On Mon, 2002-03-25 at 15:42, Yoshinori K. Okuji wrote: > At 20 Mar 2002 12:06:25 -0500, > Jeremy Katz wrote: > > There'd have to be a different version to handle unsigned longs, though, > > as you have to be able to handle a memory size of unsigned long > > (although people don't generally do mem=4294967296, it's perfectly valid > > to do so) > > More precisely, you have to be able to handle unsigned long long, > because Pentium Pro and later processors can use up to 64GB. So, if > you want to make GRUB complete theoretically, simple_strtoul is not > enough, either. > > But, as initrd cannot be put above 0x3C000000 anyway, GRUB doesn't > have to support such a large value. Instead, it would be enough to > check an overflow and set the upper limit to 0x3C000000 when it > occurs, right? > > Thus, I'm going to add a type of error which indicates an overflow and > use it rather than ERR_NUMBER_PARSING in safe_parse_maxint.
Sounds fine to me... I saw the CVS commit go by my mail, so I'll try to get a chance to build a CVS snapshot to test a bit in the next day or so. Thanks, Jeremy _______________________________________________ Bug-grub mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-grub
