Short-term: Unify unicode support on front-end code.

[mauvecow]

For DxLib in particular!

• Split up any string functions into Windows style FunctionA and FunctionW.
• Use #ifdef UNICODE to determine which one is selected via inline functions, NOT macros. This negates plain C support, but that's fine.
• Convert on the internal back-end as needed.
• Remove DXCHAR on the back-end, we explicitly support ucs16 or utf8 and that is it.

This is pretty easy and mostly just time consuming.

[mauvecow]

Working on this currently. Some notes:

• I've had to clean up the back-end first so everything going into it is UTF-8, period. No more DXCHAR cluttering up the program flow. This has been completed.
• Main headers have some macros in place to create static DXINLINE functions which will automatically choose the correct one. No defines needed.
• vsscanf hook still has some x86-specific code when building under Windows. I'd like to remove this.
• Long-term we want to detach from SDL, so we're going to have to add in custom replacements for vsscanf and vsprintf most likely. They will need specific wide-char versions too, most likely, as the semantics are slightly different.

[mauvecow]

Committed WIP to branch textoverhaul.

Need to implement a new vsnprintf and a new vsscanf, both of which are encoding-aware, and this may take some time.