Re: [exim] How to choose between mbox and maildir? was: lock…

Góra strony
Delete this message
Reply to this message
Autor: W B Hacker
Data:  
Dla: exim users
Temat: Re: [exim] How to choose between mbox and maildir? was: lock file
Christian Schmidt wrote:

> Hello Philip,
>
> Philip Hazel, 22.04.2006 (d.m.y):
>
>
>>On Fri, 21 Apr 2006, Christian Schmidt wrote:
>>
>>
>>>How can I make exim (or better: the appendfile transport) decide if a
>>>mail shall be dropped into an existing mbox file or a maildir?
>>
>>The file_format option of the appendfile transport allows for different
>>formats of single-file mailbox. You might be able to coerce it into
>>doing the right thing, but then again it might not work. I suspect
>>you'll have to make the decision at routing time (e.g. by checking for
>>the existence of appropriate files/directories).
>
>
> Thanks, Philip.
>
> But as I already mentioned, I haven't thought of problems occurring
> with the IMAP/POP server - so I'd choose a one-step-migration solution
> when it comes to migrate to maildir.
>
> Regards,
> Christian
>


You *can* do it. Dovecot-pgsql or -msql are easiest, as the
location and type can be user-specific.

- Something like:


user_query = SELECT pg_mailroot||'/'||pg_login_name||'/'
||pg_mailstore AS mail, pg_uid AS uid, pg_gid AS gid FROM
mailprof WHERE pg_recv_login||'@'||pg_domain_id ='%u' AND
pg_active LIMIT 1

Where:

mailprof(ile) is the user@domain preferences table

pg_<variable> are the DB attributes, reasonably self-documenting

pg_active is a boolean flag for active/suspended accounts

'mail' is Dovecots variable for user-specific storage

%u is Dovecot's variable for <user@domain>

The concatenation builds a path to the mailstore location and
type, i.e:

/data/mail/domain/user/Maildir or /data/mail/domain/user/mbox


Courier-IMAP/POP can be handled in a similar manner.

And the same effect can be had with lighter-weight lookups than
SQL - CDB, for example.

That said, as with an abcessed tooth, a one-step migration is
probably more easily managed.

Acute pain, rather than chronic... ;-)

Bill