AbstractChar type

[stevengj]

As was briefly discussed last month, it would be nice to have Char <: AbstractChar, and to have Base functions take AbstractChar arguments where appropriate.

I think this can be added in 1.1 since it would not be breaking?

[StefanKarpinski]

It is non-breaking. I may get around to adding this during the alpha period (since it's non-breaking, that should hopefully be ok), since it would be nice to support non-UTF-8 encodings without requiring transcoding to Char. @Keno also mentioned that he wanted to take a crack at this, but I think other things took priority.

[Keno]

I have most of a branch of this because I wanted to make sure there was nothing breaking required, but once I got to that point, I stopped shorting of cleaning up, writing tests, documenting, etc.

[StefanKarpinski]

Would be good to have a PR in case anyone wants to do the rest.

[stevengj]

Bump @Keno.

[stevengj]

I'll have a PR shortly.