Kernel status messages should be sent before & after each shell request

[takluyver]

We currently only send them on execution. I think the spec wasn't clear when we first wrote the code, but this was decided in message spec version 5:
http://jupyter-client.readthedocs.io/en/latest/messaging.html#kernel-status

This would be a relatively easy task if someone is looking to start contributing.

From jupyter/services#139

[jankatins]

@parente, @takluyver Isn't this closed with the merge of #312?

[takluyver]

Yep, thanks.

[enigmata]

@takluyver @JanSchulz any outlook on a new release incorporating this fix? we are using jupyter-js-services against R kernels and this makes it rather troublesome to use.

[jankatins]

IMO we could send IRdisplay to CRAN and then IRKernel. Maybe implement IRkernel/IRdisplay#28 first (at least set 'jupyter.rich_display' in IRkernel), but the rest seems ok to ignore for a release.

[dockter-84]

Yes, I'm interested in incorporating the fix, too.

[enigmata]

@JanSchulz seems good. so how does this progress to closing on a new release? need to vote somewhere and slip bitcoins to somebody? ;-)

[enigmata]

@flying-sheep any comment on getting a release?

[flying-sheep]

i seem to be a bit inactive in the last weeks, right? 😐

well, @JanSchulz has created milestones for both the kernel and IRdisplay. i’ll see if i get IRdisplay submitted to CRAN this weekend

[enigmata]

@flying-sheep how did things work out on the weekend? Are we rolling?

[flying-sheep]

sorry, no

btw.: i’d like to get IRkernel/IRdisplay#19 fixed first.

[enigmata]

@flying-sheep @JanSchulz so now we're waiting on this one? IRkernel/IRdisplay#20

[flying-sheep]

i think we’re basically waiting on IRkernel/IRdisplay#33 and #368.

it’s important to get this nailed down before we publish (clarify the packages’ area of responsibility)