Am Samstag, den 29.08.2009, 01:55 +0200 schrieb Robert Millan:
> On Fri, Aug 28, 2009 at 07:58:39PM +0200, Felix Zielcke wrote:
> > +#else /* ! HAVE_REALPATH */
> > + grub_util_warn ("grub-mkrelpath might not work on your OS correctly.");
> > + /* make relative path absolute. */
> > + if (*path != '/')
> > + {
> > + len = 1024;
> > + buf2 = xmalloc (len);
> > + do
> > + {
> > + buf2 = getcwd (buf2, len);
> > + if (buf2 == NULL)
> > + {
> > + if (errno != ERANGE)
> > + grub_util_error ("can not get current working directory");
> > + else
> > + len *= 2;
> > + buf2 = xrealloc (buf2, len);
> > + }
> > + } while (buf2 == NULL);
> > + buf = xmalloc (strlen (path) + strlen (buf2) + 1);
> > + strcpy (buf, buf2);
> > + strcat (buf, "/");
> > + strcat (buf, path);
> > + }
> > + else
> > + buf = strdup (path);
> > +#endif /* ! HAVE_REALPATH */
>
> Please can you leave this part out? realpath() is POSIX, so it should be
> present in all systems we support, and if it isn't, we should be using a
> complete implementation from Gnulib instead, but we don't need to worry
> about this untill/unless someone reports it as a problem.
Ok, but should I then check in configure.ac that realpath is required or
something else?
Or should I just assume that realpath is always avaible?
> > + p = strrchr (buf, '/');
> > + if (p == NULL)
> > + grub_util_error ("FIXME: no / in buf.
> > (make_system_path_relative_to_its_root)");
>
> Does this ever happen?
As Vladimir said, it shouldn't ever happen, but I thought it would be
better to check for this explicitly instead of core dumping in that
case.
--
Felix Zielcke
Proud Debian Maintainer
_______________________________________________
Grub-devel mailing list
Grub-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/grub-devel
