Yury Selivanov

Actually, uvloop can block only if you supply a `preexec_fn` argument to `loop.subprocess_shell()` or `loop.subprocess_exec()`. Could you please share your test?

> @1st1 any plan about this? I still haven't received any **clear answer** to https://github.com/MagicStack/uvloop/issues/118#issuecomment-391923149. I know that uvloop is blocking when `preexec_fn` is used, but almost nobody uses it....

Great, I'm working on a release right now, and will likely merge #268 or a variation of it.

> Here is a simple test to reproduce the behavior. It's relying on slowfs. Nice, that's great! Thanks for the slowfs tip, I didn't know about it before. > would...

https://github.com/MagicStack/uvloop/pull/160 is the first attempt at implementing the new API. The main thing I don't like about it is that it's designed around having various `_begin` and `_end` callbacks in...

> So summary, providing uvloop its own system to create and handle spans might create some resistence to be adopted by some tracers. I agree, that will be a problem....

> I like zipkin approach to spans since API is minimalistic and very powerful, here I created tracer for asyncio with aiohttp instrumentation (https://github.com/aio-libs/aiozipkin). I have implementation of spans/tracers/context etc....

Rebased agains the current master branch.

> Lots of people are looking for it. Maybe, but there's no good reason for that. Do you have a good reason? Vanilla asyncio works perfectly fine on Windows, there's...

@ahartikainen > I'am happy to help, but not sure where to start. A good start would be to rebase this PR on top of the latest master branch. I'd then...