[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian Installer GRUB test image available


i wrote:
> > If a file CD1/System/Library/CoreServices/BootX had existed on hard disk,
> > the pathspec "CD1" would have overwritten the previously inserted IsoNode
> > object which held the instruction to copy from "grub.chrp".

John Paul Adrian Glaubitz wrote:
> But there is no file CD1/System/Library/CoreServices/BootX on the harddisk,
> the whole folder structure below /System is solely created for GRUB

Then it's ok.

> >  xorriso -as mkisofs ...no.-graft-points.and.no.long.pathspecs... CD1

> Why no -graft-points and -long-pathspecs?

In the normal course of ISO definition by mkisofs or xorriso one would
first give the most general pathspec (here "CD1") and then those which
make detail adjustments to its realm inside the ISO:

    -graft-points \
    CD1 \
t/grub/powerpc-ieee1275/grub.chrp \
boot/grub/powerpc.elf \

"CD1" gets mapped to ISO's "/". Then grub.chrp and powerpc.elf get
mapped to their paths, regardless what "CD1" might have put there.

I understand that you cannot achieve this in debian-cd because the
argument "CD1" stems from the generic part and all system specific
arguments get inserted before it.

So if it was important to enforce the precedence of grub.chrp and
powerpc.elf, regardless of what "CD1" contains on hard disk, it would be
necessary to put them into "CD1" already on hard disk.

No force needed, though. So it's ok.

> > Both ISOs show no "machine code" after </CHRP-BOOT> and no "xcoff" in the
> > <BOOT-SCRIPT> part. Maybe that's what is missing in comparison to real
> > life examples.

> Why shouldn't the ISO images created by GRUB and the old Debian powerpc
> stable releases not be any real world examples?

I meant the worlds of Apple, not ours.

Have a nice day :)


Reply to: