At Fri Jan 30 22:33:27 2004, Kareem Dana wrote:
>
> Hello,
>
> I use spamassassin 2.54 and just turned on bayesian filtering. I saved
> over 1000 recent spams and hams and trained the filter on those emails.
Well, you really ought to be using a later version than 2.54. The
current version is 2
Hello PieterB,
Monday, January 26, 2004, 8:03:45 AM, you wrote:
P> Is there some way to prevent spamassassin from using SA-talk messages
P> for Bayes auto_learning. My bayes filter seems to be less effective
P> since a lot of spamphrases/tokens are discussed on this list.
My method:
1) I subscri
At 08:10 AM 1/28/04 +0200, Thomas Kinghorn wrote:
My spamd is running as xadmin
xadmin 17057 1 0 Jan27 ?00:00:22 /usr/bin/perl -T -w
/usr/bin/spamd -d -a -u xadmin
Do I need to run sa-learn as xadmin
If so, I could kick myself, I have been training it while logged in as
root...
> -Original Message-
> From: Matt Kettler [mailto:[EMAIL PROTECTED]
> Sent: 27 January 2004 05:55
> To: Thomas Kinghorn; Spamassassin-Talk (E-mail)
> Subject: Re: [SAtalk] bayes question: HAM
>
>
> At 02:10 PM 1/27/04 +0200, Thomas Kinghorn wrote:
> >While using spamd -D
On Tue, Jan 27, 2004 at 12:56:39PM +, thomas fritz wrote:
> debug: bayes: expiry check keep size, 75% of max: 112500
> debug: bayes: token count: 1215351, final goal reduction size: 1102851
Ok, so you want to keep 112k, and have 1.2m, for a reduction size of 1.1m.
> debug: bayes: atime to
At 02:10 PM 1/27/04 +0200, Thomas Kinghorn wrote:
While using spamd -D, I can see the messages being learned as ham.
However, while doing a spamassasin -D --lint, it shows only 1 ham.
sa-learn --dump magic shows
[EMAIL PROTECTED] exim]# sa-learn --dump magic
I have attached the --lint debug.
An
On Tue, Jan 27, 2004 at 10:25:20PM -0500, Rick Mallett wrote:
> The issue is that bayes expiry is broken under some circumstances
> as evidenced by this listing of my bayes database area, possibly
> because its taking more than 10 minutes, or possibly because there
> is a bug in the code
Well, you
Perhaps I spoke too quickly when I suggested that the bayes
expiry code must be broken. I finally read David Lee's message
carefully and realized that the expiry code might be getting timed
out by mimedefang or sendmail and interrupted before it has time to
complete. I should have read it earlier b
The issue is that bayes expiry is broken under some circumstances
as evidenced by this listing of my bayes database area, possibly
because its taking more than 10 minutes, or possibly because there
is a bug in the code
-rw--- 1 defang defang32 Jan 26 12:19 bayes.lock
-rw--- 1
On Tue, Jan 27, 2004 at 03:57:56PM +, David Lee wrote:
> The MailScanner maintainer, Julian Field, is very responsive, and he has
> already coded up an alternative way of driving SA from MS, so that its use
> of SA can avoid auto-expire (and thereby avoid the possible multiple
> simultaneous ex
On Mon, 26 Jan 2004, Rick Mallett wrote:
> [...]
> Now I've got a much bigger problem. The expiry is starting to take
> more than 10 minutes and as a result the journal grows to max size and
> an opportunistic rebuild kills the lock file and wrecks the expiry
> operation. Here is what I observe in
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Monday 26 January 2004 08:03 am, PieterB wrote:
> Is there some way to prevent spamassassin from using SA-talk messages
> for Bayes auto_learning. My bayes filter seems to be less effective
> since a lot of spamphrases/tokens are discussed on this l
On Monday, January 26, 2004 @ 8:03:45 AM [-0700], PieterB wrote:
> Is there some way to prevent spamassassin from using SA-talk messages
> for Bayes auto_learning. My bayes filter seems to be less effective
> since a lot of spamphrases/tokens are discussed on this list.
Do you have access to proc
On Mon, Jan 26, 2004 at 05:03:45PM +0100, PieterB is rumored to have said:
>
> Is there some way to prevent spamassassin from using SA-talk messages
> for Bayes auto_learning. My bayes filter seems to be less effective
> since a lot of spamphrases/tokens are discussed on this list.
It depends on
On January 26, 2004 11:03 am, PieterB wrote:
> Is there some way to prevent spamassassin from using SA-talk messages
> for Bayes auto_learning. My bayes filter seems to be less effective
> since a lot of spamphrases/tokens are discussed on this list.
>
> I did a "sa-learn --mbox --forget Mail/spama
On 21 Jan 2004 at 13:01, Douglas Kirkland wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> On Wednesday 21 January 2004 11:59, Jesse Regier wrote:
> > Is there a way to have Spamassassin use a seperate bayes database for
> > each local domain or group of domains?
> >
> > It would be
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Wednesday 21 January 2004 11:59, Jesse Regier wrote:
> Is there a way to have Spamassassin use a seperate bayes database for
> each local domain or group of domains?
>
> It would be nice to have something like...
>
> bayes_path = /var/spool/$DOMA
At 12:37 AM 1/21/04 -0500, Pedro Sam wrote:
My question, should
bayes ignore the habeas headers by default?
Perhaps not by default, but right now it's probably a good idea.
In general, any sudden shift of behavior from something commonly seen only
in nonspam to commonly seen in both causes troubl
Pedro Sam wrote:
I tried to remove the habeas headers mannually
You can get the same results by putting in your preference or
configuration file
bayes_ignore_header X-Habeas-SWE-1
bayes_ignore_header X-Habeas-SWE-2
bayes_ignore_header X-Habeas-SWE-3
bayes_ignore_header X-Habeas-SWE-4
bayes_ignore
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Arpi writes:
>Hi,
>
>> On Mon, Jan 19, 2004 at 03:21:06PM -0500, Larry Gilson wrote:
>> > http://useast.spamassassin.org/doc/Mail_SpamAssassin_Conf.html#learning%20op
>> > tions
>> >
>> > bayes_ignore_header header_name
>>
>> ::bangs head on wall::
Hi,
> On Mon, Jan 19, 2004 at 03:21:06PM -0500, Larry Gilson wrote:
> > http://useast.spamassassin.org/doc/Mail_SpamAssassin_Conf.html#learning%20op
> > tions
> >
> > bayes_ignore_header header_name
>
> ::bangs head on wall:: How did I miss *that*? Thanks for correcting
> my careless reading.
On Mon, Jan 19, 2004 at 08:35:28PM -0800, Justin Mason wrote:
> Actually, it works quite well. Some people get more spam than ham to
> specific To addrs, so those become spam signs -- but once a ham arrives
> at those addrs, the ham signs outweigh the To spam-sign and redeem
> the mail.
In theory
This answer does NOT really relate to bayes!
But nevertheless to a typical 'out of memory'.
On Tue, Jan 20, 2004 at 01:43:30PM +0100, Daniele Nicolodi wrote:
> I have also tryed to upgrade to sa 2.62 and convert the old db format to
> the new one but sa fill all my memory, and conseguently is kill
Thanks for clarifying Justin!
--Larry
> -Original Message-
> From: [EMAIL PROTECTED]
> Sent: Monday, January 19, 2004 11:35 PM
> To: Larry Gilson
> Cc: 'Ross Vandegrift'; [EMAIL PROTECTED]
> Subject: Re: [SAtalk] Bayes mis-learning problem
> Larry Gi
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Larry Gilson writes:
> > In a broader sense though, shouldn't fields like To: be excluded by
> > default? It seems like if I receive more than 50% spam, this is a
> > receipe for disaster. Of course, some spam won't have a valid To:
> > field, but i
> -Original Message-
> From: Ross Vandegrift [mailto:[EMAIL PROTECTED]
> Sent: Monday, January 19, 2004 4:07 PM
> To: Larry Gilson
> Cc: [EMAIL PROTECTED]
> Subject: Re: [SAtalk] Bayes mis-learning problem
>
> On Mon, Jan 19, 2004 at 03:21:06PM -0500, Larr
On Mon, Jan 19, 2004 at 03:21:06PM -0500, Larry Gilson wrote:
> http://useast.spamassassin.org/doc/Mail_SpamAssassin_Conf.html#learning%20op
> tions
>
> bayes_ignore_header header_name
::bangs head on wall:: How did I miss *that*? Thanks for correcting
my careless reading.
In a broader sense
Look at:
http://useast.spamassassin.org/doc/Mail_SpamAssassin_Conf.html#learning%20op
tions
bayes_ignore_header header_name
If you receive mail filtered by upstream mail systems, like a spam-filtering
ISP or mailing list, and that service adds new headers (as most of them do),
these headers may
On Thu, 15 Jan 2004, Billy Huddleston wrote:
> Are you all load balancing these? If so, are you doing it at the SMTP level
> or using SPAMD ? I'm getting ready to implement a 2 node system with
> qmail..and wondering if using a single SMTP machine calling spamc against a
> cluster of SA machines
On Fri, Jan 16, 2004 at 10:44:04AM -0500, Damian Gerow wrote:
> Thus spake Ed Walker ([EMAIL PROTECTED]) [15/01/04 20:04]:
> >
> > How's this holding up under heavy load with large clusters? How much
> > overhead for either the SQL server or SA?
>
> Having not written the code, and not being in a
Thus spake Ed Walker ([EMAIL PROTECTED]) [15/01/04 20:04]:
> > We've got four.. And I'm going to insert yet another shameless plug
> for the
> > SQL rewrite for the Bayes code that's currently sitting in a bug
> report.
> > 'Tis definitely multi-host safe, and 'twould make my life easier...
>
> H
hines
with qmail and spamd all running on the same machine..
Thanks, Billy
- Original Message -
From: "Damian Gerow" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, January 15, 2004 4:55 PM
Subject: Re: [SAtalk] Bayes NFS safe?
> Thus spake Rocky Ol
at 04:55:08PM -0500, Damian Gerow wrote:
> Thus spake Rocky Olsen ([EMAIL PROTECTED]) [15/01/04 16:37]:
> > I too would greatly appreciate any information - as we have 9 boxes
> > doing Spam scanning. Anyone tried this?
>
> We've got four.. And I'm going to insert yet another shameless plug
for t
On 15 Jan 2004, Rocky Olsen wrote:
> I too would greatly appreciate any information - as we have 9 boxes
> doing Spam scanning. Anyone tried this?
>
>
> On Thu, 2004-01-15 at 13:31, Mike Jackson wrote:
> > If you have multiple SA filtering boxes, is it safe to NFS-mount a partition
> > with a syst
te:
> Date: 15 Jan 2004 14:37:47 -0700
> From: Rocky Olsen <[EMAIL PROTECTED]>
> To: "[EMAIL PROTECTED]"
> <[EMAIL PROTECTED]>
> Subject: Re: [SAtalk] Bayes NFS safe?
>
> I too would greatly appreciate any information - as we have 9 boxes
> doing Spa
On Thu, Jan 15, 2004 at 04:55:08PM -0500, Damian Gerow wrote:
> Thus spake Rocky Olsen ([EMAIL PROTECTED]) [15/01/04 16:37]:
> > I too would greatly appreciate any information - as we have 9 boxes
> > doing Spam scanning. Anyone tried this?
>
> We've got four.. And I'm going to insert yet another
Thus spake Rocky Olsen ([EMAIL PROTECTED]) [15/01/04 16:37]:
> I too would greatly appreciate any information - as we have 9 boxes
> doing Spam scanning. Anyone tried this?
We've got four.. And I'm going to insert yet another shameless plug for the
SQL rewrite for the Bayes code that's currently
I too would greatly appreciate any information - as we have 9 boxes
doing Spam scanning. Anyone tried this?
On Thu, 2004-01-15 at 13:31, Mike Jackson wrote:
> If you have multiple SA filtering boxes, is it safe to NFS-mount a partition
> with a system-wide Bayesian database and share it across al
> small chastising rant: Subject lines exist for a reason, on high volume
> lists, please use them. "no subject" makes you look lazy and/or lacking in
> inteligence, usualy both.
>
Actually I was suffering from temporary (I hope) stupidity... My bad...
Sorry :)
> At 05:05 PM 1/14/2004, Christop
On January 14, 2004 12:07 pm, Paul Barbeau wrote:
> Has anyone create a button (or some other way) that i can install on a
> client outlook that will submit the email to the my bayes learning account?
> The current process to "resend the message" is above most of my users and
> button would be much
small chastising rant: Subject lines exist for a reason, on high volume
lists, please use them. "no subject" makes you look lazy and/or lacking in
inteligence, usualy both.
At 05:05 PM 1/14/2004, Christopher Tarricone wrote:
It seems to me that SpamAssassing is running as the user vpopmail so
> If you think some tokens should be "stronger" than others, please do a
> 10-fold cross-validation testing run which should *prove* that to be the
> case. We don't adopt Bayes tokenizer or combiner changes without
> such testing.
considering I have no idea how to do this or where to even be
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Chris Petersen writes:
>> See 'man sa-learn' or use
>> http://www.spamassassin.org/doc/sa-learn.html
>> http://wiki.spamassassin.org/w/BayesInSpamAssassin
>
>This doesn't say much about HOW it's used in SA, though. For instance,
>does SA bayes score
> See 'man sa-learn' or use
> http://www.spamassassin.org/doc/sa-learn.html
> http://wiki.spamassassin.org/w/BayesInSpamAssassin
This doesn't say much about HOW it's used in SA, though. For instance,
does SA bayes score URI tokens higher than it does general body tokens?
(if not, it should) Wha
check the links at
http://www.spamassassin.org/doc/Mail_SpamAssassin_Bayes.html
/robert
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Carles
Xavier Munyoz Baldó
Sent: Mittwoch, 14. Januar 2004 11:15
To: [EMAIL PROTECTED]
Subject: [SAtalk] Bayes.
Hi,
M
On Wed, Jan 14, 2004 at 11:14:51AM +0100, Carles Xavier Munyoz Bald? wrote:
> Hi,
> May someone explain me in few words what is the Bayes probability algorithm
> and how is it used in SpamAssassin ?
See 'man sa-learn' or use
http://www.spamassassin.org/doc/sa-learn.html
http://wiki.spamassassin.o
Hello Kevin,
Friday, January 9, 2004, 4:56:54 PM, you wrote:
KR> Once I have run the sa-learn routine on a group of thousands of messages, is
KR> it necessary to keep all of those messages to run the next time I run
KR> sa-learn? It takes a while for sa-learn to process thousands of message so
K
On Thu, Jan 08, 2004 at 05:51:45PM -0600, Kevin Roberts wrote:
> Is there a way to tell the bayes learning system to forget everything and
> start over from a clean corpus sa-learn session? The reason I ask is I may
> have many emails that were auto posted and learned that were learned the
> wrong
On Wed, Jan 07, 2004 at 09:09:32AM -0500, Adam Denenberg wrote:
> i am getting the following bayes error when trying to do a
> force-expire. Is there any way around this so i can expire my old
> tokens?
>
> debug: bayes: expiry check keep size, 75% of max: 75
> debug: bayes: token count: 0,
That's OK then.
What probably made it more confusing was that I put them in to a dir called
bayes.
Alan
-Original Message-
From: Matt Kettler [mailto:[EMAIL PROTECTED]
Sent: 06 January 2004 19:34
To: Alan Munday; 'Spamassassin-Talk (E-mail)'
Subject: Re: [SAtalk] Baye
At 02:12 PM 1/6/2004, Alan Munday wrote:
And testing with bayes_path /var/spool/spamassassin/bayes/bayes worked fine.
From reading the docs, I don't think this is how this is supposed to work.
But it works.
No, from reading the docs it's EXACTLY how it should work.
It's a bit confusing, since the
Since you're using bayes do you get the same errors when you try to use
sa-learn?
If this was me I'd simply run sa-learn in the perl debugger and debug
how the bayesstore is being written.
I don't think your problem is spamd specific.
On Dec 19, 2003, at 3:06 AM, Lukreme wrote:
I make sure th
>> Can anyone give me a clue as to how to resolve this?
> Looks like you need to Install BerkeleyDB first... DB::File requires
it.
> Once you get those two loaded, bayes should work for ya.
Yep, the problem was that SA was looking for /usr/local/BerkeleyDB and
the sun "addpkg" was creating it
At 04:23 PM 12/19/03 -0800, S. M. C. Butler wrote:
version.c:30:16: db.h: No such file or directory
Can anyone give me a clue as to how to resolve this?
Looks like you need to Install BerkeleyDB first... DB::File requires it.
Once you get those two loaded, bayes should work for ya.
---
On Fri, Dec 19, 2003 at 04:23:36PM -0800, S. M. C. Butler wrote:
> When I try to install DB_File using cpan I get
>
> version.c:30:16: db.h: No such file or directory
>
> Can anyone give me a clue as to how to resolve this?
Install the Berkeley DB libraries and development files. see your OS
ve
On Fri, Dec 19, 2003 at 02:06:13AM -0700, Lukreme wrote:
> Now, somehow after processing a single message spamd has set the perms
> on the bayes_journal file to --
What is "bayes_file_mode" set to? By default it's 0700.
--
Randomly Generated Tagline:
If you want to program in C, progra
25
To
Alexei Moulton <[EMAIL PROTECTED]>
cc
[EMAIL PROTECTED]
Subject
Re: [SAtalk] bayes not being
used
Alexei Moulton wrote:
>I am running exim 4.12 with exiscan and spamassassin 6.1
>
>The list of TESTS in X-Spam-Report header never includes BAYES and
also
>the bayes s
Alexei Moulton wrote:
I am running exim 4.12 with exiscan and spamassassin 6.1
The list of TESTS in X-Spam-Report header never includes BAYES and also
the bayes score from _BAYES_ is always 0.5000. From this i gather that the
bayes database is not being used?
not sure exactly what spamassassin
At 05:42 AM 12/16/03 -0800, James Nonya wrote:
Dec 16 06:39:54 gateway spamd[9846]: Cannot open bayes
databases /var/empty/.spamassassin/bayes_* R/W: lock
failed: File exists
Anyone else see this? Thanks!
lock failures reported by spamd on occasion are somewhat normal.. It
happens when spamd tri
I would guess this is normal. Think of things like Message-Id's, vs. common
words like "the" which will appear very many times.
> -Original Message-
> From: Alexander Litvinov [mailto:[EMAIL PROTECTED]
> Sent: Saturday, December 13, 2003 10:10 AM
> To: [EMAIL PROTECTED]
> Subject: [SAtalk
well ham is very site dependant (point taken especially with security
reasons), so i dont think sending ham will really be something that can
be accomplished very easily so i think sticking to spam only would be
best. By only sending spam, i would hope to achieve a large spam corpus
that someone c
Adam Denenberg wrote:
> SA List,
>
> What i want to start is a Bayes Corpus Project. I would like to be
> able to allow people to submit confirmed ham and/or spam to a large
> bayes corpus repository (or maybe just spam) where people could then
> download (or somehow do an sa-learn remotely) to
On December 11, 2003 09:10 am, Adam Denenberg wrote:
> SA List,
>
> I am writing for feedback about a new project i would like to start
> and would love feedback/help from the excellent community that has been
> built here on the SA lists.
>
> What i want to start is a Bayes Corpus Project. I w
At 09:10 AM 12/11/03 -0500, Adam Denenberg wrote:
SA List,
What i want to start is a Bayes Corpus Project. I would like to be
able to allow people to submit confirmed ham and/or spam to a large
bayes corpus repository (or maybe just spam) where people could then
download (or somehow do an sa-le
What would differentiate the proposed public corpus from the public
corpus at http://www.spamassassin.org/publiccorpus/?
-tom
> -Original Message-
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On
> Behalf Of Adam Denenberg
> Sent: Thursday, December 11, 2003 8:10 AM
> To: [EMAI
At 02:46 PM 12/7/03 -0600, Michael Satterwhite wrote:
50_scores.cf:score BAYES_80 0 0 5.300 2.862
50_scores.cf:score BAYES_90 0 0 4.027 3.002
50_scores.cf:score BAYES_99 0 0 5.200 3.008
if I'm reading this correctly more points are given for classifying a message
as 80% probable than for 90% probab
On Sun, Dec 07, 2003 at 02:46:39PM -0600, Michael Satterwhite wrote:
> Can some of you experts explain this one to me?
http://spamassassin.taint.org/faq/index.cgi?req=show&file=faq01.005.htp
--
Randomly Generated Tagline:
"The highest patriotism is not a blind acceptance of official policy, but
On Sat, 6 Dec 2003, Lukreme wrote:
> spamd[33762]: Cannot open bayes databases
> /home/user/.spamassassin/bayes_* R/O: tie failed: Permission denied
> spamd[33762]: processing message
> <[EMAIL PROTECTED]> for kremels:5003.
> spamd[33762]: clean message (0.8/5.0) for user:5003 in 0.2 seconds,
> 55
On Wed, Dec 03, 2003 at 05:07:25PM -0800, Justin Mason wrote:
> I think the implementation of "each" in DB_File must be inefficient,
> creating an in-memory list. It might be a good idea to check out
> DB_File's built-in iterator methods, now that we only support that
> db module. Haven't check
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Theo Van Dinter writes:
>The expiry should, theoretically, not take a lot of memory since it
>really only needs to have 1 token in memory at a time while it copies
>between database files. An expiry forces a journal sync though, so an
>expiry run wil
On Wed, Dec 03, 2003 at 04:11:48PM -0600, Smart,Dan wrote:
> So if I set the Journal to a really big number, like 10MB, then the
> once-per-day opportunistic sync rule should kick in and sync once per day.
>
> So a sync takes as much or more memory than an expiration run?
Yeah, fun question: The
So if I set the Journal to a really big number, like 10MB, then the
once-per-day opportunistic sync rule should kick in and sync once per day.
So a sync takes as much or more memory than an expiration run?
<>
| -Original Message-
| From: Theo Van Dinter [mailto:[EMAIL PROTECTED]
| Sent
On Wed, Dec 03, 2003 at 01:47:55PM -0600, Smart,Dan wrote:
> Is there a reason that the journal needs to be synced more than once per
> day? Why is the default set to 100KB?
The default was reasonable for most users. There's no need, really,
to sync more than once a day, but a large journal mean
ct: Re: [SAtalk] Bayes Expiry limits and Journal size
|
| [EMAIL PROTECTED] (Justin Mason) writes:
|
| > Smart,Dan writes:
| >> I'm running Bayes site wide with about 45,000 messages per
| day. Is
| >> there any harm in increasing the bayes_expiry_max_db_size and
| >&g
Hello Scott,
Tuesday, December 2, 2003, 2:36:00 PM, you wrote:
SR> 2. My spam directory is up to 564 MB. Should I be worried about size
SR> at all? Eventual corruption? Is there a way to perform periodic DB
SR> maintenance? Is it ok to delete spam received say a month ago, if it
SR> has been
At 05:36 PM 12/2/2003, Scott Renda wrote:
I interpret them as someone who isn't using bayes right... ya need to train
some ham messages my friend :)
0.000 0 2 0 non-token data: bayes db version
Bayes database version 2.
0.000 0 49175 0 non-to
Yevgeniy Miretskiy <[EMAIL PROTECTED]> writes:
> sa-learn stopped learning messages. Debugging shows that it can
> successfully tie Bayes db, extracts tokens, etc, but never actually
> writes data to the database.
>
> No matter what bayes_expiry_max_db_size is set to (I tried anything from 100K t
Well... I never say never, but I switched into the correct directory,
checked with -V and it looks good.
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
>
> Jack Gostl writes:
> >
> >I need some help here guys.
> >
> >I have updated my Db to BerkeleyDB.4.2, I have reinstalled DB_file from
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Jack Gostl writes:
>
>I need some help here guys.
>
>I have updated my Db to BerkeleyDB.4.2, I have reinstalled DB_file from
>CPAN, then I wiped out my Bayes databases and started to rebuild from my
>spam corpus.
>
>When I run db_verify I still get th
On Tue, Dec 02, 2003 at 01:06:55PM -0600, Bob Apthorpe wrote:
> > 1. Can anybody tell me how to read the values obtained from sa-learn --dump?
> > The results of my dump are below:
> >
> > 0.000 0 2 0 non-token data: bayes db version
> > 0.000 0 49175
Hi,
On Tue, 2 Dec 2003, Scott Renda wrote:
> I am running SA 2.60 with Postfix, it has simply done an amazing job since
> implemented. I did have several questions though.
>
> 1. Can anybody tell me how to read the values obtained from sa-learn --dump?
> The results of my dump are below:
>
> 0.0
> -Original Message-
> From: Dan [mailto:[EMAIL PROTECTED]
> Sent: Monday, December 01, 2003 14:22
> To: [EMAIL PROTECTED]; Matt Kettler
> Subject: Re: [SAtalk] bayes
>
>
> I took a peek, and I guess my question is this.. How do I
> train on what
> is
?
Thanks
Dan
- Original Message -
From: "Matt Kettler" <[EMAIL PROTECTED]>
To: "Dan" <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Monday, December 01, 2003 10:55 AM
Subject: Re: [SAtalk] bayes
> At 01:31 PM 12/1/2003, Dan wrote:
> >Does anyone k
At 01:31 PM 12/1/2003, Dan wrote:
Does anyone know where I can find more info on bayes? I have it turned on in
my local.cf. But Im sure there is more to setting it up! ha ha ha ha
You need to manually train it.. see man sa-learn.
(yes, SA does autolearn some email, but you can't rely entirely on
On Mon, Dec 01, 2003 at 10:31:56AM -0800, Dan wrote:
> Does anyone know where I can find more info on bayes? I have it turned on in
> my local.cf. But Im sure there is more to setting it up! ha ha ha ha
"man sa-learn" may give you some help. :)
--
Randomly Generated Tagline:
"When all else fails
On Sat, 29 Nov 2003, Gorm Jensen wrote:
> Bayes is working very well for me, but I am concerned about poisoning the
> database with extraneous, obfuscating words that many spam messages contain.
>
> A few postings to this list say that there is no problem, but I don't want
> to spoil a good thing
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sunday 30 November 2003 06:34 pm, you wrote:
> PS to brook.. webmedic? Converted street person? or just a catchy
> name?
lol well i used to do medical work but no I started a computer store called
mobile pc medic and when i started my website ab
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sunday 30 November 2003 04:19 pm, you wrote:
> This is an old email address. I've been watching the spams slowly
> increase. Too many newsgroup posts, too much online ordering, too much
> buying/selling on ebay. Who knows. Until a few weeks ago, the
Hi Jack,
My Grandfathers name was Jack... great name :)
> -Original Message-
> From: [EMAIL PROTECTED]
[mailto:spamassassin-
> [EMAIL PROTECTED] On Behalf Of Jack Gostl
> Sent: Sunday, November 30, 2003 7:19 PM
> To: Brook Humphrey
> Cc: [EMAIL PROTECTED]
> Subjec
> On Sunday 30 November 2003 02:42 pm, you wrote:
> > So, yes, I could rebuild, something that takes 30-45 mintues,
> > but last time it took several tries. Worse is that I wind up with a
> > significant number of false negatives before I even discover the problem.
> >
> > Besides, this SHOULDN'T
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sunday 30 November 2003 02:42 pm, you wrote:
> So, yes, I could rebuild, something that takes 30-45 mintues,
> but last time it took several tries. Worse is that I wind up with a
> significant number of false negatives before I even discover the pro
> > Last time this happened I had to rebuild the databases. This time I
> > thought I'd wait a bit to hear from people before I "destroyed" the
> > evidence.
> >
> > Really is a very annoying problem with a first class product.
> cant say in your caase but the only time I have ever had this is whe
Nope... no such luck.
On Sun, 30 Nov 2003, Pedro Sam wrote:
> On November 30, 2003 07:55 am, Jack Gostl wrote:
> > Well... it happened again. My Bayes files are trashed. No explanation. All
> > values in "sa-learn --dump magic" come back zero. No messages of interest
> > in the debug output of
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sunday 30 November 2003 04:55 am, Jack Gostl wrote:
> Last time this happened I had to rebuild the databases. This time I
> thought I'd wait a bit to hear from people before I "destroyed" the
> evidence.
>
> Really is a very annoying problem with a
On November 30, 2003 07:55 am, Jack Gostl wrote:
> Well... it happened again. My Bayes files are trashed. No explanation. All
> values in "sa-learn --dump magic" come back zero. No messages of interest
> in the debug output of either the --dump or the --rebuild. At least none
> that I can tell. Lo
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hello Gorm,
Saturday, November 29, 2003, 1:52:13 PM, you wrote:
GJ> Bayes is working very well for me, but I am concerned about poisoning
GJ> the database with extraneous, obfuscating words that many spam
GJ> messages contain.
I used to be concerned
Gorm Jensen said:
> Bayes is working very well for me, but I am concerned about poisoning the
> database with extraneous, obfuscating words that many spam messages
> contain.
>
> A few postings to this list say that there is no problem, but I don't want
> to spoil a good thing. Are there some rule
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Wednesday 26 November 2003 01:21 pm, Nathan Triplett wrote:
> I'm running spamassassin 2.60 on mandrake 9.2 using kmail. Just recently I
> collected 200 spam and ham mails and after looking to check if the bayes
> filter started working I found tha
fritz
---
+ Basta Ikaw Lord
-Original Message-
From: JC [mailto:[EMAIL PROTECTED]
Sent: Tuesday, November 25, 2003 9:16 PM
To: [EMAIL PROTECTED]
Subject: RE: [SAtalk] bayes learning
Do you have spamassassin running as root? If not, log on as the user that
spamassassin runs as, and t
> Yes, that size seems way out of line. It should be using about 30~50
> bytes per token, assuming typical token size.
> According to your 'non-token data: ntokens' that bayes_toks file should
> be using about 5~6 Mbytes; unless something is whacko, or you have some
> -very- large tokens in there.
1 - 100 of 358 matches
Mail list logo
