Consider Upgrading Python 3.5+ Asyncio

[slick666]

Python 3.5 incorporated updated the the asyncio.

PEP 492 - Coroutines with async and await syntax

Lets consider upgrading to the latest version of python asyncio and syntax going forward and possible back-porting these updates into the current code.

[pypingou]

While a good idea, especially when looking into the future, this would make it harder to run ircb on more stable/server distro such as RHEL or CentOS

[slick666]

@pypingou, While I agree I believe we're already in a 3.4+ situation with the codebase as it is. With RHEL/Centos only supporting 3.3+ I don't see us in any worse state. In other words if 3.4 why not 3.5?

[rtnpro]

@slick666 I agree with @pypingou on this. I think we need to consider compatibility with stable/server distro like RHEL and CentOS. The reason we had to move to Python3.5 was due to a dependency on https://github.com/gnarlychicken/aiohttp_auth. We might need to either use a forked version of the above library with supports Python 3.4/3.3.

[slick666]

I'm of two minds about this. I understand the conundrum and agree with the direction. RHEL/Centos has always had poor python support and while it's ubiquity is undeniable I'm skeptical that we can ever fit onto a RHEL7 box with a python asyncio based application. Python 3.4/3.5 are supported on Ubuntu's latest LTS which offers some hope I know its not something that everyone will embrace. (My team at work wouldn't.) I'm all for making this an platform agnostic and flexible as possible but I don't want to give us impossible goals.