Including cached messages automatically?

Including cached messages automatically?

[Kai Großjohann]

We use an IMAP server with shared folders.  Ticking a message means it
is ticked for everyone.  We would like to use the tick mark as an
indication `this is a todo item for everybody'.

Caching an article is local to a user (I think), so we would like to
use cache marks as an indication for personal todo items (in shared
groups).

Hence, I do (add-hook 'gnus-summary-generate-hook
'gnus-summary-limit-include-cached), but this has strange effects.
The cached articles get inserted into the summary buffer twice, once
at the beginning and once at the end, and the order of articles is
reversed for one of them.

Am I doing something wrong?  Is this the wrong hook?  Does the
position of the cached articles depend on the position of point when
the function is called?  Strange...

kai
-- 
This gubblick contains many nonsklarkish English flutzpahs,
but the overall pluggandisp can be glorked from context. -- David Moser

Re: Including cached messages automatically?

[Simon Josefsson]

Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

> Hence, I do (add-hook 'gnus-summary-generate-hook
> 'gnus-summary-limit-include-cached), but this has strange effects.

`g-s-limit-include-cached' require a summary buffer to work, and the
summary buffer hasn't been generated yet when that hook is called.

(add-hook 'gnus-summary-prepared-hook 'gnus-summary-limit-include-cached)

Re: Including cached messages automatically?

[Kai Großjohann]

Simon Josefsson <jas@pdc.kth.se> writes:

> Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:
> 
> > Hence, I do (add-hook 'gnus-summary-generate-hook
> > 'gnus-summary-limit-include-cached), but this has strange effects.
> 
> `g-s-limit-include-cached' require a summary buffer to work, and the
> summary buffer hasn't been generated yet when that hook is called.
> 
> (add-hook 'gnus-summary-prepared-hook 'gnus-summary-limit-include-cached)

Hm.  I can't reproduce the strange `doubled message' behavior.  I
think it might have to do with messages which are both ticked and
cached -- maybe that's why they are included twice.  I'll check soon.

But the above line has a different problem: when entering a group with
SPC (auto-select-first is on), and if there are cached messages before
the first unread one, then point will be on the wrong message.  As an
example, suppose that there are some cached messages, then two old
(ancient) ones, and then comes the first unread message.  Then the
first unread message will be shown but point will be on the third line
in the summary buffer.  (Apparently because it would have been on the
third line if it weren't for the cached messages.)

Entering a group with RET, then using `.' to select the first message,
works fine.

PS: Why do I have messages which are both ticked and cached?  Well,
that's because I switched from gnus-use-cache equal t to
gnus-use-cache equal passive a short time ago.

kai
-- 
This gubblick contains many nonsklarkish English flutzpahs,
but the overall pluggandisp can be glorked from context. -- David Moser

Re: Including cached messages automatically?

[Kai Großjohann]

Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

> Hm.  I can't reproduce the strange `doubled message' behavior.  I
> think it might have to do with messages which are both ticked and
> cached -- maybe that's why they are included twice.  I'll check soon.

I couldn't reproduce this.  Hm.  Oh, well, maybe that means something
strange had happened before...

kai
-- 
This gubblick contains many nonsklarkish English flutzpahs,
but the overall pluggandisp can be glorked from context. -- David Moser

Re: Including cached messages automatically?

[Simon Josefsson]

Kai Großjohann <Kai.Grossjohann@CS.Uni-Dortmund.DE> writes:

> Hm.  I can't reproduce the strange `doubled message' behavior.  I
> think it might have to do with messages which are both ticked and
> cached -- maybe that's why they are included twice.  I'll check
> soon.

I did see the same thing, and edebugging explained it; it insert the
cached articles, and later the regular summary buffer stuff inserts a
second copy. I think.

It does have the unwanted cursor-movement you mentioned though. This
is just a shot in the dark, but does this work better?

(add-hook 'gnus-summary-prepared-hook 
	  (lambda ()
	    (save-excursion
	      (gnus-summary-limit-include-cached))))

Re: Including cached messages automatically?

[Kai Großjohann]

Simon Josefsson <jas@pdc.kth.se> writes:

> It does have the unwanted cursor-movement you mentioned though. This
> is just a shot in the dark, but does this work better?
> 
> (add-hook 'gnus-summary-prepared-hook 
> 	  (lambda ()
> 	    (save-excursion
> 	      (gnus-summary-limit-include-cached))))

Wonderful!  Thanks a bundle.

kai
-- 
because I couldn't think of a good beginning.

Re: Including cached messages automatically?

[Kai Großjohann]

Thanks to Simon, I now use the following code:

(add-hook 'gnus-summary-prepared-hook
          (lambda ()
            (save-excursion (gnus-summary-limit-include-cached))))

There's still a problem, though: threading is wrong.  For example,
have a look at the following summary buffer (gnu.emacs.gnus):

,-----
| O  [21-Nov,  1k: Rich Churcher  ] Altered headers in drafts
| O* [07-Oct,  0b: Henrik Holm    ] Re: Footnotes [Was: Re: Colors in the article buffer]
| O* [04-Oct,  0b: BrYan P. Johnso] gnus-pers.el -- an alternative to gnus-posting-styles
| O* [08-Sep,  0b: ralph.schleiche] Re: email attachments
| O* [09-Sep,  0b: Rob Browning   ] Re: does sendmail makes sent messages disappear in gnus?
| R      [21-Nov,  1k: Jochem Huhmann ] 
| R          [21-Nov,  1k: Rich Churcher  ] 
| R              [21-Nov,  1k: Jochem Huhmann ] 
`-----

When I entered the group (with SPC, auto-select-first is on), the
first message by Jochem was shown.  I then hit `^' on that message,
and the first message by Rich was added to the buffer.

Please note that the three messages marked `R' are not children of the
message by Rob.  Instead, they are children of the first message by
Rich!

Does this mean that there is a bug, or is it a pilot error (not at all
unlikely)? 

kai
-- 
because I couldn't think of a good beginning.

Re: Including cached messages automatically?

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> Please note that the three messages marked `R' are not children of the
> message by Rob.  Instead, they are children of the first message by
> Rich!

If you `j NUMBER', where NUMBER is the article number of the article
by Rob, does the same thing happen?  (That is, if you do this without
including the cached articles.)

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen