Robert Millan <[EMAIL PROTECTED]> writes: > On Thu, Jan 24, 2008 at 10:19:55PM +0100, Yoshinori K. Okuji wrote: >> On Thursday 24 January 2008 19:34, Pavel Roskin wrote: >> > On Thu, 2008-01-24 at 20:08 +0200, Vesa Jääskeläinen wrote: >> > > Previous behavior was working correctly. You have to handle >> > > errorcodes >> > > at some point and that means when error is handled it is zeroed (or >> > > GRUB_ERR_NONE). So code is in callee where that loop was. >> > >> > I suggest that we never set grub_errno to 0 (except the initialization). >> > That would match the standard errno behavior: >> > >> > http://www.opengroup.org/onlinepubs/009695399/functions/errno.html >> >> Marco is right. As you pointed out, our error handling is different from >> errno >> on Unix, but this is intentional, because I stole the model from GRUB Legacy >> and Parted. > > Could you explain how is it supposed to work? There were clearly two bugs, > which I "fixed" in grub_disk_open first, and in grub_file_open. The solution > can be wrong, but the bugs still existed.
Hopefully my previous mail makes that clear? Were these previous problems committed? What is the bug? -- Marco _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel