Re: [abnf-discuss] FW: ABNF/RFC7405/ Update: EBNF syntactic exception

[Paul Overell <[e-mail]>]

On 12/07/2022 14:50, Jacob Friedman wrote:
> Good question. Referring to ISO 14977, exception takes the precedence. I believe that would result in (a \ b) (c \ d), much like a / b c / d resulting in (a / b) (c / d).

In ABNF concatenation binds tighter than alternation so

a / b c / d means a / (b c) / d not (a / b) (c / d)

>
> The second is optional to align EBNF's possible handling of empty terms. Better alternatives are welcome.
>
> ---
>
> ; Any repetition of 42s
> a = *(%d42)
> ; Any repetition of 42s except for 0 repetitions (empty) b = a \

If there is a need for an explicit empty then I would prefer something 
visible.

In ABNF any 0element already means empty.    So 0<> might do the trick, 
if a bit cryptic.

b = a \ 0<>

or better

empty = 0<>

b = a \ empty


Regards

-- 
Paul Overell