Add function aliases for float operators, and clean up other numeric functions

[mbrubeck]

Here's a patch to add functions to std::float that match the ones in std::int and std::uint.

Also included are some patches to clean up signatures and documentation for related functions, and remove a couple of redundant functions.

[brson]

Integrated. Thanks!

[jckarter]

For floats, 0.0 should be positive and -0.0 negative, no?

[brson]

That would seem to be the case. Thanks.

[brson]

mbrubeck, do you want to try to implement that?

[mbrubeck]

mbrubeck/rust@5fa7e614c88 changes float::positive and float::negative to be true for 0.0 and -0.0, respectively. I have no strong opinion about whether this is an improvement or not.

Note that this makes float::positive return the same thing as float::nonnegative, and float::negative return the same as float::nonpositive. Is it worth keeping the redundant functions around for consistency with std::int, or is it just confusing?

[jckarter]

You may want to have nonnegative and nonpositive also be false for nan instead of being strictly !positive or !negative.

[mbrubeck]

I tried to fix NaN behavior in mbrubeck/rust@af645e82c1, but I ran into Issue #1083.

[mbrubeck]

Fixed +0, -0, and NaN behavior in 45d7777 and 7e064de.