fix multiple writers race on windows

[jabdoa2]

We found a race between two poll_writers which can happen in our test harness. I guess this will also happen on real windows machines. This will trigger the assert Write buffer should not be empty in _write_ready. Therefore, might have to backport it.

[Hedda]

@rob-smallshire Any plans for a quick bug-fix release? E.g. 0.5.2 or similar if using semantic versioning https://semver.org or just 0.6?

[jabdoa2]

We would like a 0.5.1 as well @rob-smallshire @zsquareplusc. Would that be possible?

[puddly]

I would also like to 👍 releasing 0.5.1. Serial writes just do not occur on Windows 10 with the 0.5 release. You have to either downgrade pyserial-asyncio to 0.4 or manually install the master branch via Git.

[Hedda]

@zsquareplusc @rob-smallshire @septatrix @wolfmanjm Can you please publish bug-fix release to resolve these Windows issues?

[hojo0590]

dear maintainers - what needs to be done in order for you to release a new version including this fix? do you want any other issues fixed before? or is this project abandoned (which would be really sad)?

[rob-smallshire]

@hojo0590 Only @zsquareplusc is in a position to make releases. I have and continue to merge fixes, but I cannot release to PyPI.

[jabdoa2]

If you need an additional maintainer I would be more than happy to help. I currently maintain a few pypi packages already (pinball related).

[Hedda]

Only @zsquareplusc is in a position to make releases. I have and continue to merge fixes, but I cannot release to PyPI.

@bufferoverflow please see all the requests for a new version PyPI release of pyserial-asyncio in #69 and #70

As you are on the pySerial-team, do you also have admin privileges to assign release permission to additional devs as maintainers?

https://pypi.org/project/pyserial-asyncio/