Re: [Json] Description of parsers

[Stefan Drees <[e-mail]>]

On 2013-06-10 12:27 CEST, Carsten Bormann wrote:
> On Jun 9, 2013, at 22:38, Paul Hoffman ... wrote:
>
>>> duplicate names in objects
>
> That doesn't mean anything, until it has been defined.

I think here the contxt is confined to parsers and is a summarizing 
description with focus on the parsers. I understand the job of the 
section 2 to introduce the vocabulary. So the name is the first part of 
the ordered pair (name, value).

>
> Is "foofoo" a duplicate name?

not in itself, but maybe if some other "foofoo" in the same object and 
nesting level comes along or preceeded it.

>
> Or is it just not allowed to write
> { "foo" "foo" : "bar" }
> ?

here the parser is already out, as this is invalid JSON.


> Or is it that all maps (objects) must have unique names, so
> [{"foo": "bar"}, {"foo": "baz"}]
> is an instance of a duplicate name?

we should talk about names of a single object and nesting level.
So no.

> Or maybe this is about the names within an object, so
> {"foo": {"foo": "bar"}}
> is an instance of a duplicate name?

No. ditto.

> ->
> Of the name/value pairs that comprise a single object, no two
> [shall/should/whatever] have the same name.
> For the purposes of this comparison, two names are the same if they
> are the same sequence of characters after processing the strings (2.5).


As long as we do not "invent" something, but are all sure, that the 
"common" knowledge where it fills a real gap has also been "shared by 
all known implementers" knowledge :-) ...

> [RFC 4627 doesn't quite define what that is, either, but one can
> complete section 2.5 with common knowledge such as that the quotes
> aren't meant to be part of the decoded string.] ...

How do the existing validators and parsers decide that {"a" "a":"b"}
is invalid maybe the ABNF rules :-?)
"""
          object = begin-object [ member *( value-separator member ) ]
                 end-object
          member = string name-separator value
[...]
          string = quotation-mark *char quotation-mark
          char = unescaped /
                 escape (
                     %x22 /          ; "    quotation mark  U+0022
                     %x5C /          ; \    reverse solidus U+005C
                     %x2F /          ; /    solidus         U+002F
                     %x62 /          ; b    backspace       U+0008
                     %x66 /          ; f    form feed       U+000C
                     %x6E /          ; n    line feed       U+000A
                     %x72 /          ; r    carriage return U+000D
                     %x74 /          ; t    tab             U+0009
                     %x75 4HEXDIG )  ; uXXXX                U+XXXX

          escape = %x5C              ; \

          quotation-mark = %x22      ; "

          unescaped = %x20-21 / %x23-5B / %x5D-10FFFF

"""
where
     "a" "a"
seems not to match the rule for a string in rule member.

All the best,

Stefan.