|
|
 | | From: | Ville Herva | | Subject: | Using a signal to commit changes to mailbox and exit | | Date: | 31 Dec 2004 22:49:51 GMT |
|
|
 | I assume this has been brought up before, but I failed to find it in the
archives.
Every once in a while I run into a situation, where I have mutt running
elsewhere holding changes to a mailbox. I would like to exit that mutt
instance so that it commits the changes to be able to start a new mutt
instance using the mailbox. (kill -INT exits the mutt without commiting the
changes to the mailbox.)
I usually run mutt only under screen (so please don't suggest that :), but
sometimes (rarely) I don't (for whatever reason, forgetting to do so being
most likely one) - and then these situations come up. Slightly more common
case is where screen goes into a state where I can no longer attach to it.
In that case, I'd obviously prefer to be able to exits the mutt's under it
gracefully.
It would be nice if there was a signal (perhaps kill -USR1 ?) upon which
mutt would commit changes and exit.
Alternative suggestions welcomed. Ones that work without patching mutt get
bonus :).
-- v --
v@iki.fi
|
|
| | From: | Sven Guckes | | Subject: | Re: Using a signal to commit changes to mailbox and exit | | Date: | 1 Jan 2005 13:06:07 GMT |
|
|
| * Ville Herva [2004-12-31]:
> Every once in a while I run into a situation, where I have
> mutt running elsewhere holding changes to a mailbox. I would
> like to exit that mutt instance so that it commits the changes
> to be able to start a new mutt instance using the mailbox.
'q' (quit)
> (kill -INT exits the mutt without commiting the changes to the mailbox.)
duh
> I usually run mutt only under screen (so please don't suggest
> that :), but sometimes (rarely) I don't (for whatever reason,
> forgetting to do so being most likely one) - and then these
> situations come up. Slightly more common case is where screen
> goes into a state where I can no longer attach to it.
> In that case, I'd obviously prefer to be able
> to exits the mutt's under it gracefully.
this is not a mutt problem.
> It would be nice if there was a signal (perhaps kill -USR1 ?)
> upon which mutt would commit changes and exit.
and you'd also like signals not only for "exit" and "quit",
but also for all the other commands, right? think again!
> Alternative suggestions welcomed.
> Ones that work without patching mutt get bonus :).
killall -9 mutt
but you'd better hide that from the RSPCA..
Sven
|
|
| | From: | Ville Herva | | Subject: | Re: Using a signal to commit changes to mailbox and exit | | Date: | 1 Jan 2005 15:51:29 GMT |
|
|
| In article <2005-01-01T13-02-49@guckes.net>, Sven Guckes wrote:
> * Ville Herva [2004-12-31]:
> > Every once in a while I run into a situation, where I have
> > mutt running elsewhere holding changes to a mailbox. I would
> > like to exit that mutt instance so that it commits the changes
> > to be able to start a new mutt instance using the mailbox.
>
> 'q' (quit)
Yes, but how do I do that if my screen session is malfunctional or the mutt
instance was not started under mutt in the first place?
> > I usually run mutt only under screen (so please don't suggest
> > that :), but sometimes (rarely) I don't (for whatever reason,
> > forgetting to do so being most likely one) - and then these
> > situations come up. Slightly more common case is where screen
> > goes into a state where I can no longer attach to it.
> > In that case, I'd obviously prefer to be able
> > to exits the mutt's under it gracefully.
>
> this is not a mutt problem.
>
> > It would be nice if there was a signal (perhaps kill -USR1 ?)
> > upon which mutt would commit changes and exit.
>
> and you'd also like signals not only for "exit" and "quit",
> but also for all the other commands, right? think again!
I don't think other actions need a signal.
Thinking further, this might be sensible default action for HUP signal.
This is what pine does when it gets a HUP signal: it commits the changes so
that they are not lost. In fact, my colleague (being a Pine fan) makes fun
of me when I lose my changes on HUP. (Preferably this might be
configurable.)
> killall -9 mutt
Thank you very much for your valuable insights.
-- v --
v@iki.fi
|
|
| | From: | Sven Guckes | | Subject: | Re: Using a signal to commit changes to mailbox and exit | | Date: | 1 Jan 2005 20:02:33 GMT |
|
|
| * Ville Herva [2005-01-01]:
> Sven Guckes wrote:
>> * Ville Herva [2004-12-31]:
>> > Every once in a while I run into a situation, where I have
>> > mutt running elsewhere holding changes to a mailbox. I would
>> > like to exit that mutt instance so that it commits the changes
>> > to be able to start a new mutt instance using the mailbox.
>>
>> 'q' (quit)
>
> Yes, but how do I do that if my screen session is malfunctional
!echo $STY
but why add code for mutt that?
> or the mutt instance was not started under mutt in the first place?
huh?
>> > It would be nice if there was a signal (perhaps kill -USR1 ?)
>> > upon which mutt would commit changes and exit.
>>
>> and you'd also like signals not only for "exit" and "quit",
>> but also for all the other commands, right? think again!
>
> I don't think other actions need a signal.
however, once you add actions for signals,
other users will add requests for more.
"why doesn't it fetch my mail when
i give it the 'kick' signal?" etc
> Thinking further, this might be sensible
> default action for HUP signal.
> (Preferably this might be configurable.)
i sense a patch coming up..
> This is what pine does when it gets a HUP signal:
> it commits the changes so that they are not lost.
> In fact, my colleague (being a Pine fan)
> makes fun of me when I lose my changes on HUP.
well... "use pine!"
Sven
|
|
|
