Linux-Development-Sys Digest #303, Volume #6 Tue, 19 Jan 99 05:14:27 EST
Contents:
Thoughts on operating systems; why the Hurd should be continued (Emile van Bergen)
Re: moving linux to different partition (Andreas Schwab)
Re: 2.2.0pre7ac patches (N1ho)
Re: How to run Windows Applications on Linux (Mr. Bill)
Re: - deprecated - why? (Martin Maney)
Re: disheartened gnome developer (steve mcadams)
Re: How to run Windows Applications on Linux (Craig Eaton)
Re: Secuity hole with perl (suidperl) and nosuid mounts on Linux (Peter Samuelson)
Re: K6-2 300 Problem (Marcos Silva)
Re: disheartened gnome developer (Toon Moene)
db question ("Thierry BUCCO")
Re: Linux Sound Engine (Peter Samuelson)
- A FREE down load of Linux Real Time debugging tool (demo version) ("Henry
Mclachlan")
----------------------------------------------------------------------------
From: Emile van Bergen <[EMAIL PROTECTED]>
Crossposted-To: comp.unix.misc
Subject: Thoughts on operating systems; why the Hurd should be continued
Date: Tue, 19 Jan 1999 02:41:50 +0100
Hello,
Let me first introduce this post by stating the reasons why I'd like a
discussion on this topic.
-
One of my main interests in software in general is the operating system.
Why? Because I think that it defines a significant piece of the world in
which programmers (such as myself) live in when writing software, in
which operators live in when maintaining a site, and last but not least,
the environment in which users live when using other software tools.
It is exacly for that reason, that I am also some or less emotionally
attached to the issue. And it is not me alone; the love and hate
expressed in the comp.*.advocacy groups resembles religion more than
science. This is not so strange as it may seem at first; after all, who
doesn't care about what world we (users, operators, programmers) live
in?
But I am not interested in provoking yet another discussion about which
of the currently available operating system is best suited for what kind
of task, who likes what better and why.
Am I getting to the point? Not yet.
Being an engineer, one of my main goals when designing a piece of
software is technical elegance. I can really see aestetic value in a
clear designed and efficient tool. Odd as it may seem to some (yes, I do
have a life, house, fiancee, so don't worry ;-), I'm sure there are many
of you who would agree. To others it may even be obvious!
And it is precisely this property I seek when looking for a tool to use
myself. Because apart from the idea that for something to have this
quality is a virtue by itself (l'art pour l'art), it has occured to me
many times that a tool which has it, has probably also other qualities
that make it more pleasurable to work with. I can't prove this
hypothesis, nor do I have any idea why this may be the case (well,
perhaps the attitude of the engineer who designed it plays a role; see
also the excellent book 'Zen and the art of motorcycle maintenance' by
Robert M. Pirsig ;-), but I really do think that technical elegance and
those other qualities go hand in hand.
In my opninion, this applies to operating systems as well as to any
other tool. And because it also provides the environment the tools I
create will have to interact with, I want the operating system I use as
a foundation for my tools to have this techical elegancy, or even
'beauty' if you will!
So far so good. I'm getting there.
-
In the following part, I would like to divide the operating system in
two: the first part, which provides the user's environment (i.e. the
shell (graphical or not) and the other tools such as editors), and the
second part, which provides services for applications interact with.
The first part is actually worth a discussion in itself; here, however,
I'd like to focus on the second part. (Personally I like the GNU tools
a lot. I think many of them have some or more of this quality, and are
indeed a pleasure to use).
For the second part, I am more or less accustomed to the unix
environment. It may not always be _that_ nice, but most of the times I
am likely to be more comfortable with some unixy system than with most
other systems. (With X being an exception in that the functionality
presented to end users may be very nice, it is not quite that nice to
write software for).
But I think most currently available systems leave much to be desired if
you would look at the technical elegancy of their core parts; most *nix
systems have one very big kernel, a system library, some other libraries
and that's about it. It is not that powerful in that it enables you to
do many different things the kernel builders never thought of!
Unix's concept of small tools is so incredibly powerful to the user
familiar with them; why isn't this same philosophy extended to the part
of the kernel?
I really think this should be done, and I think the Hurd could be an
example how.
-
There is another microkernel based operating system, QNX, in which many
different components play a role in providing a more or less unix-like
environment to applications, and I think it really has some of this
technical elegance I'm looking for.
But it's owning company (yes, it is proprietary software) has apparently
decided it is satisfied with the embedded market; users aren't
encouraged to help it grow.
There is no source code, its components are rather poorly documented and
to create a full GNU development environment for it, parts of it would
need to be reverse-engineered (I talked to J.C. Michot about this, who
partly ported GCC to it), which is forbidden by the authors, etc.
So it is in no alternative in the long run, no matter how much I
appreciated some of its concepts when writing software for it.
-
And because I see no other movements in this direction, the Hurd is
currently my only hope that there will be an operating system which has
this technical elegance and is built of components which may be
combined, enhanced or to which may be added more components.
I sincerely hope that not all free software developers who want to help
build an operating system are going to spend all their time and energy
on enhancing a traditional kernel like Linux. Some day, its concept will
reach its limitations.
Because is a concept such as the Hurd's which will help designers and
programmers to be more creative and find more pleasure in their work!
-
Best regards, and thanks for reading,
Emile van Bergen (preferred e-mail address: [EMAIL PROTECTED])
------------------------------
From: Andreas Schwab <[EMAIL PROTECTED]>
Subject: Re: moving linux to different partition
Date: 13 Jan 1999 12:50:11 +0100
Tristan Wibberley <[EMAIL PROTECTED]> writes:
|> linuxsys:/# cp -aR `cat flist` /mnt
`-a' alread implies `-R', so you don't need that. Also, you can use
`cp -ax . /mnt' which will already omit mount points.
--
Andreas Schwab "And now for something
[EMAIL PROTECTED] completely different"
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (N1ho)
Subject: Re: 2.2.0pre7ac patches
Date: 19 Jan 1999 04:35:52 GMT
[EMAIL PROTECTED] complains:
>>> One of my favorite pet peeves: >>> people who recommend patching
>>> with -p0 instead of -p1.
Hmmm - guess you might want to take it up with
Linus - it's been in the README for every
kernel I've built in the development stream for the last 3-4 years.
>>> Perhaps most people use
>>> /usr/src/linux, but this is impossible if
>>> (like me) you have more than one
>>> source tree sitting around.
Well, there you have it - read the README
and you'll discover a nifty little feature of
Linux (and UNIX in general) called a
symbolic link. See the man page for 'ln' and
read about its '-s' switch. On my system,
/usr/src/linux points to the latest development
kernel, but the original linux-2.2.0pre7 is
there along with linux-2.2.0pre5 and a
linux-2.0.30something that came with Red Hat,
which I haven't used since I installed it.
So, instead of creating a tree named 'linux',
create a tree with a meaningful name and:
ln -s linux-2.2.0premeaningful linux
gzip -dc linux-2.2.0premeaningful.tar.gz | tar xvf -
to initially populate the tree, then:
cd linux
bzip2 ../patch-xxx.bz2 | patch -p1 or -p0
or whatever you want to do.
--
======
Spam (except for Hormel products) is forwarded on to [EMAIL PROTECTED] and the
originating ISP for disposal by their legal teams. (AOL-internal spam to
TOSemail1)
------------------------------
From: [EMAIL PROTECTED] (Mr. Bill)
Crossposted-To:
comp.os.linux.advocacy,comp.os.linux.development,comp.os.linux.development.apps
Subject: Re: How to run Windows Applications on Linux
Date: Tue, 19 Jan 1999 02:08:22 GMT
I have been screwing around with WINE for awhile to no avail. The
installation and configuration documents leave much to be desired. I
installed the RPM version and am trying to figure out just what I need
to do next.
My understanding is that I don't need a DOS partition at all which is
the way I have it (dedicated computer to linux). I believe I need to
modify the wine.ini file but what am I to modify it to? and what else
do I need to do. Basically, I have just installed WINE and the rest
of my time I have spent reading the docs to no avail.
Please CC: me during posting so as I don't miss the message and
repost.
Any help would be appreciated.
Mr. Bill
[EMAIL PROTECTED]
>On Mon, 18 Jan 1999 02:14:12 GMT, [EMAIL PROTECTED] wrote:
>On 15 Jan 1999 02:17:24 GMT, [EMAIL PROTECTED] (Jim Richardson)
>wrote:
>
>
>>I for one have been unsuccessful in getting wine to do anything, I am sure
>>that there is something, somewhere, I am doing wrong. I can't
>>even get it to play tictactoe... Getting wine up to play the few games I
>>use in windows, would enable me to delete windows entirely from the disk.
>> (back to to the docs...)
>
>Believe it or not i'm posting this reply from FreeAgent 32 running
>under Linux :) With the last couple of wine releases I couldn't get
>much of anything to run, not even calc.exe. With the latest release
>I've even gotten Unreal to run with sound (albiet a little staticy)!
>Check out the latest release, its beginning to shape up nicely!
>
>Matt
>
------------------------------
From: Martin Maney <[EMAIL PROTECTED]>
Subject: Re: - deprecated - why?
Date: 19 Jan 1999 02:16:47 GMT
Matthew Hannigan <[EMAIL PROTECTED]> wrote:
> Tristan Wibberley <[EMAIL PROTECTED]> wrote:
>>That's insane! All the options should have '-' prefixed.
> Tell that to the authors of tar and BSD ps.
Well, tar anyway. I was looking through the web-accessed man pages at the
FreeBSD site, and the oldest one that shows a leading hyphen, NOT documented
as optional, was from the BSD Net/2 release, dated June 1991. The 386BSD
pages weren't responding in the time I was willing to wait, so the next data
point I found, and the most recent one that said the hyphen was unnecessary,
was the 2.9.1 BSD man page (no date given, but that would be mid-ish
eighties, no?).
It looks like this "wonderful" scheme is relying on behavior that is about a
decade old as documented performance, anyway.
> As it is, it gives us a very nice way to have compatibility.
Very nice if you prefer Sys V or your Berkeley roots go back over a decade;
not so very nice if you learned to use BSD-like ps in more recent time. (As
I did. I never heard, nor cared, that the program actually still accepted
the old syntax: that was NOT the documented performance, so I've never
relied on it.)
Granted: if you really MUST combine the two different behaviors into one
this is the only semi-justifiable way to do so without renaming one or
something else obtrusive, but very nice it certainly ain't. Ugly kluge
would be more accurate. The best thing I can see about it is that I haven't
written ps into many locally-maintained scripts - at least not that I can
think of.
------------------------------
From: [EMAIL PROTECTED] (steve mcadams)
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: disheartened gnome developer
Date: Mon, 18 Jan 1999 23:41:49 GMT
[Snipped for brevity, quoted material marked with ">"]
On Mon, 18 Jan 1999 16:52:41 GMT, [EMAIL PROTECTED] wrote:
>I find it a bit disturbing that for some people on the "linux community" good
>will is not good enough, and only total submission is (I'm not talking about
>you here)
Kind of like the alpha-wolf thing where the only way not to have your
throat torn out is to lay on your back and present it submissively and
hope for the best? -steve
========================================================
Tools for programmers: http://www.codetools.com/showcase
------------------------------
From: Craig Eaton <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.advocacy,comp.os.linux.development,comp.os.linux.development.apps
Subject: Re: How to run Windows Applications on Linux
Date: Mon, 18 Jan 1999 20:42:05 -0500
Jim Richardson wrote:
> maybe diablo, and...) and one custom app for work. (weird terminal emulator
> for the as400) But I have not gotten any of the games to run. Warcraft
> starts, access' the cd, and then just hangs there eating cpu cycles.
If I remeber right, there is a Linux program out there that does the screen
at a time terminal emulation that AS400's like to see. I don't remember what
its called exactly, but I know SuSE includes it. It's got a name similar to
X3270, but where the number after the X is the type of emulation. :) Don't
know if its any good, but it might save you having to monkey with wine.
--
Craig Eaton
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Crossposted-To: comp.os.linux.misc,comp.lang.perl.misc
Subject: Re: Secuity hole with perl (suidperl) and nosuid mounts on Linux
Date: 18 Jan 1999 19:12:40 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Oleg Mercader <[EMAIL PROTECTED]>]
> If you haven't an exec permition in your /etc/fstab for a floppy
> mounting point you can't execute any from a floppy and if you copy a
> file to a harddisk you lost your suid.
All you need is an entry for a floppy which includes the "user" option
and some type other than "msdos" or "vfat" or whatever ("auto" works)
so that a user can mount an ext2 (or minix, or xia, etc) floppy.
The option "noexec" does not affect this, since the kernel *already*
has a check for this hole if you execute the script directly. (It will
refuse to run it, with EPERM.) You have to invoke the script with
perl /mnt/fd0/script.pl
or some such, so that *perl* is calling the shots, not the kernel.
That is what makes this a hole -- a bug in suidperl. suidperl is
supposed to emulate the kernel permission checks, and it just isn't
clever enough to emulate *this particular* check.
Let it be noted that allowing physical access to an important machine
is asking for trouble already, and allowing users to do things like
mount floppies and cdroms is maybe even a little more careless, so a
truly paranoid admin wouldn't have been hit by this bug anyway.
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: Marcos Silva <[EMAIL PROTECTED]>
Subject: Re: K6-2 300 Problem
Date: Mon, 18 Jan 1999 23:52:11 -0200
I 'm running linux now. When I start with dos and execute loadlin form HD it
start.
"loadlin vmlinuz root=/dev/hda5".
If i put the vmlinuz in the floppy disk and boot my computer by floppy.....
LOCK.
Marco Vannini wrote:
> I have a K6-2 300 oc. 333 (asus txp4) with RH 5.0 (kernel 2.0.32) and all
> works fine!!
--
Marcos Ferreira da Silva
Centrio Universitario do Triangulo
Uberlandia - MG
Brazil
------------------------------
From: Toon Moene <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: disheartened gnome developer
Date: Tue, 12 Jan 1999 20:44:57 +0100
Frank Sweetser wrote:
>
> [EMAIL PROTECTED] (Christopher B. Browne) writes:
> > You miss the consideration that the third line of that comment shows
> > a *really big stick* that strongly discourages Red Hat Software from
> > doing a proprietary release.
> > Were RHS to do so, there would be a *dramatic* reaction within hours.
> perfect case in point: the recent X licensing fiasco. TOG announced that
> the next version of X they released would be under a much more commercial
> friendly, and much less open source, license. within days, hundreds of
> developers, led primarily by the XFree86 people, and with vocal support
> from such free software leaders, were ready to fork off from X11R6.3, and
> leave TOG in the dust. realizing they were about to slit their throat, TOG
> backed off and canceled the license change.
Ah, but the *most* interesting issue in that series of events was that
no part of said software was covered by the GPL.
<speculation mode=heretic>
It might be that in the (perhaps not even so) long run we do not _need_
the GPL anymore to protect software from hoarding.
</speculation>
--
Toon Moene ([EMAIL PROTECTED])
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: [EMAIL PROTECTED]; egcs: [EMAIL PROTECTED]
------------------------------
From: "Thierry BUCCO" <[EMAIL PROTECTED]>
Subject: db question
Date: Mon, 18 Jan 1999 21:39:36 +0100
Hi,
I want to create a db file, with bekerley db library,so :
1- I create the db file : db = dbopen("/root/dbfile", O_CREAT | O_RDWR,
0666, DB_HASH, NULL);
it's ok the file is created.
2- I want to put one record (name + phone)
DBT key,value;
char *name="thierry\0"
char *phone="03304\0"
key.data=strlen(name)+1;
key.size=strlen(phone)+1;
db->put(db, &key, &value, 0);
when i edit this file, i see that the record is here, but i can't open
it with Exce.
Is there an incompatible file with excel ???????????? or is it just me ?
Thanks a lot - Thierry FRANCE
------------------------------
From: [EMAIL PROTECTED] (Peter Samuelson)
Subject: Re: Linux Sound Engine
Date: 18 Jan 1999 21:05:00 -0600
Reply-To: Peter Samuelson <[EMAIL PROTECTED]>
[Peter Steiner <[EMAIL PROTECTED]>]
> However I'd like to support MMX and AFAIK that cannot be used in
> kernelspace since it would conflict with userlevel programs using the
> FPU.
"Cannot" is a strong word. You do have to save and restore the FPU
registers yourself, however. And, as Ingo Molnar found out some time
ago when he wrote a patch to do IP checksums with MMX, the time to save
and restore the FPU regs can eat up much of the benefit of MMX speed.
I do not know much of anything about kernel internals ... but perhaps
there's a way to have the mixer run *between* user-level contexts,
i.e. after the FPU is saved from one context and before it is restored
in the other. Then you get the save/restore for free. You might look
into this. Is this something like what a BH does?
...But probably, if it were really that easy, Ingo would have done it
this way for the IP checksums (and RAID5 checksums, I believe).
--
Peter Samuelson
<sampo.creighton.edu!psamuels>
------------------------------
From: "Henry Mclachlan" <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,aus.computers.linux,comp.os.linux.development.apps,comp.realtime
Subject: - A FREE down load of Linux Real Time debugging tool (demo version)
Date: Tue, 19 Jan 1999 20:03:23 +1100
TARGET is a real time system diagnostic and software development tool that
may be used by C and C++ software developers to assist them in development,
test and monitoring of software running in real time distributed systems.
Please see our web page (www.ozemail.com.au/~ecat/) for a free down load of
TARGET demo version.
------------------------------
** 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
******************************
