possible bug in the way drafts works

possible bug in the way drafts works

[Harry Putnam]

Sorry this is so incomplete, but I haven't been able to reproduce this
at will, and so haven't nailed down the circumstances that cause this.

We had a debate a while back about how files should be auto saved to
drafts.  This debate was brought on by a change there that caused
messages to show up there even after being mailed.

I'm still seeing that behavior on occasion.  When I try to cause it on
purpose it doesn't happen.  But on a regular basis I find messages in
there that have already been mailed.  Not every message by any means
but fairly often, like several times per week.

Right now I have a message in drafts that I recall having
manually put there with C-c C-d then later going to it to finish it
and mailing it (I'm fairly sure) and there is a copy in my outgoing
mail archive. I don't want to possibley resend the same thing,
especially since it is part of a discussion with emacs-bugs people.
Don't want to look like a dork with those tecnicians...

Seems that both things should not be possible, at least not routinely.

If a message has appeared in the gcc destination it should not also be
in drafts... right?

My setup is somewhat more complicated by the fact that I'm running
unplugged, and never (or rarely go plugged) (Still network connected
though).  Outgoing/incoming is done with a batch fetch/send by cron in
5 minute time periods.  So with some code swiped from Lars long ago, I
never need to go plugged, just run M-x `lars-fetch-news' periodically.

Not sure what bearing that might have on drafts.  I'd think none, but
it does mean that queue gets a work out.

Re: possible bug in the way drafts works

[Lars Magne Ingebrigtsen]

Harry Putnam <reader@newsguy.com> writes:

> I'm still seeing that behavior on occasion.  When I try to cause it on
> purpose it doesn't happen.  But on a regular basis I find messages in
> there that have already been mailed.  Not every message by any means
> but fairly often, like several times per week.

One thing that could cause this is if you send a message, and then go
back to that message buffer and do some further editing.  Then it'll
be saved to the drafts group again.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: possible bug in the way drafts works

[Paul Jarc]

Lars Magne Ingebrigtsen <larsi@gnus.org> wrote:
> if you send a message, and then go back to that message buffer and
> do some further editing.  Then it'll be saved to the drafts group
> again.

In that case, could Gnus mark the draft article as 'sent or some such?


paul

Re: possible bug in the way drafts works

[Lars Magne Ingebrigtsen]

prj@po.cwru.edu (Paul Jarc) writes:

> Lars Magne Ingebrigtsen <larsi@gnus.org> wrote:
>> if you send a message, and then go back to that message buffer and
>> do some further editing.  Then it'll be saved to the drafts group
>> again.
>
> In that case, could Gnus mark the draft article as 'sent or some such?

Hm, yes...  After sending the message, I could add to the write-hook
thing a function that will save a header saying that it has already
been sent.  Like

  X-Message-Sent: yes

Or something like that.  And then the message code could check that.
Would that cover all the cases?  Hm...

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen