Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!njin!princeton!udel!mmdf
From: jnall%FSU.BITNET@cornellc.cit.cornell.edu (John Nall 904-644-5241)
Newsgroups: comp.os.minix
Subject: Still problems with Turbo C & Minix
Message-ID: <15841@louie.udel.EDU>
Date: 19 May 89 18:59:30 GMT
Sender: mmdf@udel.EDU
Lines: 27

HELP!  I'm back on this project of trying to get Minix 1.4a to compile
under Turbo C 2.0, and have about decided that the differences between
Minix 1.3 and 1.4a are not the problem.  The problem seems to be that
the disks produced with "build" under Turbo C and the disks produced
by "build" under Minix are written differently.  Comparing them, the
first 9 sectors (this is a 360K disk) are the same (i.e., sectors 0-8).
But sector 9 on the Minix disk is the same as sector 18 on the Turbo C
produced disk.  Since the part of bootblok which reads the sectors in
has not changed between Minix and the Turbo C version, there is an obvious
problem.  Sure enough, trying to boot a disk made under Turbo C bombs.  The
bootblok program reads the stuff in, since it doesn't care what is in the
sectors.  But when it jumps to fsck at the end, there is garbage there.

Does this ring a bell with anyone?  Am I overlooking anything?  Is anyone
using the Turbo C material which Deborah Mullen sent out?  If so, and if it
is working, I would be grateful for some advice.  I'm doing this on an AT,
but using 360K disks.  Have tried it on both drives (and on two different
AT's).  Stepping through bootblok on the load, it appears to know that it
has a PC disk (i.e., it uses pc sector-vector, rather than the at one).

John Nall
Computer Science Department
Florida State University
Tallahassee, Florida 32306
904-644-7043
(nall@nu.cs.fsu.edu  -- Internet/Arpa)
(jnall@rai.fsu.edu  --  Bitnet)