Limiting, catching up and undo behaviour

Limiting, catching up and undo behaviour

[Raymond Scholz]

Hi!

Catching up articles after limiting the summary (`/ t') used to catch
up _all_ unread articles of that group regardless of visible or not.
I've now upgraded to Emacs 21 and now only visible articles are
catched up.  The doc string says:

,----
| c runs the command gnus-summary-catchup-and-exit
|    which is an interactive compiled Lisp function in `gnus-sum'.
| (gnus-summary-catchup-and-exit &optional ALL QUIETLY)
| 
| Mark all unread articles in this group as read, then exit.
| If prefix argument ALL is non-nil, all articles are marked as read.
| If QUIETLY is non-nil, no questions will be asked.
`----

Why did that behaviour change?

Maybe this is related:

After catching up I say `C-_' in the group buffer and for every second
try Gnus produces the following backtrace:

,----
| Debugger entered--Lisp error: (void-variable group)
|   (gnus-request-set-mark group (list (list ... ... ...)))
|   (progn (gnus-info-set-marks (quote ...) (quote ...) t) (gnus-info-set-read (quote ...) (quote ...)) (gnus-get-unread-articles-in-group (quote ...) (gnus-active "de.comm.infosystems.www.authoring.misc")) (gnus-request-set-mark group (list ...)) (gnus-group-update-group "de.comm.infosystems.www.authoring.misc" t))
|   (lambda nil (progn (gnus-info-set-marks ... ... t) (gnus-info-set-read ... ...) (gnus-get-unread-articles-in-group ... ...) (gnus-request-set-mark group ...) (gnus-group-update-group "de.comm.infosystems.www.authoring.misc" t)))()
|   gnus-undo(1)
|   call-interactively(gnus-undo)
`----

Cheers, Ray
-- 
WYSIWYG is a step backwards. Human labor is used to do that which the
computer can do better.                         (Andrew S. Tanenbaum)

Re: Limiting, catching up and undo behaviour

[Simon Josefsson]

Raymond Scholz <ray-2001@zonix.de> writes:

> Catching up articles after limiting the summary (`/ t') used to catch
> up _all_ unread articles of that group regardless of visible or not.
> I've now upgraded to Emacs 21 and now only visible articles are
> catched up.

Is it related to using Emacs 21?  It seems you are using Oort, so the
behaviour should be the same regardless of Emacs used.  Maybe
recompiling Gnus with Emacs 21 helps?  Do you have any lisp-shadows?

> After catching up I say `C-_' in the group buffer and for every second
> try Gnus produces the following backtrace:
>
> ,----
> | Debugger entered--Lisp error: (void-variable group)
> |   (gnus-request-set-mark group (list (list ... ... ...)))

It is not related, but I fixed this.

Re: Limiting, catching up and undo behaviour

[Raymond Scholz]

Simon Josefsson <jas@extundo.com> writes:

> Is it related to using Emacs 21?  It seems you are using Oort, so the
> behaviour should be the same regardless of Emacs used.  

I just reproduced the behaviour with an old Emacs 20.7 and identical
~/.gnus.  So it is surely not an Emacs 21 issue.

[later...]

Wow, I've traced it down to a recent change in my .gnus.  I thought,
I had used adaptive scoring all the time  but I didn't.  So I added 

(setq gnus-use-adaptive-scoring t)

to my .gnus.  With adaptive scoring turned _on_, only visible unread
articles are being catched up.  Huh?

Cheers, Ray
-- 
Sigmentation fault, core dumped.

Re: Limiting, catching up and undo behaviour

[Raymond Scholz]

Raymond Scholz <ray-2001@zonix.de> writes:

> With adaptive scoring turned _on_, only visible unread articles are
> being catched up.  Huh?

By (cluelessly) looking at the code it turns out to be a feature.  It
even makes sense, because catching up would alter the score of
invisible articles with adaptive scoring turned on, which contradicts
the DWIM principle Gnus is following up ;)

Cheers, Ray
-- 
Hfr fgebat rapelcgvba!