On Thu, 14 Mar 2002, Rob McMillin wrote:
> Is it too much to assume that eight-bit characters in the e-mail part
> of an address is a sign of junk? I get a lot of Asian spam in this
> form, but I understand Unicode domains are on their way, so it will
> now be possible for me to receive mail from a domain I can't possibly
> type in to my computer.
The format of those has not been standardized yet, certain large domain
marketing groups opinions to the contrary.
> I characterize this as follows:
>
> header TO_8BIT To =~ /[\x80-\xff]/
> describe TO_8BIT Addressee has 8 bit characters
> score TO_8BIT 5
This isn't a legal DNS name, though. It would seem reasonable to match
it but, er, are you /really/ getting 8-bit characters in the headers?
I don't see raw 8-bit encoded information. I see it encoded to 7-bit, at
least in headers, pretty much exclusively.
> Another test I would propose:
>
> header TO_ADMIN_ACCT To =~ /(?:webmaster|root|postmaster)\@/
> describe TO_ADMIN_ACCT Sent to an administrative account aliased to me
> score TO_ADMIN_ACCT 1.5
>
> I get bunches of spam sent to these addresses, none of which are me,
> all of which are widely known to exist.
Er, that would really suck because, like, some people on the list
actually do get mail sent to those addresses. Like me, for example.
It's also worth noting that filtering on "postmaster" and "root" is
/strongly/ discouraged for very good reasons -- they can be the only way
to get a real human contact.
Daniel
--
Bargaining is essential to the life of the world;
but nobody has ever claimed that it is an enobling process.
-- Agnes Repplier
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
