Re: IMAP-problems

[Darko Obradovic <dobradovic gmx de>]

On Thu, 17 Apr 2003 18:12:39 -0400
Peter Bloomfield <PeterBloomfield@bellsouth.net> wrote:

> This is my (probably confused) sense of how IMAP namespaces work (see 
> RFC 2342 for the real scoop).  Your IMAP server can offer you three 
> types of namespace: personal, other users' personal spaces, and shared. 
> There can be zero, one, or more of each type.  The output in your bug 
> report shows that you have one personal namespace, `INBOX', and one 
> shared, `shared':
> 
> 0x8282bb0 s: * NAMESPACE (("INBOX." ".")) NIL (("shared." "."))
> 
> So I'm guessing that your mailbox list looks something like:
> 
> Name
>      INBOX
>          queue
>          draft
>          ...
>          lists
>      shared

exactly, but shared doesn't appear, and that's ok. it's just a namespace, no mailbox. What I learned so far, you shouldn't think of a namespace as a folder immediately.
 
> The way Balsa shows these folders and mailboxes seems natural to me. - 
> You want a top-level entry for the server, to distinguish it from any 
> other IMAP folders, and from local folders;
> - It surely must have an entry immediately beneath the top level for 
> each namespace exposed by the server; and
> - It surely must have a folder tree attached to each namespace, 
> reflecting the folders with names like `INBOX.queue'.
> 
> So how do these other MUAs show this tree?

Ok, with the point of having multiple private namespaces and so on, you led me to the decisive point, which I didn't realize up to now: an IMAP-account may have multiple private namespaces, but on a mail-client I usually only want to access one of them. For me that's INBOX. So I should and can select INBOX as the prefix for the IMAP-folder (I had to do that in Sylpheed to make it work at all, now I know why...). But then Balsa comes up with a new problem: It includes the whole original INBOX-tree in that, along with INBOX's top-level-directories. INBOX itself is a mailbox of course, but the namespace at the same time. Balsa interprets this as being a folder, Sylpheed and many other don't, and I *think* that's correct.

Attached is the output when scanning the tree with prefix INBOX. But from the output, With a given prefix, there's no NAMESPACE-query done, that's ok. But I can't see why it still includes INBOX into the tree. I isn't mentioned in the scan (or I can't see it), it should be a top-level-mailbox, but it is the whole tree as I get it with no prefix, so all folders besides INBOX are duplicated.
I know all this is pretty confusing, should I make some screenshots? *g*

> The mailbox tree may not look exactly like the directory tree that the 
> IMAP server is using as a mail store, but that seems to be an issue 
> with the server, not with the way Balsa displays what the server is 
> delivering.

sure, the server may handle it in whatever way it likes. Actually, "lists" isn't a folder at the filesystem, but all these mailboxes have names of "lists.*".

bye,

Darko

scan-with-prefix-INBOX.txt