Nick Landsberg wrote:

As Eric [below] describes there are some times 'hidden' problems

Eric wrote:-
Yes ! Like I couldn't get a 'Make' to run because the editor was replacing
the compulsory <tab> token by multiple spaces.
What you see is NOT always what you get.
I couldn't SEE anything wrong with the bash-script which reported errors.
Next time I'll remember to examine the 'hex' of the text - easy with
mc, which I always use. For me linux is intolerable [compared to my
'proper' OS: native-oberon] without mc, cut-n-paste &
command-repeat !
---------
But to get back to the original thead:
the OP was telling how by using the bash-script to step the boundry(s)
of a partition [who's entry in the first 512 byte MBR had been lost] one
could recover lost PRIMARY partitions.

I can confirm that I did this manually - perhaps just lucky ?
He uses 'stepping CYLINDERS' which is the granularity available
from fdisk.

Q1. what are smaller: sectors or cylinders ?
Q2. Is it possible that by some other means of partition eg.
fdisk select Mb or Kb size instead of 'cylinder boundry' OF
the first FAT partition made by DOS:fdisk
that an 'inter-cylinder' partion boundry exists.
Ie. will stepping cylinders while testing guarantee to find the
partition boundry ?

== Chris Glur.