Add Hsluv support

[masonium]

Round three!

This finally implements the HSLuv color space as Hsluv. A few notes.

• I didn't implement Uniform generation for Hsluv because I wasn't convinced that the naive version would be anything close to correct. I think you could do something like rejection sample in Lchuv (based on if, after converting that to Xyz, it were in-gamut) and then convert the result to Hsluv, but that's a very different approach than anything I see currently implemented.
• I also added macros for implementing core::ops::Add and related traits, since they're pretty repetitive. I also added tests to confirm that all of the arithmetic operations are implemented (though there is no correctness checking).
• I kept l as the field name, rather than lightness, since it is the same as the Lchuv l.

Closes #112

[Ogeon]

Nice! Good call with the macros. I think they could be used for the other spaces as well. As for the random sampling, I'm pretty sure you can simply use the same method as "regular" HSL uses, which is to sample a bi-cone, where the saturation becomes slimmer towards min and max lightness. There's already code for it in the HSL module. It's going to be its own space in a way, since it's warped from Luv's perspective, but I think that's fine. It's meant to be a more balanced alternative to the naive sampling (which anyone can do on their own), but not necessarily perfect in every way.

[Ogeon]

I always forget the details, but I'm pretty sure #[macro_export] is unnecessary for internal macros, but the order the modules are declared in the lib.rs file matters. Macros and macro modules have to come first. I would suggest putting this module inside the macros module that's already in here.

[masonium]

Ah, I bet the ordering was tripping me up on macros. I'll take your suggestion and just move them to macros.rs, though. I'll also do a bi-cone style implementation for Hsluv.

[Ogeon]

Nice, thank you! It looks to me like this is ready to be merged. I just need to make the no_std_test work again. Looks like it requires __aeabi_unwind_cpp_pr0 now, all of a sudden. Could be that some dependency changed.

[Ogeon]

It should hopefully be fixed now in the master branch 🤞, so you need to merge it in here or rebase your branch. Whatever you are more comfortable with. 🙂

[masonium]

I re-based the branch on top of master.

[Ogeon]

Nice and the tests passed this time, so it seem to be all good! Thank you for all of this! I'll just take the liberty to add a "closing" line to the description, so that goes into the machinery.

[Ogeon]

bors r+

[bors]

Build succeeded:

• continuous-integration/travis-ci/push