Fix: Remove default jupyter state vars that triggers requests to unconfigured server on 0.17.x (#338)

[MarcosVn]

Fixes one scenario reported in #338.

[echarles]

Thank you @MarcosVn. JupyterLab considers you need a server, so the bootstrap part has always been tricky.

Could you double check in various scenario (server, nbmodel without server, bringing the servicemanager or the kernel, notebook, cell, terminal, jupyterlabapp, jupyter context, jupyterreacttheme) and confirm there is no negative impact before I merge (just use the examples in webpackconfig).

Thank you again!

[MarcosVn]

Hey @echarles, I tested the following components:

• ./src/app/App
• ./src/examples/Cell
• ./src/examples/Cells
• ./src/examples/CellLite
• ./src/examples/Console
• ./src/examples/ConsoleLite
• ./src/examples/Deno
• ./src/examples/JupyterLabApp
• ./src/examples/JupyterLabHeadlessApp
• ./src/examples/Kernels
• ./src/examples/Notebook
• ./src/examples/NotebookKernelChange
• ./src/examples/NotebookLite
• ./src/examples/NotebookNbFormat
• ./src/examples/NotebookTheme
• ./src/examples/NotebookThemeColorMode
• ./src/examples/Output
• ./src/examples/Outputs
• ./src/examples/Terminal
• ./src/examples/Viewer

All of them worked well without any apparent inconsistencies caused by the fix for this issue.

However, some of them showed problems that occur even in the HEAD of version 0.17.x, meaning they are unrelated to the changes in this issue.

Some cases include:

• ./src/examples/Deno -> (File Load Error for 10.2_Polar DataFrames.ipynb);
• ./src/examples/JupyterLabApp -> Did not load the ipynb files correctly for use;
• ./src/examples/JupyterLabHeadlessApp -> Did not load the ipynb files correctly for use;
• ./src/examples/NotebookLite -> Could not find "ping.ipynb".

Since these issues occur even outside of this PR, I believe we can ignore them, right?

One more thing I noticed is that in the 0.17.x branch, we ended up bringing in the Kernel state for Cell and Output (#282) evolutions, which I personally hadn't used or tested until now. Is there anything about these changes that I should keep in mind or be concerned about? (such as any subsequent fixes)

[echarles]

LGTM Thx @MarcosVn

[echarles]

Great. I have opened #342 (will look at it asap).

Merging this PR. Thx again @MarcosVn

[MarcosVn]

Hey, @echarles, would it be possible for you to publish version 0.17.2 with this fix? Thanks in advance as well!

[echarles]

Just published 0.17.2 https://www.npmjs.com/package/@datalayer/jupyter-react/v/0.17.2

Curious, any reason to stick to 0.17.x and not migrate to latest?

[MarcosVn]

Just published 0.17.2 https://www.npmjs.com/package/@datalayer/jupyter-react/v/0.17.2

Thank you, @echarles

Curious, any reason to stick to 0.17.x and not migrate to latest?

The reason was mainly due to the fact that we extensively use the cell-related features in our component, and the evolution: "Kernel state for Cell and Output (#282)" or other similar evolution (I really don't remember 100%, sorry) was introduced after we had already tested and developed based on a previous version.

In other words, the idea was to better understand what changed in #282 before using the latest version to avoid risks in our app.

• Note: I briefly tried using a more recent version (0.18.x>) here as well and experienced some issues with the server connection. I can try again and report exactly what happened if it still occurs.

That said, I noticed that even our 0.17.x branch ended up integrating #282, so my initial argument is no longer valid. However, in our tests, version 0.17.x seems quite stable.

Taking advantage of the context, @echarles: Were there any relevant adjustments or changes made after #282 that I should be concerned about?

[echarles]

Thx for sharing context.

Taking advantage of the context, @echarles: Were there any relevant adjustments or changes made after #282 that I should be concerned about?

It is hard to tell without knowledge of your case and usage. This said, the big changes are before and are moving from Redux to Zustand, and add a useJupyter on top of the Jupyter React Context.

At some point staying on 0.17 will limit you for the new big features, happy to help resolving any issue you would report with latest main.