Port fixes from node-16 branch to node 17

[seanmakesgames]

No description provided.

[seanmakesgames]

@lloeki@SamSaffron
Failures on this run:

• 2 in 2.7
• 1 in ruby test 3.1:
  https://github.com/seanmakesgames/libv8-node/actions/runs/3990378667/jobs/6846869586

In file included from ../deps/v8/src/heap/cppgc/prefinalizer-handler.cc:5:
../deps/v8/src/heap/cppgc/prefinalizer-handler.h: In member function 'size_t cppgc::internal::PreFinalizerHandler::ExtractBytesAllocatedInPrefinalizers()':
../deps/v8/src/heap/cppgc/prefinalizer-handler.h:32:17: error: 'exchange' is not a member of 'std'

[Fayti1703]

2.7 failures look like ye olde bundler libc detection problem.

[Fayti1703]

482ac9c fixes the Ruby test in 3.1 musl. I expect that change to be specific to this version of node.

[seanmakesgames]

@lloeki@SamSaffron we are green.

(except for the two 2.7 tests which are failing the same way as the 16 branch)

[lloeki]

2.7 failures look like ye olde bundler libc detection problem.

this should be fixable by updating rubygems+bundler. if not then there may be an upstream regression.

[seanmakesgames]

2.7 failures look like ye olde bundler libc detection problem.

this should be fixable by updating rubygems+bundler. if not then there may be an upstream regression.

Are you doing this? Where do we update this?
I was never quite sure on which version pair for those- Is this a change in build.yml? I don't see a bundler version specified in there-

Maybe you can just make the change so we can see where that is controlled.

[lloeki]

Yeah I can make the change later today

[lloeki]

I pushed a bunch of commits to node-16 that makes CI green, one of them being about pinning rubygems. Probably these should be cherry-picked over here too.

[Fayti1703]

(Note to self: Primary commits are 0fad627 & 4db6396)

[Fayti1703]

I've dropped the change in ext/libv8-node/extconf.rb from 4db6396, since this branch doesn't yet have the Truffleruby dummy Makefile generation in.

[Fayti1703]

CI is almost entirely green, with a failure on darwin test due to the GC test: https://github.com/seanmakesgames/libv8-node/actions/runs/4009651664/jobs/6887267190

Looks like a transient failure, though. Probably something to do with how the V8 GC was feeling that run.
Reran with no changes and it's green: https://github.com/seanmakesgames/libv8-node/actions/runs/4009651664/jobs/6892966850

[seanmakesgames]

@SamSaffron@lloeki
CI is 100% green here:
https://github.com/seanmakesgames/libv8-node/actions/runs/4009651664

Can we get a merge?

[lloeki]

Done!

[SamSaffron]

wow looking amazing! how far are we from a new release here? Should we move 17 branch to main ? it is weird that the cutting edge is tucked away in a branch.

[seanmakesgames]

Completely agree @SamSaffron - though are we stopping at 17 or going to 18?

[SamSaffron]

would love a go at 18 ❤️

[seanmakesgames]

Looks like there's already a branch for 18; we'll do another rebase

[seanmakesgames]

Was just about to make a PR from 18 to master, but I'd love for us to change master to main beforehand -- Can one of you make that change? @SamSaffron@lloeki

[lloeki]

On this project I don't merge into master, so:

• Updated master to point to last node-18 commit
• Created node-19 branch

I'd love for us to change master to main

This is a separate matter, please open a dedicated issue for that (one on libv8-node and one on mini_racer)