Test: ipynb file support out of the box

[joyceerhl]

Refs: #128495

• web @weinand
• macOS @Tyriar
• windows @greazer
• linux @TylerLeonhardt

Complexity: 3

Authors: @roblourens @joyceerhl @DonJayamanne @rebornix

Create Issue

---

This iteration we added out of the box support for rendering *.ipynb files, meaning that users should no longer need to install the Jupyter extension to view *.ipynb files in a notebook UI.

Tip: You can start VS Code with --extensions-dir /tmp/codeexts --user-data-dir /tmp/codedata to try the out of the box experience in a clean environment.

Test:

• With no extensions installed, open an *.ipynb file. The native notebooks UI should be rendered
• A notification to install recommended extensions for *.ipynb should appear (you may need the clean user-data-dir to see this)
• Run button and "Select Kernel" should be rendered, user should be guided to install kernel extensions
• If the notebook contains any outputs that are not rendered by default VS Code renderers, user should be guided to install renderer extensions
  • example: https://github.com/microsoft/vscode-jupyter/blob/main/src/test/datascience/notebook/standard_widgets.ipynb
  • Click the ... next to an output, click Choose Output Mimetype, you should see one or more options marked as "renderer not available", and an option available to "Install additional renderers from the marketplace"
• Install the Jupyter extension. Dev tools should not contain an error about duplicate serializer registrations
• Try opening a bunch of *.ipynbs from GitHub and verify that we handle them well with and without the Jupyter extension installed
• Test editing them, saving them, and test hot exit with autosave disabled