On Wed, 2008-07-02 at 16:22 +0200, Robert Millan wrote: > I don't think this is necessary. First, let's take for granted that our code > is in every situation smart enough not to crash when a filesystem isn't > readable (this should always be the case, since we might occasionaly be asked > to read corrupt filesystems).
Good point. > If a flag means "access to the filesystem isn't deterministic, and grub-probe > might be able to do things that real GRUB won't", then we're in a situation in > which we'd like grub-probe to be conservative _but_ real GRUB to be > best-effort. I think this means an internal switch to tell fs probes whether > to be conservative or not. We could even use #ifdef GRUB_UTIL so the flag > checking stuff doesn't make real GRUB fatter. Another good point. We should not let users install GRUB on a filesystem that may eventually become inaccessible. Still, we probably want grub-fstest and grub-emu to be best effort to be able to debug compatibility problems. -- Regards, Pavel Roskin _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel