Stateful entry points

[fjarri]

Stacked on top of #67

This PR makes EntryPoint types stateful - that is, this removes the need for EntryPoint::Inputs because the EntryPoint-implementing type becomes the inputs itself. Fixes #64.

Benefits:

• No need to expose Round types, only the EntryPoint impl can be public
• Protocol writers can write custom constructors (basically what we do in synedrion in constructors.rs) with additional consistency checks or convenience methods

TODO:

• naming?
• Redundancy with node IDs. We get the node ID from the Session level (because it's given a Signer), but we also may have an implicit ID in the inputs (e.g. a key share). How do we ensure these two are the same? Should we? Option 1: pass id to EntryPoint::make_round(), and return an error if it's not the one expected based on other inputs. Option 2: add an EntryPoint::my_id() method, and check in Session::new() that the returned ID is the same as in Signer. Option 3: keep them entirely disconnected.
• Resolve Determining that a round may produce a result #66?

[coveralls]

Pull Request Test Coverage Report for Build 11847505816

Details

• 95 of 101 (94.06%) changed or added relevant lines in 8 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage increased (+0.8%) to 69.779%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
manul/src/session/transcript.rs	4	5	80.0%
manul/src/combinators/chain.rs	25	30	83.33%

Files with Coverage Reduction	New Missed Lines	%
manul/src/combinators/misbehave.rs	1	91.88%

Totals
Change from base Build 11823609419:	0.8%
Covered Lines:	1771
Relevant Lines:	2538

---

💛 - Coveralls