Add into_std methods to structs with from_std methods

[stevenroose]

It seems useful to have those.

[Thomasdezeeuw]

I don't think this is a good idea since the type in standard lib are generally consider blocking, while Mio's types are generally considered non-blocking.

[stevenroose]

I don't think this is a good idea since the type in standard lib are generally consider blocking, while Mio's types are generally considered non-blocking.

Well the same is true for from_std, right? And there's a disclaimer about that on from_std, so I also added a disclaimer about that on into_std. I can make the disclaimer a bit more explicit about the non-blocking state?

[stevenroose]

About the CI failure, it seems like it's unrelated. I don't get it locally, but I occasionally get other failures.

[Thomasdezeeuw]

Yeah it's a known problem: #1376, I've restarted the CI.

[Thomasdezeeuw]

Well the same is true for from_std, right?

That's true.

And there's a disclaimer about that on from_std, so I also added a disclaimer about that on into_std. I can make the disclaimer a bit more explicit about the non-blocking state?

The problem with going Mio -> std, is that Mio keeps some additional state (on Windows) that will get lost. For everything to guarantee to work correctly you would need to deregister before calling into_std. And register when doing the inverse again, i.e. from_std.

Can you add two tests for each type:

1. A tests that creates a Mio TcpStream/UdpSocket/etc. register it with Poll. Create an event for it (accept the stream/send a packet etc.). Next deregister it and call into_std. Then check that we get no events for the type (you can use expect_no_events for this).
2. Basically the same as above, create Mio type, register, create event, deregister and call into_std. But then call from_std and register it again. After that poll and we would expect to get an event (but I'm not 100% sure this will be the case).

[Thomasdezeeuw]

@stevenroose do you plan on continuing this?

[Thomasdezeeuw]

Closing this due to inactivity.