Re: [exim] Exim 4.94: address suffixes and local delivery to…

Αρχική Σελίδα
Delete this message
Reply to this message
Συντάκτης: ivanov
Ημερομηνία:  
Προς: ivanov via Exim-users
Αντικείμενο: Re: [exim] Exim 4.94: address suffixes and local delivery to subdirectories in maildir mailboxes
Hello again!

I apologize for my persistence, but I would like to know if it is possible with new Exim to create automatically new subdirectories for address suffixes in maildir mailboxes. Or it doesn't work because any destination path should be de-tainted via lookup, and lookup for non-existing directories isn't possible.

Maybe I don't understand something about de-tainting. There is no fresh documentation on my native language and maybe I missed some important things when I read English text.

It is important to me to know should I spend my time trying to find the solution of the issue with Exim or I should look to third-party utilities. Maybe it would be better for me to set up local delivery with Dovecot and Sieve rules.

--
Best Regards,
Ivanov


On Thu, Oct 01, 2020 at 08:37:28AM +0300, ivanov via Exim-users wrote:
> Hello! I try to use Exim 4.94 as my post server and I want to use address suffixes and local delivery to subdirectories in maildir mailboxes.
>
> I've used this decision from the official documentation (ch. 26.5):
>
> directory = /var/mail/$local_part_data\
>            ${if eq{$local_part_suffix}{}{}\
>            {/.${substr_1:$local_part_suffix}}}
> maildirfolder_create_regex = /\.[^/]+$

>
> In my case I use a mail directory inside a home directory and the path '/home/$local_part_data/.maildir' instead of '/var/mail/$local_part_data'.
>
> After that I've got the error "Tainted '/home/user/.maildir/.test' (file or directory name for local_delivery transport) not permitted".
>
> I'm a newbie in Exim and confused about this error. I understand that tainted data may not be used for a directory name, but I don't know how to specify in this case a path to directories that should be created from local part suffixes.
>
> If you know a possible solution for this case, tell me, please.