that'd ensure that it's 100% on par with the original implementation right? surely better than an AI rewrite
BTW: I've worked with c2rust output, I know it's shit, but surely they'd do better than my agents since they have a lot more resources and better models than I do