On Wed, May 01, 2024 at 02:31:47PM -0700, Darrick J. Wong wrote: > Hmm. @objchange==true in the _hard() case means that we already > converted the sf dir to ino8 format, right? So that's why we don't need > to increment i8count? Yes. It is a little weird to be honest and given the final state after the series I wonder if i8count manipulation should just move outside the helper.
