W3C home > Mailing lists > Public > ietf-http-wg@w3.org > October to December 2011

Re: #231: Considerations for new headers

From: Julian Reschke <julian.reschke@gmx.de>
Date: Mon, 10 Oct 2011 10:26:31 +0200
Message-ID: <4E92AC37.4010504@gmx.de>
To: Mark Nottingham <mnot@mnot.net>
CC: httpbis Group <ietf-http-wg@w3.org>
On 2011-08-28 17:08, Julian Reschke wrote:
> On 2011-08-26 17:57, Julian Reschke wrote:
>> On 2011-07-01 07:05, Mark Nottingham wrote:
>>> Revised proposal:
>>> ...
>>
>> +1 - we may want to add more things or tune, but we should get this in.
>>
>> Proposed patch:
>> <http://trac.tools.ietf.org/wg/httpbis/trac/attachment/ticket/231/231.diff>
>>
>> ...
>
> ...applied (minus some typos) as
> <http://trac.tools.ietf.org/wg/httpbis/trac/changeset/1417>.
> ...

There recently were some interesting discussion in other Working Groups, 
namely:

websec - implications of not using the list syntax when the header field 
*does* occur multiple times (Origin, Strict-Transport-Security)

hybi - when using params, whether it makes sense to disallow one of the 
two notations on the wire (token vs quoted-string)

oauth - how to I18N parameters (there's a proposal to use JSON escapes 
inside quoted-string, which IMHO conflicts with the definition in HTTPbis).

So it's good that we have started work on this (header field syntax 
recommendations), but we probably need more:

>    o  Whether the field is a single value, or whether it can be a list
>       (delimited by commas; see Section 3.2 of [Part1]).

add...: "If it does not use the list syntax, how to treat messages where 
the header field occurs multiple times (a sensible default would be to 
ignore the header field, but this may not always be the right choice)."

Also, we probably should have a single place that defines the "param" 
ABNF construct, like

   param = token OWS "=" OWS ( token / quoted-string )

or maybe

   param = token OWS [ "=" OWS ( token / quoted-string ) ]

and then clarify that predefined params never ever should parse 
different from extension parameters; for instance, they should always 
allow both notations.

See also <http://trac.tools.ietf.org/wg/httpbis/trac/ticket/307> and 
<http://trac.tools.ietf.org/wg/httpbis/trac/wiki/HeaderFieldTypes>.

Best regards, Julian
Received on Monday, 10 October 2011 08:27:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 06:51:48 GMT