Add lint to detect manual slice copies

[marcusklaas]

This PR replaces #1937.

Addresses issue #1831.

I see there are still some formatting issues. This PR has been formatted with the latest rustfmt (stable). I am receptive to instructions on how to fix these.

[oli-obk]

This PR has been formatted with the latest rustfmt (stable).

It looks like you actually ran both (first nightly, then stable).

I am receptive to instructions on how to fix these.

So I applied the method that I'd desire in situations like these directly to the PR:

1. A commit doing all the formatting without touching the files in any non-formatting way
2. A commit adding tests showing the false positives
3. A commit with the fixes and the updated tests

This makes it really easy to see the changes you did in the last commit, since the ui-tests don't show any line change noise

[oli-obk]

Did you ever address

After a more detailed inspection of the documentation on clone_from_slice, I think the current lint suggestions may cause panics. The destination slice must apparently be exactly as long as the source. Also, it may be preferable to use copy_from_slice when T: Copy, although in my experience clone_from_slice also compiles to a memcpy in release builds.

in the previous PR? This PR looks good to me with the suggestion fixed to produce the same semantics as the loop.

[marcusklaas]

Thanks for the commits (I didn't know that was possible!) and the guidance!

I did address the panic situation. The destination slice should now always be as long as the source slice, although I did not make a separate commit for it.

[oli-obk]

Great!

Another lint that will have a positive impact on many projects. So much cool stuff in clippy lately

[marcusklaas]

👍

[Manishearth]

Wow this is a pretty sweet lint