XLIST, special-use mailboxes

[Paul van der Vlis]

Hello,

Does Cyrus IMAP support RFC 6154 about special-use mailboxes?
https://tools.ietf.org/html/rfc6154

I read that it only wass supported in Cyrus version 2.4 ?
I am using version 2.5.10 (from Debian 9).

Can I put this in imapd.conf for all users?
xlist-archive: Archives
xlist-drafts: Drafts
xlist-sent: Sent
xlist-spam: Spam
xlist-trash: Trash

Is it only used on new created folders or always?

Do I need "specialusealways: 1" or something like that?

What about client-support? Is it much used in clients?

With regards,
Paul van der Vlis

[Bron Gondwana]

This post might be inappropriate. Click to display it.

[Anatoli]

This post might be inappropriate. Click to display it.

[Paul van der Vlis]

Post by Anatoli
Paul,
To add on top of what Bron said, xlist was removed in 2.5 but added as a
(somewhat?) new implementation in 3.0, so you should install the newest
version of Cyrus to take advantage of this feature.

I am using Cyrus 2.5.10, and I prefer using the Debian packages.
There is no 3.0 version in Debian at the moment.

So xlist is removed in 2.5, and the new implementation is not there?
So no support in Cyrus 2.5.10.

Post by Anatoli
In 3.0 xlist is part of autocreate functionality, i.e. an appropriate
flag is set on each newly created folder once a new mailbox is created
(AFAIK, it won't work on existing folders).

When I understand it well, this is only used while autocreating and not
for excisting folders. And customers can change this themselve in some
clients.

I am using Sogo webmail, and it lookslike it can change the special-use
of a folder.

Post by Anatoli

What about client-support?  Is it much used in clients?

From my experience, it's rather well supported. Some clients from time
to time (depending on the vendors, versions, locales, etc.) fail to
apply these flags to some of the folders, but most of the time it's very
useful.

Thanks for your help.

With regards,
Paul van der Vlis

[ellie timoney]

Hi Paul,

Post by Paul van der Vlis
So xlist is removed in 2.5, and the new implementation is not there?
So no support in Cyrus 2.5.10.

So, there's a few parts being mixed up here.

* The RFC 6154 IMAP LIST Extension for Special-Use Mailboxes is supported in 2.5. My understanding is that this is the standardisation (into the LIST, CREATE, etc commands) of the functionality previously known as "XLIST".

* The XLIST command continues to work in Cyrus 2.5, but it's basically just a synonym for LIST. Speculatively, modern clients are probably using the RFC 6154 LIST extensions these days, but we continue to accept the XLIST spelling for backward compatibility to older, pre-6154 clients.

* The `specialusealways: 1` setting exists to make Cyrus always include special-use information in its LIST responses. The default is to only include it when it's requested. Enabling this setting is not strictly standard compliant, but might be helpful if your users have buggy clients that want special-use information but do not ask for it.

* The `xlist-attribute: mailbox` settings in imapd.conf no longer exist in Cyrus 2.5. This was used by the 2.4 XLIST implementation to provide special-use information. Since RFC 6154 support was added in 2.5, special-use attributes are first class attributes on mailboxes, and are no longer faked in this way.

* The `xlist-attribute: mailbox` settings have reappeared in Cyrus 3.0 as a way of providing information to the autocreate mechanism, such that autocreated mailboxes could have their special-use attribute applied during creation. The choice to re-use the "xlist-" name was to mainly to aid migration from 2.4 systems to 3.0, but it does cause confusion! The 3.0 version of the "xlist-" names may be deprecated and renamed to something like "autocreate-specialuse-" in a future major release, but this has not yet happened.

Generally, to avoid confusion, I'd suggest referring to the special-use attributes and support for them as "special-use attributes", or "RFC 6154". The "XLIST" name referred to a specific pre-standardisation implementation of a similar idea that no longer exists.

Cheers,

ellie