On Mon, Mar 17, 2025 at 7:14 AM Roberto E. Vargas Caballero <k...@shike2.net> wrote: > > Hi, > > Quoth Tavian Barnes <taviana...@tavianator.com>: > > @@ -524,7 +525,10 @@ get_perm_arg(char *argv[], union extra *extra) > > else > > p->exact = 1; > > > > - p->mode = parsemode(*argv, 0, 0); > > + mask = umask(0); > > + umask(mask); > > + > > + p->mode = parsemode(*argv, 0, mask); > > > > I don't think this is correct. The results of -perm should not depend > of the process umask. Why do you think -perm should use umask(2)?
POSIX bug 1392 clarified this: https://www.austingroupbugs.net/view.php?id=1392
