Linux-Misc Digest #340, Volume #19                Sat, 6 Mar 99 23:13:08 EST

Contents:
  Re: Can Linux use 36-bit Xeon addressing? (Johan Kullstam)
  Re: Stupid Newbie tricks (Add this to list) (Jim Shaffer, Jr.)
  Re: Microkernels are an abstraction inversion (Anthony Ord)
  Re: Move to RH 5.2 Linux (John Hasler)
  Re: netcape + freshmeat.org then crash? (Jon Sundquist)
  comments after #!/bin/sh ? (Mike Coffin)
  PCMCIA configuration HELP!! (Raul)
  Help - Procmail rulesets ("David C.")
  Login timeouts in runlvl. 5 (John David Bowden)
  xdm problems - can't shutdown ([EMAIL PROTECTED])
  QT build problems ("Geoff Thornton")
  Re: Need to capture the 'raw' mouse under X (Kalle Olavi Niemitalo)
  Re: More bad news for NT ("Jon Wiest")
  mount win95 16 (LBA) (Rulecoyote)
  A question about MP3's and Linux (J Crowtz)

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

Crossposted-To: comp.os.linux,comp.os.linux.hardware
Subject: Re: Can Linux use 36-bit Xeon addressing?
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 06 Mar 1999 17:48:00 -0500

Robert Krawitz <[EMAIL PROTECTED]> writes:

> "David A. Frantz" <[EMAIL PROTECTED]> writes:
> 
> > I think you hit on one problem at the very least and that is if you want to
> > get away from I386 you have only one other mass produced platform and that
> > is Apples Mac.    When dealling with software I do not think binary
> > compatablity is s big deal for Linux users.    After all if you want to run
> > something you can just recompile it.
> 
> For better or for worse, the latter statement is not true, and it's
> becoming less and less true.  If your goal is to run Oracle, Informix,
> DB/2, Sybase, or what have you, binary compatibility is essential.

true.  but to exploit a 36 bit address space in order to use more than
4 GB on a xeon, would require recompiling those applications.  so your
program falls into two buckets:

1) something which you can port or find a port of for linux alpha or
   sparc solaris  &c.

2) something which exists only in binary format on i386 linux.

now in case 1, you can use the 64 bit processor.  in case 2, you
can't, but neither can you exploit over 4 GB.

therefore, there is no point in trying to make a far pointer memory
model for linux on the x86.  should you need big memory, just use
a 64 bit platform.

-- 
                                           J o h a n  K u l l s t a m
                                           [[EMAIL PROTECTED]]
                                              Don't Fear the Penguin!

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

From: [EMAIL PROTECTED] (Jim Shaffer, Jr.)
Crossposted-To: comp.os.linux.x
Subject: Re: Stupid Newbie tricks (Add this to list)
Date: Sat, 06 Mar 1999 00:11:14 GMT
Reply-To: [EMAIL PROTECTED]

On Fri, 05 Mar 1999 15:10:06 GMT, [EMAIL PROTECTED] (Pete Tolen) wrote:

>Issuing the 'shutdown now' command from an xterm window while running
>X.

I don't get it.  On my system, it does exactly what I'd expect it to do: shut
down the system.  Why exit X manually?

-- 
home page: http://woodstock.csrlink.net/~jshaffer

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

From: [EMAIL PROTECTED] (Anthony Ord)
Crossposted-To: gnu.misc.discuss,comp.os.linux.advocacy
Subject: Re: Microkernels are an abstraction inversion
Date: Sat, 06 Mar 1999 15:56:25 GMT

On Thu, 04 Mar 1999 00:40:56 GMT, [EMAIL PROTECTED] (Christopher
Browne) wrote:

>On 02 Mar 1999 22:41:35 -0700, Craig Kelley <[EMAIL PROTECTED]> wrote:
>>[EMAIL PROTECTED] (Christopher Browne) writes:
<snip>
>>Or, just design it to use a flat file (which can be arbitrarily
>>interpreted by the OS to mean anything from a disk partition to a
>>clustered, mirrored, interstellar network).  Ever heard of KISS?
>
>[Warning: Possible Misunderstanding gap!]
>
>I think you've changed topics to that of "How should a DBMS be
>implemented?"
>
>Which is a pretty gaping change of topics. 
>
>If so, I *don't* agree that "just using a flat file" is going to cut it;
>SQL may be ugly, and there may be decent isomorphisms to get *some*
>queries to map to requesting data from what *look* like flat files.
>
>But flat files *don't* give you O(log_p N) (where N is the number of
>records, and p is the number of records that fit on a page) access time
>like B-Trees do, or O(1) like hash tables provide.  

The thing is, they look like flat files, but they needn't be. You
could have a record for a sale, with the customer number, name,
address etc in. You change the customer number, and lo!-and-behold!
the name & address change in this "flat file" database.

After all, that's what queries are there for, to convert multiple
tables into one flat file, which is then printed out to a strict
deadline to be forgotten about.

>And fast, robust multi-user access to large amounts of data *does*
>legitimately require the use of schemes that will involve a database
>manager server.  KISS is what gets people things like MS Access, 

No. SUCK is what gets people stuff like Access.

>which
>sticks a whole database into a single "flat file," but does so quite
>incredibly badly from the perspectives of locking, updating, and
>robustness. 

Sucky implementation, doesn't make it a bad idea. There are problems
with pseudo flat files, but you need to think a little deeper to
discover them.

<snip>
>>  o printing
>>  o video
>>  o serial i/o
>>  o disk storage
>>  o kernel A
>>  o kernel B
>>  o etc.
>>
>>Well, what do we have here?  A quasi-monolithic kernel.  There isn't
>>much difference between the above and a completely modularized Linux
>>kernel.  Pushing functions from space to space does not necessarily
>>benefit you in any way.  If you can't trust kernel code, then you
>>can't trust kernel code: uK advocactes always assume that the uK is
>>pure.
>
>I look at it *somewhat* differently.
>
>What we get after all this is that services run as "servers" or
>"daemons" or some such thing. 
>
>Those that are performance critical will tend to get pushed to the
>"lowest level," whether that be by integration into the monolithic
>kernel, or by making them into uK "servers" that amount to about the
>same thing. 
>
>[Which may represent that we're in "violent agreement" as opposed to
>other forms thereof :-).]
>
>But I certainly agree that the more functionality that you push into
>roles that *look* like kernel roles, the more you need to trust them. 
>
>For instance, if, in Hurd, you have a shell script written in Bash that
>Does Something Important to interface a pseudo-filesystem to the
>filesystem interfaces, then you have to trust all the components
>involved as if they were kernel components.  Any bugs in any of 'em
>insinuate themselves into the overall system as if they were kernel
>bugs. 

I would disagree with that. I think that the real criterion has to be
"Would I lose system integrity if this failed?" Your shell script
above may fall into this category, or it might be well-written and
stateless, and you can restart it without problems. It depends - as
they say.

Regards

Anthony
-- 
=========================================
| And when our worlds                   |
| They fall apart                       |
| When the walls come tumbling in       |
| Though we may deserve it              |
| It will be worth it  - Depeche Mode   |
=========================================

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

From: John Hasler <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: Re: Move to RH 5.2 Linux
Date: Sun, 7 Mar 1999 01:20:21 GMT

Bill Voight writes:
> I'm the Unix SA at Amtrak HQ.  We have a national network of mostly NT4.0
> servers.  I've been asked to write a paper on the implications of moving
> to RH5.2 on both servers and desktops for a major component of our
> organization.

I suggest that you give serious consideration to Debian (www.debian.org),
especially for the servers.
-- 
John Hasler                This posting is in the public domain.
[EMAIL PROTECTED]            Do with it what you will.
Dancing Horse Hill         Make money from it if you can; I don't mind.
Elmwood, Wisconsin         Do not send email advertisements to this address.

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

Date: Sat, 06 Mar 1999 17:50:44 +0000
From: Jon Sundquist <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: netcape + freshmeat.org then crash?

Richard Griswold wrote:
> 
> Wulin Suo wrote:
> >
> > Can someone please tell why when I visit the freshmeat.net with Netscape
> > it crashes all the time?
> > I had that trouble when I use netscape 4.08, and got worse after I
> > installed netscape 4.5.
> > When I check the log file it says "bus error".    So far this site is
> > the only place that crashes my
> > navigator.

If nothing else works, try nuking your ~/.netscape directory and then
rebuild it by restarting netscape.  Sometimes stuff accumulates there
(how's that for a technical explanation?).  Anyways, it solved a bus
error problem for me once, although I have no idea why (and I tried
everything else that's been suggested before I stumbled on this
approach).

Jon Sundquist

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

From: Mike Coffin <[EMAIL PROTECTED]>
Subject: comments after #!/bin/sh ?
Date: 06 Mar 1999 17:48:36 -0800

In a shell script, is there a way to put a comment on the first line? 
In Solaris, I can do this:

    #!/bin/sh  --  # this is really -*- perl -*-
    eval "exec perl -S $0 ${1+\"@\"}"
       if $never_true;
    # more perl here ...

However, in linux (Red Hat 5.2, if that matters), executing this file
apparently passes the string "# this is really -*- perl -*-" to
/bin/sh, which thinks it is supposed to be a file name.  I can see the
logic in this, but I can't see how to do what I want to do, which is 
to put "perl" somewhere in the first line of the script.

-mike


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

From: Raul <[EMAIL PROTECTED]>
Subject: PCMCIA configuration HELP!!
Date: Sat, 06 Mar 1999 21:24:26 -0500

     Hi! I am having very serious trouble trying to get the Linux PCMCIA
services to work. After three days trying I am having thoughts of giving
up and forgetting about Linux (yes, I am fairly new to this OS, decided
to give it a try fed up of blue screens and system hang-ups, but this
doesn't look much better so far; at least the installation is being
extremely unfriendly).

     I just installed Mandrake 5.3 (it is basically Red Hat 5.2 with the
KDE desktop) from a CD-ROM in my -clone- laptop, and found the first
problem when booting up: the system locks up (needing a hard reboot)
when loading the PCMCIA modules, even without any cards in the slots.
The rescue disk wouldn't work either...

     After several more attempts I decided to wipe everything and
install again from scratch but with the PCMCIA service disabled. This
time everything went fine and I could log in, but just to find the same
problems when trying to activate the PCMCIA manually: I am able to
install the pcmcia_core module, but when I try to install the i82365 the
system locks up some times (and some others it doesn't!). By the way, my
PCMCIA controller is an O2 Micro OZ 6832 supported by the i82365 module.
When the computer doesn't lock up, I finally install the ds module. The
log messages show that the probe has found the CardBus controller and
probed a memory region, everything seems normal. After this I execute
the cardmgr, but apparently (log file again) it doesn't find any
sockets. There is no response to the insertion of cards in the PCMCIA
slots. Anyway, I only get to this point one out of five or six times.

     That is the story so far... The symptoms don't match the ones
described in the PCMCIA How-To, but I am fairly new to Linux and don't
feel very comfortable with it yet, so I might be missing something very
obvious. This is very discouraging, my PCMCIA cards have been working
fine for several months now under WIN98 and took virtually no time for
the system to identify them. I'd really appreciate it if somebody could
give me detailed information on how to proceed. Thanks!


                                  - Raul -


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

From: "David C." <[EMAIL PROTECTED]>
Crossposted-To: alt.kill.spammers,comp.mail.sendmail
Subject: Help - Procmail rulesets
Date: Fri, 5 Mar 1999 20:03:09 -0500

=====BEGIN PGP SIGNED MESSAGE=====

Howdy all,

I have the following "autoresponder" on my .procmailrc:

:0
* !^FROM_DAEMON
* !^From [EMAIL PROTECTED]
* !^Subject:.*Re:
* ^Subject:.pgp
| (/usr/bin/formail -r ; /bin/cat /master/david/docs/david.asc)|
/usr/lib/sendmail -oi -t

So far, it works half great. If it receives a mail with "pgp" on the subject, it
will email back my PGP keys. Now, What I am looking is to keep the incoming mail
also -it is not doing it now-. How can this be acomplished with procmail?

Please, email me directly. Remove the "no" and the "spam" from my address.

Thanks in advance,


- -
David Collantes, Senior Systems Administrator
College of Business Administration, UCF
Orlando, FL, U.S.A.


=====BEGIN PGP SIGNATURE=====
Version: 6.0.2ckt Get key http://www.netbros.com/pgp/david.asc or
Comment: email <[EMAIL PROTECTED]> with "pgp" (w/o "") on subject
Comment: KeyID: 0x5321A84F
Comment: Fingerprint: E088 574B 15AA 4460  AC0E 6D91 E29B 03A9

iQEVAwUBNuB+vMuBDkNTIahPAQHihAgAlYyaPp185kqDVmdH7re1pVoWIPDP0jxN
E1TKxn/R6Qu5+YGXaIjGHD3PmoPsXRKUhJ7SZ78tiosS7tVT3JkQ22TvXijjW56g
JRMMIVLoUR55h19Ggrlh5ejLujttoylrbM/PaU98c/n/TXsLzpdC8c1FFaGC3sgy
IDhHqeqQjgMLO8egQ40hiFupuybqmyQHW5M/hAoBChmUDJqmyb/2zlmwdBoUhu9G
t7U7uRHO++cnPx6ZzLXCG1l5btjFkSlv4qOHXT2rwaSjUL+B6wfYgT434ge4/mg+
Z2NJKjwtBOpoQzalOO2p5JD7vGL26ItWCIda27gpTZ3HFTxj4PTGKg==
=7uRv
=====END PGP SIGNATURE=====




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

From: [EMAIL PROTECTED] (John David Bowden)
Crossposted-To: nf.comp.linux,comp.os.linux.setup
Subject: Login timeouts in runlvl. 5
Date: 6 Mar 1999 00:30:38 GMT

        I've got this problem which is driving me nuts.  All my logins
(root and otherwise, X and console) timeout in run-level 5.  I've been
using wdm instead of xdm, but switching back doesn't help.  I've tried
clearing /var/log but that won't help (how big is too big for /var
anyway?).  I've looked at /etc/rc.d/rc3.d and ./rc5.d to see if anything
different starts up - they're the same.

        Any help would be great.

John

--
Clones are people two.
******************************************************************************
John Bowden                             Memorial University of Newfoundland 
[EMAIL PROTECTED] 
******************************************************************************

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

From: [EMAIL PROTECTED]
Subject: xdm problems - can't shutdown
Date: Sun, 07 Mar 1999 02:38:44 GMT

Linux: Red Hat 5.1
Computer: 486dx2/66, 20MB ram

When I activate xdm it works properly, except that I can not shutdown the
computer. Logged in as root, I type shutdown -h now in a command shell and the
following messages appear:
(loosely paraphrased)
Broadcasting ... (my memory fails me right now)
Starting shutdown procedure ...

then the command prompt comes up again.

halt freezes the computer, and after turning the power off and on the disk
needs to be checked in the boot up procedure.

init 0  starts the shutdown procudure, but seems to get stuck at shutting down
gpm mouse services.
init 3 has no effect (I heard that this would put me back in regular mode so
that I could exit X back to the text based interface).

So, if I run xdm then how do I shut down the computer?
How can I turn off xdm once it is activated without rebooting the computer?

This is my first installation of linux and I have not made any weird
modifications... the system should be close to original.

Thanks in advance.

Stephen

============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

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

From: "Geoff Thornton" <[EMAIL PROTECTED]>
Subject: QT build problems
Date: Sat, 6 Mar 1999 10:49:16 +1000

Has anyone else had a problem building qt-1.42 with egcs-1.1.1 and with gcc
2.8.1 and gcc 2.7.23 ? during the make I get the following error displayed,
indicating MOC which is built during the install has segment faulted:

/usr/local/qt/bin/moc dialogs/qfiledialog.h -o dialogs/moc_gfiledialog.cpp

make [1]: ***[dialogs/moc_qfiledialog.cpp] Segmentation fault (core dumped)

make [1]: ***[deleting file 'dialogs/moc_qfiledialog.cpp']

make [1]: leaving directory 'usr/local/qt/src'

make [1]: ***[src] Error 2

I have read troll techs FAQ and it said to try optimisation during the make
to -O0, but this didn't help, anyone got any ideas or solutions?




Regards Geoff :)
==================
[EMAIL PROTECTED]



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

From: Kalle Olavi Niemitalo <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: Need to capture the 'raw' mouse under X
Date: 01 Mar 1999 12:54:46 +0200

[Followup-To .development.apps only]

"G. Georgiev" <[EMAIL PROTECTED]> writes:

>       I need to read the raw mouse data from /dev/ttyS0 while under X
> for a signature-capture purposes.

GPM has a repeater mode where it reads the mouse and copies the
data to a pipe (or was it a socket?) which X can be configured to
read.  Your program could do something similar, or you could
patch GPM to pass the signatures to your program.

This would make your program harder to install, though.

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

From: "Jon Wiest" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.linux.advocacy,alt.linux
Subject: Re: More bad news for NT
Date: Sat, 6 Mar 1999 22:02:45 -0600

Jason Clifford wrote in message ...
>There is nothing in the Windows 95 interface that was not present in the
>Apple Mac interface in 1991 and earlier. The GUI itself was invented and


Yes, and?  This is such a tired old rant.  Just what is your point?  Are you
against widespread adoption of a GUI?  Does it bother you that someone made
money using a public idea?  Better get ready to shit all over RedHat.  With
all the deals they've made with Compaq, Dell and IBM you can bet they are
going to make it big.  Linux will never be the same.

Or maybe you think MS should offer Mac some royalties.  Of course, then
Xerox would be the most fabulously rich company in the world.

Did McDonalds dream up the first hamburger?  Who made the first taco chip?
Should no one do anything ever again because someone else did it first??

Get a grip.

Jon




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

From: [EMAIL PROTECTED] (Rulecoyote)
Subject: mount win95 16 (LBA)
Date: 7 Mar 1999 03:59:28 GMT

I thought I had my answer to how to mount my windows partition but when I used
the command mount -t vfat and so on and several versions thereof I got the
message that vfat was not supported by the kernel. What next?
    
                                        oldcoyote

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

From: [EMAIL PROTECTED] (J Crowtz)
Subject: A question about MP3's and Linux
Date: Sat, 06 Mar 99 01:45:47 GMT

Hello all,

 Having recently picked up an old SoundBlaster16 sound card, I am going 
to install it in my computer so I can listen to some music.  Mainly, I 
would like to play MP3s.  However, my CPU is a 486SX, running at 25mhz.  
Though I have Redhat 5.0 on my system, and I know there are many MP3 
players for Linux, will I still be able to play MP3s with no problems?
(I also have 20 megs of ram and 16 megs swap space, if that makes any 
difference.)

please respond to:      [EMAIL PROTECTED]

Thank you for your time, 

- Jason

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


** 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
******************************

Reply via email to