I though about that way.

I suppose it's just:

def fetch_values(*indexes, &block)
  indexes.map { |i| fetch(i, &block) }
end

Since it's a new method, clearly it's not a bottleneck in anyone's app. We could make a primitive call for fetch if we don't want to push an extra frame. I don't have a strong opinion though, implementing in C is fine too (though we'd probably want to convert to Ruby again in the future for YJIT? 🤷🏻‍♀️)

#11557

Yeah, my concern isn't so much performance but more about potentially yielding at an unexpected point or some other very subtle behavior issues. And also just consistency with Hash#fetch_values which it's modeled on.

But yeah, this method doesn't exactly cry hotspot.

But yeah, this method doesn't exactly cry hotspot.

The good part about implementing it from the beginning in Ruby is that if people complain about speed we can rewrite in C (or tell them to use YJIT). 😂

The commit message is wrong, s/-/_/

Corrected.

Merged #11557 instead