Linux-Hardware Digest #120, Volume #10 Thu, 29 Apr 99 13:13:33 EDT
Contents:
Re: Programmers are gods (Craig Kelley)
Re: ATI Xpert 98 AGP problems in Xwindows (Neil Steadman)
ATI Mach64 tiny HOWTO (Neil Steadman)
Re: X Windows problems ([EMAIL PROTECTED])
3COM V90 Modem Problem ("Anil Rao")
Re: ps/2 mouse (**Nick Brown)
Re: SB 16 *PCI* ("Jose M. Urena")
Re: Q: ABIT vs ASUS vs GIGABYTE for linux (Jay Ribak)
Re: halt -p not working??? help (Timothy Riehle)
Re: Programmers are gods (westprog)
Re: Programmers are gods (westprog)
Re: Programmers are gods (Chris Costello)
Re: SMP : How do I know if it's working ok. Use top? (Christophe Kumsta)
USB status question... (Stephen Quattlebaum)
----------------------------------------------------------------------------
Crossposted-To: comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Programmers are gods
From: Craig Kelley <[EMAIL PROTECTED]>
Date: 29 Apr 1999 06:39:16 -0600
westprog <[EMAIL PROTECTED]> writes:
> In the context of a modern language, the runtime error would probably be an
> exception, catchable at a higher block, but that isn't Dijkstra's concept.
>
> C, it isn't.
Strangely enough, running Perl with 'use diagnostics' and 'use strict'
will get you that type of behavior:
my $var;
{ some block here}
foo($var);
It will complain in $var is still undef at that point.
Of course, you throw out most type-checking with perl...
(But then again Perl is smart enough to know what type you *meant*
to use) :-)
--
The wheel is turning but the hamster is dead.
Craig Kelley -- [EMAIL PROTECTED]
http://www.isu.edu/~kellcrai finger [EMAIL PROTECTED] for PGP block
------------------------------
From: [EMAIL PROTECTED] (Neil Steadman)
Subject: Re: ATI Xpert 98 AGP problems in Xwindows
Date: Tue, 27 Apr 1999 16:12:18 +0059
Reply-To: [EMAIL PROTECTED]
On Mon, 26 Apr 1999 13:26:36 -0400, Robert August Vincent II wrote:
>
>I'm confused. I've got an ATI Xpert@Play 98 AGP, and it's strictly 2D.
>I'm using the Mach64 X server with generic ATI chipset support.
AFAIK your ATI board has it's own 3D accelerator. I've played GLQuake &
Quake 2 under Win95 using the Xpert98 OpenGL extensions. Granted it's slower
than Voodoos, but Quake2 was playable and looked great!
>Also have a 3D add-on board (Diamond Monster3d, IIRC) and got Mesa-GL and
>Quake-III up with a little tweaking. (had to download and recompile Mesa).
>
>Runs up to 1280x1024 at 32bpp, and 800x600 in 3D mode
Which is the proof I needed that a voodoo accelerator wouldn't clash with
the Xpert98, but I'd still rather get the ATI running 3D than buy another card.
--
How many shopping days till Xmas?
------------------------------
From: [EMAIL PROTECTED] (Neil Steadman)
Subject: ATI Mach64 tiny HOWTO
Date: Tue, 27 Apr 1999 16:47:47 +0059
Reply-To: [EMAIL PROTECTED]
In an effort to remove the countless postings on:
"How do I get my ATI... to work?"
I offer the following for Mach64 based ATI users. (Though I reckon
this will work with other programmable clockchip cards).
I'll assume you have XFree86-Mach64 installed.
I'll assume you have Xconfigurator.
Run it as root.
I'll assume the first thing it says is that it probed for the right card.
The next screen is crucial. You must set your monitor for the correct
Horizontal / Vertical rates. If your monitor isn't listed select
'custom' and choose whatever looks the nearest to what you've got.
Then select the correct Hz / Vert rates if asked.
If you don't know your monitor rates, check the back of the monitor and
do an internet search for it's make & model. You should find something.
(I used '+Sony +GDM +"17e10" +rate' in altavista).
Next you get screen configuration. Select "DON'T PROBE"
Choose the correct amount of video RAM.
At the clockchip selector select "No clockchip setting"
The next screen offers to probe for clocks. Select "SKIP"
Now choose your video modes. Since this is a test run, choose all that you
know your monitor / card can cope with. Be warned X will startup in the
lowest resolution possible. This is a good thing since that's most likely
to work correctly. You can shift up and down resolutions with CTRL+ALT+'keypad+'
and CTRL+ALT+'keypad-'. If you get in a mess, bail out with CTRL+ALT+Backspace.
Now it's working, either run Xconfigurator again and select the resolutions you
actually want, or edit /etc/X11/XF86config with your favorite editor.
If people mail me regarding video cards that work using the above, I'll add a
list of cards next time I post this, and change the Mach64 bit...
(When the 'How Do I...' count goes up again).
Good Luck
Neil.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: X Windows problems
Date: Thu, 29 Apr 1999 11:04:51 GMT
http://www.kom.auc.dk/~sonic/pres_howto.html
http://www.kom.auc.dk/~sonic/presario.html
Video and Monitor - The video card is an ATI Rage LT Pro and the monitor is a
compaq MV500. I've got it working in 1024x768, 8 bit for the compaq screen
MV500, after choosing the custom screen in a non-interlaced mode at 60 Hz.
For this select ONLY THE highest resolution wanted (1024x768) after the
custom screen was selected. (cf my XF86config file ) The key for XFree86
3.3 (RedHat 5.1 or 5.2) seems to be to put ChipId 0x4754 and ChipRev 0x01 in
the Device section of your XF86Config file.
In article <7f6134$im2$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> I'm trying to install Redhat Linux 5.2 on my Compaq Presario 1670, which has
> an AMD K6 350Mhz Processor, 64MB RAM and 8MB of VRAM. The video card is an
> ATI Technologies RAGE LT Pro 2+, which is listed at www.xfree86.org as using
> the same server (driver?) as other ATI cards in the mach64 series.
>
> But the exact card isn't listed during Xconfigurator. And I can't run X
> Windows properly. The command startx does get me to X, but the screen is
> split in half, and there's no echo in the command box and various other
> problems.
>
> I tried installin it as a generic VGA16 and changing the vertical and
> horizontal refresh rates.
>
> I would really appreciate some help.
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
>
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: "Anil Rao" <[EMAIL PROTECTED]>
Subject: 3COM V90 Modem Problem
Date: Wed, 28 Apr 1999 07:48:14 -0700
Reply-To: "Anil Rao" <[EMAIL PROTECTED]>
I am running RedHat Linux 5.2.
I am having a problem getting Linux to recognize my modem. I tried to
configure my network using linuxconf and netcfg, but in all cases I was not
able to dial out. I tried devices /dev/cua0, /dev/cua1, /dev/cua2, and
/dev/cua3.
I have a 3COM V90 Fax/Modem. It's plug and play, but it is NOT a winmodem.
I was under the impression that Linux supported att non-winmodems.
If anyone can tell me the steps I need to take to get my modem recognized in
Linux, I would be highly appreciative.
Thanks in advance.
Anil
------------------------------
From: **Nick Brown <[EMAIL PROTECTED]>
Subject: Re: ps/2 mouse
Date: Thu, 29 Apr 1999 16:59:38 +0200
Reply-To: [EMAIL PROTECTED]
Have you tried with another mouse ? I've seen at least two mice which
claimed to be compatible with both the PS/2 and COM ports, but weren't
so in practice.
> Whwn linux boots the messages say that the 'aux' device has been
> detected and that the driver is being installed but try as I might
> (changing /dev links, Xfree setup etc.) I cannot get linux to use it. It
> does not even see a ps/2 mouse plugged into the extenal port. All works
> fine using the same mouse plugged into the COM port as a serial mouse.
--
===============================================================
Nick Brown, Strasbourg, France (Nick(dot)Brown(at)coe(dot)int)
Protect yourself against Word 95/97 viruses, free - check out
http://www.geocities.com/NapaValley/Vineyard/1446/atlas-t.html
===============================================================
------------------------------
From: "Jose M. Urena" <[EMAIL PROTECTED]>
Subject: Re: SB 16 *PCI*
Date: Thu, 29 Apr 1999 11:02:23 -0400
if it is PCI, it might have SB16 support or compatibility, but I do not
think it is a SB16.
when you run a dos game, do you have to load a TSR before running the
game?
if you do, then the sound card is using an emulator to give you the SB16
features
if you are only running windows software, well, the API calls are being
translated to give you the appearance of SB16
consider the SB64 PCI, it is a 16 bit sound card with the ability to
play 64 voices. Under windows 95/98/NT it gives you full SB, SBPRO, and
SB16 sound in addition to MIDI support
under DOS, you need to load the SB16 emulator before you play any game,
and since the MIDI support is not in the hardware but only through the
windows software, you end up with no midi sound under DOS
Simon Holgate wrote:
> [EMAIL PROTECTED] wrote:
> >
> > Are you sure it's SB16 ?? Most likely it's SB64 and it uses
> > Ensonique AudioPCI chipset - you need completely different
> > drivers for it under Linux.
>
> Yep, this really is an SB16 PCI. Brand new, out of the box and
> everything :)
>
> Simon
> --
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Simon Holgate, Tel: (+44) (0151) 794 4102
> Department of Earth Sciences, (Rm 114, Oceanography)
> University of Liverpool, Geology - 794 5202
> P.O. Box 147, (Rm B17, Herdman Bldg)
> Liverpool, L69 3BX E-mail: [EMAIL PROTECTED]
> ENGLAND. http://george.seos.uvic.ca/people/simon/simon.html
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
------------------------------
From: Jay Ribak <[EMAIL PROTECTED]>
Subject: Re: Q: ABIT vs ASUS vs GIGABYTE for linux
Date: Thu, 29 Apr 1999 11:06:56 -0400
I have successfully used the ASUS P2B boards with Linux. I had two of
the Gigabyte
boards and returned them--I thought they absolutely sucked. The board
layout was very
poor and it just felt like it wasn't very well constructed. The Socket
7 socket is
so close to the DIMM slots it takes an act of extreme force to get one
of the larger
fans installed on the socket. Also, the front of the board has no
mounting holes for
screws or standoffs, so you have to rig up some kind of insulated
standoff to get the
board from bouncing off of the metal case and shorting ;) In all, it
just wasn't
a very enjoyable experience with the Gigabyte. My distributor also said
that he didn't
care for those boards very much either. It seems from other responses
in this thread
that the Asus is the way to go...
Jay R.
Val Jogai wrote:
>
> I am planning to build a pentium linux box and have narrowed my choice of
> motherboards to three:
> ABIT BX6 (REV 2.0)
> ASUS P2B-F
> GIGABYTE GA-6BXE
>
> I am fearful of making the wrong choice. Could someone comment on the
> relative merits of these boards for linux?
>
> --
> Val Jogai
> jogai+AEA-infinet.com
------------------------------
From: Timothy Riehle <[EMAIL PROTECTED]>
Subject: Re: halt -p not working??? help
Date: Thu, 29 Apr 1999 10:45:03 -0500
Matthew,
I guess you didn't take my advice to do a quick web search.
AltaVista just pulled up this link for me:
http://www.diku.dk/students/grove/poweroff.html
Read it, it explains why you need to edit /etc/rc.d/init.d/halt as I
detailed earlier. linuxhq has a brief blurb mentioning that the newest
sysklogd will support the -h feature. I had the same problems when I
switched to the 2.2 kernels and this did the trick for me, YMMV.
Tim
On 29 Apr 1999, Matthew Pound wrote:
> Timothy Riehle ([EMAIL PROTECTED]) wrote:
>
> : Hi Matthew,
>
> : Is this with a 2.2 kernel? If so you probably need to change your
> : /etc/rc.d/init.d/halt script. A USENET or web search should quickly find
> : you the info you need. I (from memory here!) changed the actual command
> : used to halt the machine in the halt script to "halt -p".
>
> : Tim
>
> No, the problem isn't changing the halt script (or knowing how to
> halt a machine like the first dumbass sugested) the problem is "halt -p",
> which should powerdown my machine after shuting down, doesn't do the
> poweroff, it just halts like "shutdown -h now" does. I have recompiled
> the kernel (2.2.6) with advanced power management, and support for using
> "halt -p" but it does not do what it says it is supposed to. And, yes, i
> have this feature turned on in my bios. later.
>
------------------------------
From: westprog <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Programmers are gods
Date: Thu, 29 Apr 1999 11:57:12 GMT
In article <[EMAIL PROTECTED]>,
Craig Kelley <[EMAIL PROTECTED]> wrote:
> westprog <[EMAIL PROTECTED]> writes:
>
> > A language that required that variables be declared, typed, initialised and
> > modified in that order would catch a variety of bugs at compile time which
> > otherwise would get into the running program. In the world of C, all kinds
of
> > tools and programming disciplines have been applied to try to fix this
> > problem. Why not let the compiler do it?
>
> As much as that would be nice, it wouldn't be "C" as we know it today:
Well, that would be a major plus.
> {
> int x;
> switch (y)
> {
> case 1: x = 1;
> break;
> case 2: x = 4;
> break;
> case 3: x = 5;
> }
> foo (x);
> }
>
> That is legal, and there's know way the compiler can know if 'x' has
> been initialized.
Dijkstra's solution to this is to insist that his if/fi structure (broadly
the same kind of thing as C's switch or Pascal's case) will generate an error
if at least one branch is not executed - he would have something like:
virvar x; { variable which we are obliged to initialise }
glocon y; { constant which already has a value }
if y = 1 -> x vir int := 1;
| y = 2 -> x vir int := 4;
| y = 3 -> x vir int := 5;
fi
foo(x); { Perform an operation with the guaranteed initialised x }
This is identical to the C code above in functionality, with the single
difference that if no branches of the if are executed, a runtime error is
generated. It is not possible for foo to be called on an uninitialised
variable. This can be verified by the compiler at compile time.
N.b. Dijkstra achieves this certainty by making an explicit distinction
between assignment and initialisation.
In the context of a modern language, the runtime error would probably be an
exception, catchable at a higher block, but that isn't Dijkstra's concept.
C, it isn't.
J.
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: westprog <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Programmers are gods
Date: Thu, 29 Apr 1999 12:02:47 GMT
In article <[EMAIL PROTECTED]>,
"Clifton T. Sharp Jr." <[EMAIL PROTECTED]> wrote:
> westprog wrote:
...
> > A language that required that variables be declared, typed, initialised and
> > modified in that order would catch a variety of bugs at compile time which
> > otherwise would get into the running program. In the world of C, all kinds
of
> > tools and programming disciplines have been applied to try to fix this
> > problem. Why not let the compiler do it?
>
> It would be nice, but I doubt a compiler in any usable language could catch
> all the possible scenarios.
I think that it would be quite easy to guarantee that variables are
initialised before being used. There would still be plenty of scope for
errors - just that particular problem would never happen. There would be
nothing to stop lazy programmers routinely initialising variables to some
meaningless value.
> Besides, programmers don't want that. They
> want "features" like type overloading, so that instead of the compiler
> gronking about a function return mismatch it just calls the wrong function
> instead.
I think that programmers are changing. They have moved from C to ANSI C to
C++ to Java - each step involves a few more restrictions and a bit more
discipline.
> And even after all that, the language you describe would *still* need lots
> of documentation for standards, for obscure algorithms, for Cute Programmer
> Tricks, etc. etc.
Something I dislike about programming in the 90's is that there is very little
language neutral code published. All Windows and Unix examples are in C, (not
even C++ usually) and translating to another language is non-trivial.
J.
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (Chris Costello)
Crossposted-To: comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: Programmers are gods
Date: Thu, 29 Apr 1999 16:18:18 GMT
In article <7g9hp6$820$[EMAIL PROTECTED]>, westprog wrote:
> I think that programmers are changing. They have moved from C to ANSI C to
> C++ to Java - each step involves a few more restrictions and a bit more
> discipline.
They aren't moving to them, they're expanding to them. For
example...
#include <stdio.h>
int
main(void)
{
char helloworld[13];
memset(&helloworld, 0, sizeof(helloworld));
sprintf(helloworld, "%s %s%c%c", "Hello", "World", '!', '\n');
printf("Smokey the bear says: %s", helloworld);
exit(0);
}
Stripped C binary: -rwxrwxr-x 1 chris chris 2596 Apr 29 11:09 hw
* This file is 2596 bytes long.
Now C++:
#include <cstdio>
#include <string>
int
main(void)
{
char helloworld[13];
memset(&helloworld, 0, sizeof(helloworld));
sprintf(helloworld, "%s %s%c", "Hello", "World", '!');
cout << helloworld << endl;
exit(0);
}
Stripped C++ binary: -rwxrwxr-x 1 chris chris 32500 Apr 29 11:14 hw_cpp
* THIS file is 32,500 bytes long!
I compiled the C binary as follows:
gcc -O -o hw_c hw.c
g++ -O -o hw_cpp hw.cpp
strip hw_c
strip hw_cpp
I don't have the Java development kit installed so I can't
provide you with a file size for that.
>
> > And even after all that, the language you describe would *still* need lots
> > of documentation for standards, for obscure algorithms, for Cute Programmer
> > Tricks, etc. etc.
>
> Something I dislike about programming in the 90's is that there is very little
> language neutral code published. All Windows and Unix examples are in C, (not
> even C++ usually) and translating to another language is non-trivial.
>
> J.
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
--
Chris Costello
There is no problem that, when programmed just right, isn't more complicated.
------------------------------
From: Christophe Kumsta <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.periphs.scsi
Subject: Re: SMP : How do I know if it's working ok. Use top?
Date: Thu, 29 Apr 1999 12:17:36 +0000
Look at : http://i5.ima.umn.edu/~klee/linux-smp.shtml
for patched xosview .
( I use it on a dual-PIII and can see correctly IRQs and CPU Load on
both processors
with RedHat5.2 and kernel v2.2.0 with RT-Linux v2D patch.)
laurent collot wrote:
>
> Curt Timmerman wrote:
>
> > If you run xosview, it will show the load and interrupts on both CPU's.
> >
>
> xosview is broken with newer 2.2.x SMP kernels since the proc reporting changed
> -> endless loop...
>
> Laurent Collot
-- KUMSTA Christophe
-- <[EMAIL PROTECTED]>
-- real-time system developper
-- RT-Linux (Use the source luke)
------------------------------
Date: Thu, 29 Apr 1999 10:52:57 -0500
From: Stephen Quattlebaum <[EMAIL PROTECTED]>
Subject: USB status question...
First: Does anyone know of the status of the USB support for Linux? I
have a USB Zip drive and I'm dying for the day that I can use it in Linux.
Thanks.
****
Stephen Quattlebaum
[EMAIL PROTECTED]
Death multitasks. Someday, your timeslice will be up.
The opinions expressed by me are not necessarily the opinions of the
University of Alabama or it's faculty/administration (but they should be).
------------------------------
** 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.hardware) 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-Hardware Digest
******************************
