Path: utzoo!attcan!uunet!snorkelwacker!usc!wuarchive!udel!mmdf
From: XBR4D75G%DDATHD21.BITNET@cunyvm.cuny.edu
Newsgroups: comp.os.minix
Subject: Problems with version 1.5.3
Message-ID: <11869@nigel.udel.EDU>
Date: 22 Feb 90 15:49:43 GMT
Sender: mmdf@udel.EDU
Lines: 35

I am running MINIX 1.5.0 on an XT- and an AT-clone since 20-JAN-90
without major problems. After the official release end december last year,
I only applied 4 fixes sent to the net, related to the routines

         peekpoke.c  portio.c  signal.c   getpwent.c

My crcs do not match with the official ones and I had to apply the
patches with the -l option because I am following the MINIX developement
for more than 2 years from the BITNET-side. Nevertheless I was conviced
until yesterday, that my system, running on two different PC's, is very close
to the official release (only the new kermit version had to be cross-compiled
because asld got out of memory).

When I applied the new patches to generate version 1.5.3, I ran into the
following problems:
 a) patches to directories include, minix, sys, tools, fs, mm and lib/....
    were applied without problems
 b) compilation and generation of the new library without problems
 c) compiling and linking of kernel mm fs and tools without problems
 d) when I started the system, loading the root file system from floppy,
    the commands in /etc/rc were executed, but init could not start
    login!!!!! Changing execle(LOGIN......) in init.c to the old
    library-call execn(LOGIN1) ; execn(LOGIN2); generated a running
    version of 1.5.3, but I think this is not the intesion of the
    authors. After this hack, I was able to mount and use my hard-disk
    partition. Nevertheless this intermediate version is not able to
    load the root-file system directly from hard-disk (illegal inodes
    detected by th FS-system).

Has anyone similar problem with 1.5.3 and perhaps some solutions to
generate a fairly stable system?

Thanks in advance

    Hans-Juergen Knobloch