Remove Install, Update and Uninstall buttons from the extension manager UI

[jtpio]

References

Incremental step towards #10554 and #11336.

To improve and unify the user experience around installing extensions, which is at the moment confusing.

And help reduce the number of "JupyterLab failed to build" issues as mentioned in the <details> section of #10554 (comment).

Code changes

This PR:

• removes the Install, Update and Uninstall buttons in the extension manager
• makes it possible to Enable and Disable prebuilt extensions

User-facing changes

Before

After

Backwards-incompatible changes

None (UI only)

[jupyterlab-probot]

Thanks for making a pull request to jupyterlab!
To try out this branch on binder, follow this link:

[jtpio]

If we decide to move forward with this for 4.0, we can then open follow-up PRs to remove more code handling the source extensions (installing, updating, uninstalling)

[fcollonval]

Thanks @jtpio for starting this.

removes the Install, Update and Uninstall buttons in the extension manager

I'm unsure we want to remove the buttons per se. The target is to provide an API that an extension could fulfill as extension provider. So I would suggest that the provider should only care about listing/installing/uninstalling extensions with the UI provided by the core extension.

makes it possible to Enable and Disable prebuilt extensions

❤️

[jtpio]

The target is to provide an API that an extension could fulfill as extension provider

That would indeed be the best long term goal. But since it's unclear if anyone is working on this and the extension manager for 4.0, this PR proposes to solve the more pressing issue of being able to install source extensions and confusing end users. Which seems to have also been a recurring topic at the weekly meetings (for example: jupyterlab/frontends-team-compass#128 (comment))

[fcollonval]

Thanks for this one. Your argument makes sense.

[github-actions]

Benchmark report

The execution time (in milliseconds) are grouped by test file, test type and browser.
For each case, the following values are computed: min <- [1st quartile - median - 3rd quartile] -> max.

The mean relative comparison is computed with 95% confidence.

Results table

Test file	large_code_notebook	large_md_notebook
open
chromium
actual	3826 <- [3919 - 3956 - 4037] -> 4301	2539 <- [2608 - 2678 - 2741] -> 3131
expected	3741 <- [3831 - 3874 - 3916] -> 4198	2469 <- [2569 - 2592 - 2625] -> 2822
Mean relative change	2.4% ± 0.6%	3.3% ± 0.9%
switch-from
chromium
actual	634 <- [695 - 724 - 761] -> 1057	449 <- [477 - 486 - 498] -> 661
expected	563 <- [673 - 702 - 853] -> 931	443 <- [469 - 479 - 490] -> 580
Mean relative change	1.7% ± 3.5%	2.1% ± 1.6%
switch-to
chromium
actual	311 <- [371 - 378 - 385] -> 438	256 <- [276 - 284 - 293] -> 318
expected	311 <- [366 - 374 - 383] -> 499	253 <- [274 - 282 - 288] -> 304
Mean relative change	0.6% ± 1.7%	1.4% ± 1.1%
close
chromium
actual	551 <- [947 - 967 - 986] -> 1084	460 <- [485 - 497 - 509] -> 583
expected	549 <- [927 - 945 - 957] -> 989	456 <- [473 - 485 - 495] -> 510
Mean relative change	3.0% ± 3.4%	3.3% ± 1.1%

Changes are computed with expected as reference.