On Sun, 2009-05-17 at 22:50 -0400, Matt Kettler wrote:
> Pieter De Wit wrote:
> > Here is the weird part - some mail comes up as spam when I hand it over to
> > spamd, but if I take the archive file and strip the header (EHLO etc) out,
> > save it it gets a MUCH lower score - I am talking 5.7/5.0 for spamd and
> > 1.x/5.0 for spamc.
> >
> > By spamd I mean connect to the port send the headers CHECK SPAMC/1.2\r\n
> > (and another line I think) and then sending the mail where as spamc is "cat
> > /archive_file | spamc -c"
You're not sending a User header? spamc does. So it is quite likely
spamd (which actually is processing the mail in both cases) will scan on
behalf of different users. Which will result in different Bayes and AWL
results, possibly even different rules' scores and other per-user conf,
unless you are using a strictly site-wide setup.
> > I guess my question is - is what I am sending to spamd correct ?
Did you have a close look at the spamd protocol docs? Did you ever
pretend to be the spamd server yourself, looking at what spamc sends it?
> Maybe. Maybe not. Are you sending the "Content-Length:" header after the
> check?
>
> To really make any decent comparison you'd want to look at the
> X-Spam-Status headers that each generates.
>
> The difference in score alone doesn't tell us much, however, the rules
> that matched could tell us quite a lot.
Definitely.
> If in doubt, post both headers to the list and we'll make some
> observations/suggestions (or at least I will, but it is likely others
> will chime in too).
/me too :)
--
char *t="\10pse\0r\0dtu...@ghno\x4e\xc8\x79\xf4\xab\x51\x8a\x10\xf4\xf4\xc4";
main(){ char h,m=h=*t++,*x=t+2*h,c,i,l=*x,s=0; for (i=0;i<l;i++){ i%8? c<<=1:
(c=*++x); c&128 && (s+=h); if (!(h>>=1)||!t[s+h]){ putchar(t[s]);h=m;s=0; }}}
