Benoit Daloze

Thank you for the report. The relevant part of the Java stacktrace: ``` Caused by: java.lang.NullPointerException at org.truffleruby.core.hash.library.PackedHashStoreLibrary.getHashed(PackedHashStoreLibrary.java:75) at org.truffleruby.core.hash.library.PackedHashStoreLibrary$Set.set(PackedHashStoreLibrary.java:207) at org.truffleruby.core.hash.library.PackedHashStoreLibraryGen$HashStoreLibraryExports$Cached.set(PackedHashStoreLibraryGen.java:107) at org.truffleruby.core.hash.HashNodes$SetIndexNode.set(HashNodes.java:238) at org.truffleruby.core.hash.HashNodesFactory$SetIndexNodeFactory$SetIndexNodeGen.execute(HashNodesFactory.java:981) at org.truffleruby.language.RubyCoreMethodRootNode.execute(RubyCoreMethodRootNode.java:48) at org.graalvm.compiler.truffle.runtime.OptimizedCallTarget.executeRootNode(OptimizedCallTarget.java:656)...

Thanks for the report and reproducer. The initial error here seems the `Assertion` from the first line. And that's then followed by `abort()` and that seems to fail because it's...

> There is a global lock for C extensions by default so that should prevent running any C extension code in parallel. I meant TruffleRuby has that (by default). And...

The failure in macOS CI is unrelated: https://github.com/rvm/rvm/actions/runs/6084945174/job/16507919460?pr=5387 That will need to be fixed first.

So test/fast/* passes now: https://github.com/rvm/rvm/actions/runs/8349160490/job/22852712175?pr=5387 But for some reason truffleruby_comment_test fails without any output. I wonder if maybe the previous CI steps delete `~/.rvm/scripts/rvm` or cause that script to exit...

https://github.com/eregon/rvm/actions/runs/8358105456/job/22878671570 So `~/.rvm/scripts/rvm` exists but: ``` Run source ~/.rvm/scripts/rvm && true Error: Process completed with exit code 1. ``` So it seems somehow the fast tests break rvm somehow, to...

This would be very useful for us too, we want to exclude PR from some bot author.

I thought about this too in the past, but it means there is no way to express "Ruby >= 3 < 4" then, i.e. Ruby 3.x. I'm kind of negative...

I filed https://github.com/actions/toolkit/issues/1320, if that's done that would be enough.

Another idea is maybe we could warn when we get `3` as the input. And then we could say use `ruby-3` if you really want `CRuby >= 3 < 4`,...