nnagent

nnagent

[Greg Stark]

This sounds pretty cool. 

Can i read my mail groups through nnagent?

That may sound pretty silly, but imagine i keep my mail on the network
filesystem, but occasionally i want to check my mail on my laptop. So i want
something that can keep and manage a cache of messages from the mail group.

I'm not really sure how this would work for reading new mail, should it accept
the new mail into the mail group then "download" it into the nnagent group,
for disconnected operation? Then i can't read my new mail unless i have at
least temporary access to the mail group. It would be really nice to just
accept the mail from the spool or pop server into nnagent and mark them
specially in the nnagent group to be uploaded the next time we're connected.

This gets into some sticky questions about the two .newsrc.eld's managing the
mail groups on disk, but maybe there's a way to use the slave code to get that
right? 

greg

Re: nnagent

[Lars Magne Ingebrigtsen]

Greg Stark <gsstark@mit.edu> writes:

> Can i read my mail groups through nnagent?

Yes; any Gnus backend can be agentized.  (Even nnweb, I think.)

> I'm not really sure how this would work for reading new mail, should
> it accept the new mail into the mail group then "download" it into
> the nnagent group, for disconnected operation? Then i can't read my
> new mail unless i have at least temporary access to the mail
> group. It would be really nice to just accept the mail from the
> spool or pop server into nnagent and mark them specially in the
> nnagent group to be uploaded the next time we're connected.

Uhm -- from the pop server, into the agent, and from there into nnml?
Wouldn't that be a waste of perfectly good bauds?

I think they way to do this is to write a script that tells nnml (on
the other side of the line) to snarf in mail.  Then the Gnus agent on
this side can fetch them over the line.

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

Re: nnagent

[Greg Stark]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> Greg Stark <gsstark@mit.edu> writes:
> > Can i read my mail groups through nnagent?
> Uhm -- from the pop server, into the agent, and from there into nnml?
> Wouldn't that be a waste of perfectly good bauds?
> 
> I think they way to do this is to write a script that tells nnml (on
> the other side of the line) to snarf in mail.  Then the Gnus agent on
> this side can fetch them over the line.

Well, imagine i have set the mail to forward somewhere convenient, or i don't
have access to run a script unattended on the server. 

I think what i'm asking for is support for accept-article in general. This
would let new mail articles to be accepted, it would let me move or copy
articles into an nnagent group, and have the articles be uploaded to the real
groups when possible. It would also be possible to support other write
operations like  editting articles for example.

I suppose this is one of those things that would be nice, but you ought to
implement the basic reading operations first :)

greg

Re: nnagent

[Lars Magne Ingebrigtsen]

Greg Stark <gsstark@mit.edu> writes:

> Well, imagine i have set the mail to forward somewhere convenient,
> or i don't have access to run a script unattended on the server.

If you have the mail forwarded somewhere, surely you must have a way
of fetching the mail from that somewhere to wherever you want it to be
online?

Uhm.  Perhaps if the "wherever" is a Windoze box, that might be a
problem... 

> I think what i'm asking for is support for accept-article in general. This
> would let new mail articles to be accepted, it would let me move or copy
> articles into an nnagent group, and have the articles be uploaded to the real
> groups when possible.

I think this is beyond the scope of offline support.  Writing scripts
to do these things shouldn't be difficult, though -- the Gnus agent
stores the articles in nnml (ie., nnspool) directories, so the scripts
just have to traverse them, pack up new messages and ship them off to
wherever.

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

Re: nnagent

[Lars Magne Ingebrigtsen]

Stainless Steel Rat <ratinox@peorth.gweep.net> writes:

> The way that agent becomes a win is to queue up outgoing mail
> locally (which smtpmail and feedmail already accomplish) assuming
> that it will be made to do this.

Yes, it will.  We'll also have an "outgoing" group where you can do
message queue maintenance (eg., cancel, editing, remove (temporarily)
from queue, etc).  

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

Re: nnagent

[Steinar Bang]

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

> Stainless Steel Rat <ratinox@peorth.gweep.net> writes:
>> The way that agent becomes a win is to queue up outgoing mail
>> locally (which smtpmail and feedmail already accomplish) assuming
>> that it will be made to do this.

> Yes, it will.  We'll also have an "outgoing" group where you can do
> message queue maintenance (eg., cancel, editing, remove (temporarily)
> from queue, etc).  

Kewl.  Maybe it will be possible to use it always, with a 10 minute
demon to shuffle it out?  That could keep those pesky cancels and
supersedes from reaching anybody's eyes...

nndraft (was: Re: nnagent)

[Per Abrahamsen]

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

> Yes, it will.  We'll also have an "outgoing" group where you can do
> message queue maintenance (eg., cancel, editing, remove (temporarily)
> from queue, etc).  

This sounds very much like the old nndraft stuff, which I miss very
much.  Will it be possible to use this part of the functionality
without the rest?

Re: nnagent

[William M. Perry]

Stainless Steel Rat <ratinox@peorth.gweep.net> writes:

> -----BEGIN PGP SIGNED MESSAGE-----
> 
> >>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
> 
> Lars> Uhm -- from the pop server, into the agent, and from there into nnml?
> Lars> Wouldn't that be a waste of perfectly good bauds?
> 
> It seems to be a waste to me.  The only rational way to use a POP server
> is to download everything from it to a local storage area, and every POP
> mail client that I know of -- including Forte's Agent -- operates on that
> premise.  The way that agent becomes a win is to queue up outgoing mail
> locally (which smtpmail and feedmail already accomplish) assuming that it
> will be made to do this.

  SPRY Mail has an option to only use what they call a 'remote inbox'.
They just download the header info from the messages and show the threaded
display of messages, then retrieve them one at a time as you try to read
them.  Sort of a poor-man's imap.

  It would be kind of nice to have a studly pop3 package that would allow
you to build this functionality into a backend.  but I'd rather see just
straight IMAP support. :)

  I solve the whole work/home laptop problem by just NFS mounting my home
directory from my laptop and symlinking my .gnus.el and .newsrc.eld files
over as well.  My .emacs has a few differences between work and home, and
lets my .gnus.el know what course of action to take.  Sort of like the
'profiles' in SPRY Mail.  I have it check different mail accounts, news
servers, and use different outgoing delivery methods (smtp vs. sendmail)
depending on whether I am at work, on the bus, or at home.  Works pretty
well.

-Bill P.

Re: nnagent

[Lars Magne Ingebrigtsen]

Steinar Bang <sb@metis.no> writes:

> > Yes, it will.  We'll also have an "outgoing" group where you can do
> > message queue maintenance (eg., cancel, editing, remove (temporarily)
> > from queue, etc).  
> 
> Kewl.  Maybe it will be possible to use it always, with a 10 minute
> demon to shuffle it out?  That could keep those pesky cancels and
> supersedes from reaching anybody's eyes...

That would certainly be nice.

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

Re: nndraft (was: Re: nnagent)

[Lars Magne Ingebrigtsen]

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

> > Yes, it will.  We'll also have an "outgoing" group where you can do
> > message queue maintenance (eg., cancel, editing, remove (temporarily)
> > from queue, etc).  
> 
> This sounds very much like the old nndraft stuff, which I miss very
> much.  Will it be possible to use this part of the functionality
> without the rest?

I think so.  I'm leaning towards reintroducing nndraft with nnagent in
mind, but it should be easy enough to make it a general thing.  One
would set `message-post-method' (and so on) to `nndraft-request-post',
which will put it in the nndraft group, and then one posts the
messages for real from that group.  Such a scheme should be useful
even when not using the agent.

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

Re: nndraft (was: Re: nnagent)

[Sudish Joseph]

Lars Magne Ingebrigtsen writes:
> I think so.  I'm leaning towards reintroducing nndraft with nnagent in
> mind, but it should be easy enough to make it a general thing.  One

Yay!

> would set `message-post-method' (and so on) to `nndraft-request-post',
> which will put it in the nndraft group, and then one posts the
> messages for real from that group.  Such a scheme should be useful
> even when not using the agent.

IIRC, the old method let you save the draft via a keystroke from
mail-mode (or was it message?).  It'd be cool to have this as an
option to channeling all posts via such a backend.

-Sudish

Re: nnagent

[Steinar Bang]

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

> Steinar Bang <sb@metis.no> writes:
>> > Yes, it will.  We'll also have an "outgoing" group where you can do
>> > message queue maintenance (eg., cancel, editing, remove (temporarily)
>> > from queue, etc).  

>> Kewl.  Maybe it will be possible to use it always, with a 10 minute
>> demon to shuffle it out?  That could keep those pesky cancels and
>> supersedes from reaching anybody's eyes...

> That would certainly be nice.

It would even let you do cancel on mail groups for mailing lists.  A
feature I've occasionally wished for...:-/

Re: nnagent

[Kai Grossjohann]

>>>>> Lars Magne Ingebrigtsen writes:

  Greg> [...] it would let me move or copy articles into an nnagent
  Greg> group, and have the articles be uploaded to the real groups
  Greg> when possible.

  Lars> I think this is beyond the scope of offline support. [...]

But it would be *very* nice to have this, nevertheless!  Using
mirror.pl to keep two nnml trees in sync is not very satisfying,
somehow...

kai
-- 
`A wizard.  A ----ing wizard.  I *hate* ----ing wizards!'  `Well,
don't --- him, then,' one of his henchmen said, effortlessly
pronouncing a row of dashes.  (Terry Pratchett)

Re: nnagent

[Steinar Bang]

>>>>> Steinar Bang <sb@metis.no>:

>>>>> Lars Magne Ingebrigtsen <larsi@gnus.org>:
>> Steinar Bang <sb@metis.no> writes:
>>> > Yes, it will.  We'll also have an "outgoing" group where you can do
>>> > message queue maintenance (eg., cancel, editing, remove (temporarily)
>>> > from queue, etc).  

>>> Kewl.  Maybe it will be possible to use it always, with a 10 minute
>>> demon to shuffle it out?  That could keep those pesky cancels and
>>> supersedes from reaching anybody's eyes...

>> That would certainly be nice.

> It would even let you do cancel on mail groups for mailing lists.  A
> feature I've occasionally wished for...:-/

Hm... to actually *discover* that I've done something that needs
correcting, I need to have copies of the outgoing messages take the
normal routes into their respective mail groups.

Ie. delayed sending, but spooling of private copies.  Today I do this
with a Bcc, but I guess/hope/belive, that you can do this within Gnus
itself...? 

Re: nndraft (was: Re: nnagent)

[Lars Magne Ingebrigtsen]

Sudish Joseph <sj@eng.mindspring.net> writes:

> IIRC, the old method let you save the draft via a keystroke from
> mail-mode (or was it message?).

It was probably mail-mode.

> It'd be cool to have this as an option to channeling all posts via
> such a backend.

Yes, there will also be support in Message for saving directly to the
draft group.

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

Re: nnagent

[Lars Magne Ingebrigtsen]

Steinar Bang <sb@metis.no> writes:

> Hm... to actually *discover* that I've done something that needs
> correcting, I need to have copies of the outgoing messages take the
> normal routes into their respective mail groups.

I usually discover that I've mistyped 50% of the words and forgot what
I was meaning to say two seconds after I've send a mail.  :-)

> Ie. delayed sending, but spooling of private copies.  Today I do this
> with a Bcc, but I guess/hope/belive, that you can do this within Gnus
> itself...? 

I never use Bcc -- I use Gcc.

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

Re: nnagent

[Steinar Bang]

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

> Steinar Bang <sb@metis.no> writes:
>> Hm... to actually *discover* that I've done something that needs
>> correcting, I need to have copies of the outgoing messages take the
>> normal routes into their respective mail groups.

> I usually discover that I've mistyped 50% of the words and forgot what
> I was meaning to say two seconds after I've send a mail.  :-)

That's the usual state of things for me, as well.  But this last
suggestion was prompted by discovering an error in a message I had
sent to a mailing list.

>> Ie. delayed sending, but spooling of private copies.  Today I do this
>> with a Bcc, but I guess/hope/belive, that you can do this within Gnus
>> itself...? 

> I never use Bcc -- I use Gcc.

I kindof like having the messages I archive been through the same MTA
massaging, as the outgoing message (more or less).  I may be fooling
myself with this, however...

Re: nnagent

[Per Abrahamsen]

Steinar Bang <sb@metis.no> writes:

> > I never use Bcc -- I use Gcc.
> 
> I kindof like having the messages I archive been through the same MTA
> massaging, as the outgoing message (more or less).  I may be fooling
> myself with this, however...

I'd like to use the exactly same rules for splitting outgoing mail as
I use for incoming mail, i.e. nnmail-split-methods (which is set to
nnmail-split-fancy).  I haven't been able to figure out how to do this
with GCC.

Re: nnagent

[Kai Grossjohann]

>>>>> Per Abrahamsen writes:

  Per> I'd like to use the exactly same rules for splitting outgoing
  Per> mail as I use for incoming mail, i.e. nnmail-split-methods
  Per> (which is set to nnmail-split-fancy).  I haven't been able to
  Per> figure out how to do this with GCC.

Well, I suppose for this case, Gcc is the wrong tool.  Using Bcc means
you don't have to do anything but xbiff beeps every time you send a
mail.  You can circumvent that with Fcc and adding the file you write
to to nnmail-spool-file.

Maybe Gcc'ing to a `write only' nnfolder group and putting that in
nnmail-spool-file would also work?  But I'd be afraid to lose mail
that way, unless Lars says it's okay.

kai
-- 
Some people say that language is for communication.
I think communcation is in spite of language.

Re: nnagent

[Steinar Bang]

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

> Steinar Bang <sb@metis.no> writes:

>> > I never use Bcc -- I use Gcc.
>> 
>> I kindof like having the messages I archive been through the same MTA
>> massaging, as the outgoing message (more or less).  I may be fooling
>> myself with this, however...

> I'd like to use the exactly same rules for splitting outgoing mail as
> I use for incoming mail,

So do I.

> i.e. nnmail-split-methods (which is set to nnmail-split-fancy).  I
> haven't been able to figure out how to do this with GCC.

If so, that rules the current gcc out for me, as well.  What I *can*
do to support offline composing, and sending of mail, is to let
sendmail queue it, and deliver Bcc'd stuff directly to /var/mail/sb on
my SPARCbook.

Hm... nope... that rules out the cancel/supersede functionality...

Do we need a New Cc Method (tm)...?

Re: nnagent

[Kai Grossjohann]

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

  Per> i.e. nnmail-split-methods (which is set to nnmail-split-fancy).  I
  Per> haven't been able to figure out how to do this with GCC.

>>>>> Steinar Bang writes:

  Steinar> If so, that rules the current gcc out for me, as well.
  Steinar> What I *can* do to support offline composing, and sending
  Steinar> of mail, is to let sendmail queue it, and deliver Bcc'd
  Steinar> stuff directly to /var/mail/sb on my SPARCbook.

You might just want to Fcc the messages to a file and include the file
in nnmail-spool-file (which may also be a list).  Or doesn't this work
for you because of some reason?

kai
-- 
Some people say that language is for communication.
I think communcation is in spite of language.

Re: nnagent

[Steinar Bang]

>>>>> Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de>:

>>>>> Per Abrahamsen <abraham@dina.kvl.dk>:
Per> i.e. nnmail-split-methods (which is set to nnmail-split-fancy).  I
Per> haven't been able to figure out how to do this with GCC.

>>>>> Steinar Bang writes:

Steinar> If so, that rules the current gcc out for me, as well.
Steinar> What I *can* do to support offline composing, and sending
Steinar> of mail, is to let sendmail queue it, and deliver Bcc'd
Steinar> stuff directly to /var/mail/sb on my SPARCbook.

> You might just want to Fcc the messages to a file and include the file
> in nnmail-spool-file (which may also be a list).  Or doesn't this work
> for you because of some reason?

Nah.  I guess it would work.  But the effects would be the same as for
Bcc, if the "delayed send" doesn't look at the Fcc or Bcc before it's
actually sent (we're talking about the possibility to "cancel" or
"supersede" (quotes intentional) articles posted on a mail group,
especially a mail group devoted to a mailing list).


- Steinar