Spam Management When Using nnimap

Spam Management When Using nnimap

[Jake Colman]

I've read the Info manual on spam management but am not sure how to apply
these techniques when using nnimap.

I'm running XEmacs and Gnus on my company's exterprise server.  This is not
my own machine where I can do things my own way.  We company uses MS Exchange
as the mail server and I use nnimap in gnus to read my email.

I downloaded and installed spamassasson locally, using IMAPassaasin as daemon
to preprocess and filter my imap inbox before gnus ever saw it.  This was
working quite well but I've been asked by my IS group terminate the
IMAPassaasin daemon (they think the daemon takes up too many resources).

So what now?  Can the techniques discussed in the Info manual be applied to
an imap situation?  If so, how? 

Thanks!

...Jake

-- 
Jake Colman                     

Principia Partners LLC                  Phone: (201) 209-2467
Harborside Financial Center               Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                  www.principiapartners.com

Re: Spam Management When Using nnimap

[Chris Green]

Jake Colman <colman@ppllc.com> writes:

> So what now?  Can the techniques discussed in the Info manual be applied to
> an imap situation?  If so, how? 
>

http://lifelogs.com/spam/spam.html seems to be the best tutorial for
this.

<ramble>
What I've not come to grips with yet is how to handle the concept of
an "incoming split folder" with multiple independent readers.

It used to make more sense to me when /var/spool/mail/cmg was the
incoming data and I wouldn't read it until it was filed away
somewhere.  Now gnus+spam.el+nnimap wants that model with IMAP but my
other clients just want to peek at the INBOX.
</ramble>
-- 
Chris Green <cmg@sourcefire.com>
Chicken's thinkin'

Re: Spam Management When Using nnimap

[Jake Colman]

>>>>> "CG" == Chris Green <cmg@sourcefire.com> writes:

   CG> http://lifelogs.com/spam/spam.html seems to be the best tutorial for
   CG> this.

But where is the determination of what constitutes a spam email?  Unless I
missed something, there is nothing in that tutorial that shows me how/what to
use as hook in gnus so that messages are automatically clasified as spam or
ham.

-- 
Jake Colman                     

Principia Partners LLC                  Phone: (201) 209-2467
Harborside Financial Center               Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                  www.principiapartners.com

Re: Spam Management When Using nnimap

[David Z Maze]

Jake Colman <colman@ppllc.com> writes:

>>>>>> "CG" == Chris Green <cmg@sourcefire.com> writes:
>
>    CG> http://lifelogs.com/spam/spam.html seems to be the best tutorial for
>    CG> this.
>
> But where is the determination of what constitutes a spam email?  Unless I
> missed something, there is nothing in that tutorial that shows me how/what to
> use as hook in gnus so that messages are automatically clasified as spam or
> ham.

Those are the various spam-use-* flags.  If you wanted to add your own
classifier, you'd need to add it to spam-list-of-checks.  The check
functions appear to be called with no parameters with the current
buffer narrowed to the headers of the article in question, and return
t if the article is ham, spam-split-group (or another group name) if
the article is spam, and nil for undecided.

-- 
David Maze             dmaze@mit.edu          http://www.mit.edu/~dmaze/
"Theoretical politics is interesting.  Politicking should be illegal."
	-- Abra Mitchell

Re: Spam Management When Using nnimap

[Jake Colman]

So has anyone here on the list come up with a solution that can pass email
that is read from an imap server through spamassasson on the fly?  And then
process the results?

>>>>> "DZM" == David Z Maze <David> writes:

   DZM> Those are the various spam-use-* flags.  If you wanted to add your own
   DZM> classifier, you'd need to add it to spam-list-of-checks.  The check
   DZM> functions appear to be called with no parameters with the current
   DZM> buffer narrowed to the headers of the article in question, and return
   DZM> t if the article is ham, spam-split-group (or another group name) if
   DZM> the article is spam, and nil for undecided.


-- 
Jake Colman                     

Principia Partners LLC                  Phone: (201) 209-2467
Harborside Financial Center               Fax: (201) 946-0320
902 Plaza Two                          E-mail: colman@ppllc.com
Jersey City, NJ 07311                  www.principiapartners.com

Re: Spam Management When Using nnimap

[David Abrahams]

Jake Colman <colman@ppllc.com> writes:

> So has anyone here on the list come up with a solution that can pass email
> that is read from an imap server through spamassasson on the fly?  And then
> process the results?

FWIW, you might want to look at my solution, which is working nicely
for me:

    http://www.boost-consulting.com/writing/server-side.html

Good luck,
-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: Spam Management When Using nnimap

[Ted Zlatanov]

On Thu, 15 May 2003, colman@ppllc.com wrote:
> I've read the Info manual on spam management but am not sure how to
> apply these techniques when using nnimap.
> 
> I'm running XEmacs and Gnus on my company's exterprise server.  This
> is not my own machine where I can do things my own way.  We company
> uses MS Exchange as the mail server and I use nnimap in gnus to read
> my email.
> 
> I downloaded and installed spamassasson locally, using IMAPassaasin
> as daemon to preprocess and filter my imap inbox before gnus ever
> saw it.  This was working quite well but I've been asked by my IS
> group terminate the IMAPassaasin daemon (they think the daemon takes
> up too many resources).
> 
> So what now?  Can the techniques discussed in the Info manual be
> applied to an imap situation?  If so, how?

This is exactly how I use spam.el, since I prefer to split on the
client side.

SpamAssassin is not added yet as an official check, but feel free to
do it, following the instructions in the manual for adding a new
backend, or look at the bogofilter backend.  It's really very simple.

I don't use SA, but I can implement this if enough people need it - it
will be faster if a SA user can write and test the code though.

Ted

Re: Spam Management When Using nnimap

[Ted Zlatanov]

On Thu, 15 May 2003, cmg@sourcefire.com wrote:
> http://lifelogs.com/spam/spam.html seems to be the best tutorial for
> this.

I haven't updated the tutorial since January, I think I'll wait until
version 2 of spam.el, which will track and undo article ham/spam
registration and do batch registration.  Not enough time to do
incremental updates there.

> <ramble>
> What I've not come to grips with yet is how to handle the concept of
> an "incoming split folder" with multiple independent readers.
> 
> It used to make more sense to me when /var/spool/mail/cmg was the
> incoming data and I wouldn't read it until it was filed away
> somewhere.  Now gnus+spam.el+nnimap wants that model with IMAP but
> my other clients just want to peek at the INBOX.  </ramble>

I'm not sure I understand what you mean, sorry.

Ted

Re: Spam Management When Using nnimap

[Ted Zlatanov]

On Fri, 16 May 2003, dmaze@MIT.EDU wrote:
> Those are the various spam-use-* flags.  If you wanted to add your
> own classifier, you'd need to add it to spam-list-of-checks.  

I thought the manual was pretty clear about this.  Maybe I need a
"starting from scratch with spam.el" section.

> The check functions appear to be called with no parameters with the
> current buffer narrowed to the headers of the article in question,
> and return t if the article is ham, spam-split-group (or another
> group name) if the article is spam, and nil for undecided.

For the statistical checks (see the recently added
spam-list-of-statistical-checks) we widen to the whole article when we
do the check.

Ted