The code in mh.tcl re-writes the .mh_sequences files directly.
You could wack it to use "exec mark" instead, but exec is expensive.
Richard Coleman said:
It think this is an nmh problem (any objection).
I'm now running on nmh-0.21-pre2.
I'm on sunos.
I got bold...I didn't back up my .mh_sequences file.
This was not my inbox.
After a commit in exmh 2.0, I got a unseen sequence with one entry
(I should have several hundred unread messages...
My unseen sequence is still being eaten...
Hmm... it could be exmh rather than nmh. Could somebody from the exmh
developer camp tell me if exmh ever directly writes the .mh_sequences
file? Or does it call `mark' to do this.
I use plenty of unseen sequences, and I haven't noticed any
problems. But I usually use nmh at the command line.
If exmh doesn't ever write this file, I'll review the relevant
code in nmh and see if I can track it down.
--
Richard Coleman
coleman(_at_)math(_dot_)gatech(_dot_)edu
-- Brent Welch Sun Labs brent(_dot_)welch(_at_)sun(_dot_)com