> Any chance of this being elaborated further?
I am pretty sure this is vague on purpose, and it should be like that, because comparing functions for equality will either need to expose implementation details or have to be underspecified.
What is specified is already pretty strong, mandating, for example, that JIT recompilation (if any) should behave so that the function "values" taken before and after recompilation compare equal.
Cheers,
V.
