12.07.2011 12:14, Julian Reschke wrote:
On 2011-07-12 11:09, Mykyta Yevstifeyev wrote:
...
Section 5.2.2, bullet 3, sub-bullet 4. When defining the ABNF for a
header, the header name should be included in it as well. So the first
line should be:
...
Why?
There is the following formulation:
The 'Foo' headers takes the form of <foo-header> ABNF rules below:
foo-header = *(APHA/DIGIT)
It should say: "The 'Foo' header field's value takes the form..."
This will eliminate the problem. Currently we have:
The ABNF of this header is defined as follows:
not its entity.
will result in the message headers like:
Upgrade: TLS/1.2
Connection: Upgrade
gfr134
and "gfr134" will be the 'Foo' header. "foo-header = "Foo:"
*(APHA/DIGIT)" will result in valid:
Upgrade: TLS/1.2
Connection: Upgrade
Foo: gfr134
See also eg. RFC 3282, RFC 2616.
Have a look at the HTTPbis drafts.
They should also be clear about whether they mean the header field or
the header field's entity.
[ . . . ]
That being said, it might be a good idea to revisit the choice of
syntax, or at least to clarify the LWS situation.
The document may reference the httpbis-p1 where the <n>#<m>rule
extension will be described for valid ABNF. See
http://tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-15#section-1.2.1
It could, but my guess is that HyBi doesn't want to wait for HTTPbis.
That's up to them.
Mykyta
...
Best regards, Julian
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf