Having trouble running nbclient 0.3.0 about nbclient

[golf-player]

In this case, I don't see the issues as being specific to async vs sync. These are attribute errors, and the kernel manager hasn't even been created yet.

As for the cases where things don't shutdown, that appears to be intentional.

Agreed, it would make sense to not worry about 0.3.0 as things work as expected in 0.2.0

[choldgraf]

many thanks for the patch @golf-player 👍

[golf-player]

CC: @kevin-bates

[kevin-bates]

Hi @golf-player - thanks for the copy. I haven't used nbclient but have experienced odd issues when mixing @gen.coroutine/yield methods with those using async/await. Are you seeing any 'method not awaited' messages or anything of that ilk buried in the logs?

Once NB 6.1 is finalized, I'll be merging the async RMK in EG. I'd like to get your changes merged prior to that, so I wonder if holding off looking into nbclient 0.3.0 would be worthwhile until the async RMK is, um, "out of the rough and in the fairway". (sorry, couldn't resist)

[davidbrochart]

When I run the above, but with reset_kc=True, I run into a lot of issues, first of which is that in execute, self.kc has not been initialized from the class before self._async_cleanup_kernel is called, so that fails and while it's handling that failure, it fails again on self.km.cleanup() since self.km doesn't exist.

We should probably check if self.kc is not None in _async_cleanup_kernel, in case self.km doesn't exist.

[golf-player]

@davidbrochart thanks for the reply. Agreed, that's probably the way to go.

If y'all are willing, I can make a pull request changing that

[davidbrochart]

That would be great!

[golf-player]

#61

not much to it, tbh.