The World of Computer Software/The World of Computer Software.iso/lnxd3-79.zip/DIGEST3.790

Internet Message Format | 1993-01-01 | 25KB

From [e-mail] Mon Dec 14 11:54:18 1992
Return-Path: <[e-mail]>
Received: from bloom-picayune (BLOOM-PICAYUNE.MIT.EDU) by montefiore.ulg.ac.be (4.1/SMI-4.1)
    id AA00732; Mon, 14 Dec 92 11:47:21 +0100
Received:  by bloom-picayune (5.57/25-eef)
    id AA09701; Mon, 14 Dec 92 02:30:37 -0500
Message-Id: <9212140730.AA09701@bloom-picayune>
From: Digestifier <[e-mail]>
To: [e-mail]
Reply-To: [e-mail]
Date:     Mon, 14 Dec 92 02:30:16 EST
Subject:  Linux-Activists Digest #790
Status: R

Linux-Activists Digest #790, Volume #3           Mon, 14 Dec 92 02:30:16 EST

Contents:
  Re: xdm and gwm (Gero Kuhlmann)
  why in the world is doc so damm slow? (Supernaw-Issen, Daniel A.)
  [BUG REPORT] swapping trouble with 0.98-pl6 ?!? (Chris Newbold)
  rz / sz cute version? (Ammo Goettsch)
  Re: PCNFSD for Linux (Ari Lemmke)
  TCP and TIOC questions (Phillip Rzewski)
  UFC-Crypt (Peter Villadsen)
  Re: [Q] How to change from Linux/minix file system to extfs ? [CONFIGS] (Joshua M Yelon)
  roaches and eyes (derrick D.goss)
  Problems with LILO 0.7 and Linux 0.99 (David W. Summers)
  Re: Can X run with only 4 megs RAM? (Spaggie)
  Re: [ANNOUNCE] linux version 0.99 is out (Chris Newbold)
  Re: Not quite bug in math stuff ([e-mail])
  Upgrade to PL6 process? (Tim Foley)
  Making a Terminal Server? (Tim Foley)

----------------------------------------------------------------------------

From: [e-mail] (Gero Kuhlmann)
Subject: Re: xdm and gwm
Date: Sun, 13 Dec 92 17:25:27 GMT

[e-mail] (John W. Linville) writes:
> 1)  When using xdm, is there anything special that I need to do to allow xdm
>     to log-in users?  I have removed the tty's from inittab and I put xdm at
>     the end of my /etc/rc file.  The login screen comes-up fine.  The problem
>     is that anytime I try to login, the login/password box disappears, only
>     to reappear without ever starting a user session.  I have made ssessreg
>     a setuid program.  Does anyone know how to make this work.  I am using
>     the latest SLS.

I had the same problem. It is that rc is started by init with /dev/tty1 setup
as stdin/stdout. But xdm returns immediately after starting because it is a
daemon putting itself into the background. That means that the rc script
returns almost immediately after calling xdm and init closes /dev/tty1. So
when xdm starts xterm that doesn't find a valid console (just do it and look
at the 'ps -ax' output: you will see that there is no console attached to
xdm and therefore couldn't be attached to xterm either). And it's not xdm
which fails but xterm which couldn't find a console. And because xterm is
called as the last process in Xsession, it returns an error to xdm which then
shows up the login window again.
But from where to start xdm then? When doing it from a normal login, xdm
tries to access the console the same way as the underlaying bash is doing.
That makes it almost impossible to enter anything at the login window. But
removing the tty from inittab isn't a solution either because then there is
no way to start xdm. To solve this problem I wrote an alias of the following
form:
        alias xdm='xdm; exec </dev/tty8; exec >/dev/tty8'

Now when starting xdm, it returns immediately and the shell attaches the
stdin and stdout to /dev/tty8 which is normally not used in the inittab.
So xdm is attached to the current controlling console and the shell is
redirected to /dev/tty8. That also means that if xdm closes the server
due to some error, you will find your previous vc with <ALT+F8>!

gero.

--
Gero Kuhlmann                   Zerberus:  [e-mail]
Donarweg 4                      SubNet:    [e-mail]
D-3000 Hannover 51              IN:        [e-mail]

------------------------------

From: [e-mail] (Supernaw-Issen, Daniel A.)
Subject: why in the world is doc so damm slow?
Date: 13 Dec 1992 23:16 CST

Got a question,
I have the SLS .98pl5 distribution with the X stuff and tried my hand at doc.
I found that it was so slow as to be completely useless.  Why?  I have 8M ram
and 16M swap.  I've run it with *NOTHING* else going, and it still is too slow
as to be useless.  any ideas?  does this beastie require a math-co?  I have a
386-33.

Daniel

signature for hire.

------------------------------

From: [e-mail] (Chris Newbold)
Subject: [BUG REPORT] swapping trouble with 0.98-pl6 ?!?
Date: 14 Dec 92 05:39:57 GMT

I have been having trouble with 0.98-pl5 and 0.98-pl6 when
running under X. The machine will suddenly lock when I try
and run commands in an xterm window. This happens 1-4
times an hour.

I can reproduce the problem almost consistently by trying
to copies of ghostview simultainously. Other commands that
frequently cause the system to hang are: tar, groff and
make/gcc.

At first I thought that this might be a problem with the
X386-S3 server. However, some experimentation proved this
theory invalid. I have this trouble under pl5 or pl6, with
X386-S3 version 0.2b or 0.3 or the X386mono server.

I stress-tested the system outside of X to see if I could
reproduce the problem, but had no luck. Though I pushed the
load average over 3 with extreme disk activity, nothing
broke. However, I was unable to use up much of the swap
space--which is what I think kills everything.

Is anybody else having troubles like this? This is perhaps
the worst bug I have disovered since the old SCSI code
would trash my disk daily...

My system:

486/33DX, EISA bus
8MB ram
AHA1542 SCSI disks

8MB swap parition

-Chris

--
>>>> Chris Newbold <<<< * "If you fool around with a thing for very long you *
University of Rochester *                 will screw it up."                 *
Disclaimer: "All warranties expire upon payment of invoice."
[e-mail] * uhura.cc.rochester.edu!ctne_ltd@uunet

------------------------------

From: [e-mail] (Ammo Goettsch)
Subject: rz / sz cute version?
Reply-To: [e-mail] (Ammo Goettsch)
Date: Mon, 14 Dec 1992 06:07:46 GMT

does anyone have a "cute" version of rz, that for example uses curses or
X to display decent download status messages, rather than the standard
one-line-per-packet messages?

Ammo Goettsch

------------------------------

From: [e-mail] (Ari Lemmke)
Subject: Re: PCNFSD for Linux
Date: Mon, 14 Dec 1992 03:46:59 GMT

In article <[e-mail]> [e-mail] (Jonathan Magid) writes:
:   In article <[e-mail]> [e-mail] (Ed Boston)  writes:
:   >Does anyone know if there is a PCNFSD available for Linux?  I have been
:   >watching some of the discussions of Linux and would like to set it up,
:   >but I would need this ability for my needs.

        If someone writes pcnfsd ;-)

:   I'll assume PCNFS is some sort of MS-DOS version of NFS...

        Yes and more .. pcnfsd handles also printing (aaargh).

:   correctly, you need to be able to use a Linux box as a NFS file server?
:   Well there is a ALPHA-BETA nfsd for linux... It seems to be pretty stable
:   and support for it is included in the SLS distribution.

        No .. I'm not sure if it will work with pcnfs, I doubt.
        Even though I have pcnfs kit (something old 2.x), I haven't
        checked does it have source code also - they say it has
        sources for pcnfsd. Then it would be only porting job,
        I'm not sure if one is allowed to distribute the code ...
        but if someone ports it to Linux, then patches could
        be distributed.

:   jem.

        arl

------------------------------

From: [e-mail] (Phillip Rzewski)
Subject: TCP and TIOC questions
Date: Mon, 14 Dec 1992 06:34:35 GMT

    I'm starting to do some fun porting, and since I haven't done too much
porting on Linux before (at least, not hard stuff) I have decided that as I run
into problems I'm going to post to c.o.l about them. I figure it's either a bug
or something that people should know about anyway. :)

    First of all. My stuff tries to use the TCP business a little. It
complained about the lack of sin_zero in /usr/include/netinet/in.h. The only
other thing that looks like it would fit instead is the __pad thing, but I was
looking at include files on other systems and they all seem to have sin_zero.
So why is Linux trying to be nonstandard here?

    Second of all. I see this kind of thing posted about a lot, but I guess we
all have to have our turn at this. My code is looking for things like TIOCGETP
and TIOCSETN. They don't appear to be declared anywhere, though linux/termios.h
certainly has a lot of defines with names that come pretty close. Are there
some equivalents from that file which I should be using in place of the
TIOCGETP and TIOCSETN? One thing that made me suspicious about whether or not
this "right" was the fact that curses.h makes a reference to TIOCGETP, which is
certainly out of place on a system that doesn't define that in any other
include file.

    Anyone wanna field these?

--
  Phillip Andrew Rzewski                      Internet: [e-mail]
             "Don't wait; while you're waiting, you're dying."
                          --- Peter Hammill

------------------------------

From: [e-mail] (Peter Villadsen)
Subject: UFC-Crypt
Date: Mon, 14 Dec 1992 05:43:09 GMT

   I recently ran the small test suite included with the ufc-crypt
package and found that it ran slower than the native crypt. For
my 386/25 (no math coprocessor), this was:
   crypt:  312.7 crypts/sec
ufccrypt:  281.9 crypt/sec

   I am wondering why this is the case? Under the Sun and RS6000,
the UFC-crypt ran much faster. I'm using 0.98pl4, gcc2.2.2d on a
386/25 with no math coprocessor. Thanks,


   Peter


------------------------------

From: [e-mail] (Joshua M Yelon)
Subject: Re: [Q] How to change from Linux/minix file system to extfs ? [CONFIGS]
Date: Mon, 14 Dec 1992 06:44:37 GMT

>> I have linux 0.98pl4, Xfree v1.1, gcc 2.2.2.7d installed on what is the
>> Linux/Minix file system. I have
>>
>>  /dev/hda2  42 Mb Linux/Minix partition
>>  /dev/hda3   7 Mb Linux Swap disk
>>
>> [Q] How do I upgrade the complete system to the Linux extfs ?

To be honest, I don't think this is a good idea just yet.  It _would_
get you long filenames.  It would also reduce your performance - the
extfs is fairly slow.  I switched over to extfs, and then was sorry
I did - I switched back.  There is a new, faster extfs in the works,
and if I were you, I'd wait for that.

But, if you must switch, here's the trick I used:  Put a tarred, zipped
copy of your hard disk on the swap partition!  Zip is usually capable
of reducing by a factor of three, so there _is_ a chance, if your hard
disk isn't full.

Incidentally, this will take up a lot of disk space unless you pipe
the tarfile directly into zip, which is possible.  Indeed, you can pipe
the resulting zipfile directly into wc, to see whether the zipfile will
fit on the swap partition, before you do anything serious at all.

Of course, don't forget to put 'unzip' on your install disks!

- Josh

------------------------------

Subject: roaches and eyes
From: [e-mail] (derrick D.goss)
Date: Sun, 13 Dec 92 11:58:03 PST

after upgrading to patch level 3 my Xeyes and Xroaches stopped working.
does anyone know the fix????

post replys oe e-mail
email to [e-mail]

------------------------------

From: [e-mail] (David W. Summers)
Subject: Problems with LILO 0.7 and Linux 0.99
Date: Mon, 14 Dec 1992 06:25:24 GMT

Hello fellow Linuxers (Linuxians? :-), I hope this isn't too long....

   Well, I just saw that 0.99 made it out so I decided to try to upgrade
from 0.98.5 to 0.98.6 and then to 0.99.

Here's the setup:
   HARDWARE:
   ---------
   486-50 AMI BIOS
   EISA Motherboard
   8 Meg Ram
   520 Meg Fujitsu 2426F (SCSI)
   AHA 1542B Controller

   Software:
   ---------
   Linux 0.98.6 and/or Linux 0.99
   LILO 0.7

   /usr/src/linux/.config:
========== BEGIN =============
#
# Automatically generated make config: don't edit
#

#
# General setup
#
CONFIG_BLK_DEV_HD = CONFIG_BLK_DEV_HD
CONFIG_TCPIP = CONFIG_TCPIP
CONFIG_MAX_16M = CONFIG_MAX_16M

#
# SCSI support
#
CONFIG_SCSI = CONFIG_SCSI

#
# SCSI support type (disk, tape, CDrom)
#
CONFIG_BLK_DEV_SD = CONFIG_BLK_DEV_SD
CONFIG_BLK_DEV_ST = CONFIG_BLK_DEV_ST
CONFIG_BLK_DEV_SR = CONFIG_BLK_DEV_SR

#
# SCSI low-level drivers
#
CONFIG_SCSI_AHA1542 = CONFIG_SCSI_AHA1542

#
# Filesystems
#
CONFIG_MINIX_FS = CONFIG_MINIX_FS
CONFIG_EXT_FS = CONFIG_EXT_FS
CONFIG_MSDOS_FS = CONFIG_MSDOS_FS
CONFIG_PROC_FS = CONFIG_PROC_FS
CONFIG_NFS_FS = CONFIG_NFS_FS
CONFIG_ISO9660_FS = CONFIG_ISO9660_FS

#
# Various character device drivers..
#
======== END =========================

   Disk partitions:
========= BEGIN =============
# /etc/lilo/disktab  -  LILO disk parameter table
#
# This table contains disk parameters for SCSI disks and non-
# standard parameters of IDE disks. Parameters in disktab
# _always_ override auto-detected disk parameters.

# Dev.  BIOS    Secs/   Heads/  Cylin-  Part.
# num.  code    track   cylin.  ders    offset

0x800   0x80    32      64      495     0       # /dev/sda
0x801   0x80    32      64      495     2048    # /dev/sda1 (OS/2 Boot Manager)
0x802   0x80    32      64      495     32      # /dev/sda2 (DOS-OS/2)
0x803   0x80    32      64      495     239616  # /dev/sda3 (Extended Partition)
0x804   0x80    32      64      495     206848  # /dev/sda4 (Linux root (EXTFS))
0x805   0x80    32      64      495     239648  # /dev/sda5 (Logical PARTITION)
0x806   0x80    32      64      495     272416  # /dev/sda6 (Logical PARTITION)
0x807   0x80    32      64      495     337952  # /dev/sda7 (Logical PARTITION)
============= END ===================

Here's the saga:
   1.) I was using LILO 0.4 with Linux 0.98.5 and everything was groovy.

   2.) I tried to upgrade to LILO 0.7 but it stops at the "LIL".  I (re)-read
       the documentation and it says that it is probably caused by a mis-matched
       /etc/lilo/disktab file but I haven't changed my disk partitioning!

   3.) I tried to boot 0.99 Image from floppy and the kernel panics with:

[previous boot stuff deleted, unless you think it is important....]
Partition table ok.
MINIX-fs magic match failed.
general protection: 0000
EIP: 0008:0000B6C4
EFLAGS: 00010202
fs: 0017
base: C00000000, Limit: 000A0000
Pid: 1, process nr: 1
86 13 85 d2 74 27 83 3a 02 74

:-(  Well, I tried looking in System.map on 0.99 and narrow it down and
I *THINK* I've narrowed it down to the 'wake_up' routine.

Could someone tell me how to narrow it down further?  Do I need to spend some
time putting 'printk' calls in 'wake_up' to try to figure out what is going on?


GOOD NEWS:  The 0.98.6 SCSI code really flies!  I'm impressed!
Also the TCP/IP code is fabulous!


Has anyone else expirienced this problem?  Am I doing something stupid or
silly?

I have read and re-read the LILO docs several times but can't figure out
what I'm doing wrong...it starts to boot.


Thanks!
   - David Summers
   ([e-mail])
--
                         "Never under-estimate the bandwidth of a station-wagon
David Summers             full of tapes, hurtling down the highway."
[e-mail]         - Tanenbaum, "Computer Networks"

------------------------------

From: [e-mail] (Spaggie)
Subject: Re: Can X run with only 4 megs RAM?
Date: Mon, 14 Dec 92 06:47:35 GMT

In article <[e-mail]> [e-mail] (W. Woody Jin) writes:
>I have been using (Xwindow + Xview3) with 0.98.2, 98.3, 98.4, 98.5
>and now 98.6, without any problem  with only 4M bytes without math processor.
>I use X stuff from SLS distribution.
>When I start X, I open 3 xterms, oclock, xeyes with olwm.
>I found that Linux uses about 2M for swap in this case.
>Xdvi, ghostview/script work just fine.
>Only problem is that compiling is so slow and you need some patience when
>moving to other window (5 sec or slightly more).
>In this case, I found that Linux uses 8M for swap (twice more than my memory).

        Strange. I have observed a problem with running X under 98.4, and 98.5,
similiar to the ones describe by the original poster.  When I changed back to
98.0, things worked again.   Seems the X server just hangs.  My system sounds
similair to yours, except I have a math co-proc. (386-40,AMD BIOS,ETEQ chipset,
4MB mem, 10MB swap, Cyrix 387-40).  Has anyone else had trouble with X under
98.x where x>1?  (by the way, it is unrelated to the video card, as I tried
3 different cards, each with the same results.

Any ideas?

Jeff

=================================================================
Jeff Sponaugle                      Purdue University
246 S. Salisbury #3                 Engineering Computer Network
West Lafayette, IN 47906            MSEE 104
[phone]                      [phone]
Internet: [e-mail]        MSEE SB50:494-6106
=================================================================



------------------------------

From: [e-mail] (Chris Newbold)
Subject: Re: [ANNOUNCE] linux version 0.99 is out
Date: 14 Dec 92 06:55:52 GMT


> In <[e-mail]> [e-mail] (Linus Torvalds) writes:
>
> >Linux version 0.99 is now available at nic.funet.fi, in the directory
> >pub/OS/Linux/PEOPLE/Linus as both full source and patches against
> >0.98.6.  It will probably show up on the other major sites soon.
>
> I download the full source from nic.funet.fi, and ran configure as follows:
>
> no kernel math
> no profiling
> no std. harddisk
>
> yes tcp/ip
> yes SCSI (disk, tape & cdrom)
>
> aha1542 driver
>
> yes minix, ext, msdos & iso9660
> no proc
>
> yes serial autoconfig
> no funny serial hardware
>
> ...and then did a make dep, make Image. It compiled w/o incident.
>
> When I booted the kernel (off a floppy, not LILO) all went well until just
> after the disk partition information was read. After getting "partition
> tables ok" the kernel dies with a NULL reference.
>
> the EIP is 8:b754, which according to the map file is inside wake_up in
> sched.c.
>
> My hardware:
>
> 486/33 DX, EISA bus
> 8MB RAM
> AHA1542-compatible SCSI host (2 disks, 1 tape, 1 cdrom attached)
>
> I had no trouble with 0.98-pl6. Any ideas what's gone wrong?
>
> -Chris


--
>>>> Chris Newbold <<<< * "If you fool around with a thing for very long you *
University of Rochester *                 will screw it up."                 *
Disclaimer: "All warranties expire upon payment of invoice."
[e-mail] * uhura.cc.rochester.edu!ctne_ltd@uunet

------------------------------

From: [e-mail]
Subject: Re: Not quite bug in math stuff
Date: Mon, 14 Dec 1992 06:31:57 GMT

In article <[e-mail]>, [e-mail] (Phillip Rzewski) writes:
>
>     I don't do much stuff with math under Linux, so I never ran into this
> before. I'm sure someone else has come across this, it's probably quite
> documented somewhere. But it took me rather by surprise.
>
>     Consider the program:
>
> #include <stdio.h>
> #include <math.h>
>
> int main(void)
> {
>   double d;
>   long l;
>
>   d = pow(2.0, 1.0);
>   printf("%1.32f\n", d);
>   l = (long) d;
>   printf("%d\n", l);
>
>   return 0;
> }
>
>     And of course the output:
>
> 1.99999999999999977795539507496869
> 1
>
>     Now, I understand how keen it is that math software speeds things up
> by using logarithms, giving us results like this. However, I know that there
> are systems out there where one doesn't notice it quite so much. The
> program I was originally working with on Linux (not like the one above :) )
> was originally something I wrote under VAX C on a VMS system, and I got
> just what I thought in the long (in other words, 2).
>
>     Of course, I was able to fix this by simply adding 0.5 to the double
> and then casting it to long. It just seems kinda icky to have to do it that
> way. I mean, this means that every person who tries to do something like I
> did above (which I would think is quite a few) would have to have this same
> weird thing happen to them and learn how to correct it.
>
>     I'm not one to make demands or anything, but the math people might
> consider fixing this?

What version of the kernel and libraries are you using?
Using the FPU emulator with 0.98.6 (I don't have a co-processor), I get
2.00000000000000000000000 [etc]
with your program.

Perhaps you are using the obsolete 'soft' math libs.

Your question does raise a valid point though. It would certainly be
nice if the most accurate answer were always given for floating point
problems. Alas, this is simply not possible. The use of finite precision
arithmetic should always be accompanied by an appropriate degree of care.
Linux is often not bad for simple problems because it runs the co-processor
at ~64 bit precision (doubles are ~53 bit precision under Linux gcc).

However, even simple programs can give results which are worse than you
might expect. Consider the following little program:

#include <stdio.h>
void main(void)
{ double d;
  d = 1.0; d = d/3.0;
  printf("%g\n", 1.0-d*3.0); }

This will give a non-zero result (compile with no optimizations).

But if you try:

#include <stdio.h>
void main(void)
{ double d, e;
  d = 1.0; e = 3.0;
  printf("%g\n", 1.0-d/3.0*e); }

Then because the intermediate result 0.333333... is kept on the co-
processor stack in this case, a different result is obtained (it
happens to be zero).

--
===============
Bill Metzenthen
Mathematics Department
Monash University
Australia

------------------------------

From: tfoley@camaro (Tim Foley)
Subject: Upgrade to PL6 process?
Date: Sun, 13 Dec 1992 18:43:41 GMT


  Someone mentioned that UUCP will break with PL6 and than PS broke with
PL4....whats the story?

  I'm still running PL1, and would like to get tcp working, so I'm ready to
take the PL6 plunge...but what things need patching?

                        1) I have the minicom PL6 Patch
                        2) Does uucp need a patch?
                        3) Does ps need a patch?
                        4) Does anything else need....

Please reply in the group as quite a few others have been asking the same
question.




--
==========================================================================
  The HeartBeat of America...Yesterdays Camaro Z28
  Call the Camaro Linux Pub-access site: 1-416-238-6550 USRobotics HST
  Note: Please, no ftpmail or mailing lists or the host gets annoyed :)

------------------------------

From: tfoley@camaro (Tim Foley)
Subject: Making a Terminal Server?
Date: Sun, 13 Dec 1992 18:54:51 GMT


  I have noticed that people are using other systems for terminal servers
with other unix systems.  Any idea how I can do this with Linux?

  I have a nice huge Motorola SYSVR3 68K machine sitting behind me, with
about 20 serial ports, but its just sitting there....doing nothing....Linux
has too many more toys...but only two com ports....

  How can I turn this monster into a terminal server for my linux box?

I take it some sort of tcp network with rlogin or something...could someone
please elaborate?

                                                TIm




--
==========================================================================
  The HeartBeat of America...Yesterdays Camaro Z28
  Call the Camaro Linux Pub-access site: 1-416-238-6550 USRobotics HST
  Note: Please, no ftpmail or mailing lists or the host gets annoyed :)

------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: [e-mail]

You can send mail to the entire list (and comp.os.linux) via:

    Internet: [e-mail]

Linux may be obtained via one of these FTP sites:
    nic.funet.fi                pub/OS/Linux
    tsx-11.mit.edu                pub/linux
    tupac-amaru.informatik.rwth-aachen.de    pub/msdos/replace

The current version of Linux is 0.98pl6 released on December 2, 1992

End of Linux-Activists Digest
******************************