Re: [Stox] Reminder WGLC on -im and -chat

Peter Saint-Andre <stpeter@stpeter.im> Mon, 09 December 2013 17:22 UTC

Message-ID: <52A5FC3E.4070704@stpeter.im>
Date: Mon, 09 Dec 2013 10:22:06 -0700
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:24.0) Gecko/20100101 Thunderbird/24.1.1
MIME-Version: 1.0
To: Dave Cridland <dave@cridland.net>, Yana Stamcheva <yana@jitsi.org>
References: <648368DC-1BD5-4D6A-B4E0-3D72B2EB3A6D@jitsi.org> <CAKHUCzwRBEDH3_zc6ESz7YYP0FAbKuZEqX3Dc+_bqV_X=fgXGA@mail.gmail.com>
In-Reply-To: <CAKHUCzwRBEDH3_zc6ESz7YYP0FAbKuZEqX3Dc+_bqV_X=fgXGA@mail.gmail.com>
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
Cc: stox@ietf.org
Subject: Re: [Stox] Reminder WGLC on -im and -chat
Precedence: list

On 12/9/13 7:22 AM, Dave Cridland wrote:
> World's shortest reviews to try and sneak past the deadline:

Hi Dave, thanks for the review. Comments inline.

> On Wed, Dec 4, 2013 at 1:17 PM, Yana Stamcheva <yana@jitsi.org
> <mailto:yana@jitsi.org>> wrote:
> 
>     draft-ietf-stox-im-05 : http://tools.ietf.org/html/draft-ietf-stox-im-05
>     <http://tools.ietf.org/html/draft-ietf-stox-im-05
>     draft-ietf-stox-chat-03>
> 
> 
> FWIW, I don't think you could actually do much of this wrong - that is,
> beyond -core (and therefore address and error mappings), I suspect
> everything else can fall into a quality-of-implementation issue - so as
> such, I think this only needs to supply guidance and wisdom rather than
> laying down the law - I'm reviewing accordingly.
> 
> This largely seems OK.
> 
> §6 appears to have interesting mangling of the characters. I think
> &amp;#xC3A9; is intended to be &#xC3A9; isn't it?

Yes. <![CDATA[]]> is your friend for escaping purposes.

> I think this is an unfortunate choice of notation, 

It's driven by the limitations of the RFC format.

> actually, because in
> the XMPP case this is valid as-is, whereas in the SIP case it needs to
> be unencoded UTF-8 

What is "unencoded UTF-8"? UTF-8 is itself an encoding. Plus, RFC 3261 says:

   SIP is a text-based protocol and uses the UTF-8 charset (RFC 2279
   [7]).

> - I don't know whether the Content-Type should also
> include a charset parameter for SIP.

Like this?

OLD
Content-Type: text/plain

NEW
Content-Type: text/plain; charset=utf-8

Given that SIP natively "uses the UTF-8 charset", it strikes me that the
charset parameter is unnecessary here.

> It may be better to show either C-style or URL escaping of the UTF-8 in
> the SIP version, or else a <U+C389>, or something. 

I'm not convinced that this is needed.

> Security Considerations should probably refer back to -core as well,
> shouldn't they?

Yes.

>     draft-ietf-stox-chat-03
>     <http://tools.ietf.org/html/draft-ietf-stox-im-05
>     draft-ietf-stox-chat-03> :
>     http://tools.ietf.org/html/draft-ietf-stox-chat-03
> 
> 
> This draft appears to consist mostly of two giant examples. :-)
> 
> However, it seems to be sufficient guidance to build an interoperable
> gateway, so I'm happy this is ready - however as with the above,
> Security Considerations should probably refer back to -im and -core.

Agreed.

Peter

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

[Stox] review of -chat (was: Re: Reminder WGLC on… Philipp Hancke
[Stox] Reminder WGLC on -im and -chat Yana Stamcheva
[Stox] review of -im (was: Re: Reminder WGLC on -… Philipp Hancke
Re: [Stox] Reminder WGLC on -im and -chat Dave Cridland
Re: [Stox] review of -im Peter Saint-Andre
Re: [Stox] review of -chat Peter Saint-Andre
Re: [Stox] Reminder WGLC on -im and -chat Peter Saint-Andre