I do read them to the end,
generally.  When one leaves an article or two it seems to prevent/lessen
the number of article marks that get lost.

At least it isn't loosing ! marks too often ^_^

hope this helps,
	Suika (who munged up a functional `update dribble buffer'
function ^_^ )
--
		wrobert2@mailhost.tcs.tulane.edu
        I am Carl of Borg.  Macross has been assimilated.

Re: Problem with nnvirtual in 0.72?

[David Moore]

"St. Suika Fenderson Roberts" <wrobert2@mailhost.tcs.tulane.edu> writes:

> I don't think this problem is in nnvirtual, for the simple reason that
> nnvirtual worked in 0.70 and didn't in 0.72, and no changes were made to
> _it_ in that span.

	Yup.  Fix in red gnus 0.75.

I always wanted to say that. Also, patch for 0.74 below for those who
don't want to wait, or don't want to down grade to 0.70.

*** ChangeLog.orig	Sun Dec  8 21:02:17 1996
--- ChangeLog	Mon Dec  9 21:02:21 1996
***************
*** 1,3 ****
--- 1,11 ----
+ Mon Dec  9 21:00:09 1996  David Moore  <dmoore@ucsd.edu>
+ 
+ 	* gnus-sum.el (gnus-summary-catchup): Out dated catchup code
+ 	removed.
+ 
+ 	* nnvirtual.el (nnvirtual-update-read-and-marked): Work around a
+ 	cache of active count in gnus-update-read-articles.
+ 
  Fri Dec  6 12:47:24 1996  Wes Hardaker  <Wesley.Hardaker@sphys.unil.ch>
  
  	* gnus-picon.el (gnus-picons-insert-face-if-exists): Don't reverse
*** gnus-sum.el.orig	Sun Dec  8 21:01:25 1996
--- gnus-sum.el	Mon Dec  9 21:01:56 1996
***************
*** 7642,7652 ****
  	    (unless to-here
  	      (setq gnus-newsgroup-unreads nil))
  	    (gnus-set-mode-line 'summary))
- 	  (let ((method (gnus-find-method-for-group gnus-newsgroup-name)))
- 	    (when (and (not to-here) (eq 'nnvirtual (car method)))
- 	      (nnvirtual-catchup-group
- 	       (gnus-group-real-name gnus-newsgroup-name)
- 	       (nth 1 method) all)))
  	  t))
      (gnus-summary-position-point)))
  
--- 7642,7647 ----
*** nnvirtual.el.orig	Mon Dec  9 20:57:10 1996
--- nnvirtual.el	Mon Dec  9 20:58:31 1996
***************
*** 401,408 ****
      ;; hits C-g, you won't leave the component groups in a half-way state.
      (gnus-atomic-progn
        ;; move (un)read
!       (while (setq entry (pop unreads))
! 	(gnus-update-read-articles (car entry) (cdr entry)))
  
        ;; clear all existing marks on the component groups
        (setq groups nnvirtual-component-groups)
--- 401,409 ----
      ;; hits C-g, you won't leave the component groups in a half-way state.
      (gnus-atomic-progn
        ;; move (un)read
!       (let ((gnus-newsgroup-active nil)) ;workaround guns-update-read-articles
! 	(while (setq entry (pop unreads))
! 	  (gnus-update-read-articles (car entry) (cdr entry))))
  
        ;; clear all existing marks on the component groups
        (setq groups nnvirtual-component-groups)

Problem with nnvirtual in 0.72?

[Jason L Tibbitts III]

I've recently started using nnvirtual to bring together several low traffic
groups from a couple of different servers.  I can't believe I didn't start
using it earlier.

Unfortunately after I upgraded to 0.72 I found that after reading all of
the articles in the group and calling gnus-group-get-new-news the unread
article count would be some large number (like 1300, I assume the sum of
(max - min article) for each group).  It seems that it's losing the list of
read articles, even though a long list shows up in the group info.

Has anyone seen this?  I'm not submitting this as a bug report because I
could be screwing something up and since I don't really know where to start
looking it would be a pretty pitiful bug report.
-- 
      Jason L. Tibbitts III - tibbs@uh.edu - 713/743-8684 - 221SR1
System Manager:  University of Houston High Performance Computing Center
                1994 PC800 "Kuroneko"      DoD# 1723

Re: Problem with nnvirtual in 0.72?

[David Moore]

Jason L Tibbitts III <tibbs@hpc.uh.edu> writes:

> Unfortunately after I upgraded to 0.72 I found that after reading all of
> the articles in the group and calling gnus-group-get-new-news the unread
> article count would be some large number (like 1300, I assume the sum of
> (max - min article) for each group).  It seems that it's losing the list of
> read articles, even though a long list shows up in the group info.
> 
> Has anyone seen this?  I'm not submitting this as a bug report because I
> could be screwing something up and since I don't really know where to start
> looking it would be a pretty pitiful bug report.

	I haven't seen it yet.  If you weren't seeing the problem in
0.70, maybe it's related to the other mysterious group count bug.
Although, you know, I do seem to vaguely recall all something like what
you just described happening to me recently.  ie, the group info seemed
correct but the display was for the whole group.  Maybe it was on an
nndir of the gnus-bugs archive, when ange-ftp got flakey?  Hah, that's
even worse bug report. :)

	Questions:
1) Do you read the component groups seperately from the virtual group?
2) If so, do you have nnvirtual-always-rescan set to t?
3) Do you have the component groups before, after, or both to the
virtual group in the *Group* buffer.



-- 
David Moore <dmoore@ucsd.edu>       | Computer Systems Lab      __o
UCSD Dept. Computer Science - 0114  | Work: (619) 534-8604    _ \<,_
La Jolla, CA 92093-0114             | Fax:  (619) 534-1445   (_)/ (_)
<URL:http://oj.egbt.org/dmoore/>    | Solo Furnace Creek 508 -- 1996!

Re: Problem with nnvirtual in 0.72?

[Matt Pharr]

David Moore <dmoore@UCSD.EDU> writes:
> 	Questions: 1) Do you read the component groups seperately from the
> virtual group?  2) If so, do you have nnvirtual-always-rescan set to t?  3)
> Do you have the component groups before, after, or both to the virtual
> group in the *Group* buffer.

I'm seeing this with 0.72, too. It doesn't _always_ happen, but it does
about 50% of the time...

1) Nope, never.
2) Nope.
3) After, and I've got at least 5 nnvirtual groups...

-matt
-- 
Matt Pharr | mmp@lux.stanford.edu | <URL:http://www-graphics.stanford.edu/~mmp>

Re: Problem with nnvirtual in 0.72?

[Han Pilmeyer]

>>>>> "David" == David Moore <dmoore@UCSD.EDU> writes:

David> 	Questions:
David> 1) Do you read the component groups seperately from the virtual group?
David> 2) If so, do you have nnvirtual-always-rescan set to t?
David> 3) Do you have the component groups before, after, or both to the
David> virtual group in the *Group* buffer.

1) Nope, never.
2) Nope.
3) Before (3 times) and both (1 time).

Cheers.
-- 
Han Pilmeyer, han@uto.dec.com, Phone: +31 30 2832671, FAX: +31 30 2832686
Digital Equipment B.V., Europalaan 44, 3526 KS Utrecht, The Netherlands
|         Answers are the easy part, questions raise the doubt   -JB          |
All opinions are those of the author, not of Digital Equipment!

Re: Problem with nnvirtual in 0.72?

[Matt Pharr]

David Moore <dmoore@UCSD.EDU> writes:
> Jason L Tibbitts III <tibbs@hpc.uh.edu> writes:
> 
> > Unfortunately after I upgraded to 0.72 I found that after reading all of
> >the articles in the group and calling gnus-group-get-new-news the unread
> >article count would be some large number (like 1300, I assume the sum of
> >(max - min article) for each group).  It seems that it's losing the list
> >of read articles, even though a long list shows up in the group info.
> ..
> > 
> > Has anyone seen this?  I'm not submitting this as a bug report because I
> >could be screwing something up and since I don't really know where to
> >start
> > looking it would be a pretty pitiful bug report.

Another little piece of data about this: I raised the level of the groups
that make up one of my virtual groups to 3 and started newsreading. When I
started gnus, all was fine--there were a handful of new messages. After I
read news in one of the virtual groups with new news and caught up, exited
and returned to the Group buffer, all of the groups that made up the
virutal group has a jazillion unread messages in them. So maybe it's a
problem with how nnvirtual handles catching up...

-matt
-- 
Matt Pharr | mmp@lux.stanford.edu | <URL:http://www-graphics.stanford.edu/~mmp>

Re: Problem with nnvirtual in 0.72?

[Steve Dunham]

Matt Pharr <mmp@graphics.stanford.edu> writes:

> David Moore <dmoore@UCSD.EDU> writes:
> > Jason L Tibbitts III <tibbs@hpc.uh.edu> writes:

> > > Unfortunately after I upgraded to 0.72 I found that after reading all of
> > >the articles in the group and calling gnus-group-get-new-news the unread
> > >article count would be some large number (like 1300, I assume the sum of
> > >(max - min article) for each group).  It seems that it's losing the list
> > >of read articles, even though a long list shows up in the group info.
> > ..

> > > Has anyone seen this?  I'm not submitting this as a bug report because I
> > >could be screwing something up and since I don't really know where to
> > >start
> > > looking it would be a pretty pitiful bug report.

> Another little piece of data about this: I raised the level of the groups
> that make up one of my virtual groups to 3 and started newsreading. When I
> started gnus, all was fine--there were a handful of new messages. After I
> read news in one of the virtual groups with new news and caught up, exited
> and returned to the Group buffer, all of the groups that made up the
> virutal group has a jazillion unread messages in them. So maybe it's a
> problem with how nnvirtual handles catching up...

This problem is still there in 0.74... I have one nnvirtual group made
up of two level 6 (unsubscribed) groups.  Whenever I catch up in the
nnvirtual group, it screws up the the read messages count in the
constituent groups.  i.e., 

 U      6207: comp.os.linux.development.apps
 U      1353: comp.os.linux.development.system
        4610: nnvirtual:comp.os.linux.development

("comp.os.linux.development.apps" 6
 ((1 . 23143))
 ((reply 28332 28811))) 

I change it to:

("comp.os.linux.development.apps" 6
 ((1 . 28811))
 ((reply 28332 28811)))

(and save with C-c C-c)
then do M-g on the virtual group and get:

 U       539: comp.os.linux.development.apps
 U      1353: comp.os.linux.development.system
        1768: nnvirtual:comp.os.linux.development

Then enter the virtual group and do a "c y" and we are back to

 U      6207: comp.os.linux.development.apps
 U      1353: comp.os.linux.development.system
        4613: nnvirtual:comp.os.linux.development

"G E" on the apps group gives:

("comp.os.linux.development.apps" 6
 ((1 . 23143))
 ((reply 28332 28811)))

Funny thing is, "c y" in the *Groups* buffer seems to work right...


Steve
dunham@gdl.msu.edu