Update locks

[jtpio]

Update the lab environment to use the final versions, and remove the rc channels.

[jtpio]

Also checking if this might fix #29 with the following link:

https://mybinder.org/v2/gh/jtpio/robotframework-jupyterlibrary/update-final?urlpath=lab/tree/docs/MAGIC.ipynb

[bollwyvl]

Thanks so much! Really sorry I haven't gotten back around to this. I see the binder has exploded... not sure why the entropy checking is no good there (it shouldn't need to solve the lock environment in binder).

Aside: I'd really be interested in your feedback on other useful keywords... if they are in the wild, #26 would be great place to put them, otherwise issue-per-keyword-family is rad.

[bollwyvl]

In lock.py, I bet if you move from conda.models.match_spec import MatchSpec inside expand_specs it will work (though might need some lint ignore juice)...

[bollwyvl]

Gah, maybe this is better: add to dodo.py:

# at some point, we'll want a scheduled excursion just for locking
if not (P.CI or P.IN_BINDER):

    def task_lock():

[jtpio]

Aside: I'd really be interested in your feedback on other useful keywords...

yeah I've been thinking on how we could use this with other clients such as voila and jupyterlab-classic. Sounds like it would make sense to discuss that in #26 for sure 👍

[bollwyvl]

jupyterlab-classic

skipping around between multiple tabs is a pretty serious pain to test, but otherwise should be alright. the ideal would be to reuse as much as possible from lab, and have a very thin set of overlay keywords. If you build a lot of new CSS classes... meh.

My real goal is to eventually be a bit cleverer, so instead of Execute JupyterLab Command, it's just Execute Command (and would know to do the right thing, because you had already Open JupyterLabed.

[bollwyvl]

conda also fails pip check because of the ruamel-yaml debacle.

[bollwyvl]

voila

sooo.... i haven't been following too closely, but is voila going to support lab3-based federated modules? the use of (or config setup to not need) unpkg is untenable for basically all of my use cases...

[bollwyvl]

i reckon if you roll back to 0f80c63 (where CI passed) and then cherry-pick 9137a2d (which got binder further) it will be ✔️

[bollwyvl]

Locally I had to fiddle with this extra step to be able to run the lab task

Yeah... it's a bit of an odd beast, and not entirely to my liking... but since we're testing a test environment... normally, i don't install geckodriver and firefox in the same env as my system-under-test.

[jtpio]

is voila going to support lab3-based federated modules?

Yes, we are planning to make the Voila frontend a lab-based frontend using 3.0 packages, so existing mime renderers and widgets should work out of the box as federated extensions. There are a couple of smaller steps before this (for example moving to TS), but in any case improving on the end-to-end testing should help catch potential issues when making such changes.

[jtpio]

i reckon if you roll back to 0f80c63 (where CI passed) and then cherry-pick 9137a2d (which got binder further) it will be heavy_check_mark

sure, will do 👍

[jtpio]

Looks like the Binder is starting now. Thanks!

[bollwyvl]

Sorry, I was getting some weird stale state in my browser, and thought some jobs were failing... opened it in another browser, and saw that it was all ✔️ ! Going to take a quick spin around the binder, and very likely merge soon thereafter. Thanks again!