Ashanti/connect state transition

[StaticallyTypedAnxiety]

Summary of changes
Changes introduced in this pull request:

• Looks like the channel was being used for concurrency, didn't implement due to rust safety guarantees
  -a bit of a translation of tipset_state in lotus
  -missing a few functions that call tipset_state
  -Changes statemanager in sync to an arc readlock due to the need of shared reference in asynchronous context

Reference issue to close (if applicable)

Closes #405

Other information and links

[austinabell]

You also need to revert the changes to the submodule you committed

[ec2]

Was there a reason why you removed the concurrent validation of block headers?

[austinabell]

Was there a reason why you removed the concurrent validation of block headers?

I was going to come back to this in a bit but might as well now, before it was spawning a blocking task which is just synconous (I didn't look too much in detail at the last PR just assumed you guys had verified it) Think it just needs to be changed to task::spawn from what it was

(I actually don't know how the FuturesUnordered handles these boxed async functions being passed in or when they are executed but I can look into it)

[StaticallyTypedAnxiety]

tually don't know how the FuturesUnordered handles these boxed async functions being passed in or when they are executed but I can look into it)

Yeah they have to be boxed because each future has a different type even though in essence it looks the same. So adding different types is according to the future trait rather than a concrete future type.

It looked like it was returning join handles with futures inside of them and using spawn_blocking so in essence didn't see much of a difference between handling them where they are vs handling them at the end where the errors are being collected.

Changed to spawn right now

[dutterbutter]

Should add some comments to the public methods if possible.