[??] delete garbage

[??] delete garbage

[Colin Rafferty]

I am constantly getting mail that I consider garbage, but my management
thinks is necessary.  I want to delete it automatically.

Currently, I get at it with nnmail-split-methods and dump it all in a
"garbage" group that has an expiration time of -1.  I then add a
function to `nnmail-post-get-new-mail-hook' that does a catchup and
expire on my garbage group.

My problem is that now I am getting a deluge of these messages.  Over
500 were waiting for me this morning, and I imagine I will receive over
a thousand more by noon.

Since I am using nnml, each time I `g'et new news, I have a hundred
little messages that say:

> Wrote /local/craffert/Mail/mail/garbage/2516
> Wrote /local/craffert/Mail/mail/garbage/2517
> Wrote /local/craffert/Mail/mail/garbage/2518

followed by a hundred more that say:

> Deleting article /local/craffert/Mail/mail/garbage/2516 in mail.garbage...
> Deleting article /local/craffert/Mail/mail/garbage/2517 in mail.garbage...
> Deleting article /local/craffert/Mail/mail/garbage/2518 in mail.garbage...

This takes too long (even on my local disk).

Does anyone know a way to trash these messages directly?  Would I have a
faster time if I used an nnfolder on that group?

Could someone help me?

-- 
Colin

Re: [??] delete garbage

[Kai Grossjohann]

>>>>> Colin Rafferty writes:

  Colin> I am constantly getting mail that I consider garbage, but my
  Colin> management thinks is necessary.  I want to delete it
  Colin> automatically. [...]

  Colin> Does anyone know a way to trash these messages directly?
  Colin> Would I have a faster time if I used an nnfolder on that
  Colin> group?

I think there's a way, but it's DANGEROUS!

You could (carefully!) set up an nnmail-split-methods that matches all
mails *except* the trash ones.  If there is no entry in
nnmail-split-methods for those mails, they will be quickly routed to
the bit bucket.  I think.

kai
-- 
Helmut Kohl, returning home from a visit to the US: "What's a
Dangaroo?" -- "Huh?  A Dangaroo?" -- "Yes, there must be many of them
-- signs saying `Dangaroos Crossing' all over the place; but I didn't
see a single one of them."

Re: [??] delete garbage

[Richard Pieri]

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

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

KG> You could (carefully!) set up an nnmail-split-methods that matches
KG> all mails *except* the trash ones.  If there is no entry in
KG> nnmail-split-methods for those mails, they will be quickly routed to
KG> the bit bucket.  I think.

A better, though untested, idea is to create a split method that matches
these messages, and make the directory where they will spool to a
symbolic link to /dev/null.

I do not know what will happen if you set the final, null string match
in nnmail-split-methods to a null string, but I think Gnus will not like
it.  Maybe Red should add a check for a symbol in the car of each split
method, something like 'garbage, and anything that matches the cdr is
"deleted with extreme prejudice".  Garbage checking should probably not
be allowed to match a null string (everything); much too dangerous, IMO.

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

iQCVAwUBMgi2Qp6VRH7BJMxHAQFS/QQAulIju4AI1a2hP3wOhEBvO3nh07AwTz+a
MF99f0TBTVAk53/qmONN0+y8/GqHLZqG+VaGhfiejEUNjbgfhi8yBDcyR0HLr1Jq
tUHR7irC3u/0PDsN2kONZzdjWGGGX7SxQWf0n1RrRkoYSv09HR7IEmuxMbZPeNgK
LGV5kAWnyWw=
=Y+Yz
-----END PGP SIGNATURE-----
-- 
Richard Pieri/Information Services \ When in doubt, cop an attitude. -A cat's
<ratinox@unilab.dfci.harvard.edu>   \ guide to life
http://www.dfci.harvard.edu/         \ 

Re: [??] delete garbage

[Colin Rafferty]

"Kai" == Kai Grossjohann writes:
"Colin" == Colin Rafferty writes:

Colin> I am constantly getting mail that I consider garbage, but my
Colin> management thinks is necessary.  I want to delete it
Colin> automatically. [...]

Colin> Does anyone know a way to trash these messages directly?
Colin> Would I have a faster time if I used an nnfolder on that
Colin> group?

Kai> I think there's a way, but it's DANGEROUS!

Kai> You could (carefully!) set up an nnmail-split-methods that matches all
Kai> mails *except* the trash ones.  If there is no entry in
Kai> nnmail-split-methods for those mails, they will be quickly routed to
Kai> the bit bucket.  I think.

I have done this.  In fact, I realized that I could do this pretty
easily by switching to `nnmail-split-fancy' (which seems faster than
nnmail-split-methods).  However, the function nnmail-split-methods
does not allow null lists.

For fancy splitting, it will post to the group list ("bogus"), and for
non-fancy splitting, it will just post to the last method (whether the
test matches or not).

By nature, I am a dangerous person.  If I want to shoot myself in the
foot, I want Gnus to help me do it (or at least let me set a variable
to allow me to do it).

I am cross-mailing this to gnus-bug.

Gnus v5.2.25; nntp 4.0; nnml 1.0; nnfolder 1.0
XEmacs 19.14 [Lucid] (sparc-sun-solaris2.4) of Sun Jun 23 1996 on xemacs

-- 
Colin

Re: [??] delete garbage

[Per Abrahamsen]

Colin Rafferty <craffert@spspme.ml.com> writes:

> For fancy splitting, it will post to the group list ("bogus"), and for
> non-fancy splitting, it will just post to the last method (whether the
> test matches or not).

If you had used a slightly older Gnus, the fancy mail split would
silently drop mail that didn't match any rule.  This was discovered by
accident by someone, and judged too dangerous a feature. 

I think ratinox' suggestion about special casing a `garbage' group
where mail will simply be dropped is better.

Re: [??] delete garbage

[Brad Howes]

>>>> Thus spake 'Richard Pieri (ratinox@unilab.dfci.harvard.edu)':

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

 KG> You could (carefully!) set up an nnmail-split-methods that matches
 KG> all mails *except* the trash ones.  If there is no entry in
 KG> nnmail-split-methods for those mails, they will be quickly routed to
 KG> the bit bucket.  I think.

 RP> A better, though untested, idea is to create a split method that matches
 RP> these messages, and make the directory where they will spool to a
 RP> symbolic link to /dev/null.
[snip]

I'm all for using elisp whenever possible, but for me its just too slow to slug
thru all of the mail I get in a 24 hour period. Situations like mine call for
procmail. With it, one can siphon off the crud from your incoming mail and save
it in a log file (which just may be /dev/null). Much faster than nnmail-split.

Brad
-- 
Brad Howes                          Motorola E-Mail ID: XBH001
EMT Development                     SMTP E-Mail: bhowes@cssun3.corp.mot.com
Motorola Corporate - MD H1780       Voice: 602 441 1522  Fax: 602 441 5455

Re: [??] delete garbage

[Jan Vroonhof]

Per Abrahamsen <abraham@dina.kvl.dk> writes:

> I think ratinox' suggestion about special casing a `garbage' group
> where mail will simply be dropped is better.

We need an 'nnblack-hole` back-end!

Jan

Re: [??] delete garbage

[François Pinard]

Colin Rafferty <craffert@spspme.ml.com> writes:

   My problem is that now I am getting a deluge of these messages.
   Over 500 were waiting for me this morning, and I imagine I will receive
   over a thousand more by noon.  Since I am using nnml, each time I
   `g'et new news, I have a hundred little messages that say:
   
   > Wrote /local/craffert/Mail/mail/garbage/2516
   > Wrote /local/craffert/Mail/mail/garbage/2517
   > Wrote /local/craffert/Mail/mail/garbage/2518
   
   followed by a hundred more that say:
   
   > Deleting article /local/craffert/Mail/mail/garbage/2516 in mail.garbage...
   > Deleting article /local/craffert/Mail/mail/garbage/2517 in mail.garbage...
   > Deleting article /local/craffert/Mail/mail/garbage/2518 in mail.garbage...
   
   This takes too long (even on my local disk).  Does anyone know a way
   to trash these messages directly?  Would I have a faster time if I
   used an nnfolder on that group?

If I had your problem, I would surely have procmail dumping these messages
even before Gnus sees them.

My setup here is: UUCP ---> sendmail ---> procmail ---> Gnus, and it
is very convenient for me.  Having sendmail calling procmail directly
reduces the overhead, and eases its usage site-wide.  I'm not asking
procmail to split my own mail spool on the local machine, as Gnus can
do it conveniently, and also because things are more simple when Gnus
gets all its spool from a single place.  However, procmail is very good
for many off-line duties, like the redirection of mail between sites,
including the /dev/null one! ;-)

-- 
François Pinard         ``Vivement GNU!''        pinard@iro.umontreal.ca
Support Programming Freedom, join our League!  Ask lpf@lpf.org for info!