A5rocks
~~Blocked on https://github.com/facebook/pyrefly/issues/43 because the `self`/`cls` parameter of decorated methods in `trio.Path` is generic (`Self`).~~ EDIT: nevermind I misunderstood, I made a new issue to track our problem! https://github.com/facebook/pyrefly/issues/507
Now that that issue is fixed, I tried this again. However, now we hit https://github.com/facebook/pyrefly/issues/1110.
Huh, then couldn't we get rid of any special casing for looking for specifically `type(x)`? i.e. couldn't we just use case 1/2 always? I imagine it's a bit strange (e.g....
IMO it's easiest to mark them as deprecated attributes and not have to worry about interface differences with pytest/aliasing/etc.
I would expect we don't need any special support in pyodide because of `trio.lowlevel.wait_task_rescheduled` (for Python awaiting JS) and the other way around is probably simple to do with plain...
OK this is incredibly annoying for other reasons too. I was looking at an old PR and I have the feeling I'm missing half of the context. I should investigate...
https://gitlab.com/gitterHQ/gitter.im/-/issues/9 !!! it's not closed?! But also this is promising: > Since a Gitter message ID is just a Mongo ObjectId, it has the [timestamp built-in](https://www.mongodb.com/docs/manual/reference/method/ObjectId/). We could have something...
Switching to Zulip or something else should be tracked by https://github.com/python-trio/trio/issues/167, I'll keep this issue only for a) decoding Gitter message IDs to a timestamp, b) setting up some sort...
@JukkaL do you have opinions on how this interacts with (maybe) redefinition? I forgot what heuristics you've mentioned using for it before.
Hmm, I'm not entirely convinced MemorySendChannel/MemoryReceiveChannel are useful descriptions of anything other than what they currently are? Like, what are you going to do, send an object synchronously using a...