auto-expiry and "d"

auto-expiry and "d"

[Michael Sperber [Mr. Preprocessor]]

The documentation is unclear if auto-expiry will affect only articles
actually really read or also those marked by
gnus-summary-mark-as-read-xxx.  The implementation says auto-expiry is
only for articles actually really read:

(defun gnus-summary-mark-as-read-forward (n)
  "Mark N articles as read forwards.
If N is negative, mark backwards instead.
The difference between N and the actual number of articles marked is
returned."
  (interactive "p")
  (gnus-summary-mark-forward n gnus-del-mark t))

This somehow is against my intuition.  If I mark an article as read
even without having actually really read it, I want all the
consequences associated with actual real reading to happen.

Bug or feature?

Cheers =8-} Mike

Re: auto-expiry and "d"

[Kai Grossjohann]

>>>>> On 02 Jul 1996 11:09:18 +0200,
>>>>> sperber@informatik.uni-tuebingen.de (Michael Sperber
>>>>> [Mr. Preprocessor]) said:

  Mike> The documentation is unclear if auto-expiry will affect only
  Mike> articles actually really read or also those marked by
  Mike> gnus-summary-mark-as-read-xxx. [...]

There is auto-expire, and then there's total-expire.  These two are
different:

auto-expire means that a message that you read is automatically marked
as expirable (`E').  (In any case, you can mark a message as expirable
by hitting `E'.)

total-expire means that messages that are marked as read (marked `r'
or `R') are treated as if they were expirable.

Both of these have their justificiation, but are also the subject of
much confusion among new users.  Please note that Gnus has started
being a news reader, and news articles behave such that they are
expired at some time by the news server.  The individual user has no
control over whether an article is expired or not, or when it is
expired.  The only thing a newsreader can do is to decide which
articles have already been seen by a user (`read' messages) and which
ones haven't (`unread' ones).

Now, for mail the situation is different:  The individual user has
control over whether messages are deleted, and the user should have
that control, too.  Therefore, a simple approach would be to let the
user mark articles as `expirable' (by hitting `E'), and letting Gnus
then delete them.  From there, you get to the current situation when
you add features.

I use the total-expire mechanism with some add-ons.  I decide between
mail groups that I treat like news groups (messages are usually read and
then disappear automatically from the disk), and those that I treat as
`important' mail (messages must be explicitly deleted to be deleted
from disk).  I do this by defining a function
kai-gnus-newsgroup-style-group like so:

(defun kai-gnus-newsgroup-style-group (groupname)
  (if (string-match "^nn\\(ml\\|db\\(\\+[a-z]*\\)?\\):" groupname)
      (if (string-match "^nn\\(ml\\|db\\(\\+[a-z]*\\)?\\):auto" groupname)
          t
        nil)
    t))

This basically says that mail groups are `important' unless their name
begins with `auto', then they are treated like news groups.  My mail
filtering mechanism puts all the mailing list mail into "nnml:auto.*"
groups.

I use this function for initially marking articles when I look at
them, like so:

(defun kai-gnus-mark-article-hook ()
  (if (and (not (or (memq gnus-current-article gnus-newsgroup-marked)
		    (memq gnus-current-article gnus-newsgroup-dormant)
		    (memq gnus-current-article gnus-newsgroup-expirable)
		    (memq gnus-current-article gnus-newsgroup-ancient)
                    (memq gnus-current-article gnus-newsgroup-reads))))
           (if (kai-gnus-newsgroup-style-group gnus-newsgroup-name)
               (gnus-summary-mark-unread-as-read)
             (gnus-summary-tick-article))))
(add-hook 'gnus-mark-article-hook 'kai-gnus-mark-article-hook)

I also use the function for defining my own functions for `n' and `p',
but that is just the icing on the cake.

You can use whichever part of this you like, or you might want to go
for a simpler mechanism.  After all, when you have such strange
behavior, you always have to be aware of what it does, otherwise you
will be quite surprised sometime...

I'm not sure if this could be made into a more general mechanism and
incorporated into Gnus.  As it stands it seems more or less like a
kludge.

kai
-- 
Life is hard and then you die.

Re: auto-expiry and "d"

[Michael Sperber [Mr. Preprocessor]]

>>>>> "kai" == Kai Grossjohann <grossjoh@ls6.informatik.uni-dortmund.de> writes:

>>>>> On 02 Jul 1996 11:09:18 +0200,
>>>>> sperber@informatik.uni-tuebingen.de (Michael Sperber
>>>>> [Mr. Preprocessor]) said:

Mike> The documentation is unclear if auto-expiry will affect only
Mike> articles actually really read or also those marked by
Mike> gnus-summary-mark-as-read-xxx. [...]

kai> There is auto-expire, and then there's total-expire.  These two are
kai> different:

kai> auto-expire means that a message that you read is automatically marked
kai> as expirable (`E').  (In any case, you can mark a message as expirable
kai> by hitting `E'.)

kai> total-expire means that messages that are marked as read (marked `r'
kai> or `R') are treated as if they were expirable.

That's not the issue, however.  I understand what total-expire and
auto-expire do, but the documentation for total-expire and auto-expire
doesn't explain what their definition of "read" is.  It could easily
be explained, and the confusion would go away.

Cheers =8-} Mike

Re: auto-expiry and "d"

[Richard Pieri]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "MS[P" == Michael Sperber [Mr Preprocessor]
>>>>> <sperber@informatik.uni-tuebingen.de> writes:

MS[P> That's not the issue, however.  I understand what total-expire and
MS[P> auto-expire do, but the documentation for total-expire and
MS[P> auto-expire doesn't explain what their definition of "read" is.
MS[P> It could easily be explained, and the confusion would go away.

- From the Gnus manual:

Read Articles
- -------------

   All the following marks mark articles as read.

`D'
     Articles that are marked as read.  They have a `r'
     (`gnus-del-mark') in the first column.  These are articles that the
     user has marked as read more or less manually.

`d'
     Articles that are actually read are marked with `R'
     (`gnus-read-mark').

`A'
     Articles that were marked as read in previous sessions are now
     "old" and marked with `O' (`gnus-ancient-mark').

`K'
     Marked as killed (`gnus-killed-mark').

`X'
     Marked as killed by kill files (`gnus-kill-file-mark').

`Y'
     Marked as read by having a too low score (`gnus-low-score-mark').

`C'
     Marked as read by a catchup (`gnus-catchup-mark').

`G'
     Canceled article (`gnus-cancelled-mark')

[...]

`E'
     You can also mark articles as "expirable" (or have them marked as
     such automatically).  That doesn't make much sense in normal
     groups, because a user does not control the expiring of news
     articles, but in mail groups, for instance, articles that are
     marked as "expirable" can be deleted by Gnus at any time.
     Expirable articles are marked with `E' (`gnus-expirable-mark').

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBMdkqF56VRH7BJMxHAQE6SwP8CV4LvGouS4q3FX77cm8zd+M8mhscIgpo
cTyoMtjhvQKuXcCWVTTo5F6YaYTjxJZGxpZ+FOHouHmu53dbwNudU6jD6TWn8Jfd
tnkevk9fZpvSHyUCufLCfHiHcvzF0uBoqcoltGYfmnncF6UxNxFF5MrecKlzqLb6
eiHVwVt3egs=
=uiZ2
-----END PGP SIGNATURE-----
-- 
Richard Pieri/Information Services \ Make your mark in the world, or at least
<ratinox@unilab.dfci.harvard.edu>   \ spray in each corner. -A cat's guide to
http://www.dfci.harvard.edu/         \ life

Re: auto-expiry and "d"

[Kai Grossjohann]

>>>>> On 02 Jul 1996 09:54:30 -0400, Richard Pieri
>>>>> <ratinox@unilab.dfci.harvard.edu> said:

  Richard> - From the Gnus manual:

  Richard> Read Articles
  Richard> - -------------

  Richard>    All the following marks mark articles as read. [...]

Maybe so.  But the point is that for auto-expirable groups, reading an
article (that is, selecting it when marked as unread) DOES NOT MARK
THE ARTICLE AS READ!  It marks the article as expirable instead.

Thus, the definition of reading an article in this context is: when an
article is unread (fresh, marked with ?\ (SPC), has never been seen
before, has all marks cleared with M-u, whatever) is selected (by
clicking mouse-2 in the summary buffer, by using `.' or `,' or `g', or
with `n', or with SPC) and thereby displayed in the article buffer.

I don't know how to express this in an understandable way. :-(

Am I wrong?
kai
-- 
Life is hard and then you die.

Re: auto-expiry and "d"

[Richard Pieri]

-----BEGIN PGP SIGNED MESSAGE-----

>>>>> "KG" == Kai Grossjohann <grossjoh@ls6.informatik.uni-dortmund.de>
>>>>> writes:

KG> Maybe so.  But the point is that for auto-expirable groups, reading an
KG> article (that is, selecting it when marked as unread) DOES NOT MARK
KG> THE ARTICLE AS READ!  It marks the article as expirable instead.

Actually, "E"xpirable is a "read" mark, just as lower-case "r"ead is a
"read" mark, which is why "E" is included in that section of the Gnus
manual (down at the bottom; maybe you missed it?).  Neither actually
mean that the message in question has actually been read by the reader,
only that they have been "marked as read" by the reader.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBMdlCT56VRH7BJMxHAQF+3gQAw9y80NfXb7y8A049fiZ9PpjLuKqm/N9H
rwFHamXyWNkmstd/FG11s585nIKYnQRYidSuxt/GnCKJZZj/n4HjtebRCZ1BV/g8
umdPjuFpd5bPIkaAkl4OBf57PRt1f1MgKRSXhesGPAXBZMCNWBStlpS5tym97lxC
mZLGGArW/Xo=
=0Grf
-----END PGP SIGNATURE-----
-- 
Richard Pieri/Information Services \ Variety is the spice of life: one day
<ratinox@unilab.dfci.harvard.edu>   \ ignore people, the next day annoy
http://www.dfci.harvard.edu/         \ them. -A cat's guide to life

Re: auto-expiry and "d"

[Darren/Torin/Who Ever...]

-----BEGIN PGP SIGNED MESSAGE-----

Kai Grossjohann, in an immanent manifestation of deity, wrote:
>There is auto-expire, and then there's total-expire.  These two are
>different:
>[...]
>total-expire means that messages that are marked as read (marked `r'
>or `R') are treated as if they were expirable.

Here's an interesting question that I've been pondering.  Is there a way
to convince Gnus to treat all articles in a mailgroup to be expirable?
I've pondered this and haven't come up with anything.  I currently have
a reaper perl script but it'd be nice to have Gnus handle it.

See, I'm on quite a few mailing lists and I don't always get to them
before they are way old.  If I could convince Gnus to go through and
just wipe out all mail messages that are older than 'nnmail-expiry-wait'
in any mailgroup with reaper-expire (or whatever) as a Group Parameter,
it would be wonderful.

Darren
- -- 
<torin@daft.com> <http://www.daft.com/~torin> <torin@debian.org> <torin@io.com>
Darren Stalder/2608 Second Ave, @282/Seattle, WA 98121-1212/USA/+1-800-921-4996
@ Do you have your clothes on? I probably don't. Take yours off. Feel better. @
@ Sysadmin, webweaver, postmaster for hire.  C/Perl/CGI programmer and tutor. @

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQCVAwUBMdm9sI4wrq++1Ls5AQGaUgP/Xb2DWZMzfJheydQzQ8av80kQKzsdOphQ
UCN7BJyRctWWCl/9sa0Zo960EqCeCADgkoAdqtw73DQSXIIbKmDJvaqTF+wwq+67
Dr37CexQwDoDxCHLPVkorNugzj5kq9Y9hU12RnKpuEA3Y/0oFbRy+qWc2XQM5BF6
GNT71oWcrXk=
=5+Ce
-----END PGP SIGNATURE-----

Re: auto-expiry and "d"

[Kai Grossjohann]

>>>>> On 02 Jul 1996 17:24:28 -0700, "Darren/Torin/Who Ever..."
>>>>> <torin@daft.com> said:

  Darren> Here's an interesting question that I've been pondering.  Is
  Darren> there a way to convince Gnus to treat all articles in a
  Darren> mailgroup to be expirable?  I've pondered this and haven't
  Darren> come up with anything.  I currently have a reaper perl
  Darren> script but it'd be nice to have Gnus handle it.

Not quite.  But you can make the group total-expire.  Then all you
need to do is to catch up the group.  All unread articles will be
marked as read, thus they will be expirable.

I think that comes close enough, don't you think?
kai
-- 
Life is hard and then you die.

Re: auto-expiry and "d"

[Darren/Torin/Who Ever...]

-----BEGIN PGP SIGNED MESSAGE-----

Kai Grossjohann, in an immanent manifestation of deity, wrote:
>Not quite.  But you can make the group total-expire.  Then all you
>need to do is to catch up the group.  All unread articles will be
>marked as read, thus they will be expirable.

The big problem with this is that I now have all this articles marked as
read that I may want to read later that aren't expired.    I guess I am
asking too much.  After all, I don't ask Gnus to expire off my unread
news articles.

Lars, would it be possible that something like this could be in Red
Gnus?  I could even wade into the Gnus source and try to do it myself
and submit the patch if you'd like.  (Once we get into the Red...)

Darren
- -- 
<torin@daft.com> <http://www.daft.com/~torin> <torin@debian.org> <torin@io.com>
Darren Stalder/2608 Second Ave, @282/Seattle, WA 98121-1212/USA/+1-800-921-4996
@ Do you have your clothes on? I probably don't. Take yours off. Feel better. @
@ Sysadmin, webweaver, postmaster for hire.  C/Perl/CGI programmer and tutor. @

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQCVAwUBMdqpuY4wrq++1Ls5AQHHqwQAoK/1Ta8mJKZBJjGrylVD0hgPhz5rSpNx
ZRgDVqySflqFLTqqrBINU3aGCmmEaRvhRyAauB8P2Hftb1VlnuKJ9OppJ3IQYLYD
z1+DIvhfVcuvGK9Kd6DwdT/U3M8+kuPfvroKHMpQtqJMBpFR3jYzRZBhOayWaWVP
eqtQxT31iOE=
=+XCG
-----END PGP SIGNATURE-----

Re: auto-expiry and "d"

[Lars Magne Ingebrigtsen]

"Darren/Torin/Who Ever..." <torin@daft.com> writes:

> The big problem with this is that I now have all this articles marked as
> read that I may want to read later that aren't expired.    I guess I am
> asking too much.  After all, I don't ask Gnus to expire off my unread
> news articles.

I'm not sure what you're asking here.  You asked for a way to expire
all read articles?  That's what `total-expire' does...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Re: auto-expiry and "d"

[Darren/Torin/Who Ever...]

-----BEGIN PGP SIGNED MESSAGE-----

Lars Magne Ingebrigtsen, in an immanent manifestation of deity, wrote:
>I'm not sure what you're asking here.  You asked for a way to expire
>all read articles?  That's what `total-expire' does...

No, what I'm looking for is a way to expire -unread- mail messages.  I'm
trying to extend news-spooling into mail even further.  I'd like there
to be a group parameter or group regexp that I can set so that mail
messages in certain groups will disappear after a time whether I've read
them or not.  This is similar to how the news spool will expire articles
whether you've read them or not.  I know this is even more dangerous
than total-expire.

For example:  There is no way that I can keep up with the cypherpunks
list.  But I enjoy looking at it sometimes.  Rather than going in and
doing a catchup every now and then, it'd be nice if mail messages older
than a week would just disappear out of the nnml directory.  Right now I
use a perl reaper script that does this at 5am.  I'd like Gnus to do it.

I'm even offerring to try coding this once we are in the Red again.
(Trying to find the time should be fun...)

Darren
- -- 
<torin@daft.com> <http://www.daft.com/~torin> <torin@debian.org> <torin@io.com>
Darren Stalder/2608 Second Ave, @282/Seattle, WA 98121-1212/USA/+1-800-921-4996
@ Do you have your clothes on? I probably don't. Take yours off. Feel better. @
@ Sysadmin, webweaver, postmaster for hire.  C/Perl/CGI programmer and tutor. @

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQCVAwUBMdt1YI4wrq++1Ls5AQG9gQQAjPDaTYYgtciHNF+n2EZTgf70zpXStw6o
06MQOEkwqxpXQndzi1XJbk40p/nKAQRk7AqCU3xE+2GTVS1NHTaMW8sFqYEmRiKl
aQjLxF1bZ8fI/yZuAAr1BlRjyBaWQ//ISjwMKn2Mb4qGfJGqSTITgthxoZAOVdYU
MD0qmGujzlo=
=EMG3
-----END PGP SIGNATURE-----

Re: auto-expiry and "d"

[Lars Magne Ingebrigtsen]

"Darren/Torin/Who Ever..." <torin@daft.com> writes:

> No, what I'm looking for is a way to expire -unread- mail messages. 

Oh, right.  Yes, ok.  I've added this to the Red Gnus todo list.  What
should the group parameter be called?  `absofucking-total-expire'?  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen

Re: auto-expiry and "d"

[Richard Pieri]

-----BEGIN PGP SIGNED MESSAGE-----

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

>> No, what I'm looking for is a way to expire -unread- mail messages. 
LMI> Oh, right.  Yes, ok.  I've added this to the Red Gnus todo list.  What
LMI> should the group parameter be called?  `absofucking-total-expire'?  :-)

This is something more than just expiry of read articles, it is the
nuking of unread articles as well... so call it 'nuke-unread'.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3
Charset: noconv

iQCVAwUBMd0U9p6VRH7BJMxHAQHSPAP/ay640IAxDKlZ6f4Mue9N+dIIHMrsltkn
cmYulRkqnFSqKkGGC2Gp4mggibL1+i6ZRAIU1Xl6Td0PgI0O7jyHmm4oSWeFYzUQ
h8AGeiTg77mPeC+lXP3JvNg7eEVWXMj6Fdd+u5Lnb/3HXI0qvoUXcuxEZvXpspSZ
+ZZC8F7ToCM=
=/7JM
-----END PGP SIGNATURE-----
-- 
Richard Pieri/Information Services \ Always give generously - a small bird or
<ratinox@unilab.dfci.harvard.edu>   \ rodent left on the bed tells them, "I
http://www.dfci.harvard.edu/         \ care". -A cat's guide to life

Re: auto-expiry and "d"

[Lars Magne Ingebrigtsen]

Richard Pieri <ratinox@unilab.dfci.harvard.edu> writes:

> This is something more than just expiry of read articles, it is the
> nuking of unread articles as well... so call it 'nuke-unread'.

Or rather `nuke-everything'.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Ingebrigtsen