On Mon, Aug 09, 2010 at 02:23:17PM -0700, Chip Camden wrote: > Quoth Chris G on Monday, 09 August 2010: > > On Mon, Aug 09, 2010 at 09:20:05PM +0100, Chris G wrote: > > [snip my mutt/mbox/NFS issues] > > > > Can someone clarify something for me please, ignore NFS and assume I'm > > running mutt and the mail delivery agent on the same system on a local > > hard disk. > > > > If I open my inbox with mutt and leave it displaying the index of the > > inbox on the screen what should happen when a new message is delivered > > to the inbox? > > > > Should it:- > > > > 1 - Not get delivered because mutt has the inbox locked and the > > delivery agent can't open it for writing. > > > > 2 - Be delivered and appear correctly to mutt as a new message > > added at the bottom of the index listing with an 'N' beside it. > > > > 3 - Be delivered but cause mutt to complain "Mailbox was > > externally modified. Flags may be wrong." > > > > 4 - Something else. > > > > What I'm getting is 3 above which seems wrong to me, I suspect it is > > the underlying problem for my NFS issue, I just get a different error > > when viewing the mbox via NFS. > > > > I get (2), but only after keyboard input or the input timeout occurs. > Version 1.4.2.3i on FreeBSD. > Aha, thanks. That prompted me to go and try without my python delivery script (by using mutt to look at another user's inbox) and, as you say, it does 2 above. It even does 2 above when the inbox is mounted by NFS.
So my problem is down to my python delivery script which, although it's using the 'proper' python classes for mail obviously isn't behaving itself properly. How does one tell what locking method(s) is/are being used by mutt? This is on an Ubuntu 9.10 system running mutt 1.5.20. Doing 'mutt -v' returns "..... +USE_DOTLOCK +DL_STANDALONE +USE_FCNTL -USE_FLOCK....", does another process *have* to use dotlock and fcntl locking or could it use just fcntl? -- Chris Green
