richardsheridan

Status: - not sure why tox is running quart tests on 3.6 still, that pulls in an old, unsupported version of quart and I tried to exclude it - it...

https://hynek.me/articles/semver-will-not-save-you/

I brought up a different usecase for this [in](https://gitter.im/python-trio/general?at=5f774c641b444b4eafb1bc12) [gitter](https://gitter.im/python-trio/general?at=5f7776b9cfe2f9049a1c67f9) a while ago. Copied for convenience: Let's say `item` is a precious result that we definitely want to send to...

> Surely this would cause blocked senders to stay blocked until an appropriate receive, rather than being put onto the buffer? It works in your case since you're forcing an...

I had to use `hex(int(...))` oddly, but no, the pointer values do not match in the failing case.

If you don't demand the pipes to have duplex behavior, the code works as expected: ```python import trio from trio._windows_pipes import PipeReceiveStream, PipeSendStream from multiprocessing import Process, Pipe def echo(recv_pipe,...

> ... if you close a handle twice, then there's a chance that in between the handle could have been reused for some other object, and you end up closing...

You could fix it in #2241 and then merge or rebase master, what about that?

I was thinking push commits to that dependabot branch until CI comes back green. That should fix the style change in not-beta black regarding the `**` operator while keeping this...

I implemented a [`cache_scope`](https://github.com/richardsheridan/trio-parallel/blob/1.0.0a0/trio_parallel/_impl.py#L58-L118) feature for trio-parallel under the misguided assumption that a `ContextVar` would behave as a `TreeVar` does. This lead to a couple failures that are along the...