Re: stage 1

[Yoshinori K. Okuji]

On Tuesday 02 March 2004 16:58, Wolf Lammen wrote:
> But, to me, there are still questions open. IMHO code (at least
> comments) should reflect the intention of the programmer. So if we
> intend not to use floppy probing (which I understand), why do we keep
> it in the code, even call it under strange circumstances?

I don't remember why. It was some years ago. You might be able to get 
some information from the archive of this list.

> In what way is the format of stage 1 fixed, so that we cannot modify
> or, at least, extend it? In other words, what depends on stage 1 so
> strongly, that we have to leave it as it is?

If you change the offset of any variable in stage1, it changes the 
install command. This makes it impossible to use a different version of 
GRUB to install GRUB. This is not strictly required, but is quite 
convenient in some cases.

Another reason is that any change could be risky. Even if you think your 
change is subtle, it can affect stupid BIOSes. Indeed, I know a BIOS 
which becomes unbootable just by chaning data in the MBR (not the 
code!).

Current stage1 is tested well. Probably it's not perfect, but I'd like 
to keep it intact, unless you have a critical reason.

If you improve it, please do it in GRUB 2.

Okuji