src: clean up MultiIsolatePlatform interface

[addaleax]

• Since this was introduced, V8 has effectively started requiring
  that the platform knows of the Isolate* before we (or an embedder)
  create our IsolateData structure; therefore, (un)registering it
  from the IsolateData constructor/destructor doesn’t make much
  sense anymore.
• Instead, we can require that the register/unregister functions
  are only called once, simplifying the implementation a bit.
• Add a callback that we can use to know when the platform has
  cleaned up its resources associated with a given Isolate.
  In particular, this means that in the Worker code, we don’t need
  to rely on what are essentially guesses about the number of event
  loop turns that we need in order to have everything cleaned up.

(Semver-major because of the ABI/API changes.)

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• documentation is changed or added
• commit message follows commit guidelines

[nodejs-github-bot]

@addaleax build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/2734/pipeline

[addaleax]

CI: https://ci.nodejs.org/job/node-test-pull-request/21201/

[addaleax]

Resume CI: https://ci.nodejs.org/job/node-test-pull-request/21230/ (:heavy_check_mark:)

[addaleax]

This would need another @nodejs/tsc review because it’s technically semver-major.

[targos]

I approve the change but can't review it. Does that count?

[mcollina]

LGTM

[addaleax]

@targos If it’s because of lack of clarity about what this patch does (on any level), I’m happy to talk it through with you or somebody else?

Also, I’m noticing that I did some similar things to what this PR had in #26525 … I think I like those changes better, and I want to avoid conflicts because this PR is semver-major, so I undid the changes to NewIsolate() here.

CI: https://ci.nodejs.org/job/node-test-pull-request/21371/

[targos]

@addaleax

If it’s because of lack of clarity about what this patch does (on any level), I’m happy to talk it through with you or somebody else?

No, it's just that I'm not good enough with C++ to understand everything that you did. I get what is semver-major in the PR and I'm fine with that.

[joyeecheung]

I remember we usually use std::list for these?

[addaleax]

Uh… is there any reason for that? Unless we need referential integrity or expect a lot of overhead from copying large entries, std::vector should be the choice with lower runtime + memory overhead … I think we should switch some of the others over to std::vector.

In any case, the expected length of the ShutdownCbList here is 1, so std::list should definitely not be necessary.

[addaleax]

Landed in 6d9aa73