Matt Yule-Bennett

Subclassing is nice but tricky to do in this case because `MethodProxy` is quite buried. You'd have to subclass `ServiceProxy` as well, and you'd have to do it for each...

This is likely to be an issue with python 3.7 and eventlet. Which versions of eventlet and requests were you using? Can you reproduce by making a `requests.post` inside an...

This is a bug in eventlet and unfortunately there's no easy fix if the [workaround](https://github.com/nameko/nameko/issues/642#issuecomment-506950873) in #642 isn't possible for you. The good news is that @daviskirk is working on...

Holy cow @daviskirk! This is a huge step forwards 🙌 > How should the compatibility layer be implemented? Currently the backend is chosen at import time. I think it should...

> There's a nameko.utils.concurrency package. Should this be moved into the new nameko.concurrency package? Yes I think so. I would quite like everything in this package to go away eventually...

This is amazing work @daviskirk. There are a couple of tiny things left, and then is this ready to go? A sensible next step would probably be to make a...

I can't reproduce this. CI runs fine against pytest 6.2.5 -- https://github.com/nameko/nameko/pull/729

Yep, it is exactly eventlet/eventlet#420. The signature of `process_request` has changed in Eventlet master. I'm not sure what the best way to work around it is yet. It's interesting that...

Thanks @notpeter. I will release a hotfix

Branch coverage is at 99% 😮 I've added an allowed failure to travis in https://github.com/onefinestay/nameko/pull/372