[Standards] rfc3921bis: multiple items in roster set

[Peter Saint-Andre]

I'm reviewing a bunch of old messages about rfc3921bis, so expect a few
posts to the list. Here is the first one...

Currently, some clients include multiple items in a roster set to
complete certain use cases, for example to rename a group:

<iq type=set' id='rs1'>
<query xmlns='jabber:iq:roster'>
<item jid='contact1 (AT) domain1 (DOT) tld'>
<group>NewGroupName</group>
</item>
<item jid='contact2 (AT) domain2 (DOT) tld'>
<group>NewGroupName</group>
</item>
</query>
</iq>

However, Rule 1 in Section 2.1.3 of rfc3921bis states:

"The <query/> element MUST contain one and only one <item/> element."

Would it break anything to allow multiple items in a roster set?

Peter

--
Peter Saint-Andre
https://stpeter.im/

[Michal 'vorner' Vaner]

Hello

On Thu, Jun 05, 2008 at 03:03:16PM -0600, Peter Saint-Andre wrote:
> I'm reviewing a bunch of old messages about rfc3921bis, so expect a few
> posts to the list. Here is the first one...
>
> Currently, some clients include multiple items in a roster set to
> complete certain use cases, for example to rename a group:
>
> <iq type=set' id='rs1'>
> <query xmlns='jabber:iq:roster'>
> <item jid='contact1 (AT) domain1 (DOT) tld'>
> <group>NewGroupName</group>
> </item>
> <item jid='contact2 (AT) domain2 (DOT) tld'>
> <group>NewGroupName</group>
> </item>
> </query>
> </iq>
>
> However, Rule 1 in Section 2.1.3 of rfc3921bis states:
>
> "The <query/> element MUST contain one and only one <item/> element."
>
> Would it break anything to allow multiple items in a roster set?

What do you return, if one fails and one succeeds? Or they work as a
transaction? (none or all)

With regards

--
Please enter password:

Michal 'vorner' Vaner

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)

iEYEABECAAYFAkhI8vkACgkQ7/oWwynB3bL5AwCggWiSlJxn9eqrBDDbt8cFryKI
vQwAn1oU9Ajn2PTXi0y0fGgjRjYkr7Cg
=5yOO
-----END PGP SIGNATURE-----

[Dave Cridland]

On Thu Jun 5 22:03:16 2008, Peter Saint-Andre wrote:
> However, Rule 1 in Section 2.1.3 of rfc3921bis states:
>
> "The <query/> element MUST contain one and only one <item/>
> element."
>
> Would it break anything to allow multiple items in a roster set?

Our code will reject that, with bad-request, prior to doing anything.
I'd be surprised if we're the only ones.

Of course, we could change that, but it's not entirely
straightforward - obviously we'd want to maintain atomicity.

From a protocol standpoint, I think we need a feature or server
capability to signal to clients that they may do this, unless it
happens that we're the only ones not allowing it.

Dave.
--
Dave Cridland - mailto:dave (AT) cridland (DOT) net - xmpp:dwd (AT) dave (DOT) cridland.net
- acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
- http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade

[Peter Saint-Andre]

On 06/06/2008 2:39 AM, Dave Cridland wrote:
> On Thu Jun 5 22:03:16 2008, Peter Saint-Andre wrote:
>> However, Rule 1 in Section 2.1.3 of rfc3921bis states:
>>
>> "The <query/> element MUST contain one and only one <item/> element."
>>
>> Would it break anything to allow multiple items in a roster set?
>
> Our code will reject that, with bad-request, prior to doing anything.
> I'd be surprised if we're the only ones.
>
> Of course, we could change that, but it's not entirely straightforward -
> obviously we'd want to maintain atomicity.
>
> From a protocol standpoint, I think we need a feature or server
> capability to signal to clients that they may do this, unless it happens
> that we're the only ones not allowing it.

I had a note about this from several years ago that one client did this,
but I didn't have any reports about server support. I have no objections
to disallowing this behavior.

Peter

--
Peter Saint-Andre
https://stpeter.im/