On Fri, Nov 24, 2006 at 05:27:08PM +0100, Jan-Pieter Cornet wrote:
> That said, after a little looking around, there's both a libspamc.so
> available (without documentation?), and a protocol description in
> the SA distribution in spamd/PROTOCOL. So it should be possible to
> create a standalone pe
On Fri, Nov 24, 2006 at 01:00:54PM +0100, Steffen Kaiser wrote:
> I was just wondering, because you shell-out one program per scan now. I
> was hoping that there is another way to connect to spamd.
Well, it's only a small C program, and it's likely in cache, not
executed via the shell but directl
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Fri, 24 Nov 2006, Jan-Pieter Cornet wrote:
I was just wondering, because you shell-out one program per scan now. I
was hoping that there is another way to connect to spamd.
Bye,
- --
Steffen Kaiser
-BEGIN PGP SIGNATURE-
Version: GnuP
On Fri, Nov 24, 2006 at 09:31:05AM +0100, Steffen Kaiser wrote:
> Do you use spamc to connect to spamd or have you re-implemented the
> protocol in Perl?
I use spamc. It's a pretty small binary. The spamc-spamd protocol
isn't defined, where the spamc interface is properly defined, so I
didn't wan
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Thu, 23 Nov 2006, Jan-Pieter Cornet wrote:
Do you use spamc to connect to spamd or have you re-implemented the
protocol in Perl?
Bye,
- --
Steffen Kaiser
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.1 (GNU/Linux)
iQEVAwUBRWatzOgJIbZtwg
was no benefit to using spamd...but we sure see one.
- Original Message -
From: "Jan-Pieter Cornet" <[EMAIL PROTECTED]>
To:
Sent: Thursday, November 23, 2006 11:12 AM
Subject: Re: spamd vs. builtin SA (was Re: [Mimedefang] Patch: adding custom
headers for SpamAssassin)
Kevin A. McGrail wrote:
Not sure what kind of headers Kevin (and others?) are adding, but it
might
be worth it ot check if plugins could be used in more cases.
Don't rub salt. I'm still learning how to write a plug-in for SA. I
think getting two or three done will be my New Year's Resolutio
On Thu, Nov 23, 2006 at 10:00:01AM -0500, David F. Skoll wrote:
> > But, given the fact that we just switched to using spamd instead of
> > the builtin Mail::SpamAssassin modules, we don't really need support
> > builtin to mimedefang at the moment either...
>
> Just curious: Why the switch? Do y
Not sure what kind of headers Kevin (and others?) are adding, but it
might
be worth it ot check if plugins could be used in more cases.
Don't rub salt. I'm still learning how to write a plug-in for SA. I think
getting two or three done will be my New Year's Resolution ;-)
Happy Thanksgivin
Jan-Pieter Cornet wrote:
> But, given the fact that we just switched to using spamd instead of
> the builtin Mail::SpamAssassin modules, we don't really need support
> builtin to mimedefang at the moment either...
Just curious: Why the switch? Do you see better performance?
Regards,
David.
___
On Thu, Nov 23, 2006 at 01:46:17PM +0100, Jonas Eckerman wrote:
> >I'll look at some compromise solution. :-)
>
> Not sure what kind of headers Kevin (and others?) are adding, but it might
> be worth it ot check if plugins could be used in more cases.
We're currently adding an extra header to sp
David F. Skoll wrote:
Jonas Eckerman wrote:
add headers to the temprary message created for SpamAssassin,
[...]
I'll look at some compromise solution. :-)
Just wanted to say that with my own SpamAssassin plugins for using p0f data and
checking sender validity and John Rudd's Botnet plu
On Friday, September 22, 2006 10:04 PM +0200 Jan-Pieter Cornet
<[EMAIL PROTECTED]> wrote:
Already done, see http://www.xs4all.nl/~johnpc/mimedefang-modular/
It would be useful to note there which version of the example filter it
emulates, perhaps with a copy of that filter included. When a n
Jan-Pieter Cornet wrote:
[...]
> As for an object-oriented MIMEDefang... I'm not convinced that is
> a good idea. OO does add overhead to perl -- certainly, calling
> an object method, compared to accessing a global variable, is a very
> costly operation... but of course you need to see that in t
Interestingly, we recently made a change almost exactly like the
one Jonas posted, using a global array. So I can confirm that
there certainly is demand for the feature :)
On Fri, Sep 22, 2006 at 10:59:08AM -0700, Kenneth Porter wrote:
> >An object oriented MIMEDefang seems logical.
> >And nice. :
--On Friday, September 22, 2006 6:07 PM +0200 Jonas Eckerman
<[EMAIL PROTECTED]> wrote:
An object oriented MIMEDefang seems logical.
And nice. :-)
What I look forward to is a system that makes upgrades easier, by allowing
me to make minor tweaks that are easy to carry forward to future examp
David F. Skoll wrote:
I'd rather not add yet another global variable.
I can understand that.
The reasons I added it as a global variable was
a) it was simple and
b) it fits the seeming philosophy of how mimedefang.pl is written.
A more clean way would be as an optional argument (giving a ref
Kevin A. McGrail wrote:
Here's what I do:
#IN ORDER TO HAVE A HEADER TEST IN SPAMASSASSIN REACT TO DATA THAT MD
CREATES, WE HAVE TO EDIT THE INPUTMSG FILE PRIOR TO
This is exactly what I wanted toi avoid. For two reasons:
1: It requires more work since there's more file reading and writing
sub change_header_immediately {
[...]
#IN ORDER TO HAVE A HEADER TEST IN SPAMASSASSIN REACT TO DATA THAT MD
CREATES, WE HAVE TO EDIT THE INPUTMSG FILE PRIOR TO
Yikes! Holy overhead, Batman!
I'll look at some compromise solution. :-)
Thanks!
regards,
KAM
__
Kevin A. McGrail wrote:
> sub change_header_immediately {
[...]
> #IN ORDER TO HAVE A HEADER TEST IN SPAMASSASSIN REACT TO DATA THAT MD
> CREATES, WE HAVE TO EDIT THE INPUTMSG FILE PRIOR TO
Yikes! Holy overhead, Batman!
I'll look at some compromise solution. :-)
Regards,
David.
_
Here's what I do:
sub change_header_immediately {
my (%params) = @_;
&append_header_immediately(%params, change=>1);
#md_syslog('warning', "Changing $params{'header'} immediately");
}
sub append_header_immediately {
#IN ORDER TO HAVE A HEADER TEST IN SPAMASSASSIN REACT TO DATA THAT MD
CRE
Jonas Eckerman wrote:
> It's just a simple way (a global array) to add headers to the temprary
> message created for SpamAssassin, so that these headers may be included
> in rules and in the bayes automagic. The change to mimedefang.pl is small.
I'd rather not add yet another global variable. A
Hi!
I posted this before, but never got any response. Not sure if that was because
my idea was bad or uninteresting or because noone noticed it. :-)
It's just a simple way (a global array) to add headers to the temprary message
created for SpamAssassin, so that these headers may be included in
23 matches
Mail list logo