enable abstract interpreter to constant prop even if method source is generated

[jrevels]

I made this change entirely because @vtjnash dreamed that he had already made me make this change, and was surprised to find that this was not the case in his wakeful reality

[Keno]

I made this change entirely because @vtjnash dreamed that he had already made me make this change, and was surprised to find that this was not the case in his wakeful reality

More common than you think. I dreamed last night that @yuyichao had left a review comment on my #28251 PR asking for an LLVM-level test case of the new feature, but that comment is nowhere to be found.

[vtjnash]

Maybe it's a good thing it hadn't translated into reality. I'm really not sure that this is a good idea.

[jrevels]

@nanosoldier runbenchmarks(ALL, vs = ":master")

[nanosoldier]

Your benchmark job has completed - possible performance regressions were detected. A full report can be found here. cc @ararslan

[vtjnash]

we might want to do this if the Method spoofing bit is set in the CodeInfo, but we'll revisit that when we run into it