Re: draft-ietf-httpbis-client-hints-06

On 2018-08-04 23:15, Ilya Grigorik wrote:
> Hi Kari. Thanks for the feedback.
> 
> On Wed, Jul 25, 2018 at 9:14 AM HTTP Working Group <ietf-http-wg@w3.org 
> <mailto:ietf-http-wg@w3.org>> wrote:
> 
>     |   If DPR occurs in a message more than once, the last value overrides
>     |   all previous occurrences. 
> 
>     So is that saying that if there is several DPR values on
>     DPR Header Field, last value is used ?
> 
> 
> Yes, that's the intent.
> 
>     If that in interpreted only that if last DPR header field is used,
>     and header field is ignored if it does not fit
> 
>         DPR = 1*DIGIT [ "." 1*DIGIT ]
> 
>     syntax, result is different is recipient combines multiple
>     DPR headers.
> 
> 
> The intent was to specify that last valid value (regardless if it is 
> within a combined or separate header) is the one that should be used by 
> the client. My read of the current wording is that it matches this 
> intent, but I'll defer to the experts here on how to best word this if 
> that's not the case.
> 
> Mark, Julian: any guidance on this one?
> 
>     |   If <X> occurs in a message more than once, the last value
>     |   overrides all previous occurrences.
> 
>     So it should mention that on these header fields also last value
>     is used when values are ',' -separated.
> 
> 
> My understanding is that this is implicit, but if that's not the case, 
> happy to rework the wording.
> ...

No, it's not implicit, so yes, a clarification would be good.

FWIW, related issue in http-core: 
<https://github.com/httpwg/http-core/issues/111>

Best regards, Julian

Received on Sunday, 5 August 2018 07:47:28 UTC