On Sun, Oct 9, 2016 at 8:06 AM Naoki INADA <songofacandy@gmail.com> wrote:

Hi.

I've just pushed C implementation of asyncio.Future [1].

It will be included in Python 3.6b2.

It may have rough edge.  But I hope it will be battle tested

before Python 3.6.0.

Please help us by running your tests on Python 3.6b2, and

report issues you found.

If you have asyncio based project, and it uses Travis-CI,

please add "nightly" to your .travis.cnf [2].

Feedback about performance difference is also welcome.

Thanks,

[1]: https://hg.python.org/cpython/rev/678424183b38

[2] https://docs.travis-ci.com/user/languages/python/#Choosing-Python-versions-to-test-against

Hi,

I have setup a 3.6 build on Travis of our project GNS3 and it seem to fail.

https://travis-ci.org/GNS3/gns3-server/builds/167703118

To be honest I didn't dig the issue and could be wrong usage of asyncio (or ugly mock in tests). But this could perhaps interest you.

I think the most interresting test is:

https://travis-ci.org/GNS3/gns3-server/jobs/167702085#L2270

Because we test a websocket client server connection.

Thanks a lot for your work