Character/Byte frequency counter
This is a C clone of a Go program (https://github.com/robpike/freq), which itself was a clone of an old C program. The Go program works on full unicode, whereas the original only worked on bytes/ASCII.
I wanted to see how painful it would be to make a C version that worked on unicode
like the Go version without using external libs like ICU. Turns out... it's
not so bad if you are on Linux or OS X, where wchar_t is 4 bytes. On Windows--even
cygwin--you get a 2-byte wchar_t. The upside is that the code still compiles and works,
but you only get the 16-bit BMP style of unicode, and any combining chars will just
be in the list as-is.
I'm sad to say that as of 2016, C11 unicode still isn't supported correctly (as fas as
I understand it) by compilers. On cygwin you get __STD_ISO_10646__ defined
even though wchar_t is too small. Cygwin and OS X clang both seem to leave out
the uchar.h header, but at least in both cases __STD_UTF_32__ was not defined.