Re: [Python-Dev] Avoiding CPython performance regressions