В Wed, 29 Jul 2015 07:51:44 +0100 John Lane <g...@jelmail.com> пишет:
> On 29/07/15 04:08, Andrei Borzenkov wrote: > > I still believe that generally ignoring hyphens for every future crypto > > implementation is wrong. In future we simply should avoid mangling > > UUID. So this should be restricted to LUKS only, where the problem > > exists. > > > > > Andrei I modified this in response to your earlier feedback so that it > doesn't modify the user-supplied data and doesn't have any effect beyond > the three places I mentioned. That not what I mean. This patch ignores hyphens in UUID for any current and future crypto backends. This means that it cannot distinguish between 11-122 and 111-22 as UUID. And we cannot be sure we never meet such backend. And GELI as far as I can tell does not actually use hyphens. We already made mistake of mangling native UUID, I'd rather avoid continue to do it. Can we restrict this to LUKS only? > These have no negative impact on uuid > values used elesewere. If you don't want to support this fair enough - > this can be dropped without impacting the other patches. > > > _______________________________________________ > Grub-devel mailing list > Grub-devel@gnu.org > https://lists.gnu.org/mailman/listinfo/grub-devel _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel
