[Tools-discuss] Re: [ABNF] Legal or not legal to have literal numbers?

[Frank Ellermann <nobody at xyzzy.claranet.de>]

Stephane Bortzmeyer wrote:
 
> Really attached, now :-) along with what I regard as a 
> "fixed" version.

> Part 1.2   Name: img-urn.abnf
>            Type: Plain Text (text/plain)

The author apparently wanted ( "0" DIGIT ) instead of
( 0 DIGIT ), so far it's clear: 

> Part 1.3   Name: img-urn-fixed.abnf
>            Type: Plain Text (text/plain)

Another question is if a legal ( 0DIGIT ) could be also
written as ( 0 DIGIT ).  In RFC 3986 they have 0<pchar>
for an empty path.  Bill's parser says "error", it does
not like a 0 in an nRule (4234 3.7).

With 1DIGIT or 1<pchar> or similar it's fine.  If I add
a space as in ( 1 DIGIT ) the parser says again "error".

The "ABNF in ABNF" section 4 offers no <c-wsp> in its

|         repetition     =  [repeat] element

Just in case, there's no trailing <c-wsp> in <repeat>
and no leading <c-wsp> in <element>.  A <repeat> 0 is
okay, maybe Bill could reduce the error to a warning.

In your list of ABNF validators posted before:

|* Jutta Degener's abnfgen accepts it,
|* Bill Fenner's ABNF Parser refuses it,
|* my Eustathius program refuses it,
|* ABNF Validator accepts it.

What is the "ABNF Validator" if it's not Bill's tool ?
And what's "abnfgen" ?

Frank



_______________________________________________
Tools-discuss mailing list
Tools-discuss at ietf.org
https://www1.ietf.org/mailman/listinfo/tools-discuss