Re: [Nmh-workers] refile handling of corrupt .mh_sequences

[Paul Fox]

hi mark --

address@hidden wrote:
 > 
 > I frequently have corrupt .mh_sequences files, most likely due
 > to interaction between procmail (using rcvstore) and claws-mail
 > (which updates .mh_sequences, but seems to ignore rcvstore locks,
 > unsurprisingly).
 > 
 > Regardless of the cause, I often end up with a .mh_sequences file that
 > has a line break between message numbers and looks like:

i have trouble dredging up sympathy for bugs that result from ignoring
a locking protocol.  :-)

i guess i avoid this by delivering from procmail to mbox files
(maildir would be fine too -- just never switched over), for which the
locking protocol is well known and understood.

when i want to read mail, i run a script that loops across all
(or optionally just a subset) of those mbox files, inc'ing into
corresponding mh folders.  the inc'ing script could be run from cron
if one wanted something closer to true asynchrony, i suppose.

paul
----------------------
 paul fox, address@hidden (arlington, ma, where it's 35.6 degrees)