Port newlib hypotf

[ChiefMilesEdgeworth]

Closes #137

• Ports the newlib version of hypotf into the main module
• Moves the old code into a submodule musl
• Adds tests for hypotf

[ChiefMilesEdgeworth]

Actually, this has a test error on the cargo test --features musl-reference-tests test.
This has to do with the new directory added, and I'm not sure what I could do to fix that. I'll dig around, but if someone knows where the issue is I'll accept the help

[ChiefMilesEdgeworth]

The test still fails, but now it's a docker issue, so I have no idea.

[ChiefMilesEdgeworth]

Here's the error:

error[E0463]: can't find crate for `std`
  |
  = note: the `x86_64-unknown-linux-musl` target may not be installed

error: aborting due to previous error

For more information about this error, try `rustc --explain E0463`.
thread 'main' panicked at 'assertion failed: status.success()', build.rs:344:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[alexcrichton]

Thanks! If we switch the implementations then there's no need to keep the old implementation. Is this implementation better than the old one? (e.g. faster, more accurate, etc)

[ChiefMilesEdgeworth]

The reason for porting this implementation as I understand it was to avoid using f64s. I'm not sure how to benchmark the speed differences, but the theory is that this will run significantly faster on architectures without a builtin representation for f64 types. I think most of the discussion on this was in the issue this PR closes, #137