Re: [Editorial Errata Reported] RFC7231 (4224)

On 2015-01-10 17:45, Bjoern Hoehrmann wrote:
> * Julian Reschke wrote:
>>> Notes
>>> -----
>>> The HTTP/1.1 RFCs define rules imported across documents using
>>> prose rules for all rules except this one. This is an error
>>> because RFC7231 does not mean to re-define the production rule.
>>> ...
>> The text is as intended; Appendix C has:
>>> The rules below are defined in [RFC7230]:
>>>    BWS           = <BWS, see [RFC7230], Section 3.2.3>
>>>    OWS           = <OWS, see [RFC7230], Section 3.2.3>
>>>    RWS           = <RWS, see [RFC7230], Section 3.2.3>
>>>    URI-reference = <URI-reference, see [RFC7230], Section 2.7>
>>>    absolute-URI  = <absolute-URI, see [RFC7230], Section 2.7>
>>>    comment       = <comment, see [RFC7230], Section 3.2.6>
>>>    field-name    = <comment, see [RFC7230], Section 3.2>
>>>    partial-URI   = <partial-URI, see [RFC7230], Section 2.7>
>>>    quoted-string = <quoted-string, see [RFC7230], Section 3.2.6>
>>>    token         = <token, see [RFC7230], Section 3.2.6>
> Appendix C should also have the corrected line
>       method = <method, see [RFC7230], Section 3.1.1>
> because the error is not importing `method`.

Again, it is not an error. The spec is correct the way it is. I see your 
preference to reference instead of copying the rule (and I might 
actually agree with it).

> If `method` in a `request-line` is actually something different from
> `method` in an `Allow` header, then RFC7231 should call `method` by a
> different non-terminal name. As it is, you cannot make a single ABNF

It is not different.

> file for all HTTP/1.1 ABNF because `method` is defined twice and there
> may be differences between them explained only in prose.

We can, and we did. It's here: 

The ABNF is just that; the ABNF. It may be true that the prose in the 
different parts of the spec add more explanations and constraints, but 
that has zero effect on the ABNF itself.

Best regards, Julian

Received on Saturday, 10 January 2015 17:05:53 UTC