Add a test to simulate integration into built-in actors about filecoin HOT 2 OPEN

I'm not too concerned with how changes here interact with the builtin actors runtime, I'm more concerned with how changes to this library might impact downstream users.

Specifically, I want to make sure that improvements to the abstractions in this library actually improve the usability of these abstractions in practice. Toy examples in this library will help, but the only sure-fire way to validate usability is to try out changes in the builtin-actors repo (for the moment, at least).

For example, the change where Token::wrap started taking the ActorRuntime by reference instead of by-value snuck into #190 but ended up causing a somewhat painful migration downstream because the ActorRuntime now needed a place to live on the stack.

Basically, I'm asking for something exploratory. Before merging a PR with significant API changes:

1. Checkout a copy of the builtin actors.
2. Patch in the change to this library.
3. Try fixing any compilation errors.

from filecoin.

I think we're saying a similar thing? The built-in actors are a downstream user. My proposal aims to detect something like that ActorRuntime-needs-the-stack issue earlier by having it emerge in code in this repo, which is much quicker to discover and so a better design environment. I'm not saying we should never also try an actual integration, but we can make it less part of tight dev cycle.s

from filecoin.