On 10/16/19 4:11 PM, Bill Cole wrote:
> On 16 Oct 2019, at 8:44, Giovanni Bechis wrote:
> 
>> I have lot of emails like this one (redacted):
>> https://pastebin.com/v5NCRK9d
>> and I would like to write a rule that matches the "=0D" that appears on some 
>> lines,
> 
> Are you sure?
> 
> That's a QP-encoded carriage return. I would expect a lot of them in both 
> spam and ham.
> 
it is a part of a meta-experiment

>> any hints ?
> 
> You could try matching it as '\r' in a rawbody rule, but I'm not sure that 
> would work. If it does, you probably want '\r[^\n]' to exclude CRLFs, but you 
> should test that carefully
> 
> What should work better is to use a 'full' rule and look for the undecoded 
> '=D0', probably with '=0D(?!=0A)' or even '=0D(?!(=0A|=?$)'
> 
> e.g.:
> 
> full QP_BARE_CR /=0D(?!(=0A|=?$))/m
> 
> CAVEAT: not well tested...
> 
Seems to work, more tests later.
thanks
 Giovanni

Reply via email to