Variable size fixed point numbers

[shishioko]

Implement fixed point decimal types, while giving the programmer the optional ability to define their own ratios of bit allocation, such as follows:

x32 a = 123.456; // standard fixed point of 32 bits
x0 b = 12.34; //system size fixed point per issue #28 
x8.24 c = 7.12345; //8 bit number, 24 bit fraction

[Zffu]

This will be added with the next batch of types before 1.0.0.