expiry in virtual groups

expiry in virtual groups

[Dan Christensen]

I have two groups nnfolder:Inbox and nnfolder:Sent, and a virtual
group nnvirtual:Both which encompasses the previous two.  None of the
groups use total-expire or auto-expire;  I manually mark articles with
`E'.  When gnus-group-expire-all-groups is run, or when
gnus-group-expire-articles is run on the virtual group, the E mark
disappears from all articles in the two component groups which were
marked as expirable, but which weren't old enough to be deleted.  Old
expirable articles are deleted as they should be.

This didn't happen until recently.  I am currently running gnus
from cvs.  The latest lisp/Changelog entry is dated 2000-01-09 22:52:35.
My previous gnus version was probably a few weeks old.

Any idea how to fix this?

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: expiry in virtual groups

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> I have two groups nnfolder:Inbox and nnfolder:Sent, and a virtual
> group nnvirtual:Both which encompasses the previous two.  None of the
> groups use total-expire or auto-expire;  I manually mark articles with
> `E'.  When gnus-group-expire-all-groups is run, or when
> gnus-group-expire-articles is run on the virtual group, the E mark
> disappears from all articles in the two component groups which were
> marked as expirable, but which weren't old enough to be deleted.  Old
> expirable articles are deleted as they should be.

I seem to be unable to reproduce this.  Is this bug still present in
Gnus? 

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

Re: expiry in virtual groups

[Dan Christensen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Dan Christensen <jdc@jhu.edu> writes:
> 
> > I have two groups nnfolder:Inbox and nnfolder:Sent, and a virtual
> > group nnvirtual:Both which encompasses the previous two.  None of the
> > groups use total-expire or auto-expire;  I manually mark articles with
> > `E'.  When gnus-group-expire-all-groups is run, or when
> > gnus-group-expire-articles is run on the virtual group, the E mark
> > disappears from all articles in the two component groups which were
> > marked as expirable, but which weren't old enough to be deleted.  Old
> > expirable articles are deleted as they should be.
> 
> I seem to be unable to reproduce this.  Is this bug still present in
> Gnus? 

It still happens consistently to me, with the 12:44 EDT version of
Gnus from cvs.  I tested it with a different nnfolder group, the only
member of a virtual group, and it happened there too.  In this second
case, the nnfolder group used auto-expire.

Dan

-- 
Dan Christensen
jdc@math.jhu.edu

Re: expiry in virtual groups

[Lars Magne Ingebrigtsen]

Dan Christensen <jdc@jhu.edu> writes:

> > > I have two groups nnfolder:Inbox and nnfolder:Sent, and a virtual
> > > group nnvirtual:Both which encompasses the previous two.  None of the
> > > groups use total-expire or auto-expire;  I manually mark articles with
> > > `E'.  When gnus-group-expire-all-groups is run, or when
> > > gnus-group-expire-articles is run on the virtual group, the E mark
> > > disappears from all articles in the two component groups which were
> > > marked as expirable, but which weren't old enough to be deleted.  Old
> > > expirable articles are deleted as they should be.
> > 
> > I seem to be unable to reproduce this.  Is this bug still present in
> > Gnus? 
> 
> It still happens consistently to me, with the 12:44 EDT version of
> Gnus from cvs.  I tested it with a different nnfolder group, the only
> member of a virtual group, and it happened there too. 

Are other marks lost as well, or is it just the `E' mark?

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

Re: expiry in virtual groups

[Paul Jarc]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> > > > the E mark disappears from all articles in the two component
> > > > groups which were marked as expirable, but which weren't old
> > > > enough to be deleted.
...
> Are other marks lost as well, or is it just the `E' mark?

FWIW, I have the same problem with my nnmaildir on Gnus 5.7.  Only E
is lost for me.


paul

Re: expiry in virtual groups

[Lars Magne Ingebrigtsen]

prj@po.cwru.edu (Paul Jarc) writes:

> > > > > the E mark disappears from all articles in the two component
> > > > > groups which were marked as expirable, but which weren't old
> > > > > enough to be deleted.
> ...
> > Are other marks lost as well, or is it just the `E' mark?
> 
> FWIW, I have the same problem with my nnmaildir on Gnus 5.7.  Only E
> is lost for me.

I think I've located and fixed this bug.  Could you check the version
in CVS in a few minuted to see whether things start working?

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

Re: expiry in virtual groups

[Paul Jarc]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> prj@po.cwru.edu (Paul Jarc) writes:
> > > > > > the E mark disappears from all articles in the two component
> > > > > > groups which were marked as expirable, but which weren't old
> > > > > > enough to be deleted.
> > ...
> > > Are other marks lost as well, or is it just the `E' mark?
> > 
> > FWIW, I have the same problem with my nnmaildir on Gnus 5.7.  Only E
> > is lost for me.
> 
> I think I've located and fixed this bug.  Could you check the version
> in CVS in a few minuted to see whether things start working?

I have no experience with CVS, so I haven't done this.  But the new
version of nnmaildir seems to have fixed it (I haven't touched Gnus),
so it was probably at least partly my (i.e., nnmaildir's) fault (in my
case, anyway, but not for the OP's).


paul

Re: expiry in virtual groups

[Dan Christensen]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> prj@po.cwru.edu (Paul Jarc) writes:
> 
> > > > > > the E mark disappears from all articles in the two component
> > > > > > groups which were marked as expirable, but which weren't old
> > > > > > enough to be deleted.
> > ...
> > > Are other marks lost as well, or is it just the `E' mark?
> > 
> > FWIW, I have the same problem with my nnmaildir on Gnus 5.7.  Only E
> > is lost for me.
> 
> I think I've located and fixed this bug.  Could you check the version
> in CVS in a few minuted to see whether things start working?

I finally got around to testing this, and it seems to work fine now.

[The problem was that E marks disappeared on articles in component
groups of a virtual group when gnus-group-expire-articles was run
on the virtual group and the articles weren't old enough to be 
expired.]

Dan

-- 
Dan Christensen
jdc@julian.uwo.ca

Re: Expiry in virtual groups

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > Could you check whether `nnvirtual-request-expire-articles' was
> > called, and if it was, what did it do?
> 
> As far as I can tell, it doesn't get called at all.

Hmm.  I've looked at the code that I think is supposed to do the
expiration, and I can't find any checks that would disable it when the
group is a virtual group...

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

Re: Expiry in virtual groups

[Stainless Steel Rat]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* Justin Sheehy <justin@linus.mitre.org>  on Mon, 19 Apr 1999
| Heh.  That's the only way that I expire.  I turn off the exit-hook to
| do expiry, which makes exiting groups much quicker.  About once a week
| I will C-c M-C-x before leaving my desk for the day.

I wrote my own exit hook function:

(defun rat:gnus-exit-gnus-hook ()
  (and (equal send-mail-function 'smtpmail-queue-it)
       (smtpmail-query-send-queued-mail-now))
  (cond ((equal gnus-select-method '(nnml ""))
         (if (not (boundp 'rat:gnus-group-expired-mail-this-session))
             (progn
               (gnus-group-expire-all-groups)
               (setq rat:gnus-group-expired-mail-this-session t))))
	(t
	 (if (not (boundp 'rat:gnus-group-expired-news-this-session))
             (progn
               (and (fboundp 'gnus-agent-expire) (gnus-agent-expire))
               (setq rat:gnus-group-expired-news-this-session t))))
	)
  (mc-deactivate-passwd nil)
  (kill-buffer (get-buffer-create "*Compile-Log*")))

In order,

* If I am using stmpmailq, query to send queued mail on exit.
* Expire nnml groups only if I am reading mail (I use separate newsrcs and
  backends for mail and news), then set a flag so I do not expire again
  that session.  I should probably use car, but the hack works as-is for me.
* Expire agentized articles only if the agent is being used, and set a flag 
  so that I do not expire again this session.
* Forcibly unset my GPG passphrase.
* Delete the *Compile-Log* buffer (I use gnus-compile frequently).
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.5 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE3G1R6gl+vIlSVSNkRAi5HAKDDx+hL5UnN51Of/snLrpwoJqs6GwCeI3k4
K2vaEiCfJvYjVHSQtsyd0mI=
=4Knw
-----END PGP SIGNATURE-----

-- 
Rat <ratinox@peorth.gweep.net>    \ Warning: pregnant women, the elderly, and
Minion of Nathan - Nathan says Hi! \ children under 10 should avoid prolonged
PGP Key: at a key server near you!  \ exposure to Happy Fun Ball.

Re: Expiry in virtual groups

[Justin Sheehy]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > Can't you just `C-c M-C-x' once in a while, like I do?  :-)
> 
> I can, but it's sort of yucky to have to do that.  I've never needed
> to do it so far.

Heh.  That's the only way that I expire.  I turn off the exit-hook to
do expiry, which makes exiting groups much quicker.  About once a week
I will C-c M-C-x before leaving my desk for the day.

Expiry never takes up my time.   :-)

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: Expiry in virtual groups

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Hrvoje Niksic <hniksic@srce.hr> writes:
> 
> > I tried exiting the nnvirtual group, and I don't think expiry really
> > happened.  At the least the familiar messages were missing.
> 
> Could you check whether `nnvirtual-request-expire-articles' was
> called, and if it was, what did it do?

As far as I can tell, it doesn't get called at all.

Re: Expiry in virtual groups

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> I tried exiting the nnvirtual group, and I don't think expiry really
> happened.  At the least the familiar messages were missing.

Could you check whether `nnvirtual-request-expire-articles' was
called, and if it was, what did it do?

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

Re: Expiry in virtual groups

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> I've now created a `nnvirtual-request-expire-articles' which simply
> runs the expiry process on all its component groups.  Could you test
> it after 0.83 has been released?  I don't have any nnvirtual groups
> that have expirable component groups...

I tried exiting the nnvirtual group, and I don't think expiry really
happened.  At the least the familiar messages were missing.

Re: Expiry in virtual groups

[Matt Pharr]

Hrvoje Niksic <hniksic@srce.hr> writes:
> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> 
> > Can't you just `C-c M-C-x' once in a while, like I do?  :-)
> 
> I can, but it's sort of yucky to have to do that.  I've never needed
> to do it so far.

One of my favorites in my .gnus.el file is the following, which I got off
of the ding list eons ago.  I forget who suggested it (Per A, maybe?).
Anyway, exiting mail groups is faster since it doesn't try to do any expiry 
then, instead periodically doing it as a demon.

(remove-hook 'gnus-summary-prepare-exit-hook 'gnus-summary-expire-articles)
(gnus-demon-add-handler 'gnus-group-expire-all-groups 30 30)

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

Re: Expiry in virtual groups

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > But `gnus-summary-expire-articles' takes into account group props
> > and stuff, and we don't want the group props from the virtual group
> > to adhere to the component groups, do we?
> 
> We don't.  But then again, maybe it really /should/ be an
> nnvirtual-specific kludge.

I've now created a `nnvirtual-request-expire-articles' which simply
runs the expiry process on all its component groups.  Could you test
it after 0.83 has been released?  I don't have any nnvirtual groups
that have expirable component groups...

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

Re: Expiry in virtual groups

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> It is not the case.  Expiry is done from
> `gnus-summary-prepare-exit-hook' (which defaults to
> `gnus-summary-expire-articles').  The obvious kludge would be to
> have the latter function also expire component groups if the current
> group is a virtual group.

I don't think that would be a kludge, it's exactly the right solution.
If you feel it breaks the Gnus-vs-backends abstraction, you could make 
a way for backends to specify some sort of expiry parameters.

> But `gnus-summary-expire-articles' takes into account group props
> and stuff, and we don't want the group props from the virtual group
> to adhere to the component groups, do we?

We don't.  But then again, maybe it really /should/ be an
nnvirtual-specific kludge.

> Can't you just `C-c M-C-x' once in a while, like I do?  :-)

I can, but it's sort of yucky to have to do that.  I've never needed
to do it so far.

Re: Expiry in virtual groups

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> Now, the question: each of these mail groups has its expiry stuff set.
> But I never enter any of them anymore because I read only the virtual
> one.  So when will the expiry process take place?  Ideally, when I
> exit the virtual groups, all the component groups should be run
> through expiry.  This might be the case; I just want to check that it
> is.

It is not the case.  Expiry is done from
`gnus-summary-prepare-exit-hook' (which defaults to
`gnus-summary-expire-articles').  The obvious kludge would be to have
the latter function also expire component groups if the current group
is a virtual group.

Uhm.  But where to do it...  I could add an
nnvirtual-request-expire-articles function, which could do its thing.
Hm.  Er.  Yes.  Probably.  But `gnus-summary-expire-articles' takes
into account group props and stuff, and we don't want the group props
from the virtual group to adhere to the component groups, do we?  That 
could be potentially catastrophic.

Hm.

Can't you just `C-c M-C-x' once in a while, like I do?  :-)

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

Expiry in virtual groups

[Hrvoje Niksic]

I'm using nnvirtual for the first time and I'm incredibly satisfied.
Specifically, I split two different things into two different
newsgroups, and then read them in a single virtual group (because the
articles in one are followups on those in the other).  The nicety of
this approach is that I can still view them separately, if I wish.

Now, the question: each of these mail groups has its expiry stuff set.
But I never enter any of them anymore because I read only the virtual
one.  So when will the expiry process take place?  Ideally, when I
exit the virtual groups, all the component groups should be run
through expiry.  This might be the case; I just want to check that it
is.