Precision of float/double values during output

[Hunter-Github]

This is a wonderful library, am trying it now, and have noticed that double values are truncated at output to 6 digits.

A possible solution could be adding a printf format specifier in parens after the type argument:

-r double(.12f)

Would be awesome to be able to print exact hexadecimal contents of the returned value since most IEEE 754 to decimal string conversions will lose information.

[taviso]

Hmm, good idea. I'll try to prototype some code today.

[taviso]

I think writing it in hex is the correct solution, but doing it by default makes it harder to interact with the numbers in bash. What about having a rawdouble:/rawfloat: type for when you don't want to lose precision?

[Hunter-Github]

Fine by me, thanks!

[cemeyer]

Hm, not sure why Github didn't auto-tag this issue from b862cd0d1753cb0bcab35088b672a5ce3bf74c4b. It's related. Maybe sufficient for your needs? Did you have something else in mind, or are you aware of any way it is deficient? Thanks.

[taviso]

Thanks, looks good to me, I can imagine how someone wants to chain a bunch of libm operations without losing precision until the final operation...perfect.