Update to libv8-node 22.5.1.0

[lloeki]

No description provided.

[tisba]

macOS 11 is no longer available in GHA. I have no idea why they hang, instead of failing right away. Fixed in #307.

Also, something seems to be off, as you can see https://github.com/rubyjs/mini_racer/pull/307/checks the build on main is fine 🤔

[lloeki]

As we saw with Sam, the failing test appears to be from V8 optimising an empty function out. I'll fix that by making it do something (same way I did for another similar test).

Also I fat fingered LIBV8_VERSION in version.rb :(

And since I want to fix this one too rubyjs/libv8-node#36 I'll certainly do a .1 release (which will be automatically picked up given the on-purpose libv8-node version requirement in mini_racer) once they're done.

[tisba]

I'm not sure why the test_timeout uses a Promise to begin with, but the timeout is triggered (correctly IMO), by just while(true) {}. At least in my testing, this makes the test pass (tested on aarch64 and Darwin 23.5.0 arm64):

diff --git i/test/mini_racer_test.rb w/test/mini_racer_test.rb
index 6b139fd..f1fd1b9 100644
--- i/test/mini_racer_test.rb
+++ w/test/mini_racer_test.rb
@@ -1044,12 +1044,8 @@ def test_timeout
     context = MiniRacer::Context.new(timeout: 500, max_memory: 20_000_000)
     assert_raises(MiniRacer::ScriptTerminatedError) do
       context.eval <<~JS
-        var doit = async() => {
-        while (true)
-          await new Promise(resolve => resolve())
-        }
-        doit();
-        JS
+        while(true) {}
+      JS
     end
   end

[SamSaffron]

Cool made #308

dropping Ruby 3.0 as well while at it.