Кунщиков Константин Владимирович wrote:
>
>
> >Konstantin Kunshchikov wrote:
> >>
> >> Hi folks!
> >>
> >> We currently testing server altnamespace code with or without
> >unixhiersep
> >> option.
> >>
> >> I think OE can't support subscription to non-existent folders, and
> >> in that case refuses to show subfolders under "Shared Folders" or
> >"Other Users".
> >> You can view subfolders only if push "Go To" button instead of "Ok" in
> >the
> >> "Imap Folders" section.
> >>
> >> Ok, this is a hack, but is it possible to make this folders ("Shared
> >Folders"
> >> and "Other Users" virtually existent?
> >
> >This appears to be a problem with OE. Netscape Messenger and Outlook
> >both behave correctly in my testing, and I'm *fairly* certain that the
> >server is returning the correct info.
>
> >Ken
>
> Yes, this is a problem with OE. Netscape Messanger works fine as expected.
> But this may be also an architecture thing. "Shared Folders" and "Other
> Users"
> are in general virtual.
> Its not possible to view ACL's of this folders with cyradm.
The same can be said for "user" in the standard namespace. "user" is
the prefix for the other users' namespace, just like "Other Users" and
"Shared Folders" are prefixes for the respective namespaces. Even if OE
doesn't support the NAMESPACE extension, a LIST command returns the
\Noselect flag for the prefixes, which tells the client that these are
not selectable mailboxes.
> Other things may also cause a problem.
> For example, is it possible to drag mail in to the "Shared Folders"?
It shouldn't be, and the server tells the client so as described above.
> As I undestood, altnamespace has 2 major features:
>
> 1. Users subfolders at the same level of that user Inbox.
> 2. "Shared Folders" and "Other Users" prefixes.
>
> I think making 1 and 2 separate with different switches in imapd.conf would
> be
> generally good idea, and solve the problem with OE too.
1 and 2 can not be made mutually exclusive, either you get both
(altnamespace) or neither (standard namespace).
IMHO, the architecture is correct (similar to other commercial variants
of Cyrus) and the problem lies completely with OE.
Ken
--
Kenneth Murchison Oceana Matrix Ltd.
Software Engineer 21 Princeton Place
716-662-8973 x26 Orchard Park, NY 14127
--PGP Public Key-- http://www.oceana.com/~ken/ksm.pgp
