Sieve filtering shared mailboxes (bb+xxx)

Paolo Negri p_negri at
Wed Jan 26 02:51:48 EST 2005

You have perfectly reason, infact i have many "publicmanager" so i can 
delegate the sieve script administration to a capable user and forget 
about it.
Example: if you have 50+ public folders relative to 10 departements, you 
have to find 10 users, one for each departement as admin of the 
filtering script.
At the end you'll have 10 special users: salespublicmanager 
prpublicmanager finanacepublicmanager hrpublicmanager etc. etc.
You can set up the squirrelmail web front end to sieve to make things 
easier for your users.
You can give to publicmanagers also the "lcr" permission to managers and 
delegate also the decisions about internal structure of public folders.

Don't know if this helps, but that's how i managed the problem.


Gilles Bruno wrote:
> I also though about this workaround, halas there are so many shared 
> folders (50+) that this solution would become quickly hard to maintain :/
> Apparently you're facing the same problem : it seems there aren't any 
> ways to filter shared mailboxes "out of the box" ( no joke here ;) )
> For example if you want your "publicmanager" to filter 2 separate shared 
> folders, you'll face this dilemna :
>  . (make your MTA aliases bla and foo -> publicmanager)
>  . users of "bb+bla" want emails coming from XX at YY to be posted (moved)
>    under bb+bla.XX (a simple "fileinto" so far)
>  . at the same time users of "bb+foo" want emails also coming from XX at YY
>    to be posted under bb+foo.XX
> you'll have to make/manage many cases in publicmanager's sieve script
> The other drawback of this solution is that *you* have to manage the 
> whole script ("parts" of it can't be delegated)...
> Furthermore, *IIRC* sieve scripts can't be "nested" (if I understand 
> RFC3028  § 2.10.7, sieve implementation must support at least fifteen 
> levels of nested test list), for example  :
>   IF <condition A> then
>       IF <condition B> then
>            action
>       FI
>   ELSE
>       ....
>   FI
> doesn't seem applicable (no test inside another test) - which of course 
> (if my assumption about nested tests is true) tend to make the script 
> more "complicated" :/
> I'll dig further...
> Thanx paolo
> Best regards,
Cyrus Home Page:
Cyrus Wiki/FAQ:
List Archives/Info:

More information about the Info-cyrus mailing list