RE: Multiple Alt-Svc parameters of the same type

Hi Kari,

Thanks for spotting the obvious mistake on my part. Your "ma" examples much better illustrate my thoughts on this.

Lucas

> -----Original Message-----
> From: hurtta@hurtta09lk.keh.iki.fi [mailto:hurtta@hurtta09lk.keh.iki.fi] On
> Behalf Of Kari Hurtta
> Sent: 16 January 2017 18:18
> To: Lucas Pardue <Lucas.Pardue@bbc.co.uk>
> Cc: HTTP working group mailing list <ietf-http-wg@w3.org>; Kari Hurtta
> <hurtta-ietf@elmme-mailer.org>
> Subject: Re: Multiple Alt-Svc parameters of the same type
> 
> > For example, is it OK for a server to generate multiple “persist”
> > parameters perhaps with different values, how should a client deal
> > with that? Perhaps I’m worrying too much.
> 
> Bad example, I think.
> 
> https://tools.ietf.org/html/rfc7838#section-3.1

> 
> |   This specification only defines a single value for "persist".
> |   Clients MUST ignore "persist" parameters with values other than "1".
> 
> 
> I think that
>  persist=1
> and
>  persist=1;persist=1
> 
> are equivalent. And all other persists values are ignored.
> 
> ( Although it also make sense treat that same way than unknown
>   parameter values. )
> 
> 
> But betetr question is, what client should do
> when there is several "ma" paramater values.
> 
> https://tools.ietf.org/html/rfc7838#section-3.1

> 
> |   Syntax:
> |
> |   ma = delta-seconds; see [RFC7234], Section 1.2.1
> |
> |   The delta-seconds value indicates the number of seconds since the
> |   response was generated for which the alternative service is
> |   considered fresh.
> 
> There is only
> 
> |   This specification defines two parameters: "ma" and "persist",
> |   defined in Section 3.1.  Unknown parameters MUST be ignored.  That
> |   is, the values (alt-value) they appear in MUST be processed as if the
> |   unknown parameter was not present.
> 
> Because this is not defined all of these make sense
> 
> ∙ Pick first value
> ∙ Pick last value
> ∙ Pick smallest "ma" value
> ∙ Pick largest "ma" value
> ∙ Treate that as unknown parameter
> ∙ Consider alternative service as invalid because
>   max-age is unclear
> 
> / Kari Hurtta
>