You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Ilan Aisic <il...@pointer.co.il> on 2004/02/25 11:36:08 UTC

Flase Positives on Hebrew Emails (because Hebrew is considered "raw illegal characters")

 Hi Everyone,
 
   In /etc/mail/spamassassin/local.df I have the following line:
 
      ok_languages en he
 
  This supposedly allows English and Hebrew. 
  However, I get lots of false positive on Hebrew letters because it wrongly
identifies Hebrew letters as "raw illegal characters" and gives them high
scores.
  You can see 7 points are contributed to the scroe just from these 2 lines
copied from the reports:
 
       4.3 FROM_ILLEGAL_CHARS     From contains too many raw illegal
characters
       2.7 SUBJ_ILLEGAL_CHARS     Subject contains too many raw illegal
characters

 Anyone can advise on this?
 
--
Ilan Aisic
Pointer Software Systems, Ltd.
Tel: +972-4-993-7330;  Fax: +972-4-993-5332, Cell: +972-64-815-533
 

RE: Flase Positives on Hebrew Emails (because Hebrew is considered "raw illegal characters")

Posted by Ilan Aisic <il...@pointer.co.il>.
OK thanks for enlightening me.
Indeed there are plenty of other Hebrew email messages that don't trigger
these rules.
I guess I'll have to educate some of the violators in here.

--ilan


> -----Original Message-----
> From: David B Funk [mailto:dbfunk@engineering.uiowa.edu] 
> Sent: Wednesday, February 25, 2004 1:24 PM
> To: Ilan Aisic
> Cc: spamassassin-users@incubator.apache.org
> Subject: Re: Flase Positives on Hebrew Emails (because Hebrew 
> is considered "raw illegal characters")
> 
> 
> On Wed, 25 Feb 2004, Ilan Aisic wrote:
> 
> >  Hi Everyone,
> >
> >    In /etc/mail/spamassassin/local.df I have the following line:
> >
> >       ok_languages en he
> >
> >   This supposedly allows English and Hebrew.
> >   However, I get lots of false positive on Hebrew letters 
> because it 
> > wrongly identifies Hebrew letters as "raw illegal characters" and 
> > gives them high scores.
> >   You can see 7 points are contributed to the scroe just 
> from these 2 
> > lines copied from the reports:
> >
> >        4.3 FROM_ILLEGAL_CHARS     From contains too many raw illegal
> > characters
> >        2.7 SUBJ_ILLEGAL_CHARS     Subject contains too many 
> raw illegal
> > characters
> >
> >  Anyone can advise on this?
> 
> No that isn't a false positive, that's an appropriate hit 
> against a bad e-mail client that is violating internet standards.
> 
> Internet standard RFC-2822 (section 2.2) unequivocally states 
> that you MUST use only 7-bit characters in e-mail HEADERS. If 
> you want to represent non-7-bit characters in a header (such 
> as 'From' or 'Subject') you must use some kind of encoding 
> (such as 'QP' or Base64), not the "raw" data.
> 
> Good e-mail client programs follow RFC standards and would 
> not generate such messages. Spammers usually are not 
> concerned with following standards and are more likely to 
> generate such garbage.
> 
> The "ok_languages" option sets the types of languages that 
> you will accept, after they're decoded. (IE what kinds of 
> character sets can be encoded).
> 
> Thus if somebody were to send you a message with an encoded 
> Korean subject your SA would score against that.
> 
> 
> -- 
> Dave Funk                                  University of Iowa
> <dbfunk (at) engineering.uiowa.edu>        College of Engineering
> 319/335-5751   FAX: 319/384-0549           1256 Seamans Center
> Sys_admin/Postmaster/cell_admin            Iowa City, IA 52242-1527
> #include <std_disclaimer.h>
> Better is not better, 'standard' is better. B{
> 


Re: Flase Positives on Hebrew Emails (because Hebrew is considered "raw illegal characters")

Posted by David B Funk <db...@engineering.uiowa.edu>.
On Wed, 25 Feb 2004, Ilan Aisic wrote:

>  Hi Everyone,
>
>    In /etc/mail/spamassassin/local.df I have the following line:
>
>       ok_languages en he
>
>   This supposedly allows English and Hebrew.
>   However, I get lots of false positive on Hebrew letters because it wrongly
> identifies Hebrew letters as "raw illegal characters" and gives them high
> scores.
>   You can see 7 points are contributed to the scroe just from these 2 lines
> copied from the reports:
>
>        4.3 FROM_ILLEGAL_CHARS     From contains too many raw illegal
> characters
>        2.7 SUBJ_ILLEGAL_CHARS     Subject contains too many raw illegal
> characters
>
>  Anyone can advise on this?

No that isn't a false positive, that's an appropriate hit against a
bad e-mail client that is violating internet standards.

Internet standard RFC-2822 (section 2.2) unequivocally states that you
MUST use only 7-bit characters in e-mail HEADERS. If you want to represent
non-7-bit characters in a header (such as 'From' or 'Subject') you must
use some kind of encoding (such as 'QP' or Base64), not the "raw" data.

Good e-mail client programs follow RFC standards and would not generate
such messages. Spammers usually are not concerned with following
standards and are more likely to generate such garbage.

The "ok_languages" option sets the types of languages that you will
accept, after they're decoded. (IE what kinds of character sets
can be encoded).

Thus if somebody were to send you a message with an encoded Korean
subject your SA would score against that.


-- 
Dave Funk                                  University of Iowa
<dbfunk (at) engineering.uiowa.edu>        College of Engineering
319/335-5751   FAX: 319/384-0549           1256 Seamans Center
Sys_admin/Postmaster/cell_admin            Iowa City, IA 52242-1527
#include <std_disclaimer.h>
Better is not better, 'standard' is better. B{

Re: Flase Positives on Hebrew Emails (because Hebrew is considered "raw illegal characters")

Posted by Graham Murray <gr...@gmurray.org.uk>.
"Ilan Aisic" <il...@pointer.co.il> writes:

>   However, I get lots of false positive on Hebrew letters because it
> wrongly identifies Hebrew letters as "raw illegal characters" and
> gives them high scores.  You can see 7 points are contributed to the
> scroe just from these 2 lines copied from the reports:
>  
>        4.3 FROM_ILLEGAL_CHARS     From contains too many raw illegal characters
>        2.7 SUBJ_ILLEGAL_CHARS     Subject contains too many raw illegal characters
>  Anyone can advise on this?

Yes, they are illegal characters in the From and Subject
lines. Non-ASCII characters in header line should be encoded using
RFC2047.