dattier(_at_)wwa(_dot_)com (David W. Tamkin) writes:
Philip Guenther revised his suggestion for Ken Marsh,
| Okay, how about this instead:
|
| # preset or unset some variables
| R TOCOUNT=0 CCCOUNT=0
|
| # set $R to be Resent- iff there are any headers that indicate this message
| # as having been resent.
| :0
| * ^Resent-(From|Date|To|Cc|Message-Id):
| { R="Resent-" }
How about just this?
:0
* ^Resent-
{ R=Resent- }
Well, a "Resent-Foo: bar" header shouldn't cause use to look for
Resent-To: or Resent-Cc: headers, as "Resent-Foo:" is ignored by the
MTA. I believe the ones listed above to be the only ones that force
the MTA to process Resent-To/Cc/etc instead of To/Cc/etc
...
You're right, Philip: if there are Resent- headers, then those without
"Resent-" don't matter and shouldn't be counted. We should include
Apparently-(Resent-)To: as well, though.
There is no "Apparently-Resent-To:" (or "Resent-Apparently-To:") header.
Arguably, an "Apparently-To:" header should only be counted iff none of
the four headers (Resent-)To: and (Resent-)Cc: exist. The following
recipe, to be added after the set I gave previously should do it:
# Check for "Apparently-To:" headers. There should only be one address
# per header. The 'E' flag is extraneous.
:0E
* ! ^(Resent-)?(To|Cc):
* 1^1 Apparently-To:
* -19^0
{ EXITCODE=77 HOST }
[Curiously, I've never seen commas in Apparently-(Resent-)To: lines.
Whenever there is more than one apparent recipient, they all get their own
fields.]
I think you're right, and have depended on it above. Since
Apparently-To: was a pure sendmailism, this should be mostly
dependable.
Philip Guenther