Floating-point NaN or Infinity values should be allowed as a feature …

[Dani-Hub]

…#209

Introduce 'allowSpecialFloats' for readers and 'useSpecialFloats' for writers, use consistent macro snprintf definition for writers and readers, provide new unit tests for #209

[cdunn2001]

I like this change. It seems good as-is. But I do have one idea: NaN/Infinity/-Infinity are just conventions, right? They could be any 3 strings. We could use something easier to parse, like _NaN/_PosInfinity/_NegInfinity. But I guess this is really for Javascript, yes?

[Dani-Hub]

They are "just" conventions, but they are very popular ones, as described in the rationale of my issue #209. Please don't change these values lightheartedly, since doing that would break compatibility to gson, json5, simplejson, and probably several other json libraries.

[BillyDonahue]

FYI: I did a little more testing and found a parsing bug.
Dani-Hub#2

[Dani-Hub]

Thanks for bug-finding test, Billy! The revised implementation should honor you test cases now properly.

[cdunn2001]

I'll try to cherry-pick this to 0.y.z too, but that's probably not a big deal.