Do a test pass (and subsequent fix pass) of unicode scenarios

[blakeohare]

Concerns:

• Python 2.x vs 3.x unicode support
• JavaScript across browsers, reading/writing to local storage
• C#/Java uses UTF16 internally (I think?)

FileIO implementations for Python and JS ignore the encoding parameter right now. That is bad.

Crayon should universally behave as though it is using UTF32 internally (whether it actually does is an implementation detail), read files by correctly selecting the encoding, dropping BOMs when reading in text mode, and honor encodings on writing files with FileIO such that reading the bytes back will produce identical and accurate results.

This would make a great unit test.

[blakeohare]

Not a Preview Release blocker, though.

[blakeohare]

Ideally these should be added to the new FileIOTest demo project. Write files in each of the formats with unicode and read back as binary and vice-versa.