On Tue, Apr 04, 2023 at 02:21:35PM -0700, Linus Torvalds wrote: > Fair enough. That works. I still think it should be made to be > US-ASCII only, in order to not have any strange oddities. > > Do you really have to support that "pseudo-latin1" thing? If it's > literally just a "xfs_hashprep()" function, and you arbitrarily pick > one random function, why make it be a known-broken one? Because that's the one that has been used for 15 years as no one would do this for new code?
