"Ashutosh S. Rajekar" wrote: (shortend)
> I am working on the change in the GRUB partition syntax, to
> correspond with the BSD-style slice nomenclature. I have only changed the
> syntax, i.e.
> (hd1,0) is entered as /dev/hd1s1,
Shouldn't it be hd1s0?
> (hd0,2,a) is entered as /dev/hd0s2a,
> (hd0,0) is entered as /dev/hd0s0.
> part. 1 |/dev/hd0s1 | (hd0,0) | /dev/hd0s0
> part. 3, BSD 'a' slice |/dev/hd0s3a| (hd0,2,a) | /dev/hd0s2a
> part. 1 |/dev/hd2s1 | (hd1,0) | /dev/hd1s0
> part. 4 |/dev/hd2s4 | (hd1,3) | /dev/hd1s3
> All the 'should be' items are due to the fact that there is no checking
> for CD-ROM drives connected as IDE drives. If we were to add that, then
> the conversion to the BSD-slice syntax would be complete. The change in the
> partition names is very easy, we just have to add 1 to the
> current_partition, so I haven't included it as yet. But 'current_drive' is
> problematic.
> All root partitions have to be given with a preceeding '/dev/'. All
> kernels are specified with a kernel=/filename, without the preceeding
> "/dev/....." argument, this simplifies the code to a great extent.
Hi,
why change that at all? I considered the GRUB style the most rational. I
always wondered, why hds are 0-indexed and partitions 0-indexed.
I could follow the argument, that the user should not have to learn a
new scheme, but then you should implement the bsd-scheme complemte, this
means 1-indexing the partition names (if I got you right, this is you
intention).
I don't know, how *BSD behaves in this case, but you should make sure,
that devices keep their names, even if I add a new one (and don't change
the old one, of course). I recently lost 2 GB of data because of such a
thing.
Ben
