You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Han Boetes <hb...@utelisys.com> on 2011/01/24 11:13:08 UTC

Not spam but spamassassin adds 50 hitpoints

Hi,

A client of ours does bussiness with Atos Origin and their mail-server
automatically adds this text attachement:

http://boetes.org/~han/disclaimer.txt

Admittely, this is not a pretty attachement but it's not spam. How can I
explain to spamassassin it isn't spam?


I already asked Atos Origin to replace this disclaimer with

http://boetes.org/~han/disclaimer2.txt

Which is the same text but without the encoding errors.


Met vriendelijke groet,


Han Boetes
-- 
Utelisys


Re: Not spam but spamassassin adds 50 hitpoints

Posted by Martin Hepworth <ma...@gmail.com>.
Han

you don't mention which version of SA you're running with your MailScanner
implementation. Might be time to upgrade if you're running something that's
reasonably old.

-- 
Martin Hepworth
Oxford, UK


On 24 January 2011 13:54, Han Boetes <hb...@utelisys.com> wrote:

> On 01/24/2011 02:22 PM, Karsten Bräckelmann wrote:
> > The above is the *only* stock SA rule. The overall score of that ham
> > should be 0.8, way below the default threshold. All remaining rules
> > shown below are your local rules, responsible for the false positive.
> >
> > Hint: A rule that matches "mangled short English word" tends to be not
> > suitable for any other language.
>
> I started looking and there are a lot of rules from
>
> http://www.rulesemporium.com/
>
> Which is down. Also the date on those rules is 2004! That can't be
> effective anymore.
>
> Thanks for your help.
>
>
> Met vriendelijke groet,
>
>
> Han Boetes
> --
> Utelisys
>
>

Re: Not spam but spamassassin adds 50 hitpoints

Posted by Karsten Bräckelmann <gu...@rudersport.de>.
On Mon, 2011-01-24 at 14:54 +0100, Han Boetes wrote:
> On 01/24/2011 02:22 PM, Karsten Bräckelmann wrote:
> > The above is the *only* stock SA rule. The overall score of that ham
> > should be 0.8, way below the default threshold. All remaining rules
> > shown below are your local rules, responsible for the false positive.
> >
> > Hint: A rule that matches "mangled short English word" tends to be not
> > suitable for any other language.
> 
> I started looking and there are a lot of rules from
> 
> http://www.rulesemporium.com/
> 
> Which is down. Also the date on those rules is 2004! That can't be
> effective anymore.

Indeed. SARE, the SpamAssassin Rules Emporium is EOL'ed!

Some of their rules have been committed to stock SA long ago. Anything
else was either not effective, dangerous, or for ancient versions only.
Or both.


-- 
char *t="\10pse\0r\0dtu\0.@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; }}}


Re: Not spam but spamassassin adds 50 hitpoints

Posted by Han Boetes <hb...@utelisys.com>.
On 01/24/2011 02:22 PM, Karsten Bräckelmann wrote:
> The above is the *only* stock SA rule. The overall score of that ham
> should be 0.8, way below the default threshold. All remaining rules
> shown below are your local rules, responsible for the false positive.
>
> Hint: A rule that matches "mangled short English word" tends to be not
> suitable for any other language.

I started looking and there are a lot of rules from

http://www.rulesemporium.com/

Which is down. Also the date on those rules is 2004! That can't be
effective anymore.

Thanks for your help.


Met vriendelijke groet,


Han Boetes
--
Utelisys


Re: Not spam but spamassassin adds 50 hitpoints

Posted by Karsten Bräckelmann <gu...@rudersport.de>.
On Mon, 2011-01-24 at 13:59 +0100, Han Boetes wrote:
> On 01/24/2011 11:59 AM, Karsten Bräckelmann wrote:
> > Anyway, there's nothing we can do, unless you provide the full SA
> > headers at least. Maybe even a life sample, with those headers of
> > course.
> 
> I just sent myself a short test message with the afformentioned
> attachement and I get this SA report from MailScanner. Is that what
> you'd like to see? If not how do I get it?

Not the full report, but the rule names and scores. Yep, that's
sufficient.

> 0.80 BAYES_50

The above is the *only* stock SA rule. The overall score of that ham
should be 0.8, way below the default threshold. All remaining rules
shown below are your local rules, responsible for the false positive.

Hint: A rule that matches "mangled short English word" tends to be not
suitable for any other language.

> 2.30 MANGLED_BEST mangled best
> 2.30 MANGLED_CNFDTL mangled confidential
> 2.30 MANGLED_DELIVY mangled delivery
> 2.30 MANGLED_FORM mangled form
> 2.30 MANGLED_FREE mangled free
> 2.30 MANGLED_FROM mangled from
> 2.30 MANGLED_GOOD mangled good
> 2.30 MANGLED_MEDS mangled med(s)
> 2.30 MANGLED_MEN mangled men
> 2.30 MANGLED_NAIL mangled nail
> 2.30 MANGLED_OFF mangled off
> 2.30 MANGLED_PLEASE mangled please
> 2.30 MANGLED_WORKS mangled work(s)
> 2.30 MANGLED_YOUR mangled your
> 1.50 MIME_ASCII0 Message body contains ASCII-0 character
> 0.50 MIME_BAD_LINEBREAK Message body with fishy line breaks
> 0.20 MR_NOT_ATTRIBUTED_IP Beta rule: an non-attributed IPv4 found in headers
> 0.50 NO_RDNS Sending MTA has no reverse DNS (Postfix variant)
> 0.50 ONE_WORD_SUBJECT Subject with only one lower-case word

-- 
char *t="\10pse\0r\0dtu\0.@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; }}}


Re: Not spam but spamassassin adds 50 hitpoints

Posted by RW <rw...@googlemail.com>.
On Mon, 24 Jan 2011 13:59:48 +0100
Han Boetes <hb...@utelisys.com> wrote:

> On 01/24/2011 11:59 AM, Karsten Bräckelmann wrote:
> > Anyway, there's nothing we can do, unless you provide the full SA
> > headers at least. Maybe even a life sample, with those headers of
> > course.
> 
> I just sent myself a short test message with the afformentioned
> attachement and I get this SA report from MailScanner. Is that what
> you'd like to see? If not how do I get it?
> 
> 0.80 BAYES_50
> 2.30 MANGLED_BEST mangled best
> 2.30 MANGLED_CNFDTL mangled confidential
> 2.30 MANGLED_DELIVY mangled delivery
> 2.30 MANGLED_FORM mangled form
> ...

The MANGLED_* rules are non-standard, if you wish to keep them I would
suggest rewriting them as meta-rules. 

Re: Not spam but spamassassin adds 50 hitpoints

Posted by Han Boetes <hb...@utelisys.com>.
On 01/24/2011 11:59 AM, Karsten Bräckelmann wrote:
> Anyway, there's nothing we can do, unless you provide the full SA
> headers at least. Maybe even a life sample, with those headers of
> course.

I just sent myself a short test message with the afformentioned
attachement and I get this SA report from MailScanner. Is that what
you'd like to see? If not how do I get it?

0.80 BAYES_50
2.30 MANGLED_BEST mangled best
2.30 MANGLED_CNFDTL mangled confidential
2.30 MANGLED_DELIVY mangled delivery
2.30 MANGLED_FORM mangled form
2.30 MANGLED_FREE mangled free
2.30 MANGLED_FROM mangled from
2.30 MANGLED_GOOD mangled good
2.30 MANGLED_MEDS mangled med(s)
2.30 MANGLED_MEN mangled men
2.30 MANGLED_NAIL mangled nail
2.30 MANGLED_OFF mangled off
2.30 MANGLED_PLEASE mangled please
2.30 MANGLED_WORKS mangled work(s)
2.30 MANGLED_YOUR mangled your
1.50 MIME_ASCII0 Message body contains ASCII-0 character
0.50 MIME_BAD_LINEBREAK Message body with fishy line breaks
0.20 MR_NOT_ATTRIBUTED_IP Beta rule: an non-attributed IPv4 found in headers
0.50 NO_RDNS Sending MTA has no reverse DNS (Postfix variant)
0.50 ONE_WORD_SUBJECT Subject with only one lower-case word


Met vriendelijke groet,


Han Boetes
--
Utelisys


Re: Not spam but spamassassin adds 50 hitpoints

Posted by Karsten Bräckelmann <gu...@rudersport.de>.
On Mon, 2011-01-24 at 11:13 +0100, Han Boetes wrote:
> A client of ours does bussiness with Atos Origin and their mail-server
> automatically adds this text attachement:

Rather unlikely that any text on its own will result in a score of 50...

> Admittely, this is not a pretty attachement but it's not spam. How can I
> explain to spamassassin it isn't spam?

Anyway, there's nothing we can do, unless you provide the full SA
headers at least. Maybe even a life sample, with those headers of
course.


-- 
char *t="\10pse\0r\0dtu\0.@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; }}}