Actually, using -a with -S should increase the false _negative_ rate, not the
false positive rate, since it will lower the AWL score for a spammy sender to
around 5.
C
Duncan Findlay wrote:
DF> On Mon, Jun 10, 2002 at 02:34:54PM +0200, Michael Stauber wrote:
DF> > Hi Craig,
DF> >
DF> > > The -S
Hi,
Running masscheck I got the following errors:
Premature end of base64 data at
/home/java/on/CVS/spamassassin/masses/../lib/Mail/SpamAssassin/PerMsgStatus.pm line
1875, line 15606.
Premature end of base64 data at
/home/java/on/CVS/spamassassin/masses/../lib/Mail/SpamAssassin/PerMsgStatus.
On Mon, Jun 10, 2002 at 02:34:54PM +0200, Michael Stauber wrote:
> Hi Craig,
>
> > The -S flag to spamd should also help greatly in constrained hardware
> > situations.
>
> Yes, I know. It sure is a great way to cut back the load, but the scoring gets
> a little less effective and I noticed a f
On Monday 10 June 2002 07:42 am, Vivek Khera wrote:
> > "MC" == Matthew Cline <[EMAIL PROTECTED]> writes:
> MC> automatically compile Perl down to machine instructions for improved
> MC> performance. So it probably wouldn't be worth it to re-write SA in C.
> I doubt that will help anyway.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Craig R. Hughes said:
> Sean Rima wrote:
>
> SR> I thought it would not affect me too much but even with a -m 2 and
> SR> a -s 61440 I reach a load average of over 15 which cripples my
> SR> poor old mail box :)
>
> Hmm, my unde
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Craig R. Hughes spake thusly:
> The -S flag to spamd should also help greatly in constrained hardware
> situations.
>
Never thought of that, thanks Craig.
Sean
- --
Sean Rimahttp://www.tcob1.n
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Jeremy Zawodny uttered the following:
> On Mon, Jun 10, 2002 at 12:33:13AM +0100, Sean Rima wrote:
>>
>> > On Mon, Jun 10, 2002 at 12:19:20AM +0100, Sean Rima wrote:
>> >> >
>> >> > It's going to be a rather long wait, I think.
> "MC" == Matthew Cline <[EMAIL PROTECTED]> writes:
MC> automatically compile Perl down to machine instructions for improved
MC> performance. So it probably wouldn't be worth it to re-write SA in C.
I doubt that will help anyway. The majority of the time spent in SA
is in the pattern matc
Hi Craig,
> The -S flag to spamd should also help greatly in constrained hardware
> situations.
Yes, I know. It sure is a great way to cut back the load, but the scoring gets
a little less effective and I noticed a few more false positives. Can't be
avoided the way -S works, sure. So I usually
Matthew Cline wrote:
> On Sunday 09 June 2002 03:06 pm, Sean Rima wrote:
>
>
>>My ISP is also looking at SA but performance maybe a problem.
>
>
> When Perl 6 comes out, it will have a Just In Time (JIT) compiler, which will
> automatically compile Perl down to machine instructions for improv
Sean Rima wrote:
SR> I thought it would not affect me too much but even with a -m 2 and a -s
SR> 61440 I reach a load average of over 15 which cripples my poor old mail
SR> box :)
Hmm, my understanding of "load" is "number of processes in the WAIT queue",
which with -m 2 can only be 3 for SA --
The -S flag to spamd should also help greatly in constrained hardware
situations.
C
Michael Stauber wrote:
MS> Hi Sean,
MS>
MS> > My ISP is also looking at SA but performance maybe a problem.
MS>
MS> Yes, it sure can be. I've set up SA for a couple of ISPs on their Cobalt RaQ
MS> servers. Those
On Mon, Jun 10, 2002 at 12:33:13AM +0100, Sean Rima wrote:
>
> > On Mon, Jun 10, 2002 at 12:19:20AM +0100, Sean Rima wrote:
> >> >
> >> > It's going to be a rather long wait, I think. Perl 6 probably
> >> > won't be production ready for a few years.
> >> >
> >>
> >> Darn, off to the forsale N
On Sunday 09 June 2002 03:53 pm, Derrick 'dman' Hudson wrote:
> The real problem you have is not so much spamd's performance, but the
> fact that it sits idle 99% of the time, then gets slammed with 1K
> messages in a matter of seconds. Even with my Duron 750 and not much
> else happening, the b
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Jeremy Zawodny yowled:
> On Mon, Jun 10, 2002 at 12:19:20AM +0100, Sean Rima wrote:
>> >
>> > It's going to be a rather long wait, I think. Perl 6 probably
>> > won't be production ready for a few years.
>> >
>>
>> Darn, off t
On Mon, Jun 10, 2002 at 12:19:20AM +0100, Sean Rima wrote:
> >
> > It's going to be a rather long wait, I think. Perl 6 probably won't
> > be production ready for a few years.
> >
>
> Darn, off to the forsale NG's to price a new P4 methinks :)
If you could use an old Celeron 300A board and CP
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Jeremy Zawodny uttered the following:
>> >> My ISP is also looking at SA but performance maybe a problem.
>> >
>> > When Perl 6 comes out, it will have a Just In Time (JIT) compiler,
>> > which will automatically compile Perl dow
On Sun, Jun 09, 2002 at 11:45:33PM +0100, Sean Rima wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> On Sun, 9 Jun 2002, Matthew Cline stipulated:
>
> >> My ISP is also looking at SA but performance maybe a problem.
> >
> > When Perl 6 comes out, it will have a Just In Time (JIT) co
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Matthew Cline stipulated:
>> My ISP is also looking at SA but performance maybe a problem.
>
> When Perl 6 comes out, it will have a Just In Time (JIT) compiler,
> which will automatically compile Perl down to machine instruction
On Sunday 09 June 2002 03:06 pm, Sean Rima wrote:
> My ISP is also looking at SA but performance maybe a problem.
When Perl 6 comes out, it will have a Just In Time (JIT) compiler, which will
automatically compile Perl down to machine instructions for improved
performance. So it probably woul
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Mon, 10 Jun 2002, Michael Stauber stipulated:
>> My ISP is also looking at SA but performance maybe a problem.
>
> Yes, it sure can be. I've set up SA for a couple of ISPs on their
> Cobalt RaQ servers. Those RaQs are pretty outdated boxes (AMD K
Hi Sean,
> My ISP is also looking at SA but performance maybe a problem.
Yes, it sure can be. I've set up SA for a couple of ISPs on their Cobalt RaQ
servers. Those RaQs are pretty outdated boxes (AMD K6/2 with 300-450 MHz) and
performance wise they don't have much to offer.
Even with SPAMd/
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sun, 9 Jun 2002, Jeroen Scheerder told this:
>>Is there any word on the C version of SA as I am getting him by
>>performance issuses.
>
> A C version of the daemon, you mean? Or are you running "spamassassin
> -P"?
>
> Using the C client (spamc)
Sean Rima (9/6/02 22:05 +0100) [[SAtalk] C version]:
>Is there any word on the C version of SA as I am getting him by
>performance issuses.
A C version of the daemon, you mean? Or are you running "spamassassin -P"?
Using the C client (spamc) against a local spamd (perl), most messages
(about 75
24 matches
Mail list logo
