On Tue, Apr 13, 2021 at 11:37 PM Xabier Oneca -- xOneca <[email protected]> wrote: > > > + static const char *time_values ALIGN1 = > > > + /* OPT_a: */ "access\0" "atime\0" "use\0" > > > + /* OPT_m: */ "modify\0" "mtime\0"; > > > + int pos = index_in_substrings(time_values, time_arg); > > > + > > > + if (pos >= 3) { > > > + opts |= OPT_m; > > > + } else if (pos >= 0) { > > > + opts |= OPT_a; > > > + } else { > > > + //bb_error_msg("Invalid value: %s", time_arg); > > > + bb_show_usage(); > > > + } > > > > how about this instead? - > > opts |= (time_arg[0] == 'm') ? OPT_m : OPT_a; > > Your suggestion is pretty smart. But I don't know... If it were a more > popular option maybe. But I don't like it 'half baked'. > > This option being neither popular[citation needed] nor part of the > standard, if I had to vote I would reject it entirely. I'm guilty of > featuritis. Sorry. :)
My proposal would make all _valid_ uses of --time XYZ to work properly, with minimum of code. It would work incorrectly only for invalid XYZs, there are far fewer people who would care about that. _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
