Re: draft-ietf-httpbis-replay-03, "5.1. The Early-Data Header Field"

Hi Julian,

On Fri, May 11, 2018 at 01:53:41PM +0200, Julian Reschke wrote:
> "The Early-Data header field carries a single bit of information and clients
> MUST include at most one instance. Multiple instances MUST be treated as
> equivalent to a single instance by a server."
> 
> What would that mean if what's sent is:
> 
>   Early-Data: 1
>   Early-Data: 0
> 
> ?
> 
> Or
> 
>   Early-Data: 1
>   Early-Data: 0
> 
> ?

I'd argue it's similar to :

  Early-Data: 1
  Early-Data: foobar

In that only the token "1" is defined as a valid value for this header field.
At this point I think it will depend on the implementation. Some will find
"Early-Data: 1" and be happy with it. Others will look closer and will find
an invalid value, either discarding it or deciding that Early-Data doesn't
carry a valid value (eg: "Early-Data: 1, foobar").

I would personally suggest that we propose that the presence of any non-empty
value in this header field MUST be considered as a boolean "true" value and
that implementation SHOULD write 1 there. It would be safer in my opinion.

Willy