Re: WGLC review of p1-messaging (editorial stuff)

On 11/03/2012 10:39 AM, Julian Reschke wrote:
>> Removing the word "hypertext" from the last line would make this a
>> more accurate description of present-day HTTP.

> The "H" in HTTP stands for "Hypertext". As such I'd prefer to keep this,

Sure, but "it's not just for hypertext any more".

>>> 2.7.3. http and https URI Normalization and Comparison
>>
>>>     Likewise, an empty path component is equivalent to an absolute
>>>     path of "/", so the normal form is to provide a path of "/"
>>>     instead.
>>
>> Except that an empty path is not equivalent to a path of "/" in an
>> OPTIONS request sent to a proxy...
> 
> That's the "authority-form", right? That's not an HTTP(s) URI anyway.

No, not authority-form. 5.3 says:

>    For example, the request
> 
>      OPTIONS http://www.example.org:8001 HTTP/1.1
> 
>    would be forwarded by the final proxy as
> 
>      OPTIONS * HTTP/1.1
>      Host: www.example.org:8001

but "OPTIONS http://www.example.org:8001/ HTTP/1.1" would be forwarded
as "OPTIONS / HTTP/1.1". So the empty-path and "/"-path forms are not
equivalent in this one case.

(although now that you mention it, using authority-form here instead of
absolute-form-minus-"/" would have made a lot more sense...)

>>> A.2. Changes from RFC 2616
>>
>> We probably want to double-check these sections against the issue
>> tracker before final publication, but one particular thing that stuck
>> out to me as missing is the addition of the CONNECT rule and removal
> 
> You mean the inclusion of CONNECT? We already mention that in P2. Should
> we mention here as well?
> 
>> of the multipart/byteranges rule from Section 3.3.3.

Sorry, both parts of that sentence were referring to Section 3.3.3; the
rule about CONNECT responses in 3.3.3 is new since 2616.

-- Dan

Received on Saturday, 3 November 2012 22:55:05 UTC