Hello!
El jue, 19-09-2002 a las 07:37, Adrian von Bidder escribió:
...
> I think these names are not really clear enough. default/mail-transport
> and default/mail-program? default/mail/mta and default/mail/mua?
Ooops! Let's leave some room for the mta-switch./etc/default/mta ;-)
There is one convergence point for MTA, MDA and MUA, the system inbox.
To not complicate the proposal and tackle the first priority PITA when
setting up a system, my proposal
(www.magma.com.ni/~jorge/proposal/delivery.html) just and only addressed
this one, others can follow at any time.
The name "defaultdelivery" reflected this purpose (in my opinion), but
from the start I thought, that this file could be used in an extensible
form for MDA, MUA purposes, so Miquels proposal to call it "mailbox"
seemed more generic to me: it reflects system mailbox & not only
delivery purposes.
>
> (Unclear for me would be, at which point the mua defaults would be
> invoked. When creating the user? On MUA installation? I guess the values
> from mail/mua should propagate to etc/skel if default configurations for
> some popular muas are given?
The later, in principle. When somebody installs a Debian system s/he
should be able to select: Maildir, mbox, MH, mbx as default user
mailbox, and all MUA, MTA, MDA etc. packages should respect this choice
when beeing installed.
Creation of /etc/skel files should be limited as much as possible, as
well as configuration effort at user creation time.
However, some programs need them, e.g. Qmail won't deliver to a user if
"./Maildir/" is the default delivery method, and the user has no
~/Maildir/ directory in the correct format and no other delivery method
specified by a .qmail file. So putting a Maildir into /etc/skel is
necessary. However, standard unix mbox delivery states, that if a
/var/mail/<user> file does not exist, it is silently created, so when
using this method nothing has to be done in /etc/skel.
Also take into account, that gdm and login need to know about the users
choice of mail delivery, because they set respective environment
variables.
Regards,
Jorge-León
