Micah Geisel

Yes, that was exactly my thought... extracting `MakeMakeFile` to its own file, and `mkmf` would just load that file and include the module. I think it would probably be nice...

Top-level namespace stuff is weird. This works: ``` ruby require 'uninclude' require 'mkmf' Object.send :uninclude, MakeMakefile ``` After which you can call methods directly on `MakeMakefile`. FWIW.

@rwaskiewicz Very interested in seeing this resolved, too. I just spent a week porting one of our web components to Stencil (which was a huge improvement), and was very surprised...

I was bashing my face against a wall for hours before I found this issue. For me, trying to attach a path relative to working dir caused the _Chrome browser...

> Ferrum now requires absolute path and Cuprite will auto-expand local paths. I dived through the stack trace, and I'm not seeing this, even on current master. This is where...

I tried today, and was surprised to find that the binary wasn't compiled with yjit. ``` 15:09 $ rvm install 3.2 Searching for binary rubies, this might take some time....

Yes, to _compile_ Ruby with YJIT from source, rust is required. But rust isn't required to _run_ Ruby with YJIT, and rvm prefers to install precompiled binaries when it can....

Bumping because still not available in prebuilt binaries for 3.3

@mathieujobin Ah, it looks like the situation has changed! ``` micah@yupa:~$ rvm install ruby-3.3.0 Found remote file https://rubies.travis-ci.org/ubuntu/22.04/x86_64/ruby-3.3.0.tar.bz2 [snip] micah@yupa:~$ ruby --yjit -v ruby 3.3.0 (2023-12-25 revision 5124f9ac75) +YJIT [x86_64-linux]...

Wow, really interesting. My kneejerk hypothesis is that this is a bug somewhere upstream, but that's honestly of little value. :) If you can reduce this down to some kind...