UTF-8 streams will be butchered by lexer about klaxon

[fboldog]

Could you provide a problematic example? I saw what is the problem, but a real world example, always better. ;)

[ssuukk]

Surely, but not earlier than Monday evening. If I can find any API that doesn't require my private key... BTW - thanks for klaxon, it helps keep my code neat ;)

[PaperTsar]

The issue still persists. "ú" and "\u00FA" should produce the same output when parsed. The former gets mangled. (Great API btw:)

[cbeust]

I would really like to see a concrete example where Klaxon fails, then I can add it as a regression test and fix it.

Thanks!

[VladRassokhin]

Problem lies in Lexer#nextChar as it assumes that all chars are single-byte, which is false for example for é (part of your name, Cédric) as that symbol encoded using two bytes C3 and A9.
Simple test to reproduce is to add next method to TestRFC7159 class.

    fun unicodeUnescaped() {
        jsonEquals("{\"a\":\"é\"}", "{\"a\":\"é\"}")
    }