When I suggested this code to Brock Rozen,
| > :0
| > * !^X-Backup: Disabled
| > {
| > :0Wc: # $BACKPFILE$LOCKEXT will do just fine as the local lockfile
| > * < 500001
| > $BACKPFILE
| > :0EWhc:
| > $BACKPFILE
| >
| > :0
| > { EXITCODE = 75 HOST }
| > }
He asked,
| This last section worries me -- is seems like it will run every time a
| piece of mail comes through. If so, then one piece of mail should snowball
| -- as it will get sent back to sendmail each time, no?
Yes. Sorry. It was a case of half-rumped editing: originally the brace-
launching recipe orkfayed a clone, and the two saves to $BACKPFILE did not
have their own `c' flags, so the clone would have fallen to EXITCODE=75 HOST
only if the save to $BACKPFILE failed. But the parent procmail would still
deliver the message according to later recipes, and we didn't want that, so
I got rid of the clone and put `c' flags onto the saves to $BACKPFILE.
Thanks for catching it, Brock. Here's a fix:
:0
* !^X-Backup: Disabled
{
:0Wc: # $BACKPFILE$LOCKEXT will do just fine as the local lockfile
* < 500001
$BACKPFILE
:0EWhc:
$BACKPFILE
:0
* $ ! LASTFOLDER ?? (^^|/)$BACKPFILE^^
{ EXITCODE = 75 HOST }
}
| I'm just wondering, shouldn't it be :0e ?
An `e' flag might do instead of the test on $LASTFOLDER, but I'm not sure
exactly how that would work with the preceding `E' flag.