Update: Finish writing tests, fix endianness, simplify some code.

Update: Finish writing tests, fix endianness, simplify some code.

Now that f_char_t exists and is in use, rename f_utf_character_t to f_utf_char_t to be consistent (and simpler).

The endianness detection is incorrect.
The BIG_ENDIAN is not represented as expected.
Change the detection to use GCC's __BYTE_ORDER__ macros.
Fallback to little endian if those macros are not available.

It took a while to figure out this was the problem and I got rather confused going back and forth between byte orders.
I think the changes I made are correct, but I need additional better testing on the byte orders.