How does expiry in mail groups work?

How does expiry in mail groups work?

[Kai Grossjohann]

You all know of the NNDB backend for Gnus.  Now, I have gotten fairly
far in doing all the necessary stuff in ELisp, but expiry is still
missing.  Right now, nndb.el uses the date of the article when doing
expiry.  I don't like this -- better would be the time since the
article was marked as expirable.

How is this done in the other backends?  Is the time of the article
being marked as expirable being used there?  How do these backends
keep track of said time?

I'd be rather grateful for any kind of advice you could give me.

Thanks a lot in advance,
kai
--
There ain't no cure for the summertime blues.

Re: How does expiry in mail groups work?

[Jack Vinson]

>>>>> "KG" == Kai Grossjohann <grossjoh@charly.informatik.uni-dortmund.de> writes:

KG> How is this done in the other backends?  Is the time of the article
KG> being marked as expirable being used there?  How do these backends
KG> keep track of said time?

Lars simply uses the time the article was snarfed into Gnus as the "time."
This is not optimal either, but I think he didn't want to add yet another
line with the expiry time in it.  Maybe this is something for Lobster Gnus?

-- 
Jack Vinson                                     Off www.cern.ch...
jvinson@cheux.ecs.umass.edu                     Through gopher.umn.edu
<http://www.cis.upenn.edu/~vinson/home.html>    Around ftp.indiana.edu
Sunderland, MA                                  Nothin' but 'Net - Pete Bretz

Re: How does expiry in mail groups work?

[Lars Magne Ingebrigtsen]

Jack Vinson <jvinson@cheux.ecs.umass.edu> writes:

> >>>>> "KG" == Kai Grossjohann <grossjoh@charly.informatik.uni-dortmund.de> writes:
> 
> KG> How is this done in the other backends?  Is the time of the article
> KG> being marked as expirable being used there?  How do these backends
> KG> keep track of said time?
> 
> Lars simply uses the time the article was snarfed into Gnus as the "time."
> This is not optimal either, but I think he didn't want to add yet another
> line with the expiry time in it.  Maybe this is something for Lobster Gnus?

Not even in Petunia Gnus, I think.  Hmn.  Whee!  This is already in
September Gnus, now that I think of it.  (Sort of.)

`nn*-request-update-mark' is called whenever an article mark is set.
If nndb wishes to use it to keep a database of when articles are
marked as expirable and use that date when expiring, that's totally up
to nndb.  

I guess that most people use `total-expiry' on groups, though, so
perhaps arrival date is better in any case.

-- 
  "Yes.  The journey through the human heart 
     would have to wait until some other time."

Re: How does expiry in mail groups work?

[Joe Hildebrand]

> "Lars" == Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

   Lars> `nn*-request-update-mark' is called whenever an article mark
   Lars> is set. If nndb wishes to use it to keep a database of when
   Lars> articles are marked as expirable and use that date when
   Lars> expiring, that's totally up to nndb.

nndb could add some commands to the NNTP::tng protocol to set and
retrieve modification times.  That would be fairly trivial.

How about:

TOUCH <number> to set the date associated with an article to now.
STAT  <number> to get the last date.  In what format should the date
      come back?

-- 
Joe Hildebrand                  Fuentez Systems Concepts
hildjj@fuentez.com              Lead Software Engineer
	"Breakfast recapitulates phylogeny" - Spider Robinson

Re: How does expiry in mail groups work?

[Kai Grossjohann]

>>>>> On 12 Apr 1996 14:59:00 -0600, Joe Hildebrand
>>>>> <hildjj@fuentez.com> said:

  Joe> STAT  <number> to get the last date.  In what format should the date
  Joe>       come back?

>From ELisp, it's easiest to process a date in ELisp internal format:
two numbers, as in the documentation of current-time.  OTOH, Lars has
written useful functions to parse stuff that looks like the output
from /usr/bin/date, so that would be OK, too.

kai
--
There ain't no cure for the summertime blues.