You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Cedric Foll <ce...@ac-rouen.fr> on 2006/03/15 18:43:12 UTC

sa-update and channels

Hi,

what is "channel" in sa-updates ?
What are the channel available ?

Where can we found information about update rules processes included in
3.1.1 ?

BTW find for all the great work done on spamassassin !

Regards.


Re: sa-update and channels

Posted by Theo Van Dinter <fe...@apache.org>.
On Fri, Mar 17, 2006 at 11:15:44AM -0600, Chris Thielen wrote:
> That said, my opinion is that YES indeed this is a replacement for RDJ 
> in at least the SARE.  I'm sorta waiting for the sa-update framework to 
> gel a bit more before we tackle getting the SARE rules published as 
> channels.

FWIW, the framework for what a channel is and how updates are published and
distributed, etc, hasn't changed since sa-update was originally written 9+
months ago.  (the script has been modified a bit, but not the design of a
"channel")

> Again, I haven't taken the time to see how sa-update is actually 
> implemented so I may be off base with this comment.

I thought I had a document available which talked about how sa-update
works, but I can't find it now, so I started work on a new wiki page
(http://wiki.apache.org/spamassassin/RuleUpdates) to talk about how to
use sa-update and how to publish channels.  It's still being worked on,
so don't laugh if you look at it right now. :)

If you have any questions, feel free to ask and that'll let me add information
to the wiki doc.

-- 
Randomly Generated Tagline:
"I'll have to be careful when I walk past that thing ... It'll blow the rest
 of my hair off."                - Isadore Katz

Re: sa-update and channels

Posted by Chris Thielen <cm...@someone.dhs.org>.
Michael Monnerie wrote:
> On Mittwoch, 15. März 2006 19:32 Theo Van Dinter wrote:
>   
>> A channel is essentially a set of rules published by some
>> organization, which is accessed and downloaded via dns/http.  ie:
>>     
>
> Would that be a possible replacement for RulesDuJour? I love that 
> script, but not having to install something extra saves some energy and 
> time, and that is always appreciated *g*
>
> mfg zmi
>   
I'm the RDJ guy.  I admit I haven't been very active lately and haven't 
even looked at how sa-update is implemented. 

That said, my opinion is that YES indeed this is a replacement for RDJ 
in at least the SARE.  I'm sorta waiting for the sa-update framework to 
gel a bit more before we tackle getting the SARE rules published as 
channels.

However, RDJ may still be useful for those single rulesets here and 
there who's authors maybe don't have access to a DNS server, etc.    
Again, I haven't taken the time to see how sa-update is actually 
implemented so I may be off base with this comment.


Chris

Re: sa-update and channels

Posted by Michael Monnerie <m....@zmi.at>.
On Donnerstag, 16. März 2006 02:19 Theo Van Dinter wrote:
> On Thu, Mar 16, 2006 at 12:26:32AM +0100, Michael Monnerie wrote:
> > > A channel is essentially a set of rules published by some
> > > organization, which is accessed and downloaded via dns/http.  ie:
> >
> > Would that be a possible replacement for RulesDuJour? I love that
> > script, but not having to install something extra saves some energy
> > and time, and that is always appreciated *g*
>
> If the rules were published as a channel, sure. :)

Any SARE people to comment on this? Sounds like a welcome tool to 
propagate the use of SARE rules.

mfg zmi
-- 
// Michael Monnerie, Ing.BSc  ---   it-management Michael Monnerie
// http://zmi.at           Tel: 0660/4156531          Linux 2.6.11
// PGP Key:   "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952  F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net                 Key-ID: 0x70545879

Re: sa-update and channels

Posted by Theo Van Dinter <fe...@apache.org>.
On Thu, Mar 16, 2006 at 12:26:32AM +0100, Michael Monnerie wrote:
> > A channel is essentially a set of rules published by some
> > organization, which is accessed and downloaded via dns/http.  ie:
> 
> Would that be a possible replacement for RulesDuJour? I love that 
> script, but not having to install something extra saves some energy and 
> time, and that is always appreciated *g*

If the rules were published as a channel, sure. :)

-- 
Randomly Generated Tagline:
 Earth men are real men! 

Re: sa-update and channels

Posted by Michael Monnerie <m....@zmi.at>.
On Mittwoch, 15. März 2006 19:32 Theo Van Dinter wrote:
> A channel is essentially a set of rules published by some
> organization, which is accessed and downloaded via dns/http.  ie:

Would that be a possible replacement for RulesDuJour? I love that 
script, but not having to install something extra saves some energy and 
time, and that is always appreciated *g*

mfg zmi
-- 
// Michael Monnerie, Ing.BSc  ---   it-management Michael Monnerie
// http://zmi.at           Tel: 0660/4156531          Linux 2.6.11
// PGP Key:   "lynx -source http://zmi.at/zmi2.asc | gpg --import"
// Fingerprint: EB93 ED8A 1DCD BB6C F952  F7F4 3911 B933 7054 5879
// Keyserver: www.keyserver.net                 Key-ID: 0x70545879

Re: sa-update and channels

Posted by Theo Van Dinter <fe...@apache.org>.
On Wed, Mar 15, 2006 at 06:43:12PM +0100, Cedric Foll wrote:
> what is "channel" in sa-updates ?

A channel is essentially a set of rules published by some organization,
which is accessed and downloaded via dns/http.  ie:

sa-update checks (by default) updates.spamassassin.org to see if any updates
are available:

[25712] dbg: dns: 1.1.3.updates.spamassassin.org => 385719, parsed as 385719

so for 3.1.1, the latest update is 385719.  If there is no installed version,
or the installed version is older, the update is downloaded from one of the
published list of mirrors, lint checked, and installed if everything looks ok.

If you want to see what's happening but not install the updates at the system
level, you can do "sa-update -D --updatedir /tmp/foo --nogpg".

> What are the channel available ?

The only one that the SpamAssassin project makes available right now
is updates.spamassassin.org.  I don't know of any other channels at the
moment, though there's no reason why third parties couldn't publish them
if desired.

> Where can we found information about update rules processes included in
> 3.1.1 ?

Good question -- we're still in the process of coming up with those processes.
Generally speaking, we put in rules for development which get tested nightly
against several people's mail corpus.  At the moment, we then manually take
the rules that perform well (good spam hitrate, low/zero ham hitrate, etc,)
and move them into the 3.1 update area.  Scores are manually entered, and
after some review time, the update is published.

There's some work being done wrt a better way to set the score initially,
doing score generation runs more often to keep everything in-line with
the spam that's out there, etc.  There's also a system to make updates
happen automatically based on the nightly results, though it's not quite
ready for use with 3.1 right now (it's being used for 3.2/development).

Hopefully this helps.  :)


BTW: I really need to write up a wiki doc about sa-update.  I would swear
that I wrote up something back before 3.1.0 came out, but I can't find it
now... :(   If there's anything people want to know about sa-update or
channels, etc, please feel free to ask.  I'll probably end up writing
a document based on mails I receive so that I'm sure the questions
are answered.

-- 
Randomly Generated Tagline:
"Crystals do for new age what velvet paintings did for art."
                      - From the movie Road Scholar