Linux-Development-Sys Digest #22, Volume #8 Sat, 15 Jul 00 01:13:07 EDT
Contents:
Re: GetModuleFileName() in linux (John Reiser)
Re: problem of little indian to big indian code (Grant Edwards)
Re: problem of little indian to big indian code ([EMAIL PROTECTED])
Re: Kernel Panic Recovery ([EMAIL PROTECTED])
Re: Microsoft's new ".NET" (Nix)
Re: Install new gcc (Nix)
Re: pure virtual function. (Nix)
Re: warning on volatile ([EMAIL PROTECTED])
Re: pure virtual function. (Kaz Kylheku)
Re: [ANN] Beta testing of CW for Linux on Intel and PowerPC (Christopher Browne)
virtual time alarm? (Elvis Chen)
Raw disk access and the 'raw' program (Myke Morgan)
shm module not found (Justo M. Casablanca)
Re: Login problem with Lineo ("Phillip Soltan")
Re: A good distribution anybody? (Taketoshi Sano)
----------------------------------------------------------------------------
From: John Reiser <[EMAIL PROTECTED]>
Subject: Re: GetModuleFileName() in linux
Date: Fri, 14 Jul 2000 12:17:23 -0700
On Linux, the pathname at the time of exec is put at the fixed end
of the stack. On many common Linux x86 distributions where the stack ends
at 0xc0000000, the 32-bits at 0xbffffffc is 0, and the '\0' which terminates
the pathname at time of exec is at 0xbffffffb. Search backwards for another
'\0', and the pathname lies between.
--
------------------------------
From: [EMAIL PROTECTED] (Grant Edwards)
Subject: Re: problem of little indian to big indian code
Date: Fri, 14 Jul 2000 19:21:12 GMT
In article <dupjk8.kuq.ln@localhost>, Sian Leitch wrote:
> sgriobh bruno.patin <[EMAIL PROTECTED]>:
>
>>I have an application that creates data on another platform than Linux
>>and when we read these datas on the linux box, we have to work on the
>>order of the bytes by eight block count. Is there a mean to do the work
>>automatically by options on the opening of the file ? where can i find
>>answers on this subject ?
>
>(1) It's not "indian", but "endian", the "end" part referring to
>those people who open their boiled eggs at a particular end (yes,
>really!).
>
>(2) What does "order of the bytes by eight block count" mean?
>
>(3) Generally speaking, reversing endian-ness has to be done by
>individual programs. While it is true that X servers cater for
>different endianness of the client computer, other programs seem not
>to do so. So you are probably on you own in this. However, here is a
>procedure to reverse endianness of integers:-
If I were you, and I wanted to do portable binary stuff
(there's an oxymoron), I'd always write things in "network
order". There are already macros available to convert between
host and network byte order: ntohs(), ntohl(), htons(),
htonl(). These macros get used a _lot_ in network stacks and
are probably as optimal as you can get. Unfortunately, they're
only available for "short int" and "long int". Network geeks
don't care about floats and doubles (or long long ints,
apparently).
Way back when, Mark Williams Co. used to sell a v7 clone called
Coherent which ran on various processors (8086, 68K, etc.).
They had a patent on storing data on disk in a standard byte
order. So, if you write data to disk in network order you owe
them royalties -- perhaps it's expired. I don't know that they
ever tried to enforce it, and MWC is, I believe, out of
business anyway.
--
Grant Edwards grante Yow! YOW!!! I am having
at fun!!!
visi.com
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: problem of little indian to big indian code
Date: Fri, 14 Jul 2000 20:11:40 GMT
Grant Edwards <[EMAIL PROTECTED]> wrote:
> If I were you, and I wanted to do portable binary stuff
> (there's an oxymoron), I'd always write things in "network
> order". There are already macros available to convert between
> host and network byte order: ntohs(), ntohl(), htons(),
> htonl(). These macros get used a _lot_ in network stacks and
> are probably as optimal as you can get. Unfortunately, they're
> only available for "short int" and "long int". Network geeks
> don't care about floats and doubles (or long long ints,
> apparently).
You _could_ use something like xdr (man 3 xdr, then run screaming for
a good book on rpc ;). However, be prepared for some inflation. I
believe xdr requires four bytes per unit.
--
Matt Gauthier <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Kernel Panic Recovery
Date: Fri, 14 Jul 2000 13:10:30 -0700
On or about Fri, 14 Jul 2000 13:30:07 GMT, Justo M. Casablanca <[EMAIL PROTECTED]>
scrivened:
> Hi all, I recently placed a "mem=128M" option to linux while booting,
> and got a kernel panic error message. What I want to know is, it safe
> to hit the reset button on my PC at this point ? I did, and it rebooted
> into Linux just fine, but I need to know. If it is not OK to hit reset,
> what do I do ? I don't seem to have any control over the text console
> once the kernel panic happens.
Power-cycling is your main option. If you haven't mounted filesystems,
there is little (probably no) risk to your system.
You can compile in the magic-alt-sysrq option to your kernel which
provides some minimal functionaliity to a hosed system, including
flushing buffers, remounting disks r/o, unmounting disks, killing
processes, and restarting the sytem. This *should* allow you to close
out a system cleanly even when it pigs in a major way while running.
--
Karsten M. Self <[EMAIL PROTECTED]> http://www.netcom.com/~kmself
Evangelist, Opensales, Inc. http://www.opensales.org
What part of "Gestalt" don't you understand? Debian GNU/Linux rocks!
http://gestalt-system.sourceforge.net/ K5: http://www.kuro5hin.org
GPG fingerprint: F932 8B25 5FDD 2528 D595 DC61 3847 889F 55F2 B9B0
------------------------------
From: Nix <$}xinix{[email protected]>
Subject: Re: Microsoft's new ".NET"
Date: 14 Jul 2000 21:59:53 +0100
[desperately trying to drag this thread on-topic]
Tim Roberts <[EMAIL PROTECTED]> writes:
> Microsoft wrote the specification, but they don't write BIOSes. There's
> plenty of blame to go around.
Agreed; BIOSen are astoundingly broken, in the main, as are motherboard
designs (viz the interrupt pins on an LM78, and their
non-hooked-upness).
Linux is sometimes/often shot down by them. The 1024-cyl boot limit and
similar are design faults, but (e.g.) the propensity of systems to crash
on powerdown instead of powering down is plain and simple incompetence
on the part of the BIOS implementors. There are flat protected mode
entry points to APM; they just hadn't been tested worth a damn.
--
`I can guarantee it's no problem in my network, and if I don't get some
sleep soon, I'll guarantee it will become a problem in your network.'
--- Chris `Saundo' Saunderson deals with a late-night phone call
------------------------------
From: Nix <$}xinix{[email protected]>
Subject: Re: Install new gcc
Date: 14 Jul 2000 22:01:56 +0100
[EMAIL PROTECTED] (Grant Edwards) writes:
> If you want to know what version is being used on the current
> system do:
>
> $ gcc --version
This doesn't work with versions <=2.7.x. `gcc -v' always works (well,
back into the stone age at least).
--
`I can guarantee it's no problem in my network, and if I don't get some
sleep soon, I'll guarantee it will become a problem in your network.'
--- Chris `Saundo' Saunderson deals with a late-night phone call
------------------------------
From: Nix <$}xinix{[email protected]>
Subject: Re: pure virtual function.
Date: 14 Jul 2000 22:12:42 +0100
[EMAIL PROTECTED] (Kaz Kylheku) writes:
> GNU C++ has an extension known as signatures which provides additional
> flexibility in this regard; it allows any class which implements
> some functions having certain names and type signatures to be
> accessed through the signature, even though it's not derived from it.
This extension will very probably go away. It makes the grammar
horrible, and it provides nothing that is not already provided by
function templates (think about it: they, too, are interface enforcement
mechanisms) or abstract virtual bases with all members pure.
> >I already solved this problem. I just implemented a dummy function to
> >replace it as below:
> >
> >extern "C" int __pure_virtual()
> >{
> > /* do nothing */
> > return 0;
> >}
*sick* if you do this, it shows that you *definitely* do not understand
C, and probably not OOP :(
--
`I can guarantee it's no problem in my network, and if I don't get some
sleep soon, I'll guarantee it will become a problem in your network.'
--- Chris `Saundo' Saunderson deals with a late-night phone call
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: warning on volatile
Date: Fri, 14 Jul 2000 22:03:40 GMT
I am using the struct very often and it represents an area of memory.
I see my mistake.
Thanks
regards
jeseem
In article <[EMAIL PROTECTED]>,
Erik Max Francis <[EMAIL PROTECTED]> wrote:
> [EMAIL PROTECTED] wrote:
> >
> > Hi,
> > i have defined a struct as volatile
> > typedef volatile struct xxx{
> > int i;
> > }X, *PX;
>
> I don't believe what you're trying to do has much meaning. volatile
is
> a characteristic of variables, not types. gcc refuses to compile such
> calls with all warnings on in ANSI pedantic mode.
>
> What is it that you're trying to do?
>
> --
> Erik Max Francis / [EMAIL PROTECTED] / http://www.alcyone.com/max/
> __ San Jose, CA, US / 37 20 N 121 53 W / ICQ16063900 / &tSftDotIotE
> / \ Who'd ever think it / Such a squalid little ending
> \__/ The American/Florence, _Chess_
> Interstelen / http://www.interstelen.com/
> A multiplayer, strategic, turn-based Web game on an interstellar
scale.
>
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Subject: Re: pure virtual function.
Reply-To: [EMAIL PROTECTED]
Date: Fri, 14 Jul 2000 22:56:57 GMT
On 14 Jul 2000 22:12:42 +0100, Nix <$}xinix{[email protected]> wrote:
>[EMAIL PROTECTED] (Kaz Kylheku) writes:
>
>> GNU C++ has an extension known as signatures which provides additional
>> flexibility in this regard; it allows any class which implements
>> some functions having certain names and type signatures to be
>> accessed through the signature, even though it's not derived from it.
>
>This extension will very probably go away. It makes the grammar
>horrible, and it provides nothing that is not already provided by
>function templates (think about it: they, too, are interface enforcement
>mechanisms) or abstract virtual bases with all members pure.
I understand. I don't really see the point of signatures. The mere presence of
functions that have certain names and type signatures does not an interface
make! If I understand it correctly, if I have void Lottery::Draw() and void
Widget::Draw(), then the signatures feature lets me use the Draw function of
both classes as common interface, even though neither class declared any
conformance to such an interface. This could result in severe semantic errors.
--
#exclude <windows.h>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.powerpc
Subject: Re: [ANN] Beta testing of CW for Linux on Intel and PowerPC
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 Jul 2000 00:19:42 GMT
Centuries ago, Nostradamus foresaw a time when
[EMAIL PROTECTED] would say:
>In comp.os.linux.development.system bill davidsen <[EMAIL PROTECTED]> wrote:
>
>| | Even more to the point, "these are the only supported versions of
>| | linux". Like, get real. Give me the source openly and I'll compile it
>| | *myself* on a Debian GNULinux/PPC setup and check that it works - which
>| | effectively does some 'work' *for* the company. Fail to cater for /all/ my
>| | tastes in platform, both now and in the future, by introducing arbitrary
>| | restrictions for no good reason, and I'll not even look at it.
>|
>| I don't think there's a restriction there, simply a list of supported
>| distributions. If you want to run it on Slackware (I do!) or some other
>| distribution the only restriction is that they don't promise that it
>| will work.
>|
>| That's not a restriction, it's common sense. Back when we had SLS and
>| Slackware as the only two significant distributions you could test on
>| all of them. Not true any more. Don't read "not supported" for "can't
>| work."
>
>But why should they limit themselves to several distributions while
>leaving out a couple of the major and important ones?
- Because they didn't consider what you consider "major and important"
to be, from their perspective, "major and important"?
- Because the process of making sure the product _works_ on another
distribution platform requires a reasonable amount of effort that
they don't have staff that they feel like putting into the effort?
>Distribution dependency could be as much of a difficulty as the kind
>of OS dependency we have to deal with now. Granted a common kernel
>helps a lot, but I think commercial software developers in general
>are still poor at making things truly portable. Making something that
>works on Redhat also work on Slackware is a far cry easier than making
>it work on Windows 2000, and Slackware is still a major distribution
>with lots of users, so why not?
Does Patrick Volkerding meet with folks from Oracle, Informix, IBM,
and Dell to set up "marketing collateral?"
The _critical_ issue when dealing with organizations that are even the
_slightest_ bit "marketing-driven" is to have a _MARKETING_ portion of
your organization to go "do lunch" with people from everyone else's
"Marketing Divisions." And if you don't do that, _you don't exist._
Debian only "exists" because VA Linux Systems, Storm Linux, and Corel
Linux have had some people "doing lunch," because there are a few
celebrities (e.g. - Neal Stephenson), and because there are _some_
connections to FSF "politicking/marketing" efforts. Individually,
these "connections" aren't of much significance, but they're more than
Slackware has, in the "post-Walnut Creek" era...
--
[EMAIL PROTECTED] - <http://www.hex.net/~cbbrowne/>
"When we understand knowledge-based systems, it will be as before
-- except our fingertips will have been singed." -- Alan Perlis
------------------------------
From: Elvis Chen <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.mandrake
Subject: virtual time alarm?
Date: Fri, 14 Jul 2000 20:45:57 -0400
greetings,
I have posted previously to report that my dual-cpu system, when installed
with Mandrake 7.1 (kernel 2.2.16-9), is NOT stable under heavy load. The
same system was, however, rock stable with Mandrake 7.0. I ran some more
tests today and got a bit more clues.
I ran some in-house program that was written in C/C++. The same program
ran fine in another SMP-PC (with Mandrake 7.0). However, when I ran it
under Mandrake 7.1, sometimes it finishes, but sometimes it is terminated.
When it is terminated, I got an error message:
"virtual time alarm".
Can someone please give an insight of what this error message means?
Here is my configuration:
Gigabyte 6BXDS (dual SLOT 1, with adaptec UW SCSI)
dual celeron 366
Mandrake 7.1, clean installed with kernel updated to 2.2.16-9
ReiserFS with software RAID (stripping)
XFree86-4.0 (but my computer crashes with/without X)
any help is very much appreciated,
Elvis
------------------------------
From: [EMAIL PROTECTED] (Myke Morgan)
Subject: Raw disk access and the 'raw' program
Date: 15 Jul 2000 01:18:46 GMT
Someone in this group recently pointed me to the 'raw' program and using
it to access disk drives in 'truly raw' mode. It works great. But then I
tried the program on other machines and I've had various problems.
It works on Red Hat 6.2 with 2.2.14-5.0smp kernel. However, on RH 6.1
(2.2.12-20smp) and RH6.2 with 2.2.16 kernel it does not work.
2.2.12 says:
% raw -qa
Cannot open master raw device '/dev/raw' (No such file or directory)
2.2.16 says:
% raw -qa
Cannot open master raw device '/dev/rawctl' (No such device)
I know the 2.2.16 system is kind of flakey (not installed by me). But the
2.2.12 system seems like it should work.
The only thing I've found in the kernel source is the file
drivers/char/raw.c. This file is in the 2.2.12 source, but not in the
2.2.16 source that I have. Do I need to do something to enable the raw
access? Did I just get lucky with my first system? Possibly, but every
other 2.2.14 machine I have looked on works fine. It doesn't look like it
is a loadable module or anything...
TIA,
myke
--
I proclaim you, FOUR!
------------------------------
From: Justo M. Casablanca <[EMAIL PROTECTED]>
Subject: shm module not found
Date: Sat, 15 Jul 2000 02:30:04 GMT
During the bootup process, I get a message to the effect of "shm module
not found" or "cannot find shm module" or something like that. I am
running Red Hat 6.2, kernel 2.2.14-100. I have not performed any kernel
upgrades since installing the OS from the installation CD.
What is shm ? Do I need it ? If I don't need it, how do I stop Linux
from trying to load it at boot-time ?
Thx in advance !!
Justo M. Casablanca
[EMAIL PROTECTED]
[EMAIL PROTECTED]
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: "Phillip Soltan" <[EMAIL PROTECTED]>
Subject: Re: Login problem with Lineo
Date: Sat, 15 Jul 2000 03:27:05 GMT
You might try turning on the "suid" bit on the "login" file if it isn't
already set (chmod +s login). In any case, you can always look at the
source code to see what the problem might be. You do have the source code
don't you?
Phil Soltan
Sumit Birla wrote in message ...
>Hello all,
> Just got done making a custom boot disk but I am unable to login. The
>following is what I see on the screen -
>
>blackbox1 login: root
>login[35]: invalid password for 'root' on 'tty1'
>
>
>It doesn't even ask for a password. Does anyone know what causes this
>behavior? I have the /etc/passwd, /etc/shadow and all the /lib/libnss_*
>files in place. Do I have to do an ldconfig or something? Also, how do I
>instruct the kernel to boot up in single-user mode?
>
>Thanks.
>
>Sumit
>
>
>
------------------------------
From: Taketoshi Sano <[EMAIL PROTECTED]>
Subject: Re: A good distribution anybody?
Date: 15 Jul 2000 06:41:22 +0900
Hi.
on Sat, 8 Jul 2000 17:41:29 +0200, [EMAIL PROTECTED] (J.H.M. Dassen (Ray)) wrote:
> Of course if you always want the "latest and greatest" versions, you can run
> the Debian "unstable" tree which is available for public download.
Currently there are some CD vendors who make their CD-ROMs from the "unstable"
tree of the Debian, and some of them will do ship international. So you can
buy the CD-ROM from them as well as you can download the distribution from
the ftp site freely.
Check the web page at <http://www.debian.org/distrib/> for details.
--
Taketoshi Sano: <[EMAIL PROTECTED]>
------------------------------
** 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.development.system) 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-Development-System Digest
******************************
