Greg Cirino - Cirelle Enterprises <[EMAIL PROTECTED]> wrote:

> We use this:
> 
> header SUBJECT_ENCODED_CE  Subject:raw =~ /=\?.*\?=/i
> describe SUBJECT_ENCODED_CE        Subject encoded
> score SUBJECT_ENCODED_CE 10.7
> 
> Appears to kill any encoded subject.

And how much real mail do you lose because of it?  Do you even 
know?  Assigning 10 points to a single rule -- especially one 
with so much potential for false positives -- seems to defeat 
the whole philosophy of SpamAssassin.

Note that your rule will block lots of mail from people outside 

the US.  I hope you don't subscribe to many software-related 
discussion lists (like this one).  It will even block mail from
Americans who happen to be silly enough to use a Microsoftian 
curly apostrophe in a subject line, and that happens all the 
time in a business environment.

I use a rule for catching some similar subjects, but it's 
limited to catch only those where the encoding is unnecessary 
(the characters in the subject are all ASCII), and I score it 
low because even then it hits some legitimate mail.  Here it 
is:

header __SUBJ_ASCII  Subject !~ /[^\t -~]/
header __SUBJ_EQ_BANG  Subject =~ /=\?/
header __SUBJ_ENCODED  Subject:raw =~ /=\?/
meta L_SUBJ_GRATUITOUS_ENCODING  __SUBJ_ASCII && 
!__SUBJ_EQ_BANG && __SUBJ_ENCODED
describe L_SUBJ_GRATUITOUS_ENCODING  Subject is encoded 
unnecessarily
score L_SUBJ_GRATUITOUS_ENCODING  1

-- 
Keith C. Ivey <[EMAIL PROTECTED]>
Washington, DC

Reply via email to