Thomas Kluyver

Thanks @smurfix! I'll reply here, but if there's more specific feedback about my code, let's take it [the Gitlab merge request](https://gitlab.com/takluyver/jeepney/-/merge_requests/11) so we don't generate noise for people who're interested...

I've had another go at [implementing this](https://gitlab.com/takluyver/jeepney/-/merge_requests/11), and I wound up creating a trio 'Future'. I'm wary of implementing a primitive that trio itself doesn't provide, because I know you're...

Thanks, good to know about `AsyncDictionary`. But I think using `is_waiting()` would have the same issue I hit trying to implement that myself: I haven't proved that the waiting task...

Another point I was uncertain about trying to do this was exception types. The [ReceiveChannel interface](https://trio.readthedocs.io/en/stable/reference-io.html#trio.abc.ReceiveChannel) specifies a few different exceptions for cases like "you've closed the receiver already", "the...

It looks like this proposal relies on having a separate send task, which probably makes sense. I've so far avoided this in my [prototype](https://gitlab.com/takluyver/jeepney/-/merge_requests/11), but looking back at your initial...

Yes, I think you're right. What I'm actually using in my code is now a simpler version that doesn't use hazmat - thanks to @ntninja for contributing this. Simplified trio...

Thanks. The necessary feature is in prompt_toolkit 2.0.7?

I doubt this is something browsers expose. Our text editor component is [CodeMirror](http://codemirror.net/). If that exposes a way to control cursor-blinkitude, we could add a separate setting for it.

The first lot of failures seem to indicate that nbformat is missing a file. That doesn't look like a problem with nbval, and I don't think there's any obvious reason...

See also #194.