Linux-Misc Digest #140, Volume #27 Sun, 18 Feb 01 07:13:02 EST
Contents:
Installing Problem (Swatantra Bhargava)
Re: Where can I find rp-pppoe-2.4 ? (Shaun)
Re: Linux & Windows Shared Internet (Robert E. Meuse)
keymap or font mismatch (Mike Mcclain)
System backup, repartitio (Mike Mcclain)
Re: keymap or font mismatch (Kenneth Mokkelbost)
Re: ksh script problem: pwd works differently for ksh then linux binary file (Shai
Kedem)
Trying to install RH 7.0 on an HP (Anonym5530)
Re: su command while in a program? (Michael Heiming)
Re: I need software recommendations: (Jean-David Beyer)
Re: troubles installing gdbm (Robert Schweikert)
Re: Size of LINUX (Jean-David Beyer)
Re: Size of LINUX (Jean-David Beyer)
Re: Size of LINUX (Jean-David Beyer)
Re: Size of LINUX (Jean-David Beyer)
----------------------------------------------------------------------------
From: Swatantra Bhargava <[EMAIL PROTECTED]>
Subject: Installing Problem
Date: Sun, 18 Feb 2001 07:30:03 -0000
hi,
i want to install windows98/95 and linux (red had) on same hard disk.so
please tell me full steps for doing this(including partition how i can
make).
thanks
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: Shaun <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: Where can I find rp-pppoe-2.4 ?
Date: Sun, 18 Feb 2001 07:35:13 GMT
In article <[EMAIL PROTECTED]>, Carfield Yim
<[EMAIL PROTECTED]> wrote:
> After I upgrade to rp-pppoe-2.8.1, I can browse any web page, and I can
> successful connect even if I put the wrong password to the system. I
> would like to rollback to use older version, can anybody give me?
>
http://www.roaringpenguin.com/pppoe/ppp-2.4.0-pppoe4.tgz
(Visit http://www.roaringpenguin.com/pppoe/ if that link does not work
for some reason.) Be sure to let Roaring Penguin know if there is a
vulnerability in the newer version.
Good luck,
Shaun
--
[EMAIL PROTECTED] (remove mypants to email me) | http://www.shaunc.com/
Operations and Support, Host with the Most | http://www.hostwiththemost.net/
------------------------------
From: Robert E. Meuse <[EMAIL PROTECTED]>
Subject: Re: Linux & Windows Shared Internet
Date: Sun, 18 Feb 2001 08:54:12 +0100
Reply-To: [EMAIL PROTECTED]
On 18 Feb 2001 03:12:59 GMT, [EMAIL PROTECTED] (Dances With Crows)
wrote:
>On Sun, 18 Feb 2001 03:10:36 +0100, Oliver Battenfeld staggered into the
>Black Sun and said:
>>[EMAIL PROTECTED] schrieb:
>>> Hello. I am in dire need of help. I must know how to get a
>>> computer running Windows Me to share it's internet connection
>>> with RedHat Linux (6.0) Any (positive) input is apreciated.
>>
>>Win ME comes with a simple NAT software called "Internet Connection
>>Sharing", which can be installed via Control Panel/Software (if not
>>already).
>>
>>Check out this website for some more instructions:
>>http://www.annoyances.org/win98/features/ics.html
>>
>>If that doesn't need your requirements, you will have to spend some
>>money on software like NAT 32 or Winroute.
>>
>>The setup of your Linux machine is covered by numerous man pages
>>(ifconfig, route, etc.).
>
>IIRC, ICS works by emulating a DHCP server. Try a "pump -i eth0" on
>your RH 6.0 box. No need to reboot. See the HOWTOs at
>http://linuxdoc.org/HOWTO/ and http://linuxnewbie.org/ .
>
>This is really an arse-backwards way of doing things, the Linux machine
>should be the one connected to the Net. It can do firewall/NAT for
>Lose9x clients behind it (read the IP-Masqing HOWTO) and most likely do
>a better job.
Hereis a short script for linux to "share the internet with clients"
I have this script in my startup. (rc.local)
The Linux box has 2 network cards in it and the one that looks at the
internal network has an ip address of 10.250.5.10
IPchains must be installed to run this script.
All the network clients that want to use the internet use 10.250.5.10
as their gateway address.
Every computer that uses the server needs a seperate ip address
computer1->10.250.5.20
computer2->10.250.5.30
..
.
The rest of the client services should be configured normally as if
they where hooked directly to the internet.
#!/bin/sh
#
# This script will be executed *after* all the other init scripts.
# You can put your own initialization stuff in here if you don't
# want to do the full Sys V style init stuff.
echo "1">/proc/sys/net/ipv4/ip_forward
echo "1">/proc/sys/net/ipv4/tcp_syncookies
/sbin/ipchains -P forward DENY
/sbin/ipchains -A forward -s 10.250.5.10/24 -j MASQ
------------------------------
From: [EMAIL PROTECTED] (Mike Mcclain)
Subject: keymap or font mismatch
Date: Sun, 18 Feb 2001 08:03:52 GMT
Howdy All,
Every once in a while I see a message here where there
is a mismatch between the system of the guy that wrote the
message and mine where the message is displayed, like so:
All today{s OS, I{m sure. This is what he sent.
All today's OS, I'm sure. This is what he meant.
My question: Where is the misconfiguration?
His system or mine? What's not set right?
TIA,
MiKe
--- MultiMail/Linux v0.31
------------------------------
From: [EMAIL PROTECTED] (Mike Mcclain)
Subject: System backup, repartitio
Date: Sun, 18 Feb 2001 08:03:53 GMT
Howdy Barry,
Your approach will work, but don't use the -P switch with tar
and do use -p. I've never tried to tar a partition while running
from it, but have seen discussion of problems if you don't go to
'init s' and re-mount it read only first. I believe lilo will be
broken after your re-install since I believe lilo stores raw disk
addresses of the kernals and your re-install probably won't put the
kernel in the same place on the disk.
Let me suggest another alternative that I think will be easier
for you and more flexible. Leave hda1 as it is. Break the rest of
the drive into 1GB chunks with hda2 and hda3 as primary partitions and
hda4 as extended with logical partitions claiming the rest of the drive.
You can put linux and swap in logical partitions with no problems.
Now you want to expand RH5, so copy /usr to say hda5, mv /usr /usr.old,
mount hda5 on /usr and run it for a couple of days 'til you're confident
enough to scrub /usr.old.
Linux, FreeBSD and Solaris x86 will each install on a 1GB partition
and you'll still have free partitions to use for any of them.
Another reason to do it this way is if you decide in six months that
you would like to try FreeBSD or any other OS, they're going to want
to install on a primary partition, doing it this way you have them
available without re-partitioning again.
One last thing, I don't know what happens if you've swap on and
delete your swap partition nor what happens if you install Debian
onto a partition that RH5 thinks is swap. Might not be a problem,
might be a headache.
G'Luck,
MiKe
--- MultiMail/Linux v0.31
------------------------------
From: Kenneth Mokkelbost <[EMAIL PROTECTED]>
Subject: Re: keymap or font mismatch
Date: 18 Feb 2001 09:09:13 +0100
[EMAIL PROTECTED] (Mike Mcclain) writes:
> Howdy All,
> Every once in a while I see a message here where there
> is a mismatch between the system of the guy that wrote the
> message and mine where the message is displayed, like so:
>
> All today{s OS, I{m sure. This is what he sent.
> All today's OS, I'm sure. This is what he meant.
>
> My question: Where is the misconfiguration?
> His system or mine? What's not set right?
>
> TIA,
> MiKe
>
> --- MultiMail/Linux v0.31
Whar about the character coding? Is it ISO-8859-1 for both?
Kenneth
------------------------------
From: Shai Kedem <[EMAIL PROTECTED]>
Subject: Re: ksh script problem: pwd works differently for ksh then linux binary file
Date: Sun, 18 Feb 2001 09:30:11 -0000
I ran the test script you suggested, and got this every time:
pwd is a shell builtin
/usr/local/bin
/usr/local/bin
/usr/local/bin
I changed the first line of the script to: /bin/ksh, /usr/bin/ksh
and /usr/bin/pdksh and got the same results.
However, when I installed AT&T ksh for Linux, the test program did gave a
good result !
So I wonder if this is a bug in PDKSH, or I have a bad release. I
installed pdksh 5.2.13 running on Redhat Linux 6.2.
I can't use a higher release of pdksh which will only run on RedHat Linux
7.
any other suggestions on writing a ksh script which will simulate pwd
somehow in the right way ?
thanks for the help anyway.
Dan Mercer wrote:
>
>
> I went to the trouble of installing pdksh (I use ksh93) and
> could not reproduce your problem:
>
> #!/usr/bin/pdksh
>
> cd
> [[ -d td ]] || ln -s /usr/local/bin td
> cd td
> whence -v pwd
> pwd
> pwd -P
> echo $PWD
> $ tb
> pwd is a shell builtin
> /home/dam/td
> /usr/local/bin
> /home/dam/td
>
> So, the question is, what are you doing wrong? As you can see,
> pwd is a builtin. Now, it is possible you have ENV set and your
> ENV file (.kshrc?) is doing something screwy to cause this. To
> make sure that isn't happening, change the shebang to
>
> #!/usr/bin/pdksh -p
>
> You might want to post the script.
>
>
> --
> Dan Mercer
> [EMAIL PROTECTED]
>
> In article <[EMAIL PROTECTED]>,
> Shai Kedem <[EMAIL PROTECTED]> writes:
> >
> > Harlan Grove wrote:
> >>
> >>
> >>
> >> Shai Kedem <[EMAIL PROTECTED]> wrote in message
> >> news:[EMAIL PROTECTED]...
> >> ...
> >> >I have a pdksh installed on Linux RedHat 6.2.
> >> >From an interactive shell, when I type pwd in a directory which is a
> > link,
> >> >I get the directory name (not the one linked to), as I expect.
> >> >When /bin/pwd is run, I get the linked directory name !
> >> >from a ksh script, it seems it will always run /bin/pwd for pwd
rather
> >> >then the ksh implementation.
> >> >Any one knows how to override this problems in ksh scripts to it
will
> > run
> >> >the pwd of the ksh command ?
> >>
> >> Why bother with running pwd either as an internal command or an
external
> >> binary? If you're using ksh as your interactive shell _and_ as the
shell
> >> running your script, wouldn't the PWD environment variable suffice?
> > x=`pwd`
> >> is every bit as redundant as uuoc.
> >>
> >> I believe pwd is an alias rather than a built-in command like cd,
which
> >> would explain why it works in interactive shells but not in scripts
> > (though
> >> I believe it's possible for the script to trick the shell into
thinking
> > it's
> >> interactive).
> >>
> >>
> >
> > I tested the $PWD but it behaves the same - it will resolve the
symbolic
> > link when run from ksh script !
> > any idea how to force a command to be interpeted as alias rather then
> > executable name in ksh script ?
> >
> >
> >
> > --
> > Posted via CNET Help.com
> > http://www.help.com/
>
>
>
> Opinions expressed herein are my own and may not represent those of my
employer.
>
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: [EMAIL PROTECTED] (Anonym5530)
Date: 18 Feb 2001 10:13:46 GMT
Subject: Trying to install RH 7.0 on an HP
Hi,
I am trying to install Red Hat 7.0 onto an HP 8750C pavillion machine. I
cannot turn of the Intel 810 onboard video card, and RedHat's anaconda won't
work...nor will x. I have an ati video card installed that I would prefer not
to have to remove everytime I want to run linux.
I've used linux for many years...just have no experience with extreme technical
dificulties like this. If you have any ideas as to what I can do...besides
throw the machine away =) I'd love to hear them. If you need more
information, please feel free to let me know.
I'll try and check this ad within the next day or so
you can also get ahold of me at [EMAIL PROTECTED]
Thank you
Jon
Love your neighbor, But don't get caught!
------------------------------
Date: Sun, 18 Feb 2001 12:02:46 +0100
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: su command while in a program?
Richard James Panturis Giuly wrote:
> Is there some way to change to a different user (like su) so that you
> suddenly gain that users access priveledges while you are still in a
> program. I mean you try to access a file in a program, and it fails, so
> you just change users and you have access, without exiting the program.
>
> --
> Richard Giuly
>
> (remove animal from email address)
Hello,
I'm im doubt this would be useful, as you have to give this users pwd,
unless
you're root, and that would mean, you don't need to do it as there are no
limits
or access settings that restrict roots access on a UNIX system.
You' ll be better of using /etc/groups and adjust the group permissions.
Good luck
Michael Heiming
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: I need software recommendations:
Date: Sun, 18 Feb 2001 06:22:35 -0500
Mike Perry wrote (in part):
> I could probably never use outlook again or agent. I like a client I can
> ssh to several machines I have accounts on and read mail there without
> some clumsy GUI that gets in the way.
With Linux and openSSH I can do the same thing with Netscape, of
course. I did it once to be sure I could, and it worked just fine,
though it made me laugh.
I have one dial-up connection on this machine, that is connected
with a 100-TX LAN to my other machine. Both run Linux. My sister was
using Netscape with her account on this machine, and I wanted to
browse the internet, so I just did
ssh -X -l mylogin thismachine
and when I got connected, I just ran Netscape on this machine using
the X-server on the other machine. So we were both running our own
instance of Netscape on this machine, and my other machine was just
running as an X-Terminal. Of course, there is nothing special about
Netscape; you could use lynx, mutt, slrn, or whatever suits your
needs, and run it through ssh just as well (if not better, depending
on the speed of your connection. With 100 Megabit LAN, it was fast
enough for Netscape).
Reminded me of the middle-days (like the old days, but not quite
that old) when all machines were multi-user, before the
brain-damaged operating system came out.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:10am up 20 days, 14:37, 4 users, load average: 2.03, 2.08,
2.16
------------------------------
From: Robert Schweikert <[EMAIL PROTECTED]>
Subject: Re: troubles installing gdbm
Date: Sun, 18 Feb 2001 06:27:17 -0500
Mark,
Thank you, Fixed.
Robert
Mark Post wrote:
> On Sat, 17 Feb 2001 12:53:37 -0500, Robert Schweikert
> <[EMAIL PROTECTED]> wrote:
>
> >I get the following error when I try to do make install for gdbm.
> >
> >/bin/install -c -m 644 -o bin -g bin gdbm.h /usr/local/include/gdbm.h
> >/bin/install: invalid user `bin'
> >make: *** [install] Error 1
> >
>
> Do you have user bin defined in /etc/passwd?
>
> Mark Post
>
> Postmodern Consulting
> Information Technology and Systems Management Consulting
> To send me email, replace 'nospam' with 'home'.
--
Robert Schweikert MAY THE SOURCE BE WITH YOU
[EMAIL PROTECTED] LINUX
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Size of LINUX
Date: Sun, 18 Feb 2001 06:40:17 -0500
Harlan Grove wrote:
>
> [EMAIL PROTECTED] (Rolie Baldock) wrote:
> ...
> > . . . We engineers should have kept computers for the exclusive
> >use of engineers. Engineers used to have a code of excellence in my
> >days in the profession. The PDP-6 operating system reflected this
> >excellence.
>
> Gee, I'm a mathematician by education. Should I never have been given a
> computer?
>
> This is a fine example of one engineer's narrowmindedness. There are a
> great many things computers are good at besides quantitative analysis,
> simulation and data acquisition. Should these other computer
> applications be forbidden to all nonengineers?
>
> OK, the PDP's were before my time. FWIW, Unix originated on a PDP (10 or
> 11?), so I'd guess it qualified as tightly-coded way back then.
I first used it on a PDP-11, but I think Ken and Dennis did a
prototype on a spare PDP-7 (or 9?); certainly NOT a PDP-10. Mine was
the first one they had access to that had a memory management unit
init (theirs were PDP-11/20 if I remember correctly, and I had a
PDP-11/45.
I did not much like UNIX in those days, because what I was doing
amounted to real-time process control, and I had already written an
OS for a Honeywell DDP-224 that would do that. But the machine was
getting old, and the manufacturer could not guarantee that they
would be able to keep it running, so we got the PDP-11/45 and re-did
the rest of the system to use that instead. We first used a modified
version of UNIX in it, but it would not stay up long enough to run
our jobs (some of them took several days to run), so after fighting
with it for about a year, we switched to DEC's RSX-11D OS that
worked just fine, though it was somewhat more difficult to use at
first.
I will never forget when I forgot to lock the memory manager to
core. It swapped itself out, and since it was out, it could not swap
itself back in, nor could any other swapping take place, so the
whole thing deadlocked.
RSX-11D was sort of what would now be called a micro-kernel design.
The hardware had three modes: user, supervisor, and kernel (I do not
know if those were the right names anymore), and there was not a
whole lot that ran in kernel mode. IIRC, the device drivers, memory
manager, and GOK what else, ran in supervisor mode (a mode with more
privilege than user, but less than kernel).
> However,
> unless there was magically no space-time tradeoffs back then, it's a
> certanty that the OS would have had better response time if it had been
> allowed to occupy more RAM.
Not necessarily, but there certainly were tradeoffs. One was just
clarity vs. size. I once wrote some code where I did a CLEAR
REGISTER 7 (the PC) on a machine, rather than a JUMP 0, because the
former instruction took 2 less bytes of memory, and that was all I
had. But it sure was obscure, not a good thing. We also spelled
error messages incorrectly, or used obscure error codes, to reduce
the number of bytes required. No one does that today (I hope).
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:25am up 20 days, 14:52, 4 users, load average: 2.04, 2.07,
2.08
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Size of LINUX
Date: Sun, 18 Feb 2001 06:47:14 -0500
Harlan Grove wrote (in part):
>
> [EMAIL PROTECTED] (Rolie Baldock) wrote:
> ...
> >Are you a super fast typist? Don't you ever stop to scratch you head
> >because the problem is a little baffling. 110 baud was quite adequate
> >for SERIOUS programmers. We used BRAINS in the old days not FANCY GUIs.
> ...
>
> Maybe you SERIOUS types didn't need output because you could memorize
> the source code for entire systems AND calculate program output in your
> sleep. Kinda makes me wonder why you superhumans bothered with computers
> in the first place.
I remember the first time I tried reading the source code to the
Unix kernel. In those days, it was 6 files named something like
k1.s, k2.s, ... . You did cat *.s | as and the a.out file produced
was your new kernel. The files were not organized in any particular
way that I could figure out. In those days, the maximum file size of
UNIX was 65536 bytes, so when one filled up, they just started
another.
But I inferred that even Ken and Dennis could not keep all that in
their minds, because you would frequently find a jump in one file to
a label in another, where the first instruction at that label was a
jump to yet another label. Obviously, the person writing the former
jump remembered that the first label did what was needed, but did
not know that that label was a jump to yet the other label where it
got done. The first jump should probably have been a jump direct to
the ultimate label.
So even Dennis and Ken could not remember all the source to the UNIX
kernel back when it was only 6 small assembler files.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:40am up 20 days, 15:07, 4 users, load average: 2.06, 2.08,
2.08
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Size of LINUX
Date: Sun, 18 Feb 2001 06:55:57 -0500
Rolie Baldock wrote:
>
> Hello Jean-David,
>
> The PDP-6 had multiported memory which meant that one did not have to
> ask the processor for memory access. The memory ports were prioritized
> so the device with the highest real-time needs was connected to the
> highest priority memory port. Did it matter if the processor had to
> take the second port? NO not a jot.
>
So did my DDP-224. The memory actually had 4 ports -- sort-of, and
the "DMA" channels came in two flavors: DMA and FBC. The DMA units
did cycle stealing, sharing with the CPU. The FBC channels did not
need to steal cycles, but could run totally independently, if they
were using a different block of memory (memory was divided into 4
blocks of 4096 words each on my machine); otherwise, they stole
cycles too. Even with the DMA channels, the contention was not much
of a problem because most machine instructions took two (or more,
with multiply, divide, and floating point) cycles, only one of which
was needed for memory access. The contention was negotiated at the
hardware level, not needing interrupts for each word transferred.
That machine had two removable-pack disk drives (made by CDC), each
of which could hold 40 Megabytes. That was enormous, in those days.
My RAM on this machine over 12x the size of one of those disk
drives. The clock on that machine was about 800 KHz, and the clock
on this machine is 2x550MHz (I guess you could say).
I do prefer my present machine running dual Pentium IIIs in Linux
SMP.
Not wishing for "the good old days" ... .
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:45am up 20 days, 15:12, 4 users, load average: 2.13, 2.08,
2.08
------------------------------
From: Jean-David Beyer <[EMAIL PROTECTED]>
Subject: Re: Size of LINUX
Date: Sun, 18 Feb 2001 07:04:52 -0500
Rolie Baldock wrote:
>
> Hello Folks,
>
> I am not specifically condemning LINUX as such. I am wondering if
> LINUX was written in assembly language by a COMPETENT assembly
> language programmer(the likes of which I have not seen for a long
> time) would it be a lot smaller and faster. Putting all ones faith in
> the author of the C compiler in my view is not good software
> engineering. It simply CANNOT be proved that the compiled code is the
> tightest and fastest that can be produced. My experience with high
> level language compilers is that they are ALWAYS inefficient. I write
> assembly language code for my PC and the .COM files are always VERY
> SMALL, amazingly so sometimes. "Get last disk" uses just 651 bytes!!
> Try that in C and see how many bytes it takes.
>
I do not know for sure about the Linux OS, which I believe is
written almost entirely in C, with perhaps a trifle of assembler,
but the first versions of the UNIX OS were written entirely in
assembler by some extremely competent programmers: Ken Thompson and
Dennis Ritchie. After Dennis got the C Programming Language designed
and wrote a compiler for it that produced reasonably good code, they
converted the whole thing to C. I am certain they would not have
done that if they thought it compromised the performance very much.
Ken had worked on the MULTICS project, after all, and one of the
lessons they learned on that project, which was written in EPL (an
early version of PL/I) was that their performance problems were not
due to writing in a higher level language, and that using EPL saved
the integrety of the system because people could understand the
code, something not possible when everything was written in
assembler.
Later, when Steve Johnson wrote "The Portable C Compiler", it made
it possible to make a few trifling machine-dependent changes to the
source code, retarget the compiler, and then port the UNIX OS to
another machine architecture with relatively little effort. That was
a really big deal in those days. Before that, it was unheard of to
run the same OS on machines of different manufacturers.
--
.~. Jean-David Beyer Registered Linux User 85642.
/V\ Registered Machine 73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 6:55am up 20 days, 15:22, 4 users, load average: 2.04, 2.07,
2.08
------------------------------
** 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: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.misc.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Misc Digest
******************************
