Fix UTF16 byte order + extend utf8 & utf16 test

[Sigma42]

This fixes #40 by using ByteOrder.nativeOrder().

I changed InputEncoding.valueOfto accept StandardCharsets.UTF_16BE, StandardCharsets.UTF_16LE and StandardCharsets.UTF_16 but ignore the endianness (to stay backwards compatible with e.g. a big-endian machine).

I also changed the parseUtf8 and parseUtf16 tests to verify that their small example code are actually parsed correctly.

[ObserverOfTime]

Isn't StandardCharsets.UTF_16 enough on its own?

[Sigma42]

Not as a value for the enum variant InputEncoding.UTF_16.
Because String.getBytes(StandardCharsets.UTF_16) may return any byte order, it just includes a byte-order mark at the start. (in my testing using openjdk on linux x86 it still retuned a big-endian encoded string, just with the added BOM). And treesitter ignores the BOM.

But for the InputEncoding.valueOf it might make more sense to only allow StandardCharsets.UTF_16.
I just added StandardCharsets.UTF_16BE so it would not be a breaking change. But I guess it was probably never used anyway because it would have only worked for big-endian machines.

[ObserverOfTime]

See tree-sitter/tree-sitter#3740