mailing list archives

meli community discussions

⚠️ if something does not work as intended when interracting with the mailing lists,
reach out Github mirror Gitea repo @epilys:matrix.org

E-mail headers
From: Abhijit Menon-Sen <ams@oryx.com>
To: imap-protocol@u.washington.edu
Date: Fri, 08 Jun 2018 12:34:37 -0000
Message-ID: 20060914060431.GA15997@penne.toroid.org permalink / raw / eml / mbox
Our IMAP server uses a conventional mailbox hierarchy, where each user's
"home directory" is /users/name (that is, an empty reference argument to
LIST is intepreted as /users/name), and a user's personal mailboxes may
be accessed either as foo or /users/name/foo.

What would happen if we started interpreting an empty reference as "/"?

That is, every mailbox would have just one name. /users/foo/bar, or
/archives/imap-protocol, or whatever. Would clients be able to cope
with not being able to "CREATE Drafts" or whatever? (I'm told that
isn't possible with Cyrus anyway, so it should be OK, but I haven't
tested.)

Does anyone see a problem with such a layout?

-- ams
Reply
E-mail headers
From: mrc@CAC.Washington.EDU
To: imap-protocol@localhost
Date: Fri, 08 Jun 2018 12:34:37 -0000
Message-ID: Pine.OSX.4.64.0609132313200.368@pangtzu.panda.com permalink / raw / eml / mbox
Your change would effectively put IMAP sessions at the root.  I don't 
think that this a particularly good idea, especially given that there are 
so many clients that do a * wildcard lists.

It should work, but you may end up deciding that it was a bad idea.

-- Mark --

http://panda.com/mrc
Democracy is two wolves and a sheep deciding what to eat for lunch.
Liberty is a well-armed sheep contesting the vote.
Reply
E-mail headers
From: dave@cridland.net
To: imap-protocol@localhost
Date: Fri, 08 Jun 2018 12:34:37 -0000
Message-ID: 8267.1158220687.012298@peirce.dave.cridland.net permalink / raw / eml / mbox
On Thu Sep 14 07:04:31 2006, Abhijit Menon-Sen wrote:
> Our IMAP server uses a conventional mailbox hierarchy, where each 
> user's
> "home directory" is /users/name (that is, an empty reference 
> argument to
> LIST is intepreted as /users/name), and a user's personal mailboxes 
> may
> be accessed either as foo or /users/name/foo.
> 
> What would happen if we started interpreting an empty reference as 
> "/"?
> 
> That is, every mailbox would have just one name. /users/foo/bar, or
> /archives/imap-protocol, or whatever. Would clients be able to cope
> with not being able to "CREATE Drafts" or whatever? (I'm told that
> isn't possible with Cyrus anyway, so it should be OK, but I haven't
> tested.)
> 
> Does anyone see a problem with such a layout?

In terms of the protocol, no (although if you don't support 
NAMESPACE, there shall be much gnashing of teeth).

However, in practical terms, there are some clients (The variant of 
Outlook for Pocket PC springs to mind, although that could be an 
older version) that expect to be able to create a slew of unqualified 
mailbox names "CREATE Drafts" included. These then exhibit Strange 
And Fearsome behaviour when they can't.

Although it's true that with Cyrus IMAP, the default is that the 
user's own heirarchy begins at "INBOX.", I used to switch (with 
altnamespace) so that user's mailboxes were at the top level as a 
routine when installing servers, in order to ensure it worked with 
these broken clients.

The clients *are* against specification, of course, but the fix for 
them is also within specification, so I'd generally lean toward at 
least having an option such that they work.

Dave.
-- 
Dave Cridland - mailto:dave@cridland.net - xmpp:dwd@jabber.org
  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
  - http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade
Reply
E-mail headers
From: ams@oryx.com
To: imap-protocol@localhost
Date: Fri, 08 Jun 2018 12:34:37 -0000
Message-ID: 20060914075330.GA17958@penne.toroid.org permalink / raw / eml / mbox
At 2006-09-13 23:15:04 -0700, mrc@CAC.Washington.EDU wrote:
>
> I don't think that this a particularly good idea, especially given
> that there are so many clients that do a * wildcard lists.

Indeed. That's the only thing I could think of.
Can you see any other potential problems?

(The whole idea makes me slightly uncomfortable, which is why I'm
posting to ask if anyone can think of objections to it.)

-- ams
Reply