🚀 go-pugleaf

Every once in a while a message will be handed to sendmail that has an
unbalanced '"' in one of the headers.  This gives sendmail indigestion
because when it generates error mail, it re-parses the header and
generates *another* message.  This recursion continues until all
machines that were implicated (sender, gateway, homes of support
people) are up to the max load average the particular daemon was
configured for, and the mail queue just continues to grow.  The only
way I know to stop one of these fires is to SIGSTOP the sendmail daemon
on every machine involved, and then go and manually delete every copy
of the offending message from the mail queues before starting the
daemons back up.  This is a real inconvenience.  The remedy sometimes
has to be repeated a time or two, as reflections from remote machines
come back.

Am I missing something?  Maybe there's a magic cookie that goes in
the sendmail.cf to control this (he said, hopefully).

Richard