Re: [Json] [Technical Errata Reported] RFC7493 (6861)

To clarify, since someone asked:

There were many deployed JSON parsers by the time it entered
standardization both in the IETF and TC39. Originally, it only allowed "{"
and "[" as the leading characters. People found this inconvenient, and it
was noted that the primitives could be unambiguously expressed without a
surrounding data structure, while preserving compatibility with the
"eval()" functions in many languages.

There was wide agreement, and everyone rearranged their parser state
machines. I sincerely doubt anyone has ever written separate Unicode
routines for the case the errata mentions. I'd accept the errata but do no
further work.

thanks,
Rob

On Fri, Feb 25, 2022 at 2:59 PM Rob Sayre <sayrer@gmail.com> wrote:

> No, everything accepts them, unless they are really strict I-JSON ones,
> but I haven't seen many of these. Nothing crashes. This is not worth
> working on, imho.
>
> There's an understandable mistake here, since top-level primitives were
> added late in the game, but it doesn't matter.
>
> thanks,
> Rob
>
>
> On Fri, Feb 25, 2022 at 2:33 PM Mark Nottingham <mnot@mnot.net> wrote:
>
>> Agreed - they're not rules so much as guidelines.
>>
>> We should check with TC39, of course.
>>
>>
>> > On 26 Feb 2022, at 7:49 am, John Levine <johnl@taugh.com> wrote:
>> >
>> > It appears that Carsten Bormann  <cabo@tzi.org> said:
>> >> On 2022-02-25, at 16:56, Paul Hoffman <paul.hoffman@vpnc.org> wrote:
>> >>>
>> >>> I agree it would be stupid; it was also kinda stupid for us to have
>> missed this in the creation of I-JSON. I am
>> >> arguing agains the acceptance of this erratum from the standpoint of
>> the long-established erratum rules, not from
>> >> whether the proposed change is a good idea.
>> >>
>> >> I’m not sure I understand what you mean.
>> >>
>> >> This is clearly a case where the English words turn out not to express
>> what the consensus of the WG was (*).  That
>> >> needs to be fixed, and the fix is obvious to anyone skilled in the art.
>> >
>> > We get lots of errrata that say the text in the RFC says X but that is
>> > clearly wrong, it should say Y. We approve them without any trouble. I
>> > don't see why this one is any different.
>> >
>> > I also concur with the Tim's observation about stray surrogates. In my
>> > experience, either the code catches them and returns an error, or it
>> > doesn't and it crashes.
>> >
>> > R's,
>> > John
>> >
>> > _______________________________________________
>> > json mailing list
>> > json@ietf.org
>> > https://www.ietf.org/mailman/listinfo/json
>>
>> --
>> Mark Nottingham   https://www.mnot.net/
>>
>> _______________________________________________
>> json mailing list
>> json@ietf.org
>> https://www.ietf.org/mailman/listinfo/json
>>
>