ietf-822
[Top] [All Lists]

Re: Mailing List Last Call for 2822 update internet-draft

2007-07-30 09:07:48

[Resending with my actual subscribed address so that this gets through in a timely fashion. My apologies to all. And yes, I've re-added the pine rule to do this automatically that I lost when my laptop died...]


On Mon, 30 Jul 2007, Paul Overell wrote:
In message <20070730073355(_dot_)GL2903(_at_)apb-laptoy(_dot_)apb(_dot_)alt(_dot_)za>, Alan Barrett <apb(_at_)cequrux(_dot_)com> writes

On Mon, 30 Jul 2007, Alan Barrett wrote:
> It sure was. Comments and folding white space were allowed between
> any two tokens in RFC 822. And since the syntax for zone was:
>
>                         ("+" / "-") 4DIGIT
>
> a comment was allowed between those two. I don't know how to read
> 3.1.4 and 3.4.3 in RFC 822 and not come up with that interpretation.

Yes, you are correct.

Hang on...  "+" and "-" are not in the list of special characters that
RFC822 section 3.1.4 says are used to delimit tokens.

No, but quote marks are in the list of specials, "+" and "-" are quoted-strings.

The quote marks around the plus and minus signs are part of the 822-BNF syntax and do not make those the quoted-strings referred to by RFC 822 section 3.1.4.

(If that isn't true, then someone needs to explain domain-literal, as its definition starts with "[", but it's described in section 3.1.4 as a lexical symbol distinct from quoted-string.)


So I would say
that, given input like "+0000", the sign and the digits together make
up an "atom" per the description in section 3.1.4 and the grammar in
section 3.3.  The ( ("+" / "-") 4DIGIT ) production in the grammar in
section 5.1 is looking inside a single token, not looking at two tokens.

I agree.


Disagree, a quoted-string is a token

Please identify the quoted-strings in this example Date header field from the RFC:
        Date:     26 Aug 76 1430 EDT



So, I disagree with Pete and Paul and instead agree with Alan that RFC 822 did not permit comments or folding whitespace between the sign and the number of a numeric zone.


Philip Guenther