In message <>, Philip Guenther dixit:
>Depending on how exmh internally keeps track of the changes to be made,
>this could be worse: if it just caches the sequences and delays writing
>them back out till you commit or leave, that creates a _huge_ window
>wherein changes by rcvstore will be lost. On the other hand, if it
>instead remembers "remove 56, 57, 58 from sequence foo and add 104 and
>245 to it", then rereads the sequences file, applies the changes, and
>writes it back out, then this will indeed minimize the window of
>danger. The details on this may be infernal residences.
Looks like we have a candidate for a new NMH program:
User Commands USF(1)
NAME
usf - update sequence file
SYNOPSIS
usf sequence [+folder]
DESCRIPTION
Usf atomically updates the sequence file 'sequence' ...
etc.
Patrick O'Callaghan <[EMAIL PROTECTED]> <http://www.ldc.usb.ve/~poc>
Universidad Simon Bolivar, Caracas, Venezuela | "Errare
Tel: +58 (2) 906-3200, FAX: +58 (2) 906-3202 | uHmanum
Handle: PO22-ARIN (Postal address on request)| Est"
