RFC 8259 compatibility mode

[Harvie]

Hello,
do you think it would be possible to add compatibility mode to enforce JSON output strictly corresponding to standard of choice? eg. RFC 8259.

We have issues that json-c is generating JSON code containing NaNs, which is possible in ECMA, but not in RFC. Using null in such case might make sense.

[dota17]

1. ECMA doesn't support NaN

Number: a signed decimal number that may contain a fractional part and may use exponential E notation, but cannot include non-numbers such as NaN.
https://en.wikipedia.org/wiki/JSON

2. json-c will report error "null expected" when it parse "NaN" in STRICT mode. Maybe you use it in wrong way?

[hawicz]

While there's a strict mode when parsing json, there is generally no validation if you're working with json_object's, nor when a tree of json_object's is serialized into a json string.
I suppose one option would be to fail (crash?) if you call json_object_new_double() that way.
Or, we could add a "strict output" flag to json_object_to_json_string_ext(), but then we'd also need to add some way to indicate where in the tree it failed.
Neither one of these seems like a good option. Instead, I'd say that if you don't want NaN in the output, don't call json_object_new_double(NAN).

[dota17]

The problem has already been solved