On 2/28/02 1:44 PM, "Greg Ward" <[EMAIL PROTECTED]> wrote:

> [I suggest a new rule]
>> Here's a quick and dirty attempt:
>> 
>> header TO_REALNAME_EQ_LOCALPART    To =~ /\"?(\w+)\"?\s+<\1\@[^<>]+>/i
>> describe TO_REALNAME_EQ_LOCALPART  Real name in "To:" equals local part
>> score TO_REALNAME_EQ_LOCALPART     2.5
> 
> [Craig responds]
>> Sounds like a good rule -- even people like me who use their name as email
>> address probably won't have many people using "Craig"
>> <[EMAIL PROTECTED]>; if they're going to go to the trouble of setting
>> the real name, they'll now what my last name is.
> 
> OK, what next?  Is that regex good enough?  I doubt it -- I cobbled it
> together off-the-top-of-my-head, and consulted only the copy of RFC 822
> which I carry around in my brain (not entirely infallible).  Do you want
> me to take a crack at honing the rule, so eg. something more appropriate
> than \w+ is used, or so it doesn't match when there are mismatched
> quotes?  Or do you want to take over from here?

Greg, I think I'm going to be busy for a few days just with scoring
refinements.  If you could solidify the regex I'd appreciate it.

> I'm not really clear on how strict SA's regexes should be -- ie. is the
> goal to follow the grammar in the RFC, or to just do a good enough job
> that most spam is caught?

I think as a general guide, the rules should be RFC-compliant, unless
there's a special reason not to be.  Also, 2822 should be your guide, not
822.

C


_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to