Re: [apps-discuss] I-D Action: draft-ietf-appsawg-http-forwarded-01.txt

[Andreas Petersson <[e-mail]>]

On Tue, 27 Mar 2012 09:54:02 +0200
Julian Reschke <[e-mail]> wrote:

> On 2012-03-27 09:41, Martin Thomson wrote:
> > On 26 March 2012 17:34, Andreas Petersson<[e-mail]>  wrote:
> >> I believe there is not much left to fix in this draft.
> >> My biggest concern right now is this:
> >>
> >>     extension = 1*( ALPHA | DIGIT | "-" )
> >>     ext-value =<any OCTET except CTLs, "," and ";",
> >>       but including SP>
> >>
> >> Is that too liberal? Should recommendations be done in this draft on
> >> what character encoding should be used?
> >
> > Why not use the normal token / quoted-string constructs?
> 
> Yes, please.
> 

For ext-value I assume you mean?

> Also, it's no entirely clear why the spec depends on RFC 2616, not 
> HTTPbis (although it copies some of its style).
> 

Because HTTPbis is not an RFC yet. When we started working on this draft
it was unclear both when this draft will be finished and when HTTPbis
will be finished. If HTTPbis becomes an RFC before this one, I intend to
make adjustments to this one to refer to HTTPbis and also use ABNF as
used in HTTPbis instead of RFC2616-style BNF.

> Finally, as noted in other WGs: please do not special case the syntax of 
> certain parameters. Just define *all* of them to be
> 
>    token / quoted-string
> 
> and then define additional constrains separately (see epic discussions 
> in OAuth and WebSec).

I don't see the point in this. If you can't motivate it in one e-mail I
doubt there is a good one. :-)


 /Andreas Petersson