Paul O Bartlett wrote:
<> | The sendmail people have recently acknowledged that sendmail versions
<> | beyond 8.7.5 haven't worked properly with certain types of procmail
<> | rulesets. This problem has been corrected in 8.9.x; 8.9.1 was just
<> | released this month.
<>
<> Apparently the two recipes I recently posted were "certain types of
<> procmail rulesets" which did not work with the new sendmail/procmail
<> combination. (I do not know what the new operating systems are; I
<> suspect SunOS or Solaris.) Comments?
Yeah ... the guy who you quotes was smoking too much dope ... Sendmail
calls the local mail, which may or may not be procmail. If you've got
procmail called out of a .forward, it doesn't matter what Mlocal is 'cos
you're using the Mprog program mailer ...
The point is, one the message is handed off to procmail, the only thing
sendmail cares about is the return value, which tells it either the
message was "delivered" or not (Where the bitbucket, another program or a
file are possible values of "delivered").
As someone stated before: If some recipes work and other recipes don't
it's a procmail artifact. If _none_ work -- ie., procmail never gets
called correctly or isn't executable, or whatever -- _then_ it's a
sendmail problem.
Reto L.
--
R A Lichtensteiger rali(_at_)meitca(_dot_)com -or-
rali(_at_)world(_dot_)std(_dot_)com
http://www.meitca.com/ITA/People/rali
"Yes, you're doing things right, but are you doing the right things?"
"Nope. I'm just doing something dumb fast."