e.g. "Azide" should compare less than
"Ångstrom" = {'A', 0x030a, 'n', 'g', 's', 't', 'r', 'o', 'm' },
whereas memcmp's byte-by-byte comparison on little-endian machines was
comparing the low bytes first.
Really this ought to be a pull request against NetBSD—are you already in contact with them?
e.g. "Azide" should compare less than "Ångstrom" = {'A', 0x030a, 'n', 'g', 's', 't', 'r', 'o', 'm' }, whereas memcmp's byte-by-byte comparison on little-endian machines was comparing the low bytes first.
Really this ought to be a pull request against NetBSD—are you already in contact with them?