Implementing rand::Rand

[shingtaklam1324]

Implements #12 .

Right now it is implemented with calling rng.gen::<T>() for both the real and imaginary parts. I'm not 100% sure on the mathematical randomness of the distribution if the complex number is used as a vector and the angle is wanted.

As well as that, for the dependency on rand, I just pulled the newest version on crates.io, but it may be better to use another version. I don't have many projects that use this so I'm not sure how big of an issue this will be.

Changes:

• Optional rand feature and dependency on rand
• impl<T> rand::Rand for Complex<T>

[cuviper]

Thanks!

The distribution is an interesting question, but I think the simple way you've implemented it is the best we can do in general. If someone wants a smoother distribution of the angle, for instance, they can generate that separately and use from_polar.

We'll need to enable this feature in CI -- and if you do that for all targets you'll see that it won't work for rustc 1.8. Rand 0.4 requires rustc 1.15, and rand 0.3.22 now re-exports 0.4 too. I don't mind that this new optional feature isn't fully compatible though. I also intend to update rustc in a num-complex 0.2 bump very soon anyway.

Please do squash your commits -- you have a lot of unrelated churn. The final diff is fine, but I'd rather avoid altering the history of those other lines, affecting tools like git blame.

[shingtaklam1324]

@cuviper
I have squashed the commits as best as I can, but as a git noob, I can't seem to get it to do it properly.

As with the version of rand, I have it on = 0.3.14 right now as I can't get any other to compile on rustc 1.8.0

[cuviper]

I rebased and squashed your commits completely, then updated to 0.4 since we're now using Rust 1.15 min.

Thanks again!

bors r+

[bors]

Build succeeded

• continuous-integration/travis-ci/push