[commandlinefan]

> won't work if num is negative

I remember reading (although I can't find it now) a great analysis of all the optimizations that Javascript compilers _can't_ do because of the existence of the "eval" instruction.

[cibyr]

The extra fun thing about this is that eval has different semantics if it's assigned to a different name, in order to allow JavaScript implementations to apply extra optimizations to code that doesn't call a function literally named "eval": https://developer.mozilla.org/en-US/docs/Web/JavaScript/Refe...

Andy Wingo (of course!) has a good explanation of this: https://wingolog.org/archives/2012/01/12/javascript-eval-con...

[LoganDark]

Could this perhaps be it? https://janvitek.org/pubs/ecoop11.pdf

[astrange]

A JIT can do any optimization it wants, as long as it can deoptimize if it turns out it was wrong.

[adrianN]

You also want to prove that the „optimization“ doesn’t make things slower.

[astrange]

I don't think anyone actually bothers to do that tbh.