Nathaniel J. Smith

> Hey Nathaniel, > Thank you for writing in to GitHub Support! > We've detached python-trio/hip and made it a standalone repository. I think this means we all need to...

@tomchristie Yeah, we definitely need to figure out more about how hip/httpx relate. For me the biggest difference is the visions; as you wrote as you wrote in https://github.com/encode/httpx/issues/522: >...

Yeah, I'm aware of this style, and it's certainly possible to do. It's just that I'm not sure it will actually be a UX win? Managing tools like 'black' gets...

oh forget to say the main point: I'm totally open to doing this if it turns out to make sense. UX is the kind of thing you have to experiment...

Posy does use the magic marker file that's supposed to tell pip not to mess with its environments: https://github.com/njsmith/posy/blob/cfea84949d057664e0ec1f2f88dbccc701630ce1/src/data-files/EXTERNALLY-MANAGED

I think the issue is that when we scan through to find and run all the trio fixtures, we do that by scanning over the test's kwargs, and I guess...