Clean up Frontend Typings

[jtpio]

References

This is a continuation of #11196.

Code changes

• Allow alternate ServiceManager implementations in JupyterFrontEnd. Specifically this allows the use of a mock manager.
• Make the JupyterLab.IOptions more permissive: do not require a shell when options are given and allow a shell that meets the ILabShell interface
• Change the language used in the mock irkernel to get some diversity in the kernel data for testing purposes
• Document breaking changes in https://jupyterlab.readthedocs.io/en/latest/extension/extension_migration.html#jupyterlab-3-x-to-4-x. edit: docs preview available here: https://jupyterlab--11537.org.readthedocs.build/en/11537/extension/extension_migration.html#api-breaking-changes

User-facing changes

None

Backwards-incompatible changes

Consumers relying on the concrete ServiceManager have to be updated.

[jupyterlab-probot]

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

[jtpio]

Some tests are failing due to the major version bump, since the packages have not been published to npm yet.

[jtpio]

For reference this seems to be working fine on Binder:

Also the major version bump does not prevent the prebuilt jupyterlab-link-share extension defined here to be activated:

jupyterlab/binder/environment.yml

Line 5 in a353063

- jupyterlab-link-share=0.2

Although webpack logs some warnings in the dev tools console:

[jtpio]

The failing CI checks will require a new 4.0 alpha release to be done as a follow-up. Which I'm happy to do after the PR is merged.

This will also help other PRs that introduce breaking changes to avoid merge conflicts on many packages, since the version bump is done here in one go. And closes #11600.

cc @blink1073 if you would like to have a quick look, since this is a continuation of the original PR started in #11196, now rebased onto the latest master.

[blink1073]

Thanks for picking this up @jtpio, looks good to me!

[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	4227 <- [4320 - 4365 - 4414] -> 4707	2870 <- [2943 - 2960 - 2981] -> 3098
expected	4190 <- [4309 - 4374 - 5252] -> 6147	2752 <- [2858 - 2927 - 3104] -> 3395
Mean relative change	-6.7% ± 2.1%	-0.7% ± 1.1%
switch-from
chromium
actual	535 <- [599 - 623 - 714] -> 1029	441 <- [458 - 469 - 481] -> 610
expected	551 <- [607 - 640 - 741] -> 1104	428 <- [462 - 491 - 564] -> 619
Mean relative change	-3.9% ± 4.0%	-7.5% ± 2.1%
switch-to
chromium
actual	592 <- [641 - 656 - 674] -> 808	489 <- [512 - 528 - 537] -> 562
expected	698 <- [723 - 736 - 750] -> 862	545 <- [627 - 637 - 647] -> 673
Mean relative change	-10.7% ± 1.4%	-16.5% ± 1.0%
close
chromium
actual	533 <- [560 - 590 - 947] -> 997	433 <- [445 - 451 - 462] -> 493
expected	626 <- [676 - 694 - 756] -> 1084	474 <- [507 - 520 - 553] -> 647
Mean relative change	-5.1% ± 6.0%	-16.0% ± 1.5%

Changes are computed with expected as reference.

❗ Test metadata have changed

--- /dev/fd/63	2021-12-06 23:06:30.251544709 +0000
+++ /dev/fd/62	2021-12-06 23:06:30.251544709 +0000
@@ -4,7 +4,7 @@
     "BENCHMARK_REFERENCE": "actual"
   },
   "browsers": {
-    "chromium": "98.0.4695.0"
+    "chromium": "97.0.4666.0"
   },
   "systemInformation": {
     "cpu": {

[jtpio]

Thanks @blink1073!

Moving forward with this and making a new alpha release right after.