Florian Weimer <fwei...@redhat.com> writes: > I thought the bug is that it's not actually a hint? That DT_UNKNOWN > here leads to incorrect results in glob?
All the DT_* are hints, according to the man pages, which say that DT_UNKNOWN must be handled by the caller. So I guess it depends on whether the original developers knew this and weren't worried about accuracy, or if there were no filesystems that ever returned DT_UNKNOWN for regular files, so it was decided to always include "I don't know"'s as a "future proofing" option. Doesn't matter to me either way, of course.