Re: [feature req] Configurable behaviour after MTA failure

2001-05-19 Thread Dave Pearson

On Sat, May 19, 2001 at 02:00:10AM +0530, Joane Lispton wrote:
 Hi Dave,
 
 Sounds like you're trying too hard here. With sendmail (insert your MTA
 of choice here as I'm sure they all work in a similar way) all I simply
 need to do is type mailq and I can see what's still in the queue and
 why.
 
 Your are right, of course. Dan also pointed that out. But, using mailq
 still means that I must put mutt into the background and run some command,
 which is pretty different from knowing the Mail sent means ... well,
 exactly that! :-)

True, there's a difference in your case. I wasn't suggesting that your
method of working off line as any less valid than mine (it seems that I'm
really off-line but you have some on-line version of off-line, dial on
demand perhaps?).

As for putting mutt into the background. No need for that at all. A quick
mutt macro to run mailq would let you feel that the facility was part of
mutt itself. I run all sorts of external tools from mutt to help me manage
email.

Point here being, there are far easier and far more integrated ways of doing
this (that don't require mutt to be modified) than scanning logs. I know
this doesn't make your environment as useful as you'd like but the current
setup doesn't need to be as unwieldy as you've reported it is.

-- 
Dave Pearson:  | mutt.octet.filter - autoview octet-streams
http://www.davep.org/  | mutt.vcard.filter - autoview simple vcards
Mutt:  | muttrc2html   - muttrc - HTML utility
http://www.davep.org/mutt/ | muttrc.sl - Jed muttrc mode



Re: [feature req] Configurable behaviour after MTA failure

2001-05-18 Thread Dave Pearson

On Thu, May 17, 2001 at 11:49:15PM -0400, Walt Mankowski wrote:
 On Thu, May 17, 2001 at 12:49:27PM +0530, Joane Lispton wrote:

  So, if, e.g., the PPP link is down and I haven't noticed that, mutt will
  tell me that my email has been sent, while it is actually stored in some
  directory on my hard-drive, awaiting my next connection to the
  Internet... probably established to download the replies to the emails I
  _thought_ I had sent!
 
 I actually think queueing is a *feature* on laptops. I frequently compose
 emails while I'm disconnected (say on a train or in a coffeeshop) and
 intentionally let the MTA queue them up. When I'm back online the MTA will
 take of sending them without my worrying about it.

It's not just laptops. Any machine with a dial-up connection where that
connection is dial-on-demand really needs queueing. I've had a net
connection like this since 1995 and have been using sendmail to do the queue
handling all that time.

When I bring up the link the queue gets flushed. Nice and simple.

-- 
Dave Pearson:  | mutt.octet.filter - autoview octet-streams
http://www.davep.org/  | mutt.vcard.filter - autoview simple vcards
Mutt:  | muttrc2html   - muttrc - HTML utility
http://www.davep.org/mutt/ | muttrc.sl - Jed muttrc mode



Re: [feature req] Configurable behaviour after MTA failure

2001-05-18 Thread Dave Pearson

On Fri, May 18, 2001 at 04:43:02PM +0100, Dave Pearson wrote:

 It's not just laptops. Any machine with a dial-up connection where that
 connection is dial-on-demand really needs queueing. [SNIP]

That should have read /isn't/ dial-in-demand.

-- 
Dave Pearson:  | mutt.octet.filter - autoview octet-streams
http://www.davep.org/  | mutt.vcard.filter - autoview simple vcards
Mutt:  | muttrc2html   - muttrc - HTML utility
http://www.davep.org/mutt/ | muttrc.sl - Jed muttrc mode



Re: [feature req] Configurable behaviour after MTA failure

2001-05-18 Thread Dan Boger

On Fri, 18 May 2001, Joane Lispton wrote:
 To sum it up, when it comes to the trade-off between
 
 - doing all the queueing manually, and knowing (from within mutt) that mail 
 has been successfully relayed
 
 and
 
 - having a MTA do the queueing for me, yet, if I wish to make sure that 
 something is already at the relay host, I have to dig in log files,

wouldn't mailq still tell you the mail is in the queue, and hasn't been
flush yet?  not saying that the option not to store the fcc on failed
delivery is a bad thing, just another way to check :)

Dan




Re: [feature req] Configurable behaviour after MTA failure

2001-05-18 Thread Dave Pearson

On Fri, May 18, 2001 at 11:48:23PM +0530, Joane Lispton wrote:
 Hi Dave,
 
 When I bring up the link the queue gets flushed. Nice and simple.
 
 That's my problem with queueing MTAs: if the link goes down before the MTA
 relays the emails (e.g., my connection is accidently dropped while I am in
 mutt), I think they were flushed but they still are on my hard-drive. And
 to really make sure that mail got _sent_, I have to look at SMTP log
 files... :-(

Sounds like you're trying too hard here. With sendmail (insert your MTA of
choice here as I'm sure they all work in a similar way) all I simply need to
do is type mailq and I can see what's still in the queue and why.

 To sum it up, when it comes to the trade-off between
 
 - doing all the queueing manually, and knowing (from within mutt) that mail 
 has been successfully relayed
 
 and
 
 - having a MTA do the queueing for me, yet, if I wish to make sure that 
 something is already at the relay host, I have to dig in log files,

I've never needed to dig in the logs to see if anything has gone. I simple
check the queue. Further to that, most of my net connections are done at
timed intervals and are brought up via cron. Quite often I'm not even at my
desk when this happens. As you might imagine any kind of manual intervention
on my part, be it handling the queue by hand or reading logs, would be out
of the question.

-- 
Dave Pearson:  | mutt.octet.filter - autoview octet-streams
http://www.davep.org/  | mutt.vcard.filter - autoview simple vcards
Mutt:  | muttrc2html   - muttrc - HTML utility
http://www.davep.org/mutt/ | muttrc.sl - Jed muttrc mode