\u unicode values are not parsed correctly

[gcamp]

Seems like \u values are still not parsed correctly.

Ex : \u00e7 (can be found in https://open.tan.fr/ewp/tempsattente.json/COMB2)

I'm not entirely confortable with specs, but it seems that a big part of it is not supported. For example, the current code doesn't support surrogate pair.

We can also compare with what other json parser are doing and it is quite different : https://github.com/miloyip/rapidjson/blob/master/include/rapidjson/reader.h#L696 https://github.com/open-source-parsers/jsoncpp/blob/master/src/lib_json/json_reader.cpp#L682

[gcamp]

I can look into it, but anybody has insight?

[sunalbert]

Yes，there is a problem when parsing values like \u.