Professional Software Engineering wrote:
er, and you're going to know the collqial names of everyone in your office?
"pookie" "snookums", "Dick" instead of "Richard", "Rob" vs. "Robert", etc?
Too many people don't provide a name along with email. It's rather arbitrary.
We're pretty small so the list of names will be pretty insignificant.
Pet names
like "pookie" don't apply, as we're required to put at least our first
or last name.
The only variety which could pose a problem would be middle initials, but
that's easy enough to work around.
As for it being arbitrary, so are the individual tests that comprise
most scanners
like spamassassin. But, use the sum of the individuals and you have
something.
Since we also use spamassassin, it would be easy for me to check if a
message
has say three stars *and* a lack of a name - it's a pretty good bet it's
spam.
The lack of a name won't be an end all, be all, of course, but due to
the fact that
a large portion of the emails we get *do* have a name provided, the rare
few that
don't have them (and are legit) should be negligible.
I'm not sure the whole concept is particularly feasable. Have you saved a
corpus of spam which you feel would meet the criteria? I _archive_ all the
spam that reached the LDA (plenty is warded away during the SMTP
phase). These I feed back into my spam filters at later revisions.
I don't save spam, no. But I have been watching the spam coming through
for the last 6 months or so and have noticed this pattern.
If you set _variables_ within procmail during message handling, there's no
need to add a header - if procmail is the tool you're expecting to take
action on the message (versus passing it along to the user to let their own
MUA filters deal with it).
Good point. I never think to use variables. My knowledge of procmail's
recipes is limited at best. Scoring and matching still leave me scratching
my head. (^= (yes, I've read the procmailsc man page, and I've been
watching
this list for about two years, and still ... )
0 = success
!0 = failure.
Qool. Thanks!
there are flags to say to ignore the exit code though. You also want to
pay attention to whether you're passing in only the header, and if your
filter doesn't read in the whole header -- if it terminates after it's
found what it is looking for without reading in everything on its STDIN,
then procmail may flag an error unless told to ignore that condition.
flags: I'm familiar with the flags. That's why I asked if I should
bother with
exit codes.
passing header: of course. For a script that uses only the header, there
would be no point in passing in the body, especially on large emails.
early termination: I assumed that reading a partial header would trigger
some kind of warning or error.
Thanks for the information! At minimum, I have enough info to sandbox
some tests and see what kind of trouble I can cause. (^=
TLD
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail