Gnus and NoCeM.

Gnus and NoCeM.

[Ken McGlothlen]

Well, I'm finally playing around with NoCeM.  I've always loved this idea in
principle, but just never enabled it.  So now I have.

What I love:  Dang.  Spam and velveeta just completely disappears in handfuls.
It's not perfect---I'm afraid that's going to have to wait for the long-awaited
gnus-ai.el---but it makes reading Usenet a lot more interesting.

What I hate:  My gawd, it takes forever.  Starting Gnus this morning took well
over an hour (on a lightly loaded 200MHz Pentium FreeBSD 3.2 box with a 384K
DSL link).

Some questions:

    1.	Is there a way to run NoCeM asynchronously?
    2.	Is there a way to shorten the process---perhaps a support program to
	prefetch and preprocess the NoCeM articles into a form that Gnus can
	process a lot faster?
    3.	Is there an easy way to divide up the gnus-group-get-new-news function
	into two functions:  gnus-group-get-new-news (which obeys NoCeM) and
	gnus-group-get-new-mail (which doesn't)?  (This doesn't matter if
	either 1 or 2 can be answered in the positive.)

However, I should point out that I'm gradually incorporating my email (finally)
into Gnus, and dang, I love this thing.  Only 80 more old mailfolders and a
LOT of procmail recipes to go.  :)

							---Ken

Re: Gnus and NoCeM.

[Colin Rafferty]

Ken McGlothlen writes:

>     1.	Is there a way to run NoCeM asynchronously?

(gnus-demon-add-nocem)

Since you have a DSL link, I asume you have a full-time connection, so 
the nocem can take place all day long (as long as emacs is up).

>     3.	Is there an easy way to divide up the gnus-group-get-new-news function
> 	into two functions:  gnus-group-get-new-news (which obeys NoCeM) and
> 	gnus-group-get-new-mail (which doesn't)?  (This doesn't matter if
> 	either 1 or 2 can be answered in the positive.)

I wrote this function to do exactly that.  I bind it to C-M-g.  Note
that this only works for nnml backends (though I'm sure it can be
easily converted to other nn*).

(defun gnus-group-get-new-nnml ()
  "Get new mail only."
  (interactive)
  (let ((gnus-group-new-mail-list nil)
	(method '(nnml ""))
	(nnml-prepare-save-mail-hook
	 (cons '(lambda ()
		  (setq gnus-group-new-mail-list
			(union gnus-group-new-mail-list
			       (mapcar 'car group-art))))
	       nnml-prepare-save-mail-hook)))
    (when (gnus-check-server method)
      (gnus-request-scan nil method))
    (while gnus-group-new-mail-list
      (let ((gnus-group-marked (mapcar '(lambda (group)
					  (concat "nnml:" group))
				       gnus-group-new-mail-list)))
	(setq gnus-group-new-mail-list (cdr gnus-group-new-mail-list))
	(gnus-group-get-new-news-this-group)))
    (when gnus-goto-next-group-when-activating
      (gnus-group-next-unread-group 1 t))
    (gnus-summary-position-point)
    (run-hooks 'gnus-after-getting-new-news-hook)
    (gnus-group-list-groups)))


> However, I should point out that I'm gradually incorporating my email (finally)
> into Gnus, and dang, I love this thing.  Only 80 more old mailfolders and a
> LOT of procmail recipes to go.  :)

Don't forget that your old folders can be just viewed with nnfolder,
and you can just continute using nnfolder as your backend.

If you want to actually convert them to nnml, you can `G f' the file
in *Group*, then in *Summary*, `C-c C-s C-n', followed by `9999 #',
followed by `B c'.

Of course, first you need to convert your procmail to nnmail-split.

-- 
Colin

Re: Gnus and NoCeM.

[Ken McGlothlen]

On 15 Jul 1999, Colin Rafferty provided the following enlightenment:

| Ken McGlothlen writes:
| 
| >     1.	Is there a way to run NoCeM asynchronously?
| 
| (gnus-demon-add-nocem)

Oh.  Wow.

One question, though.  Do I still have to set gnus-use-nocem?  If I can avoid
that looong startup NoCeM check, this solution would be *perfect*.

| I wrote this function to do exactly that.  I bind it to C-M-g.  [...]

Very nice.  I'm going to have to figure out how to adapt that to the
nnfolder back-end.

One more question:  How *much* faster is nnml than nnfolder?  And is there
a convenient way of converting nnml back to nnfolder or nnmbox format?

| Don't forget that your old folders can be just viewed with nnfolder, and
| you can just continute using nnfolder as your backend.

In fact, I've been relying on that very fact.

| Of course, first you need to convert your procmail to nnmail-split.

Actually, I'd prefer to run all this through procmail at the moment; keeps all
my mail processing in one place rather than two.  :)

Re: Gnus and NoCeM.

[Colin Rafferty]

Ken McGlothlen writes:
> On 15 Jul 1999, Colin Rafferty provided the following enlightenment:
> | Ken McGlothlen writes:

> | >     1.	Is there a way to run NoCeM asynchronously?

> | (gnus-demon-add-nocem)

> One question, though.  Do I still have to set gnus-use-nocem?  

Yes.

> If I can avoid that looong startup NoCeM check, this solution would
> be *perfect*.

Since you have a DSL line, can't you just keep your emacs job running
all the time, and always have the nocem stuff being updated (while
you're not doing real work)?

> One more question:  How *much* faster is nnml than nnfolder?  

I can't quantify, but I would have to say "much".  Entering a group is 
going to be much faster, since the .overview file is much smaller and
easier to parse than a mail folder.

I imagine that reading individual articles may be a little slower
since it's opening files, but you won't have this huge folder open in
your emacs consuming all that memory.

> And is there a convenient way of converting nnml back to nnfolder or
> nnmbox format?

Nothing convenient.  Only the hand-made way that I described in my
previous message.

-- 
Colin

Re: Gnus and NoCeM.

[Ken McGlothlen]

On 15 Jul 1999, Colin Rafferty wrote:

| > If I can avoid that looong startup NoCeM check, this solution would
| > be *perfect*.
| 
| Since you have a DSL line, can't you just keep your emacs job running all
| the time, and always have the nocem stuff being updated (while you're not
| doing real work)?

In theory.  However, my DSL line goes to my home machine, and I'd like to read
email from my workplace, so there's going to be some downtime somewhere.
However, I will be trying it at home tonight.

(Interestingly enough, the gnus-demon-add-nocem call didn't seem to do much; I
just tried another "g", and am going through the same loooong process.)

Re: Gnus and NoCeM.

[David Hedbor]

Ken McGlothlen <mcglk@itchy.serv.net> writes:

> In theory.  However, my DSL line goes to my home machine, and I'd like to read
> email from my workplace, so there's going to be some downtime somewhere.
> However, I will be trying it at home tonight.

I run Gnus in XEmacs with one frame displayed at home and one at
work. Ie GNU Client rocks. :)

 
> (Interestingly enough, the gnus-demon-add-nocem call didn't seem to do much; I
> just tried another "g", and am going through the same loooong process.)

-- 
[ Below is a random fortune, which is unrelated to the above message. ]
Have a nice day!

Re: Gnus and NoCeM.

[Dan Christensen]

Colin Rafferty <craffert@ms.com> writes:
> > One more question:  How *much* faster is nnml than nnfolder?  
> 
> I can't quantify, but I would have to say "much".  Entering a group is 
> going to be much faster, since the .overview file is much smaller and
> easier to parse than a mail folder.

I've always wondered how hard it would be to add .overview files to
the nnfolder backend.  So, how hard would it be?

Dan

Re: Gnus and NoCeM.

[Justin Sheehy]

Dan Christensen <jdc@jhu.edu> writes:

> I've always wondered how hard it would be to add .overview files to
> the nnfolder backend.  So, how hard would it be?

It is not that it would be hard, but that it wouldn't buy you much of
anything.  Sure, group entry speed would increase a lot... until you
selected an article.  Then the slowdown that usually happens when
entering an nnfolder group would have to happen.

The .overview file doesn't let you get around the fact that with
nnfolder you must read and parse a file that grows with the number of
articles in the group.

There are conceivable ways that one might make it a bit faster, but
certainly just adding .overview files wouldn't get around the thing
that makes nnfolder slow.

-Justin

 

Ugh! Ack! Argh! Nggh!

[Ken McGlothlen]

Gnus has really scared me a few times.

Tonight, for example.  I've spent the last four days moving 140MB of old mail
from VM to Gnus nnml folders, writing procmail recipes, and so forth.  I still
haven't transferred my email delivery to the new system (I have over 3000
messages piled up waiting for Gnus to start running smoothly).  And tonight, I
get home to convert the last bit of mail, start up Gnus, and get

	Wrong type argument, integerp

Heart attack.  Complete heart attack.

Yet, the second time I tried it (identical circumstances---both times, I'd
started Emacs afresh), it worked fine.

But I'm a bit shaky now.

Here's another thing that concerned me.  I was dealing with some messages in my
"main" mailbox that I hadn't yet written a procmail recipe for.  I got to a
message, hit B m and moved it to a nnml folder that didn't exist.  Sure enough,
Gnus asked me if I wanted to create it, I said "yes", and the message got moved
and marked "G".

So where's the group?  Ain't there.  I even left Emacs and restarted from the
ground up, and it's not there.  And yet, it's in the ~/Mail directory; just not
in the .newsrc.eld file.  Why not?  And what happens if I create it now?  Will
it erase the old stuff?

The *promise* is that it's going to get a lot better.

Also, one question:  What's a quick way to take an nnml folder and convert it
to an mbox-compliant file so that I can reprocess it with procmail?  Or even a
slow way?

Whee!  The adventure continues.

Re: Ugh! Ack! Argh! Nggh!

[Hans de Graaff]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> > Also, one question: What's a quick way to take an nnml folder and
> > convert it to an mbox-compliant file so that I can reprocess it with
> > procmail?  Or even a slow way?
> 
> % cat nnml-folder-dir/* > mboxfile
> Or have I missed something vital?

Yes, nnml messages don't contain a From_ line at the start, but an
mbox style file would need those. Your solution would create a file
with one very large message in it.

Better to create an nnmbox folder in Gnus, and respool the messages
from the nnml folder to the nnmbox folder with B c.

Hans

Re: Ugh! Ack! Argh! Nggh!

[Michael Cook]

Hans de Graaff <graaff@xs4all.nl> writes:

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
> 
> > > Also, one question: What's a quick way to take an nnml folder and
> > > convert it to an mbox-compliant file so that I can reprocess it with
> > > procmail?  Or even a slow way?
> > 
> > % cat nnml-folder-dir/* > mboxfile
> > Or have I missed something vital?
> 
> Yes, nnml messages don't contain a From_ line at the start, but an
> mbox style file would need those. Your solution would create a file
> with one very large message in it.

I think gnus automatically detects if the mbox file is using MMDF
format.

$ cd nnml-folder-dir
$ for i in $(ls|sort -n);do echo -e '\1\1\1\1';cat $i;done>../mboxfile

Of course, that assumes that none of your messages already contains
four ^A characters alone on a line.

M.

Re: Ugh! Ack! Argh! Nggh!

[Kjetil Ødegaard]

* Jonas Steverud
|
| Ken McGlothlen <mcglk@itchy.serv.net> writes:
| 
| [...]
| 
| > Also, one question: What's a quick way to take an nnml folder and
| > convert it to an mbox-compliant file so that I can reprocess it with
| > procmail?  Or even a slow way?
| 
| % cat nnml-folder-dir/* > mboxfile
| 
| Or have I missed something vital?

Hm.  Shouldn't you just do `F' in the group buffer?

-- 
Gravity brings me down.

Re: Ugh! Ack! Argh! Nggh!

[Kai Großjohann]

Jonas Steverud <d4jonas@dtek.chalmers.se> writes:

> Ken McGlothlen <mcglk@itchy.serv.net> writes:
> 
> [...]
> 
> > Also, one question: What's a quick way to take an nnml folder and
> > convert it to an mbox-compliant file so that I can reprocess it with
> > procmail?  Or even a slow way?
> 
> % cat nnml-folder-dir/* > mboxfile
> 
> Or have I missed something vital?

I think you missed two things: the From_ lines and the file order.
The From_ lines thing is probably vital...

Another possibility would be to use Gnus: create an nnfolder group and
move all nnml messages there.  Take care that the nnml backend
(server) and the nnfolder backend (server) are not using the same
active file, though.

kai
-- 
Life is hard and then you die.

Re: Ugh! Ack! Argh! Nggh!

[Harry Putnam]

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

> Jonas Steverud <d4jonas@dtek.chalmers.se> writes:
> 
> > Ken McGlothlen <mcglk@itchy.serv.net> writes:
> > 
> > [...]
> > 
> > > Also, one question: What's a quick way to take an nnml folder and
> > > convert it to an mbox-compliant file so that I can reprocess it with
> > > procmail?  Or even a slow way?
> > 
> > % cat nnml-folder-dir/* > mboxfile
> > 
> > Or have I missed something vital?
> 
> I think you missed two things: the From_ lines and the file order.
> The From_ lines thing is probably vital...
> 
> Another possibility would be to use Gnus: create an nnfolder group and
> move all nnml messages there.  Take care that the nnml backend
> (server) and the nnfolder backend (server) are not using the same
> active file, though.


I think he just wants a spool file to run through procmail again.
In that case:  cd to Mail/nnmlgroup  then run:

sed 's/X-From-Line:/From/' `ls [0-9]*` >MBOX_OUT
NOTE: The second set of single quotes are back ticks.. not single
quotes : )

Gnus removes the "From " from the original "From line" and replaces it
with "X-From-Line:".  So let sed remove it and reinsert "From"

Test the new file with `mutt -f MBOX_OUT'. Or in gnus group buffer do `G
m f' (make nndoc group),to see if the new file is properly formatted,
if so ...  Run it back through procmail.

Re: Ugh! Ack! Argh! Nggh!

[Kai Großjohann]

Ken McGlothlen <mcglk@itchy.serv.net> writes:

> 	Wrong type argument, integerp
> 
> Heart attack.  Complete heart attack.

Type M-x toggle-debug-on-error RET then try to reproduce the problem.
The resulting backtrace will be more informative than the short message.

> So where's the group?  Ain't there.

Try hitting `F' to tell Gnus to look for it.  Maybe the group is
killed?  Type `A k' for a list of killed groups.  Maybe the group is a
zombie?  Type `A z' for a list of zombie groups.

kai
-- 
Life is hard and then you die.

Re: Ugh! Ack! Argh! Nggh!

[Ken McGlothlen]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=X-UNKNOWN, Size: 745 bytes --]



On 17 Jul 1999, Kai [iso-8859-1] Großjohann wrote:

| Type M-x toggle-debug-on-error RET then try to reproduce the problem.
| The resulting backtrace will be more informative than the short message.

Yes, of course.  I panicked, is all.  I'll do that next time.

| > So where's the group?  Ain't there.  [After creating one with a "B m" to a
| > nonexistent group]
| 
| Try hitting `F' to tell Gnus to look for it.  Maybe the group is
| killed?  Type `A k' for a list of killed groups.  Maybe the group is a
| zombie?  Type `A z' for a list of zombie groups.

None of the above works.  But subsequently creating the group does work, so
there's at least that.  I do wish it would create a new group properly in the
Group buffer, though.  :)

Re: Ugh! Ack! Argh! Nggh!

[Kai Großjohann]

Ken McGlothlen <mcglk@itchy.serv.net> writes:

> [...] I do wish it would create a new group properly in the Group
> buffer, though.  :)

`G m'.

kai
-- 
Life is hard and then you die.

Re: Ugh! Ack! Argh! Nggh!

[Jack Vinson]

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

KG> Ken McGlothlen <mcglk@itchy.serv.net> writes:
>> [...] I do wish it would create a new group properly in the Group
>> buffer, though.  :)

KG> `G m'.

But his original point was that when Gnus asked him about creating a group,
it should _create the group_.  I just 'B c' on Kai's article and answered
with a new group name.  The following messages popped into the minibuffer:

Creating mail directory w:/jmvins/private/Mail/jack_is_great/
Wrote w:/jmvins/private/Mail/jack_is_great/1

One would expect that group to be created, no?  If I look at my *Group*
buffer, it does not exist.  If I 'A z' or 'A k' it does not show up.
However, when I quit and restart gnus, the group then becomes a zombie
group.  (I didn't try to obvious 'A M' to find a matching group.)

-- 
Jack Vinson

Re: Ugh! Ack! Argh! Nggh!

[Hrvoje Niksic]

Jack Vinson <vinson@unagi.cis.upenn.edu> writes:

> But his original point was that when Gnus asked him about creating a
> group, it should _create the group_.  I just 'B c' on Kai's article
> and answered with a new group name.  The following messages popped
> into the minibuffer:
> 
> Creating mail directory w:/jmvins/private/Mail/jack_is_great/
> Wrote w:/jmvins/private/Mail/jack_is_great/1
> 
> One would expect that group to be created, no?  If I look at my
> *Group* buffer, it does not exist.

Try pressing `F' -- the group should spring into existence.

Re: Ugh! Ack! Argh! Nggh!

[Paul Stevenson]

Hrvoje Niksic <hniksic@srce.hr> writes:

> Jack Vinson <vinson@unagi.cis.upenn.edu> writes:
> 
> > But his original point was that when Gnus asked him about creating a
> > group, it should _create the group_.  I just 'B c' on Kai's article
> > and answered with a new group name.  The following messages popped
> > into the minibuffer:
> > 
> > Creating mail directory w:/jmvins/private/Mail/jack_is_great/
> > Wrote w:/jmvins/private/Mail/jack_is_great/1
> > 
> > One would expect that group to be created, no?  If I look at my
> > *Group* buffer, it does not exist.
> 
> Try pressing `F' -- the group should spring into existence.

I just tried this (actually I `B m'ed some articles from one nnml group
to another which didn't yet exist). Gnus asked me if I wanted to create
the group so I said yes, and I saw the messages like Jack
indicated. 

Now I press `F' but it does not let me see the group. `j' followed by
the new group's name does, and shows that it was killed, so I subscribe
to it.

Now the new group /looks/ like any other, but I can't enter it (can't
select group). It's an nnml group, so I do nnml-generate-nov-databases
and that thinks for a while. Now the directory ~/Mail/mynewgroup has the
messages and an .overview file but still I can't select the group. 

What magic is needed ? (I know that if I restart gnus something occurs
which lets me read the group)  As Jack suggests, shouldn't creating a
new group default to subscribing to it, too?

Paul

Re: Ugh! Ack! Argh! Nggh!

[Paul Stevenson]

Paul Stevenson <spaul@mail.phy.ornl.gov> writes:

> [tale of new nnml group woe]
> 
> What magic is needed ? 

Ah, it's to check for new news.

Re: Ugh! Ack! Argh! Nggh!

[François Pinard]

Paul Stevenson <spaul@mail.phy.ornl.gov> writes:

> What magic is needed ?

In my opinion, a real bug appeared in that area, relatively recently.

Moved articles do not always show, even when moved into already existing
group, until Gnus is restarted.  I reported this to Lars, a few weeks ago.

-- 
François Pinard   http://www.iro.umontreal.ca/~pinard

Re: Ugh! Ack! Argh! Nggh!

[Robert Bihlmeyer]

Hi,

>>>>> On 26 Jul 1999 13:45:23 -0400
>>>>> Paul Stevenson <spaul@mail.phy.ornl.gov> said:

 Paul> Now I press `F' but it does not let me see the group.

What is `gnus-subscribe-newsgroup-method' set to?

        Robbe

-- 
Robert Bihlmeyer	reads: Deutsch, English, MIME, Latin-1, NO SPAM!
<robbe@orcus.priv.at>	<http://stud2.tuwien.ac.at/~e9426626/sig.html>

Re: Ugh! Ack! Argh! Nggh!

[Paul Stevenson]

Robert Bihlmeyer <e9426626@stud4.tuwien.ac.at> writes:

> Hi,
> 
> >>>>> On 26 Jul 1999 13:45:23 -0400
> >>>>> Paul Stevenson <spaul@mail.phy.ornl.gov> said:
> 
>  Paul> Now I press `F' but it does not let me see the group.
> 
> What is `gnus-subscribe-newsgroup-method' set to?

gnus-subscribe-alphabetically

Re: Gnus and NoCeM.

[Kai Großjohann]

Justin Sheehy <dworkin@ccs.neu.edu> writes:

> There are conceivable ways that one might make it a bit faster, but
> certainly just adding .overview files wouldn't get around the thing
> that makes nnfolder slow.

One could include the message length in the .overview file, then Gnus
could add up the lengths of the messages preceding the desired one to
quickly find the right offset in the file to go to.

Would that be a workable solution?

The problem of having to load the entire nnfolder file into memory for
editing and stuff still remains, though.  Hm.

kai
-- 
Life is hard and then you die.

Re: Gnus and NoCeM.

[Sudish Joseph]

Colin Rafferty writes:
>> One question, though.  Do I still have to set gnus-use-nocem?  

> Yes.

[ Apologies if I misunderstood the question. ]

AFAIK, once the daemon is initialized, gnus-use-nocem can be nil,
allowing purely asynch operation.  Here's what I use:

;; NoCem
(setq gnus-use-nocem t
      gnus-nocem-groups '("news.lists.filters"
                          "news.admin.net-abuse.bulletins"
                          "news.admin.net-abuse.announce"))

;; daemons
(setq gnus-use-demon t)
(gnus-demon-add-handler 'gnus-demon-scan-nocem   30 10)
[...]
(gnus-demon-init)
(setq gnus-use-nocem nil)

I'm almost certain that setting gnus-use-nocem to t before setting the 
nocem demon handler is unnecessary, but I haven't been motivated to
see if I can remove it.


The only other demon I use is this one, which causes incoming mail to
be split when my XEmacs is otherwise idle.  Very useful if you get
lots of incoming mail and you do not want to wait on "g."  Note that
it assumes that all mail groups are level 2 or better.

(gnus-demon-add-handler 'sj/gnus-demon-scan-news 20 5)
(defun sj/gnus-demon-scan-news ()
  (when (gnus-alive-p)
    (save-window-excursion
      (save-excursion
        (set-buffer gnus-group-buffer)
        (gnus-group-get-new-news 2)))))


-- 
Sudish Joseph                                          MindSpring Enterprises

Re: Gnus and NoCeM.

[Justin Sheehy]

Ken McGlothlen <mcglk@itchy.serv.net> writes:

> One more question:  How *much* faster is nnml than nnfolder? 

When entering reasonbly large groups, the difference is very large.

> And is there a convenient way of converting nnml back to nnfolder or
> nnmbox format?

Sure.  Create a group of the new type, process-mark all of the
articles in the old group with M P a, and B m or B c them into the new 
group.  Voila.  :-)

-Justin

 

Re: Gnus and NoCeM.

[Andrew Hobson]

On Thu, 15 Jul 1999 12:32:31 -0700 (PDT), Ken McGlothlen <mcglk@itchy.serv.net> said:

> One question, though.  Do I still have to set gnus-use-nocem?  If I can avoid
> that looong startup NoCeM check, this solution would be *perfect*.

I don't think so.  At least, I have a similar setup and I have
gnus-use-nocem set to nil.

Drew

Re: Gnus and NoCeM.

[Ken McGlothlen]

On 16 Jul 1999, Andrew Hobson wrote:

| On Thu, 15 Jul 1999 12:32:31 -0700 (PDT), Ken McGlothlen <mcglk@itchy.serv.net> said:
| 
| > One question, though.  Do I still have to set gnus-use-nocem [with
| > gnus-demon-add-nocem]?
|
| I don't think so.  At least, I have a similar setup and I have
| gnus-use-nocem set to nil.

Well, currently I have both

	(gnus-demon-add-nocem)
	(setq gnus-use-nocem t)

in my .gnus.el file.  I haven't honestly noticed any difference---long ramp-up
time for both startup and for "g" persist.

Hm.

Re: Gnus and NoCeM.

[Jack Twilley]

>>>>> "Ken" == Ken McGlothlen <mcglk@itchy.serv.net> writes:

[...]

    Ken> Well, currently I have both

    Ken> 	(gnus-demon-add-nocem) (setq gnus-use-nocem t)

    Ken> in my .gnus.el file.  I haven't honestly noticed any
    Ken> difference---long ramp-up time for both startup and for "g"
    Ken> persist.

I had a similar setup to this.  However, "long ramp-up time" is a kind
way to describe the effect this had on my system.  I have a cablemodem
and an occasionally-wonky provider.  It took over an hour for pgnus to
grab all the NoCeM crap at startup, and the process was consuming over
a hundred megs of RAM.  In fact, several of my processes died with bus
errors while pgnus was sucking up resources.  I've disabled it for now
with the hopes that maybe an upcoming revision will work a lot better.

    Ken> Hm.

Jack.
(whee!  bricktext.)
-- 
Jack Twilley
jmt at nycap dot rr dot com
http colon slash slash jmt dot dhs dot org slash tilde jmt slash

Re: Gnus and NoCeM.

[Kai Großjohann]

Ken McGlothlen <mcglk@itchy.serv.net> writes:

> However, I should point out that I'm gradually incorporating my
> email (finally) into Gnus, and dang, I love this thing.  Only 80
> more old mailfolders and a LOT of procmail recipes to go.  :)

Gnus can interoperate well with procmail.  See the mail-sources
variable (documented in the `Mail Source Specifications' node in the
pgnus info file).

The idea is to tell procmail to write mail to ~/spool/foo.bar.in, say,
and Gnus then puts all mails from that file into the nnml:foo.bar
group (works with all backends, not just nnml).

kai
-- 
Life is hard and then you die.

Re: Gnus and NoCeM.

[Wes Hardaker]

>>>>> On Thu, 15 Jul 1999 10:59:42 -0700 (PDT), Ken McGlothlen <mcglk@itchy.serv.net> said:

Ken> What I hate: My gawd, it takes forever.  Starting Gnus this
Ken> morning took well over an hour (on a lightly loaded 200MHz
Ken> Pentium FreeBSD 3.2 box with a 384K DSL link).

I used NoCeM a year or two ago (now I have a news server that wipes
spam on their end, which is even cooler).  The long startups aren't as 
bad after the first run, right?  The first one takes the longest if I
remember.  And backgrounding the demon function, that has been pointed 
out, is the other really really good way to go.

On a side note, you can also speed it up if you turn off the pgp
verification (which could be hazardous if anyone posts a NoCeM article 
which isn't a valid one).  I didn't have pgp when I first started
using it, and there was a definite slow down afterwards.

-- 
"Ninjas aren't dangerous.  They're more afraid of you than you are of them."