ietf-822
[Top] [All Lists]

Re: Revisiting RFC 2822 grammar

2004-01-16 11:16:16

Charles Lindsey writes:
No, that was always allowed, because a dot-atom can always be preceded by CFWS. But it is a good example of how hard it can be to check some of these cases in the present syntax.

Yes.

3. The date syntax is reordered, and four-digit years seem to have disappeared from the obsolete syntax. This looks like a good idea:

I don't think 4DIGIT was ever in obs-year. Nor even in RFC 822?

Sorry, Pete is right, 2*DIGIT was on the page and my brain was seeing 2DIGIT. 2*DIGIT is perfectly okay.

7. There may be a few cases where FWS in 2822 is replced by CWFS in Bruce's grammar. It's a little hard to tell. For example, in Bruce's grammar there always is CFWS following field-name ":", I'm not sure the same holds for 2822. It may.

Not so. Look at orig-date, for example, where it is explicitly [FWS] that follows the ":".

Anyway, I think Bruce's stuff is better, because it's easier to read. In 2822 I'd need a long time to be sure where CFWS is permitted (as opposed to only FWS), in Bruce's grammar it's much easier.

Arnt