add Data.Char.Unicode module

[cdepillabout]

This commit adds a Data.Char.Unicode module, which is a combination of GHC's Data.Char and GHC.Unicode.

Data.Char.Unicode is a high-level module that uses low-level functions from Data.Char.Internal.

Data.Char.Internal is created with the script ubconfc. ubconfc is modified from a file of the same name in the GHC sources.

In GHC, ubconfc is used to create the C file WCsubst.c. WCsubst.c is called from GHC.Unicode.

I have modified ubconfc to produce PureScript code instead of C code.

[cdepillabout]

Data.Char.Internal is pretty rough around the edges. I wanted to get a PR up so I could get comments on what should be fixed/changed/etc.

Data.Char.Internal is also pretty long. It's currently over 4000 lines. Since the purescript-strings package is pulled in by so many other packages, it might make more sense to pull out this PR into a separate package. Maybe call it something like purescript-char-unicode or purescript-unicode.

[garyb]

Uh yikes, that is a big file 😄! Yeah, I was going to suggest perhaps this should be purescript-unicode rather than part of the basic strings library. Also because this module is now "difficult" to fully build (well, not really, but it won't be possible on Windows without msys or something for instance).

[paf31]

Hmm, I think I'll have to review this over the weekend 😄

One comment though - we should update the LICENSE file.

Thanks for this! I think I agree with @garyb about the separate package though.

[cdepillabout]

Pulling this out into a separate package sounds good.

Would you want the package to be in the purescript or purescript-contrib organizations? Or should I just make it a personal package?

[garyb]

It'd be great to have as at least a somewhat officially sanctioned library, so I'd be 👍 for contrib at least. I'll make a repo for it and add you there, and we can always bump it into purescript later.

[cdepillabout]

This has been released as the purescript-unicode package.