Add back ipythonhandler for backwards compatibility for notebook extensions on nbclassic

[echarles]

Add back IPythonHandler for backwards compatibility for notebook extensions on nbclassic

This is related to https://github.com/jupyter/nbclassic/pull/113

[codecov-commenter]

Codecov Report

Merging #868 (015b7dd) into 1.x (25b8011) will increase coverage by 0.06%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##              1.x     #868      +/-   ##
==========================================
+ Coverage   70.35%   70.42%   +0.06%     
==========================================
  Files          62       62              
  Lines        7603     7654      +51     
  Branches     1255     1270      +15     
==========================================
+ Hits         5349     5390      +41     
- Misses       1873     1878       +5     
- Partials      381      386       +5     

Impacted Files	Coverage Δ
jupyter_server/base/handlers.py	65.46% <100.00%> (+0.13%)	:arrow_up:
jupyter_server/services/contents/filemanager.py	72.20% <0.00%> (-0.59%)	:arrow_down:
jupyter_server/serverapp.py	65.42% <0.00%> (-0.20%)	:arrow_down:
jupyter_server/services/contents/handlers.py	86.55% <0.00%> (+1.09%)	:arrow_up:
jupyter_server/services/kernels/kernelmanager.py	80.58% <0.00%> (+1.94%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 25b8011...015b7dd. Read the comment docs.

[echarles]

I see a few CI errors. Is this expected on 1.x branch?

[blink1073]

If you merge the current branch it will fix the check-release. If you run pre-commit run --all files it will fix the lint errors. I'll backport #865 for the docs part.

[blink1073]

You merged with master instead of 1.x

[echarles]

Aargh, and I have even written it :)... Let me undo...

[blink1073]

Ah, I hadn't merged https://github.com/jupyter-server/jupyter_server/pull/871 yet, can you please merge from 1.x again?

[echarles]

I have rebased on latest 1.x. There are still 4 failing jobs which are imho unrelated.

[echarles]

Moving to draft as we are discussing this in the notebook meeting.

[blink1073]

Okay, the last remaining error is due to a beta release of tornado. We'll have to handle this in main and then backport.

[kevin-bates]

I was wondering if we need to worry about future enhancements made to JupyterHandler that we may not (?) want to support (or can support) in subclasses of the "classic" IPythonHandler?

If that's something to consider, we may want to either swap the two (i.e., rename JupyterHandler to IPythonHandler and let JupyterHandler be the alias for now until new changes occur), or, better yet, rename JupyterHandler to some kind of "base handler" and let both JupyterHandler and IPythonHandler be aliases. With the latter approach, using instanceof on a handler asking about either JupyterHandler or IPythonHandler would then be definitive since the two are siblings.

[echarles]

Thx for jumping here @kevin-bates. We have been talking about this PR during today notebook meeting and other ideas have been setup like having the IPythonHandler being added in notebook or nbclassic. These ideas still need to percolate and land softly.

We are still defining who is shimming what. It sounded to me that jupyter-server was the perfect place to host those base handlers and ensure the backwards and forwards compatibilities but have not raised consensus on that so far.

If that's something to consider, we may want to either swap the two (i.e., rename JupyterHandler to IPythonHandler and let JupyterHandler be the alias for now until new changes occur), or, better yet, rename JupyterHandler to some kind of "base handler" and let both JupyterHandler and IPythonHandler be aliases. With the latter approach, using instanceof on a handler asking about either JupyterHandler or IPythonHandler would then be definitive since the two are siblings.

Sounds good to me. We could review/confirm also our current handler hierarchy at that occasion, but this will mainly depend on the outcome of the current discussion.

[Zsailer]

Hey @echarles, we're triaging PRs in the Jupyter Server meeting today. Because there hasn't been activity on this PR in a little while, I'm going to close it. Thanks for your work here and feel free to re-open anytime if this is in-progress.