<<< Date Index >>>     <<< Thread Index >>>

Re: set sendmail_wait=-1, what happens on failure?



On Thu, Aug 25, 2005 at 11:13:19AM -0600, Bob Proulx wrote:
> Chris Green wrote:
> > I have the following in my muttrc:-
> > 
> > set sendmail="/proj/chris/bin/ssh -l chris <somewhere> /usr/lib/sendmail 
> > -oem -oi"
> > set sendmail_wait=-1
> > 
> > So, what happens when the <somewhere> system is unreachable?
> 
> I think if it were me that instead of asking that question I would
> create a script, say ~/bin/mysendmail, and I would set mutt's sendmail
> to that script.  I would not do anything with sendmail_wait and just
> let it run normally.
> 
> Now in that script you have complete control.  You can trap errors and
> handle them as you see fit.  You can save the output to a file or
> whatever you feel is useful.  If the ssh failed then keep the mail in
> a spool and keep retrying.  You can retry previous messages whenever
> there is a new message.  Lots of possibilities.
> 
> What I do is set mutt's sendmail to my own sendmail-offline script to
> spool messages in an outgoing directory.  Then when I connect to the
> internet I rsync those from my laptop to my mail relay and inject them
> into the sendmail there.  My own process is rather a hack and not
> suitable for sharing.  But perhaps it will give you ideas.
> 
I guess that's one way, but the remote system is so reliably
accessible that it seems hardly worth going to those lengths.  This is
the first time I've had a problem in a couple of years.

All I really need is something to warn me that it's broken so I know
the mail isn't getting sent.  There are then lots of workarounds - like
walking across the office!  :-)

-- 
Chris Green (chris@xxxxxxxxxxx)

    "Never ascribe to malice that which can be explained by incompetence."