Linux-Development-Sys Digest #334, Volume #6 Tue, 26 Jan 99 23:13:58 EST
Contents:
Re: How can I build a Linux system from scratch - NO distribution? (Richard Jones)
Re: Free Fortran90 + Parallel? (M Sweger)
Re: How do you read core dump on Linux? ([EMAIL PROTECTED])
Re: - deprecated - why? (bill davidsen)
Re: Why I'm dumping Linux, going back to Windblows (Todd Ostermeier)
Linux ext2 dump (Dominik Epple)
Re: How can I build a Linux system from scratch - NO distribution? (Jens Kristian
Søgaard)
Re: Can I run a DOS Device Driver in an emulator (maciej h babinski)
D-Link NIC driver, anyone? ("Aaron Dershem")
Re: Why I'm dumping Linux, going back to Windblows (Leslie Mikesell)
Re: Why I'm dumping Linux, going back to Windblows ([EMAIL PROTECTED])
Re: Why I'm dumping Linux, going back to Windblows (jedi)
Firewall/Proxy/Masquerading? ("ncc1701d")
Re: Need help with XFree (JP)
Need help with XFree (JP)
----------------------------------------------------------------------------
From: Richard Jones <[EMAIL PROTECTED]>
Subject: Re: How can I build a Linux system from scratch - NO distribution?
Date: Tue, 26 Jan 1999 14:24:45 +0000
Stefan Monnier <[EMAIL PROTECTED]>
wrote:
:>>>>> "Richard" == Richard Jones <[EMAIL PROTECTED]> writes:
:> The nameserver sees that there is no route to host, but doesn't send back a
:> negative response. It (seems to) just forget about the request and log an
:> error to syslog, so the clients still block endlessly.
: Yuck !
: Qualifies as a bug, AFAIC !
I don't think so. Things would be pretty bad
if every time there was a slight routing problem
between two nameservers on the Internet, one
of the nameservers started returning negative
responses to requests. The problem is that
named isn't really designed to work over
dial-up -- normally it would expect to fall
back to another nameserver.
Rich.
--
- Richard Jones. Linux contractor London and SE areas. -
- Very boring homepage at: http://www.annexia.demon.co.uk/ -
- You are currently the 1,991,243,100th visitor to this signature. -
- Original message content Copyright (C) 1998 Richard Jones. -
------------------------------
From: [EMAIL PROTECTED] (M Sweger)
Crossposted-To: comp.os.linux.misc,comp.os.linux.development.apps
Subject: Re: Free Fortran90 + Parallel?
Date: 26 Jan 1999 23:00:35 GMT
Reply-To: [EMAIL PROTECTED]
M Sweger ([EMAIL PROTECTED]) wrote:
: http://www.fortran.com/fortran and select free s/w
: Ratfor
: http://ww.sepwww.standford.edu/sep/prof
The above link is http://sepwww.stanford.edu/sep/prof
This link is off of the www.fortran.com/fortran/free.html list.
: --
: Mike,
: [EMAIL PROTECTED]
--
Mike,
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: How do you read core dump on Linux?
Date: Tue, 26 Jan 1999 23:05:58 GMT
In article <[EMAIL PROTECTED]>,[EMAIL PROTECTED] (Daniel R.
Grayson) wrote:
>
> gdb foo core
>
> where foo is the name of the executable file that died, and examine the core
> file with gdb commands. If you don't know the name, put any old thing there
> and gdb will tell you.
How do you read the contents of a core file if you no longer have the
executable?
Someone ran a program on one of my machines, and then left the process
running, but deleted the original file. I killed the process so that it
dumped a core file, but I'd like to examine the core file to exactly what the
process was doing.
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Subject: Re: - deprecated - why?
Date: 27 Jan 1999 00:49:24 GMT
In article <77u3tu$16c$[EMAIL PROTECTED]>,
Matthew Hannigan <[EMAIL PROTECTED]> wrote:
| In article <[EMAIL PROTECTED]>,
| Tristan Wibberley <[EMAIL PROTECTED]> wrote:
| >That's insane! All the options should have '-' prefixed.
|
| Tell that to the authors of tar and BSD ps.
|
| As it is, it gives us a very nice way to have compatibility.
AIX has one set of options with the - prefix, quite another without. In
some ways useful, I regularly use both forms depending on the desired
presentation.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"Too soon we grow old, and too late we grow smart" -Arthur Godfrey
------------------------------
From: Todd Ostermeier <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: Tue, 26 Jan 1999 18:50:11 -0600
On 26 Jan 1999, Peter Samuelson wrote:
: > why is that? 32 bit pointers w/ signed magnitude? (effectively, 31
: > bits of addressable memory space)
:
: For files, I think you're right. The lseek() system call takes a
: 32-bit signed integer for its argument, since it's possible to use
: negative offsets in SEEK_CUR ("seek from current position") mode.
I did not know that. It makes sense, though. (I'm applying my basic
knowledge of computer hardware here, rather than any low-level asm/C
coding I've done).
:
: For process address space: x86 before (I think) Pentium Pro has a 4GB
: address space. (PPro/PII can have 36-bit addressing, or 16GB of
: addressing, but Linux doesn't use it.) This space in Linux is divided
: into virtual memory and running-process address space. The default is
: a 2GB-2GB split, so that any single process can have 2GB of address
: space, as can the system as a whole.
Okay, that makes sense, too. I was trying to understand the whole idea of
a signed pointer for memory addresses, and it never dawned on me to see it
as a split of the memory space.
: > I'm assuming here that we're talking about 32-bit x86 architectures.
:
: Yes. 64-bit architectures can have both larger files and larger
: address spaces. 64-bit file sizes for 32-bit arch's is rumored to be
: in the works for Linux 2.3.x.
heh :)
________________________________
Todd Ostermeier
[EMAIL PROTECTED]
http://www.ews.uiuc.edu/~ostermer/index.html
ICQ UIN: 2253928
A-723
________________________________
------------------------------
From: Dominik Epple <[EMAIL PROTECTED]>
Subject: Linux ext2 dump
Date: Sun, 24 Jan 1999 22:28:44 +0100
Hi,
I have problems with dump, encountered on RH5.2 with the kernels 2.0.36
and 2.1.132: when dumping a file system, it says:
[amanda@euler amanda]$ /sbin/dump 0Bf 1048576 /dev/null /usr
DUMP: Date of this level 0 dump: Sun Jan 24 22:18:16 1999
DUMP: Date of last level 0 dump: the epoch
DUMP: Dumping /dev/sda7 (/usr) to /dev/null
DUMP: mapping (Pass I) [regular files]
DUMP: mapping (Pass II) [directories]
DUMP: estimated 473441 tape blocks on 0.45 tape(s).
DUMP: dumping (Pass III) [directories]
DUMP: master/slave protocol botched.
DUMP: The ENTIRE dump is aborted.
[amanda@euler amanda]$
the dump is from the RH package dump-0.3-14 .
I already have seen other people with that problem posting news, but no
one answered them. :-(
some background information: as I can see, this happens only to large
file systems and/or quite full file systems. I have this problem on 2
machines:
1. Dual PII 400, 512 Meg Ram, 122GB Hardware Raid System (not md driver,
real raid) (no, I dont have this at home, I work with it at my job ;) ),
where it does not happen to the internal system disks, but to the RAID.
2. PII 233, 96 Meg Ram, 4GB SCSI disk (IBM DCAS 4GB), where it fails to
/home, /usr, /share, /local, i.e. all file systems larger with 768MB.
The / and /var system with 128MB and 256MB are dumped correctly. Here,
this error dit not occur right from the beginning, but later, since the
file systems are somewhat full (between 13% and 61%). And it happens
with the backup user (amanda) and root as well.
Any help would be GREATLY appreciated.
dominik
------------------------------
From: [EMAIL PROTECTED] (Jens Kristian Søgaard)
Subject: Re: How can I build a Linux system from scratch - NO distribution?
Date: 24 Jan 1999 23:20:13 +0100
Richard Jones <[EMAIL PROTECTED]> writes:
> :> a system with X-Windows and various components taking a very long time to
> : This is probably a DNS problem. Setup your DNS server to reply with a
> : "nonexistent domain"-error when an internet-domain is requested
> : ( ofcourse it should reply normally when connected to the net ).
> Can you explain how to do this?
Well, I can try ;-)
If you have a local DNS server running on your Linux computer ( named
), it's quite easy. Just make sure that you have no forwarders defined
( they aren't accessible when you're not connected ). Then remove the
zone "." so that it's doesn't use any internet-servers.
Instead of removing the zone, you could empty the named.ca file.
--
Jens Kristian Søgaard,
[EMAIL PROTECTED]
------------------------------
From: maciej h babinski <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development,comp.os.linux.hardware,comp.os.linux.misc
Subject: Re: Can I run a DOS Device Driver in an emulator
Date: 26 Jan 1999 18:36:03 GMT
In comp.os.linux.hardware Norm Dresner <[EMAIL PROTECTED]> wrote:
> I might be able to boot DOS, initialize the board, and then boot Linux, but
> that stinks, especially if I have to modify the settings.
This really isn't that bad an option. If you set up a DOS boot menu
and loadlin, it works well. It's what I use to initialize my joystick
port.
> Is there any way that I can call (perhaps hack up a DOS-style
> load-device-driver-from-the-command-line program) the driver from a DOS
> emulator in Linux to do the job.
You could use dosemu and grant it privelages to whatever IO ports
your driver uses, but that could be dangerous if they're, say, the
PCI bus. If it's an ISA plugNplay board, have you tried the isapnp
tools?
------------------------------
From: "Aaron Dershem" <[EMAIL PROTECTED]>
Subject: D-Link NIC driver, anyone?
Date: Sun, 24 Jan 1999 15:13:37 -0600
Does anyone out there have a driver for a D-Link DFE-530TX Network card?
I'm running RedHat 5.2, but it won't detect it, and the tulip driver doesn't
work. If you could, e-mail it to [EMAIL PROTECTED]
Thanks!
Aaron Dershem
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: 26 Jan 1999 12:40:44 -0600
In article <78klok$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>This point have been for me something I just could never understand
>about the Unix world.
It has more to do with having the same person write the manual that
wrote the program than unix vs. anything else.
>to give a manual page for a command, that might contain n number of options,
>yet a typical command might only use m of them at a time, where m << n,
>yet, for the user to determine the correct m for each particular case,
>they must read about each option, and then try to determine the subset
>of them they need to compose to build the command.
Who is going to decide what is 'typical'?
>For example, look at cpio man page. If the author of cpio have given
>few examples showing how to use cpio to backup a tree on a file system,
>and then how to restore the result back, and an example of how to do the
>same moving one or more files, and such examples, then life will be so
>much easier.
What about the person who only uses cpio in -p mode to mirror
to other directories?
>Iam to this day, could never understand why would the authors of man pages
>refuse to show examples.
There are too many possibilities to show them all and the author of the
program wouldn't want to omit any. Remember, he just spent weeks
adding an obscure function and now you want him to leave it out
of the documentation because for you it isn't 'typical' usage.
>it looks like the only reason is lack of concern for new users, and the
>inability to put a little bit more effort in trying to make Unix a little
>easier for others to use. it is as if they want to make it harder
>by purpose.
New functions are added as the need arises and they are added into the
list of options. Should the fact that you don't happen to need one
of those options today mean that it shouldn't be given equal space
in the manual?
>may I am wrong on this, but I would really be interested in hearing from
>others who wrote man pages, why did they not put some effort in putting
>examples.
I have always thought that every complex product should have two
completely separate manuals. One should be the 'getting started'
tutorial that describes the minimum you need to know to make something
work, the other should be the reference manual that shows everything
it can do. You might want to see the tutorial section once, then
you never want to see it again - you'll want to know what options
are available, not the subset someone else happened to like. The
solution is to have someone else write the throwaway tutorial from
a perspective of minimal use since the programmer has to be concerned
with all of the options and optimal use.
Les Mikesell
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: 26 Jan 1999 15:56:45 -0800
In article <[EMAIL PROTECTED]>, John says...
>
>kna writes:
>> Iam to this day, could never understand why would the authors of man pages
>> refuse to show examples.
>
>When I write man pages I do provide examples when appropriate.
>
good. I hope more programmer learn to do this.
>> may I am wrong on this, but I would really be interested in hearing from
>> others who wrote man pages, why did they not put some effort in putting
>> examples.
>
>How many man pages have rewritten so far? What response did you receive
>from the maintainers you submitted them to?
>--
I think of a man page as PART of the system. when I modify the program,
I certainly need to update the man page at the same time to reflect
the changes.
I need to keep the program and the man page and the examples in synch.
offcourse this is more work, but this is what software is about.
If I did not care about the user, I would make a very short man page,
and take the easy way out and be dome with it.
kna
------------------------------
From: [EMAIL PROTECTED] (jedi)
Crossposted-To: alt.os.linux,comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: Why I'm dumping Linux, going back to Windblows
Date: Tue, 26 Jan 1999 18:17:29 -0800
On Wed, 27 Jan 1999 01:59:43 GMT, Christopher Browne <[EMAIL PROTECTED]> wrote:
>On Tue, 26 Jan 1999 16:06:57 -0600, Todd Ostermeier
><[EMAIL PROTECTED]> wrote:
>>On 25 Jan 1999, Gordon Scott wrote:
>>
>>: Hm, files (on x86s anyway) are limited to 2Gb -- memory might also be.
>>: I don't think it'll be *too* long before a 64bit system addresses that --
>>: there are already a few protests at the 2Gb file limit.
>>
>>This is probably way off topic, but why is that? 32 bit pointers w/
>>signed magnitude? (effectively, 31 bits of addressable memory space)
>
>There are protests because there are databases out there that are larger
>than 2GB in size. (I work with one database at work that, today, is
>57GB in size.)
So? One does not typically keep the entire database in one
file.
>
>The average *disk* sold these days is larger than 2GB, after all, and
>ext2fs supports partition sizes of on the order of 2TB, so limiting file
>size to 2GB seems a mite unreasonable.
It's a mild annoyance on the high end of things but isn't
quite to a critical point yet. It's certainly a lame excuse
to not deploy a non-trivial database on Linux.
--
Herding Humans ~ Herding Cats
Neither will do a thing unless they really want to, or |||
is coerced to the point where it will scratch your eyes out / | \
as soon as your grip slips.
In search of sane PPP docs? Try http://penguin.lvcm.com
------------------------------
From: "ncc1701d" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Firewall/Proxy/Masquerading?
Date: Tue, 26 Jan 1999 12:19:59 -0600
NNTP-Posting-Date: Tue, 26 Jan 1999 10:20:10 PDT
Hello People,
I'm hoping that someone can help me out here. I have a Red Hat Linux box on
my network along with a cable modem and an ISDN connection out to the net.
Currently I'm using a quick script I created in Linux, Cable on/off, in
order for me to easily use the cable modem from the Linux box itself. What
I would like to do is have Linux intercept all requests for web access from
itself and clients on my network and redirect them out the cable modem.
Everything else would stay normal and go out ISDN. Red Hat is the router
for my network, I have four network cards in the Linux box, 2 separating my
10/100 side of the network and 2 separating the class C address of the Cable
modem. I'm anticipating that all of my clients would have to sit in the top
half of the 192 segment(The Cable modem) in order for Linux to masquerade,
but I would really like to be able to keep my registered IP's on the clients
themselves. Is this something that proxying would accomplish? Routing
works great, I can ping machines sitting any where on my network, except the
cable modem(I think it'll only allow IP communication from 1 address,
192.168.100.2), that is why I think I have to use masquerading.
To sum it up, can Linux be setup to intercept TCP/UDP requests on specific
ports and redirect them out a different device other then it's default
gateway? But have all other requests go out the default gateway?
I have been to the Linux Documentation project and read most of the HOWTO's
on doing firewalling, proxying, and masquerading. Plus I have read the
HOWTO for IPCHAINS. The HOWTO's on the LDP use IPFWADM which is
incompatible with my kernel, so I'm only able to use IPCHAINS. When I use
the commands specified in the HOWTO's on the LDP they seem to be
incompatible with arguments for IPCHAINS. With enough playing around I'll
be able to figure out how to use IPCHAINS exactly, but I'm really looking
for someone to tell me want I want to do is even possible. That way I won't
go wasting my time trying to do something impossible to begin with. :^)
Thanks in advance!
Jason
PS If you have any links that might be of use to me please let me know :)
------------------------------
From: JP <[EMAIL PROTECTED]>
Subject: Re: Need help with XFree
Date: Tue, 26 Jan 1999 13:39:25 -0600
Loren Osborn wrote:
> JP wrote:
> >
> > I am trying to run XFree with my TNT AGP card. I set the Xconf for Riva
> > 128 and 4MB. I run XFree, and it displays at 320x320. Argh! How do I get
> > larger desktop sizes, ala 1024x768?
> > I'm a newbie, so please make any advice simple to understand =)
> >
> > --
> > JP
>
> I'm not familiar with your video card (it might not be supported) but
> have you tried to run XF86Setup? (having your video-card/monitor/mouse
> manuals --if any-- nearby is useful.) Hope that helps....
>
> Loren
How? Seriously, is this a program in Linux? I am using RedHat 5.2.
--
JP
[EMAIL PROTECTED]
Nolite te bastardes carborundorum
__________________________________
The mind and the sword are one-
if the mind is right, the sword is right.
__________________________________
------------------------------
From: JP <[EMAIL PROTECTED]>
Subject: Need help with XFree
Date: Tue, 26 Jan 1999 12:53:45 -0600
I am trying to run XFree with my TNT AGP card. I set the Xconf for Riva
128 and 4MB. I run XFree, and it displays at 320x320. Argh! How do I get
larger desktop sizes, ala 1024x768?
I'm a newbie, so please make any advice simple to understand =)
--
JP
------------------------------
** 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
******************************
