Remove libgreen and runtime abstractions

[brson]

Tracking rust-lang/rfcs#230

[aturon]

Nominating for P-High, not on 1.0 milestone.

I'm going to miss the next triage meeting, but I'd like to confirm that the team does not view this as a 1.0 milestone issue, as it does not directly impact the std::io APIs.

I do plan to carry out this implementation work hand-in-hand with API stabilization for I/O, however.

[thestinger]

It does impact the design in that many APIs were designed around the libuv-based runtime.

[aturon]

@thestinger

It does impact the design in that many APIs were designed around the libuv-based runtime.

To clarify:

• The RFC is explicitly motivated by the desire to decouple std::io from the libuv runtime model.
• However, the RFC itself (and therefore this issue) does not propose specific changes to the std::io API. It is about restructuring the implementation to enable such changes in the future.
• I think we can and should consider specific changes to std::io as part of API stabilization (or via new RFCs, when the changes are substantial enough to warrant it.). But that effort is not part of this issue or the RFC it's tracking.
• We are not generally blocking the 1.0 milestone on API stabilization effort, and do not plan to make specific promises about which APIs will be stable by the 1.0 release.
• All that said, as I proposed in my original comment, I think it probably makes sense to work on API revisions hand-in-hand with the work on this issue.

[pnkfelix]

Assigning P-high, not 1.0 milestone.