"Gregory M. Paris" <paris(_at_)dma(_dot_)isg(_dot_)mot(_dot_)com> writes:
Is customizing the config.h file hacking the source? If so, then no
need to read the rest of this.
Well, yes. The idea is that I want to use a plain procmail without having to
recompile it at all (I'm really not afraid of hacking the source, I just want
to configuration do be done after compilation. I don't like hardcoded config
parameters). Let's say that I read the rest anyway, just in case.
This is exactly how I run procmail here, avoiding lots of superfluous
automounts of people's home directories. Both .forward and .procmailrc
files are copied to a central directory and the ones in the users'
homes are never accessed by sendmail or procmail (which is the MDA here).
Sounds a lot like my setup.
Works great, except that users need to be reminded that the copies in
their home directory don't immediately go into effect.
I mostly avoid the problem by having the equivalent of your
/etc/mail/procmailrcs NFS exported and have ~/.procmailrc be a symlink pointing
to the (usually non-existant) procmailrc file. This symlink is auto-created
when adding a user and regularily checked by a cron script (which you seem to
also have, if I understand the "don't immediately go into effect" correctly).
Has anybody another solution ?
It looks like it should be possible to force procmail to exit by using
at the very end of the /etc/procmailrc, but does it work ? and does it keep
the right semantics of EXITCODE and such ?
How about having procmail-3.12 use ~/.procmailrc only if /etc/procmailrc cannot
be found ?