Re: Getting rid of serverboot. Finally.

[Roland McGrath]

> I have confirmed that the following combinations, first with your code
> and then, with my new patches, work:

Thanks!

> > [subhurd] new hurd's boot -> old boot script -> old hurd
> 
> Check.  However, I fail to see why we even want to be compatible here;

My rationale was that one might want to have an old filesystem that one
doesn't touch at all, and use its boot script and everything else with a
new boot.  Obviously, these oddball compatibility cases are not really a
priority (they will probably never happen, as you note).

> First, GNU Mach.  There are two errors in determining the
> compatibility case: first, we need to also ignore trailing newlines;

Really?  It seems to me like it must be a bug in GRUB if there are newlines
in the multiboot module strings.  But I put this change in since it can't
cause any harm.

> second, in the case where there are multiple modules, the logic was
> backwards.  

Oops.  I fixed this with my own change slightly different from yours.

> This patch also corrects the output: I have no clue what the padding with
> spaces was all about, however, it seems to me to be completely
> superfluous; the output is now consistent with other messages.

Was there a problem with the output, or is this an aesthetic change?  The
reason for the space-filling is that the messages are printed with CR sans
LF, intended for each message to overwrite the last.