Add asio's coroutine.hpp for C++ coroutine support

Add asio's coroutine.hpp for C++ coroutine support

R=ikilpatrick@chromium.org,eae@chromium.org
BUG=635619

[Ryan Sleevi, 2016-08-11 18:02:09]

rsleevi@chromium.org changed reviewers:
+ rsleevi@chromium.org

[Ryan Sleevi, 2016-08-11 18:02:10]

Drive-by: This doesn't seem allowed by
https://google.github.io/styleguide/cppguide.html#Boost

[cbiesinger, 2016-08-11 18:07:33]

On 2016/08/11 18:02:10, Ryan Sleevi (slow) wrote:
> Drive-by: This doesn't seem allowed by
> https://google.github.io/styleguide/cppguide.html#Boost

Not sure this counts as a boost library? I got it from http://think-async.com/

[Ryan Sleevi, 2016-08-11 19:52:15]

They're basically the same thing, except the non-boost version includes some of
its own support libs (e.g. threading, errors).

http://think-async.com/Asio/AsioAndBoostAsio

[ikilpatrick, 2016-08-11 20:13:49]

Description was changed from

==========
Add asio's coroutine.hpp for C++ coroutine support

R=ikilpatrick@chromium.org,eae@chromium.org
BUG=635619
==========

to

==========
Add asio's coroutine.hpp for C++ coroutine support

R=ikilpatrick@chromium.org,eae@chromium.org
BUG=635619
==========

[ikilpatrick, 2016-08-11 20:13:49]

ikilpatrick@chromium.org changed reviewers:
+ jyasskin@chromium.org

[ikilpatrick, 2016-08-11 20:17:44]

+jyasskin

Right we discussed this with jyasskin about about this last week; and this
seemed to the best library to fit our needs short of rolling our own.

In an ideal world we'd wait until c++XX has resumable functions but we don't
live in that world quite yet :)

jyasskin thoughts?

[xam, 2016-08-11 20:22:38]

On 2016/08/11 20:17:44, ikilpatrick wrote:
> +jyasskin
> 
> Right we discussed this with jyasskin about about this last week; and this
> seemed to the best library to fit our needs short of rolling our own.
> 
> In an ideal world we'd wait until c++XX has resumable functions but we don't
> live in that world quite yet :)
> 
> jyasskin thoughts?

LGTM from OSPTR

[Jeffrey Yasskin, 2016-08-11 21:10:36]

On 2016/08/11 20:17:44, ikilpatrick wrote:
> +jyasskin
> 
> Right we discussed this with jyasskin about about this last week; and this
> seemed to the best library to fit our needs short of rolling our own.
> 
> In an ideal world we'd wait until c++XX has resumable functions but we don't
> live in that world quite yet :)
> 
> jyasskin thoughts?

It's reasonable to check with our local style folks. I don't think it makes
sense to ask the Google arbiters to allow this library, since there's no google
code that wants to use it.

One reason Boost libraries are banned is that they tend to pull in the world,
but this one's standalone, so that argument against doesn't apply. The bigger
risk with this library is that we'll find something subtly confusing about the
'yield' style it encourages. That's a bit mitigated if it's compatible with the
C++ proposal that's already implemented in MSVC.