Re: draft-ietf-httpbis-http2-latest, 8.1.2.1 Request Header Fields | Re: draft-ietf-httpbis-http2-latest, 8.1.2.1 Request Header Fields | Re: draft-ietf-httpbis-http2-latest, 5.5 Extending HTTP/2

On 25/07/2014 2:09 a.m., Martin Thomson wrote:
> On 24 July 2014 06:52, Amos Jeffries wrote:
>> The recipient of that 1.1 request is supposed to assume path value of
>> "/" in these cases not "*".
> 
> The link Kari provided shows that there is an exception for OPTIONS
> requests in the absolute form that address an HTTP URI with path-empty
> form.
> 
> https://svn.tools.ietf.org/svn/wg/httpbis/specs/rfc7230.html#rfc.section.5.3.4
> 

Oh, right. I missed that. (I'm not getting Kari's mails for some reason,
just your replies).

> That seems pretty explicit to me.  Maybe you can look at the editor's
> draft and see if my latest tweaks have corrected this.
> 

I forsee a problem with the current editors draft text for proxies
relaying this in the h2->1.1 direction since the following is a valid
URL request:

 OPTIONS http://some.host* HTTP/1.1


IMHO it be more correct to say simply that :path may be omitted on
OPTIONS and represents a request for "*" asterisk-form? as opposed to a
0-length :path field which represents the path-empty case.

CONNECT requests provides precedent for omission.

Amos

Received on Thursday, 24 July 2014 15:15:18 UTC