src: introduce internal C++ SetImmediate() and remove async_hooks destroy timer

[addaleax]

Introduce env->SetImmediate(fn) and remove the handle we use for async_hooks destroys.

After #17105, a similar thing could be done for HTTP/2.

(The one test change is due to the slightly different timing for async_hooks, since timer and check handle callbacks run at different times.)

Checklist

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

Affected core subsystem(s)

src

[addaleax]

CI: https://ci.nodejs.org/job/node-test-commit/14126/
CI: https://ci.nodejs.org/job/node-test-commit/14131/

[jasnell]

i'm going to be revisiting the lifecycle of http2 sessions and streams soon after #17105 lands, good timing on this :-)

[addaleax]

Landed in 85f3e31, 5d7f5c1

[MylesBorins]

Should this be backported to v9.x-staging? If yes please follow the guide and raise a backport PR, if not let me know or add the dont-land-on label.

[jasnell]

Yes, this will need to be back ported as it is needed by http2 (or at least will be soon)

[MylesBorins]

Managed to get it to land

[jasnell]

@addaleax @mcollina ... this needs a backport to v8.x-staging in order for me to be able to successfully backport the current http2 implementation to v8.x.

I'll try to get to it this week but if one of you happen to get to it first I'll owe you one beverage of your choice :-)

[MylesBorins]

ping re: 8.x backport

Set don't land for v6.x

[addaleax]

@MylesBorins The v8.x backport for this was in #18086 (or @AndreasMadsen’s #18179)

What’s the reason for dont-land-on-v6.x? I’d consider it possible that other backports in the future may want to have SetImmediate available.

[mcollina]

@addaleax I'm for backporting on a need-to basis. This was backported to 8 because it was needed for http2. Given that we are not going to backport that, what's the need for v6? Of course, if there is something we want backported that depends on this, we should backport. I don't think there is a need to backport sooner.

[MylesBorins]

I assumed it wouldn't land on 6 as there are not currently plans to backport async hooks. Feel to change labels as appropriate though

[addaleax]

@MylesBorins The first commit here isn’t really about async_hooks, it introduces a very general internal utility – I’m okay with doing what @mcollina suggests and just backporting when it starts to get used by something else