On Sun, Mar 23, 2014 at 03:06:08PM +0100, Denys Vlasenko wrote: > On Saturday 22 March 2014 23:46, Isaac Dunham wrote: > > On Sat, Mar 22, 2014 at 08:40:48PM +0100, Harald Becker wrote: > > > Hi Isaac ! > > > > > > Your program will fail on lines starting with the word server > > > (eg. serverxyz), that is it does not check for clear word > > > boundary and gives wrong results in that case. > > > > ...which are not legitimate entries in ntp.conf. > > > > My aim is to parse a correct ntp.conf, and not cause security problems > > on incorrect ones. > > bbox has config parsing routines to avoid coding this again and again. > > How about this? > > function old new delta > add_peers - 98 +98 > packed_usage 29470 29511 +41 > ntp_init 407 428 +21 > pw_encrypt 14 27 +13 > ------------------------------------------------------------------------------ > (add/remove: 1/0 grow/shrink: 3/0 up/down: 173/0) Total: 173 bytes >
Definitely better than mine. I'm inclined to agree with Harald as to the "skipping" messages; _if_ they are included, they should happen in verbose mode only. Any idea why pw_encrypt is changing? Now, regarding the "default n" vs "default y"/whether compat should always be/is always on... I'm aware of at least two compatability options that default to the smaller less compatible version: CONFIG_MODPROBE_SMALL is certainly not more compatible (-ablD being missing in the small version.) CONFIG_EXTRA_COMPAT is default n, and that is a good thing. All the people I've see asking for the feature use distro binaries, rather than defconfig. That said, I take no position on whether to make this "default n". Thanks, Isaac Dunham _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
