Benoit Daloze

So compilation and C extension handles lifetimes don't change this. It's really weird, also the fact it does not reproduce on other machines.

@phortx Could you try to determine where is the problem by deduction (since it seems to fail reliably on your machine)? For instance you could try without the `.to_f`, without...

Are you using --core-load-path in this case or not? If yes, it's most likely the same as https://github.com/oracle/graal/issues/5125. If not, then the Source doesn't actually have a file that VSCode...

With `--core-load-path` it worked fine when we tried recently, although one needs to first find such file and they are not all in Loaded Scripts.

Potentially we could workaround by using `WarnCallback.NONE` until it's fixed in Joni, but maybe that introduces some other incompatibilities? BTW you can make a PR to Joni if you'd like,...

It got fixed on May 22, but now it needs a Joni release and us to import a newer version of Joni

I think you are seeing https://github.com/oracle/truffleruby/issues/2431. That issue has more details on what's currently done and what's missing. The problem is we do synchronize so only one thread autoloads, but...

One idea is for the constant being autoloaded to not set it yet but either store it in a new field of AutoloadConstant or maybe in the RubyConstant#value of the...

A reliable reproducer: ```ruby # main.rb autoload :C, "#{__dir__}/c.rb" $inside = false t = Thread.new { p C.new.foo } Thread.pass until $inside p C.new.foo t.join ``` ```ruby # c.rb class...

I tried an edge case, what if while loading the file we create another thread which tries to access the being-autoloaded constant? Answer: it deadlocks on CRuby: ```ruby # c.rb...