Are there any known problems with nmh file locking? I'm using nmh-0.23
on an OpenBSD system with FCNTL_LOCKING defined. The .mh_sequences
file in one folder is corrupted 3-4 times a week. The only thing
writing to that folder is rcvstore run from slocal. I've noticed that
the last messages written to the folder prior to the .mh_sequences
corruption occur with approx the same timestamp (output of cron jobs).
nmh (and MH) do not currently lock the .mh_sequences file. That's why
many people (including myself) using the directory locking in procmail.
Adding such locking in on the TODO list.
--
Richard Coleman
coleman(_at_)math(_dot_)gatech(_dot_)edu