Add a Meow_Float typedef and set it to a float

[DragoonAethis]

This allows library users to easily swap the type used by meow_fft to
something more accurate (doubles or long doubles). The default float
type is kept for backwards compatibility.

[JodiTheTigger]

Thanks for the PR. I'll check it out fully when I get time, but on the surface I don't see anything wrong with it.

[JodiTheTigger]

Trying to get it to work on the development branch when set to double but I keep getting stack smash exceptions. I'll keep looking at it.

[DragoonAethis]

Not sure what's going on... I needed this a while ago for this patchset on IGT but it wasn't merged or even reviewed in the end so it didn't go anywhere further. For reference, the usage is in this patch, maybe it'll help figuring out if it's a usage or the library problem. Sorry :/

[JodiTheTigger]

Hey - not a problem :-)

My development branch could be improved quite a bit, along with the code in general (memory safety, sanitizer builds, naming, alignment, etc)

If I can generalize the data size I might play with 16 bit floats and other stuff just for fun.

but first - double...