Linux-Misc Digest #269, Volume #19 Tue, 2 Mar 99 19:13:11 EST
Contents:
Re: FreeBSD vs LINUX (Robert Sexton)
Is linux 64-bit ready already? (B. Binegar)
Re: Killing Zombie Processes (Juergen Heinzl)
Re: Killing Zombie Processes (brian moore)
Linux SLOWER than win95? ("Raf Meeusen")
Re: FreeAgent for Linux ([EMAIL PROTECTED])
Re: Public license question (brian moore)
Re: Best Free Unix? (why FreeBSD?) (Paul Hughett)
Re: Backup software (Markus Wandel)
Re: Microkernels are an abstraction inversion (Stefan Skoglund)
Ghostscript/Ghostview binary distribution (Sergio Antoy)
Re: Linus can settle some of this (Re: Best Free Unix? (why FreeBSD?)) (drwho)
Help Working under w95 but not under RH 5.2 Linux (Rajat & Papia Goon)
Re: Anti-Virus for Linux (Jason Clifford)
Re: Restricting su (Juergen Heinzl)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Robert Sexton)
Crossposted-To: comp.unix.bsd.freebsd.misc
Subject: Re: FreeBSD vs LINUX
Date: 2 Mar 1999 15:21:09 GMT
Donn Miller ([EMAIL PROTECTED]) wrote:
: Seth Van Oort wrote:
: happen, as FreeBSD and Linux have different models for development. The
: kernel code for each OS is different, and FreeBSD would like stay with the
: legacy BSD kernel code as much possible. The best you could do is invent
: a common i386 binary type that runs on the i386 versions of FreeBSD and
: Linux.
FreeBSD already runs linux binaries. As executable formats get
hammered out, the converse will probably become true.
The UNIX would continues to inch towards portable executable formats.
--
Robert Sexton - [EMAIL PROTECTED], Cincinnati OH, USA
Kudra.com - New home of the hazardous data disposal service,
featuring our exclusive Write Only Memory(TM) Technology!
Read the Newton FAQ! <http://www.kudra.com/newton/newton-faq>
------------------------------
From: [EMAIL PROTECTED] (B. Binegar)
Subject: Is linux 64-bit ready already?
Date: Tue, 02 Mar 1999 21:07:02 GMT
Reply-To: [EMAIL PROTECTED]
I've been trying to determine which platforms will support the 64-bit version of
MapleV (a commercial software package for carrying out symbolic and graphical
mathematical calculations), since the 32-bit versions can not handle for example a
generic 40th degree polynomial in 4 variables (that requires a coefficient list with
more than 2^16 -1 entries).
I posted to sci.math.symbolic with this question and learned the following. If you
install Red Hat linux 4.2 on an alpha system, and then install MapleV for linux
it appears to be 64 bit (visit http://130.15.26.62/NewDemo/frame.html to see
examples of tensor calculations in General Relativity that were carried out in this
fashion). In fact, if you give the Maple command
> kernelopts(wordsize)
it will kick back 64 (as opposed to 32 on a intel box).
What's darn strange is that, as far as I can tell, the binary distribution of MapleV
for linux is the same for both alpha and i386 platforms.
How could this be? Is it possible that executable compiled for linux are 64-bit
ready in this sense? Or is this just a fluke? (One could perhaps explain away the
second example, if kernelopts(wordsize) justs polls the operating system, but
nevertheless, it does seem that Maple on an linux/alpha machine is capable
carrying out calculations requiring a larger word size than is available on a
linux/intel machine).
Thanks in advance, Birne Binegar
(BTW, MapleSoft would neither confirm or deny the above, they simply say that
YMMV but Digital Unix is the only platform for which *they* support a 64 bit version
of Maple).
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: Killing Zombie Processes
Date: Tue, 02 Mar 1999 20:52:39 GMT
In article <[EMAIL PROTECTED]>, David W. Robinson wrote:
>You can kill processes using "top," which has a man page explaining how.
>
>"William T. Trotter" wrote:
>
>> Now there must be a better way round this
>> mess. I can't believe that Linux requires a
>> reboot to get rid of the remnants of a program
>> that crashed.
It has nothing to do with Linux but the Unix process model, see Bach
for more (might be the best one). A zombie actually is not a process
anyway, it is *just* an entry in the process table. While it is
possible to kill a system that way I never encountered that case in
practice, even on machines with 6 - 12 months uptime ... the longest
time without a powerfailure so far 8)
Cheers,
Juergen
--
\ Real name : Jürgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
From: [EMAIL PROTECTED] (brian moore)
Subject: Re: Killing Zombie Processes
Date: 2 Mar 1999 15:32:04 GMT
On Tue, 02 Mar 1999 08:24:47 -0500,
Thomas Boggs <[EMAIL PROTECTED]> wrote:
> brian moore wrote:
>
> > First, what you're describing isn't a zombie. A zombie is a process that
> > is not merely broken, but quite dead already and you can't kill it. (The
> > parent of the zombie needs to collect the return code from its child and
> > until it does so, the zombie will hang around the process table.)
> >
> > Yours is simply a broken netscape.
> >
> > These can be slightly tricky to kill in that as of Netscape4, early in
> > the startup process Netscape forks into the web browser you see and a
> > coprocess that does DNS lookups and such.
> >
> > You need to kill both of them.
> >
> > One will ignore you (it traps the signal) and the other will nicely die
> > when killed.
> >
> > Look closer next time and you'll see '(dns helper)', the demon (as opposed
> > to daemon) spawn of Netscape.
>
> So how do you kill processes which actually are zombified? I keep getting
> zombies of Xsession and Gnome panel.
You don't. (Zombies can't be killed: they're already dead.)
If you wish to make them go away, you should either wait for sunlight or
have their parent collect their return status.
Every process on Unix collectes several values (resources used and
whatever main() returns as well as whether it was an exit or a signal
that killed it). These values must go somewhere. A zombie is merely an
entry in the process table waiting for its parent to collect these values
before it goes away.
It is the responsibility of the parent to collect these or to properly
abandon the child so that init will adopt it. If it doesn't, the
parent is broken and should be fixed or replaced.
--
Brian Moore | "The Zen nature of a spammer resembles
Sysadmin, C/Perl Hacker | a cockroach, except that the cockroach
Usenet Vandal | is higher up on the evolutionary chain."
Netscum, Bane of Elves. Peter Olson, Delphi Postmaster
------------------------------
From: "Raf Meeusen" <[EMAIL PROTECTED]>
Subject: Linux SLOWER than win95?
Date: Tue, 2 Mar 1999 16:31:15 +0100
I installed the latest Linux Mandrake (=Redhat+KDE) on a P60 with 16meg ram.
I use a swap partition of 70 Mb.
But it is much slower than my windows 95.
Is this normal?
Is there a way to speed it up?
(like recompiling the kernel)
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: FreeAgent for Linux
Date: Tue, 02 Mar 1999 23:14:07 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> I have heard rumours (from god knows where) that there is a Linux porting
> project of the Agent reader going on. Can anyone confirm this?
If not, this is the cruelest tease to which I have ever been subjected. :-)
Regards, Dustin
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (brian moore)
Crossposted-To: comp.os.linux
Subject: Re: Public license question
Date: 2 Mar 1999 23:21:56 GMT
On Tue, 02 Mar 1999 14:36:48 -0500,
Rick Onanian <[EMAIL PROTECTED]> wrote:
> brian moore wrote:
> >
> > On Tue, 02 Mar 1999 11:43:15 -0500,
> > Rick Onanian <[EMAIL PROTECTED]> wrote:
> > >
> > > As I understand it...The GPL is really not able to be enforced by
> > > anyone. I get the feeling the the FSF doesn't have the money to take
> > > anyone to court over a gpl violation...I also get the feeling that
> > > maybe noone really cares too much - the gpl is more a matter of intent
> > > than technicality. If you write a program and GPL it, do you necesarily
> > > make a registered copyright? Or do you just release it and hope or not
> > > care..
> >
> > Hint: do you think NeXT released their modifications to gcc for
> > 'objective C' because of warm fuzzy feelings?
>
> These ideas, in part, come from previous posts involving WordPerfect,
> StarOffice, and other non GPL'd programs available for Linux.
They're wrong: Corel and others can release software for Linux and
retain all the source they want. Completely legal and no one would be
able (or care to) threaten them with that.
But there is a difference between being "legal" and being "ethical":
Corel is going much farther than many companies and is aiding the WINE
project and porting the kernel to the StrongARM platform and sharing the
code for those projects. This shows they are members of the community.
> > The Free Software Foundation does have lawyers, and they do pursue
> > license violations on code they control.
>
> I did not know they even had much money. I got the feeling that they
> didn't.
No one needs to pay for lawyers.
> > That said, they don't control the Linux kernel.
>
> Or much of the other GPL'd software.
Of course not.
> Correct me if I am wrong, but to obtain an actual copyright on
> something, you have to register it with some authority? Probably,
> some government?
Nope: you -can- register if you want, but it's not required. The Berne
convention holds that anything that is copyrightable is copyrighted
unless specifically renounced. This post, for example, is copyrighted.
> I have been told that you can copyright something simply by saying
> copyright all over it, however, I don't know how that would hold in
> a court of law.
You don't even need to label it. If it is copyrightable, it is
copyrighted.
> That being said, how many parts of any given distribution of Linux
> have real copyrights? I would venture to guess that much of GPL'd
> software out there is not really copyrighted.
You'd be wrong. Please read the Copyright FAQ.
> I do not claim to know law, or programming, for that matter. I do
> not practice law; rather, I offer my opinions from a point of view
> that I believe to be [very] slightly more informed than those who
> ask.
You believe wrong.
--
Brian Moore | "The Zen nature of a spammer resembles
Sysadmin, C/Perl Hacker | a cockroach, except that the cockroach
Usenet Vandal | is higher up on the evolutionary chain."
Netscum, Bane of Elves. Peter Olson, Delphi Postmaster
------------------------------
From: [EMAIL PROTECTED] (Paul Hughett)
Crossposted-To:
comp.unix.questions,comp.unix.advocacy,comp.unix.misc,comp.unix.bsd.freebsd.misc
Subject: Re: Best Free Unix? (why FreeBSD?)
Date: 2 Mar 1999 15:46:35 GMT
Gregory L. Hansen ([EMAIL PROTECTED]) wrote:
: >>> "RSS" == Robert S Sciuk wrote on Thu, 18 Feb 1999 17:56:34 -0500:
: >
: >RSS> Stallman's vision is slightly myopic with respect to the forces
: >RSS> which drive innovation. I respect his/your/anyone's right to
: >RSS> give away whatever they like, but the GPL is akin to the
: >RSS> Bolshevik forced collectivization at the turn of the century
: >RSS> ... it takes away others rights to derive proprietary products.
: >RSS> That is something to which I cannot subscribe.
: Rights? What "rights" to derive proprietary products? Since when did
: anyone have the RIGHT to derive proprietary products from any licensed
: code? You may get permission to derive proprietary products. You might
: have to ask, and pay money, to get permission, and permission may be
: denied. The GPL tells you under what conditions you are allowed to derive
: proprietary works. The BSD license is more liberal -- if you use any
: BSD-licensed code, be sure to thank the authors for doing you the favor.
Well put. That's exactly why some of us believe that the GPL is less
free than the BSD license. You--as a developer--should choose the
license that is appropriate to your goals. If you want your code to
be free for all and sundry to use and improve on, use the BSD license.
If you want to prohibit proprietary derivatives, use the GPL.
Paul Hughett
------------------------------
From: [EMAIL PROTECTED] (Markus Wandel)
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup,comp.os.linux.x
Subject: Re: Backup software
Date: 2 Mar 1999 15:29:08 GMT
In article <[EMAIL PROTECTED]>,
Rick Knight <[EMAIL PROTECTED]> wrote:
>Can anyone recommend a backup/restore program for Linux (RH 5.2) and=20
>X? I have a Seagate STT28000n SCSI Travan tape drive GUI based program=20
>to use with it. I've tried Knox's ARKEIA but it seems to buggy and the=20
>backup utility that comes with KDE doesn't support my drive.
If you are using RH5.2 and a SCSI tape drive, you can almost certainly
get started with something like this:
tar cvf /dev/st0 /
In this case "c"reate an archive "v"erbose output, archive "f" is /dev/st0
which is the first SCSI tape drive (actual SCSI ID doesn't matter in my
experience.) Thereafter follows a list of files/directories to back up.
In this case "/" should back up everything.
Of course in Unix tools are kept basic so you can combine them in your own
way. In my case I made a basic incremental backup system out of "find"
and "tar" with a bit of other utilities like "awk" and "grep" to glue it all
together. The man page for "find" is daunting but well worth the read; this
is probably _the_ most powerful tool for backup work in Unix.
I have a 10GB disk but the "working set" that needs backing up, specifically
data which
- is not reasily reinstalled e.g. from the RH5.2 CD ROM
- is not junk e.g. /tmp/*, GIMP swap files, core files
- is not replaceable, e.g. */.netscape/cache/*, */.xvpics/*
- is not backed up in other ways, e.g. a photo library that also exists
on a CD-ROM and on my brother's computer
amounts to about 100-200MB, and that is the key to actually doing effective
backups.
If anyone wants my backup scripts -- which are very basic and not "plug and
play" foolproof but work very well for me -- they can contact me privately.
Markus
------------------------------
From: Stefan Skoglund <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss,comp.os.linux.advocacy
Subject: Re: Microkernels are an abstraction inversion
Date: Wed, 03 Mar 1999 00:13:14 +0100
Francois-Rene Rideau wrote:
> Of course. Just because =B5K is evil does not mean lack of =B5K solves =
it all.
> It's just like all evil ideas. Let's take communism, for instance.
> Just because you discard the evil idea isn't enough to make you happy;
Communism in the version created by Lenin was evil because it was
enforced
on other people with a big GUN. The mass-conscription and requisitioning
of food stuff created a very nasty famine. That was enough to break
the Russian agricultural life until this time today ie they
haven't recovered yet.
the problem with ML and the other languages which you mentioned
is that they are pretty hard to understand for someone who didn't
learn functional programming from the beginning ie started to work
with Basic,C or shudder C++.
ML doesn't really allow side-effects does it ehhh??
------------------------------
From: Sergio Antoy <[EMAIL PROTECTED]>
Subject: Ghostscript/Ghostview binary distribution
Date: Tue, 02 Mar 1999 15:54:48 -0800
Could someone please point me to the latest binary distribution of
ghostscript/ghostview for slackware?
Thanks,
Sergio
------------------------------
From: drwho <[EMAIL PROTECTED]>
Crossposted-To:
comp.unix.questions,comp.unix.advocacy,comp.unix.misc,comp.unix.bsd.freebsd.misc
Subject: Re: Linus can settle some of this (Re: Best Free Unix? (why FreeBSD?))
Date: 02 Mar 1999 17:56:01 -0600
[EMAIL PROTECTED] (Matthias Warkus) writes:
> It was the Mon, 01 Mar 1999 17:35:16 GMT...
> ..and DrBoom <[EMAIL PROTECTED]> wrote:
> > Linus, do you have any intentions of changing the terms of the Linux
> > kernel license?
> >
> > Could you do it even if you wanted to?
> >
> > What are the terms of use for the Linux brand name?
> >
> > Is it possible for you to restrict its use in any way?
>
> I doubt that Linus will bother to answer any of these questions, as he
> has probably already answered them in full, and possible in print, in
> a zillion other places.
Well, it would certainly be nice if he, or someone, at least pointed us
to the location of these "answers". But the simple fact is, I don't think
he has answered these questions, and I doubt that he will.
--
Fight email spam: http://www.cauce.org/
------------------------------
From: Rajat & Papia Goon <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.hardware,comp.os.linux.setup
Subject: Help Working under w95 but not under RH 5.2 Linux
Date: Tue, 02 Mar 1999 09:58:15 -0600
Reply-To: [EMAIL PROTECTED]
I am using SMC PCI-NIC. In auto probing it finds it irq 11 and address
0xe800 hich matches
with win95. If I run ifup it's giving delaying eth0 initialization. I am
using smc9194 module. If
I use ifconfig then
SIOCSIFBRDADDR
eth0 : interface not found
I am desperately trying to make it work.
Initially it was giving at the booting time smc9194.o: init_module the
device is busy, delaying eth0 initialization.
But now it is not giving any messages like that but it is not loading
the module. But PCI is picking up the card
correctly. Can anyone help me out from this place.
note : my card is also ne2000 compatible. Shall I try to load ne2k-pci.o
module with IRQ 8 and base address
0xe800. I haven't try this option.
Thank you in advance,
--
Rajat Goon
E-mail : [EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED]
------------------------------
From: Jason Clifford <[EMAIL PROTECTED]>
Subject: Re: Anti-Virus for Linux
Date: Tue, 2 Mar 1999 14:17:40 +0000
On Sun, 28 Feb 1999, John Thompson wrote:
> you set up that machine to monitor the traffic for potential
> viruses. And if that machine happens to run linux, you'd
> need a linux program for this job.
Amavis for scanning emails and attachments in sendmail. You can link it in
with pretty much any anti-virus solution that will run on Linux such as
AntiVir/X, McAfee, Dr. Solomon's (?) or Sophos. I have recently been
looking at the Sophos one with a view to including an evaluation with
Definite Linux and it appears to be pretty good. It also offers an option
to have clients use it as the AV engine with a much reduced client side.
I am still of the opinion that you really should have up to date
anti-virus software on any system that is liable to be subject to virus
risk as well as on the server and any gateways. Or even better, simply
don't use a platform for which literally thousands of virii have been
written due to it's absolute lack of any security features.
Jason Clifford
Definite Linux Systems
http://definite.ukpost.com/
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Subject: Re: Restricting su
Date: Tue, 02 Mar 1999 21:34:48 GMT
In article <w2eAs8OZ#[EMAIL PROTECTED]>, Michael Lee Yohe wrote:
>>How do I restrict su to only certain users? Is there a different version
>>of su available that has this feature, and if so, where would I get it.
>>Linux newbie, please be gentle. :)
>
>I suggest downloading the latest "SUDO" (Super User) package. With sudo,
>you can remove "su" from your system and let people "try" to hack away at
>it. It works like this:
[...]
Or if you've got the latest shadow package see man suauth(5), esp.
the DENY keyword.
Cheers,
Juergen
--
\ Real name : Jürgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
** 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 (and comp.os.linux.misc) via:
Internet: [EMAIL PROTECTED]
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
******************************
