Node Re-architecture.

[ostrosco]

• Instead of storing the closure within the structure and causing us to keep track of the closure, instead we pass in the actual closure when constructing the struct and the function just exists in the structure.
• To ensure ease of use when importing the crate, we added a tests folder where we can show how users of the crate would directly operate.
• Minor rustfmt updates.

[Alex-Addy]

Changes look good. My only final question would be for an example node that uses those fields.

[ostrosco]

@rfdoell I'm not sure where your comment went (I hope I didn't accidentally delete it), but here's the answer to your question.

In node/mod.rs, test_throughput, the Node3 instantiation uses the fields. I'll add an example to the docs along the lines of what @Alex-Addy was asking for with a node that uses all of the options in the macro.

[rfdoell]

Generally looks good to me. Regarding your previous comment, I found the docs after I commented, so I deleted it (forgetting that GH does push notifications on all activity :P) - I'll try and be a little more thorough in the future.

If I were to add anything, it'd me more tests and examples, but the ones that you have should be enough to get started and start making stuff happen. For @garbagetrash , I had the same confusion until I saw the doc on line 53:

/// - func: The function this node executes upon executing `call()`. The
///   function must accept a mutable reference to the node being constructed as
///   its first parameter, but if the function doesn't need to access state the
///   parameter can be safely be ignored.

After playing with it a little, it makes sense to me - this is what I got to: https://play.rust-lang.org/?gist=081728c67f1122c9da72626d66e25576&version=stable&mode=debug&edition=2015 and I think this is how we'll access member fields for nodes which maintain state.