Memory and string functions can be improved dramatically on x86 and x86-64 glibc

This is a pretty sad exchange. Fog essentially says "I'd love to donate my code" and the glibc maintainers respond that they aren't that interested unless he can provide ready-to-apply patches. Then it comes up that the framework to do dynamic runtime checks doesn't really exist. Then there follows a terse exchange as to which mailing lists should or should not be used to discuss such matters.

It looks like the discussion ended at that point. For reference, the author (Agner Fog) is the maintainer of one of the best sites on the web for assembly optimization. This sounds like a great project for a student or young programmer who wants to bridge the gap between Fog's expertise and the FSF's beaurocracy. If you are excited by fast code, this would be a way to make a difference in a hurry.