Benoit Daloze

I'm not comfortable with making `--core-as-internal` supported, it's a hack (e.g., makes the profiler ignore core methods when using the Chrome Inspector) and in fact it already doesn't work in...

@nurse @nobu What do you think of these changes? I found that these changes are enough to pass all tests, and remove all usages of `rb_define_dummy_encoding()`. So I think `rb_define_dummy_encoding()`...

@nurse Could you give me an example of these ten dummies? The fact that no test reveals the usage of those this sounds to me like it's a non-issue in...

For NKF, AFAIK all encodings listed in https://docs.ruby-lang.org/en/master/NKF.html are already built-in in Ruby, so I don't see the need for dummy encodings.

The issue here is that `$FILENAME` is a preexisting global, which when read returns the same as `ARGF.filename`. And that also fails on CRuby in such a case: ``` $...

Thank you for the report, we'll look into it. FWIW here is the error on JDK19 with more details: ``` Cannot load from object array because "this.globalContainers[id]" is null (java.lang.NullPointerException)...

This is fixed now.

I think it's very difficult/impossible to cancel ongoing interop messages for local variables. But, obviously they should be faster. Another idea would be to run that in a different thread,...

I'm unsure how we can solve this from a quick look, because the variable could be reassigned (and we have no way to know that currently I think) and the...

One idea is most (all?) usages of rb_gc_register_address()/rb_global_variable() are done during the call to the `Init_` function. So we could just collect all these addresses during that call and then...