Phantom message 14

Phantom message 14

[Eric Abrahamsen]

I'm trying to compact my Sent messages group, (nnml mail numbers
currently run from 7112 to 10300) and every time I do gnus tells me
that it can't delete message 14. There is no message 14. It starts the
compacting process, writes a blank message to 1, then says there's no
message 14.

So far as I know, the active file in this backend and the .newsrc.eld
file are the only two places where gnus could be getting information
about existing messages. I've killed emacs, made sure the relevant line
in the active file reads:

Sent 10300 7112 y

and that .newsrc.eld doesn't have a 14 in it, then restarted emacs and
gnus. Again, the complaint about message 14. After I've done this, gnus
rewrites the active file to read:

Sent 10300 14 y

I just don't know why. Where else could this 14 be coming from?
References from other messages? Hidden configuration files? Can anyone
give me a logical place to start looking, before I turn to voodoo?

Thank you,
Eric

Re: Phantom message 14

[Eric Abrahamsen]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> I'm trying to compact my Sent messages group, (nnml mail numbers
> currently run from 7112 to 10300) and every time I do gnus tells me
> that it can't delete message 14. There is no message 14. It starts the
> compacting process, writes a blank message to 1, then says there's no
> message 14.
>
> So far as I know, the active file in this backend and the .newsrc.eld
> file are the only two places where gnus could be getting information
> about existing messages. I've killed emacs, made sure the relevant line
> in the active file reads:
>
> Sent 10300 7112 y
>
> and that .newsrc.eld doesn't have a 14 in it, then restarted emacs and
> gnus. Again, the complaint about message 14. After I've done this, gnus
> rewrites the active file to read:
>
> Sent 10300 14 y
>
> I just don't know why. Where else could this 14 be coming from?
> References from other messages? Hidden configuration files? Can anyone
> give me a logical place to start looking, before I turn to voodoo?
>
> Thank you,
> Eric

Still don't know where the 14 was coming from, but after looking in the
source code for a bit I went into the Sent group, popped (14 . "14") off
the nnml-file-article-alist variable, returned to the *Group* buffer, and
successfully compacted the group without, so far as I can tell, any lost
or mis-filed messages.