Linux-Misc Digest #698, Volume #27 Mon, 23 Apr 01 15:13:04 EDT
Contents:
Re: Disabling Core dumps (Francis Litterio)
Re: What program languages (or tools) are X-apps written in ? (Elf Sternberg)
Re: problem with gnome-terminal after upgrade (David Miller)
Re: Can't get program to run from inetd (Francis Litterio)
move linux install to new HD (Steve Bui)
Re: fdopen error in libc.so.6 ("Peter T. Breuer")
automount error ("Andrew Moore")
Re: fat32 + ext2 partitions lost. How to recover? (Roberto Inzerillo)
Re: Voodoo (usual) problems ("Rich Rudnick")
Re: How about those kernel-panics :( (Michael Heiming)
Re: RAID question. ("James Moore")
Re: A Linux emulator for Linux, does this exist? (bill davidsen)
----------------------------------------------------------------------------
From: Francis Litterio <[EMAIL PROTECTED]>
Subject: Re: Disabling Core dumps
Date: Mon, 23 Apr 2001 17:10:27 GMT
"Peet Grobler" <[EMAIL PROTECTED]> writes:
> That was it. Thanks.
>
> Alumne FIB - MARC COLL CARRILLO wrote
> >> I remember there is a way to disable core dumps in Linux.
> >> How do you do it?
> >
> >Edit /etc/bashrc, and add the following:
> >
> >ulimit -c 0
/etc/bashrc is not sourced by every Bash shell at login time. Edit
/etc/profile instead -- every login Bash and Bourne shell sources
/etc/profile. But keep reading ...
If you edit /etc/profile and if every user's login shell is a shell that
sources /etc/profile, then you are OK for core dumps from processes
spawned by interactive users, but rshd or sshd don't always create login
shells, so processes spawned by those daemons can still dump core.
The same goes for processes spawned by inetd or init.
Also, the above solution fails to prevent core dumps from processes
spawned by users who have csh and tcsh (and probably others) as their
login shells. To truly cover all bases (for login shells), you have to
edit every possible login shell's /etc script, not just /etc/profile.
--
Francis Litterio
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Elf Sternberg)
Subject: Re: What program languages (or tools) are X-apps written in ?
Date: 23 Apr 2001 17:08:52 GMT
In article <9bvcvq$[EMAIL PROTECTED]>
"tvn" <[EMAIL PROTECTED]> writes:
>Hello, I am an api, mfc windows programmer and recently just try to play
>around with linux - unix. I am running rh and freebsd now. I noticed there
>are 3 categories the nix apps are divided in - console, kde , gnome and X .
>As far as I know, X is the the standard for both Linux and Unix. So I am
>wondering what programming languages are used to make these X - apps. And
>could someone explain to me further about QT and GTK because I heard about
>them but not really sure what they are for. Thanks
Okay, here's the basics. Unlike Windows, where "Windows" is the
operating system, _Linux_ is the operating system, and all it has is a
simple text console.
X is just a program that does two things: draw stuff on a
graphics-capable terminal, and take input from the keyboard and mouse.
That's it. X is a server; what that means is that any client program that
wants to draw stuff on a graphics-capable terminal can send requests to X
saying "Draw me a box with buttons and labels" and X will do it, and
when someone types on the keyboard, X will send those keystrokes and any
associated mouse actions back to the client program.
If you're used to client-server notions, it can take a while to
wrap your head around the notion that X is a middleware server where
you, the human being, are both the front-end and the back-end.
Because X is a client-server system, it's possible to run the
client program on one machine and the X server program on a different
machine. It's also possible to have "multi-headed" systems where one
computer has multiple graphics cards and three or more monitors, because
the X server is extensible and can have more than one instance of a
"screen" (that is, a display) in memory at any given time. This make X
much more powerful than Windows in some very fundamental respects. It
also makes X more secure. The price for these rich capabilities of X is
that it can require a great deal of bandwidth when using X over a
network, and can also require a great deal more knowledge.
X has a very primitive library at its core with a set of simple
things it can do-- basic drawing primitives like lines, rectangles,
fonts, color sets, and so on. This is the Xlib, and if you want to
learn how to use X, you're welcome to start with Xlib and all its
primitives. X uses a specific and unique network protocol for
communication between clients and servers, and Xlib is the library that
has the client code for communicating with X.
Xlib is written in and to entirely in C.
Over the years, several teams have written "X Toolkits." The
idea is to make programming X easier and less "primitive." The
primitives offered in Xlib are, well, primitive-- people like buttons
with three-d effects, and programmers hate having to manage each
individual pixel when drawing an image, and neither of these
capabilities is provided by Xlib.
X Toolkits, such as Motif, QT (not "KDE"), and GTK (not
"Gnome"), provide all of these additional features. Motif was the first
and was very popular, but while X's core was free, Motif was not. Linux
has a Motif clone, "Lesstif," but its importance has faded as computers
became more powerful and could afford the egregious CPU and graphics
power needed for high-end toolkits like QT and GTK.
GTK, like Motif, provides a huge collection of graphics widgets
(that's the technical term): buttons, sliders, canvases, frames,
windows, decorations, and so on, all in an event-oriented and object-
based framework that makes writing GTK applications a matter of defining
the interface and then gluing in the responses to events after the
fact. A useful tool for doing this is Glade, which allows you to
basically draw your static interfaces, much like visual basic, and then
spits out C code, to which you attach the behaviors of your application.
GTK, Motif, Lesstif, and QT are all written in C. (Regarding
QT, could someone confirm this? I have a vague memory of someone
telling me QT is written in C++. I don't know; I've migrated from Motif
on Solaris to using GTK on Linux and BSD.)
Gnome, KDE, CDE (which is a desktop environment for Motif), are
all *further* wrappers around GTK, QT, and Motif respectively; they add
additional operating-system functionality like interprocess
communication, drag-and-drop facilities, and a rich set of applications
for interacting with your X environment, your "desktop." Gnome, KDE,
and CDE are distinct from GTK and QT in that they are both libraries
*and* a collection of applications that use those libraries. But if you
want to make *more* Gnome-ready programs with all of the facilities of
Gnome, the libraries are where you look.
Ximian is a wrapper around Gnome; it provides even more intense
facilities like componentization and rapid-access to specialized file
system retrieval, as well as an abstract file access layer that allows
one to treat CDs, floppies, hard drives, network drives, FTP sites, and
even web sites as if they were all just parts of the filesystem.
Although these are all written in "C", the Unix philosophy has
led to other programming languages with ready hooks into the X
libraries. For example, one of the earliest graphics toolkits is TK,
which is an XLib-based library for the interpreted language TCL.
Although TK was originally written be used with TCL, it was so easily
extensible that the scripting languages Perl and Python both support
writing applications using the Tk Toolkit.
GTK, especially, has proven easy to extend, and there are
wrappers for it such that one can write applications in C, C++, Perl,
Python, or Java and still have access to the GTK Toolkit. Even the
Gnome libraries are accessible through these other languages. If you're
a VB programmer, I highly recommend learning Python; its syntax will be
familiar to you, but the OO facilities, metaprogramming capabilities,
and discipline of Python will make you both a smarter and a better
programmer.
Where to start? Well, one good place to start, where you'll
find immediate results, is the GTK Tutorial. Better than any of the
dozens of "How to program GTK and Gnome" books, this tutorial will walk
you through the basics of writing GTK programs in C. It will not teach
you how to use GCC, GDB, or any Unix IDE's, and it does not hold you
hand for you like some Windows wizards do. You *will* learn everything
you need to know about coding GTK from here. If you don't know how to
hack code for a Unix box, well, that's a different lecture.
GTK Tutorial
http://www.gtk.org/tutorial/
To learn about Xlib and Motif, start here:
http://csc.lsu.edu/tutorial/Xnotes/X_lecture.html
And, if you're going to get serious about programming X, there's
Kenton Lee's excellent (if a bit high-handed) "Technical X-Windows Catalog":
Technical X Windows
http://www.rahul.net/kenton/xsites.html
Elf
--
Elf M. Sternberg, rational romantic mystical cynical idealist
http://www.halcyon.com/elf/
Fast food restaurants are like gay bathhouses in San Francisco,
places where people go to engage in high-risk behaviors.
- Greg Critser
------------------------------
From: David Miller <[EMAIL PROTECTED]>
Subject: Re: problem with gnome-terminal after upgrade
Date: Mon, 23 Apr 2001 13:23:28 -0400
Alex,
No, i haven't figured it out yet. Using KDE for now.
-Dave
Alex Ramos wrote:
> David,
>
> Did you figure out what was going on with gnome-terminal? I am
> having the same difficulty.
>
> Alex.
>
> In article <[EMAIL PROTECTED]>, "David Miller"
> <[EMAIL PROTECTED]> wrote:
>
>
>> I recently upgraded from RH 7.0 to 7.1. When I try to bring up a
>> gnome-terminal I get the following error: gnome-terminal: error while
>> loading shared libraries: gnome-terminal: undefined symbol:
>> zvt_term_set_open_im eh?
>> Any help is appreciated.
>> Dave
>
------------------------------
From: Francis Litterio <[EMAIL PROTECTED]>
Subject: Re: Can't get program to run from inetd
Date: Mon, 23 Apr 2001 17:26:59 GMT
[EMAIL PROTECTED] writes:
> Apr 23 15:30:39 thorin inetd[31597]: 62010/tcp: bind: Address already in
> use
>
> However if I take the one line out of /etc/inetd.conf and restart
> inetd I still can't telnet on this port so I can't understand why
> the port is already in use.
Type "netstat -ant | grep 62010" and I'll bet there's a TCP socket in a
wait state (e.g., TIME_WAIT) on that port. This is normal -- TCP
prevents the use of a port for a short time after its socket is shutdown.
inetd could prevent this by setting the SO_REUSEADDR option on the
socket, but it seems not to be.
If you wait a minute, the socket should abandon that port, making it
useable again. If you can't wait, change the port number in
/etc/inetd.conf.
--
Francis Litterio
[EMAIL PROTECTED]
------------------------------
From: Steve Bui <[EMAIL PROTECTED]>
Subject: move linux install to new HD
Date: Mon, 23 Apr 2001 10:29:45 -0700
I have a HD that appears to be on its last leg. I would like to move the
contents of the entire HD onto a new one and be able to boot off of
that. What's the safest way to do this? Thanks.
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: fdopen error in libc.so.6
Date: Mon, 23 Apr 2001 19:19:41 +0200
Dan Yost <[EMAIL PROTECTED]> wrote:
> Hello,
> I've encountered an error that's got me stumped. I'm running Red Hat 7.0.
> All of a sudden, with *no* changes to the libraries that I know of (classic
> denial there), I've started getting errors:
> $ rpm -qa
> /usr/lib/rpm/rpmq: error while loading shared libraries:
> /usr/lib/libz.so.1: symbol fdopen, version GLIBC_2.1 not defined in file
> libc.so.6 with link time reference
Classic example of an upgrade done late at night. You are facing
version hiking. Undo some of the library changes that you think you
didn't do :-).
> $ ftp 192.168.1.1
> ftp: error while loading shared libraries: ftp: symbol fdopen, version
> GLIBC_2.1 not defined in file libc.so.6 with link time reference
Yes. I believe the error arises from libdl, orrigianlly, no?
Peetr
------------------------------
From: "Andrew Moore" <[EMAIL PROTECTED]>
Subject: automount error
Date: Mon, 23 Apr 2001 13:26:26 -0400
Hello,
I am working on porting application software from an HP-UX platform to a
RedHat linux 7.0 platform.
I am currently using kernel version 2.2.16 - 22.
When I attempt to use nfs to mount a mount point on the local node the first
attempt fails. For example, on my local node (mdw) the command
ls -alp /nfs/mdw/export returns stating that the directory does not exist.
A second call succeeds so obviously the mount is done. Future attempts to
access this directory succeed until the automounter removes the mount.
I know it seems silly to use nfs to access directories on the local node but
much of the application code does exactly that. This works fine on an HP-UX
platform.
Perhaps someone may have some insight into this problem. I am enclosing the
error message from /var/log/messages along with some configuration files.
Thank you for your assistance.
Apr 23 15:47:38 mdw automount[468]: attempting to mount entry /nfs/mdw
Apr 23 15:47:38 mdw automount[1019]: starting automounter version 3.1.6,
path = /nfs/mdw, maptype = file,
mapname = /etc/auto.nfssub
Apr 23 15:47:38 mdw automount[1019]: using kernel protocol version 3
Apr 23 15:47:38 mdw automount[1019]: attempting to mount entry
/nfs/mdw/export
Apr 23 15:47:38 mdw modprobe: modprobe: Can't locate module bind
Apr 23 15:47:38 mdw automount[1021]: >> mount: fs type bind not supported by
kernel
Contents of /etc/host.conf
order nis,hosts
Contents of auto.master
/misc /etc/auto.misc --timeout=60
/nfs /etc/auto.nfs
/home auto.home
Contents of auto.nfs
* -fstype=autofs,-Dhost=& file:/etc/auto.nfssub
Contents of auto.nfssub
* ${host}:/&
Contents of /etc/nsswitch.conf
passwd: files nisplus nis
shadow: files nisplus nis
group: files nisplus nis
hosts: nis [NOTFOUND = return UNAVAIL = continue] files [NOTFOUND =
return UNAVAILABLE = continue TRYAGAIN = return]
bootparams: nisplus [NOTFOUND=return] files
ethers: files
netmasks: files
networks: files
protocols: files nisplus nis
rpc: files
services: files nis
netgroup: files nisplus nis
publickey: nisplus
automount: files nisplus nis
aliases: files nisplus
------------------------------
From: [EMAIL PROTECTED] (Roberto Inzerillo)
Subject: Re: fat32 + ext2 partitions lost. How to recover?
Date: Mon, 23 Apr 2001 17:44:32 +0000 (UTC)
Thanks to Svend Olaf, who gave me full support in every step of the
repairing process, the problem was solved without any loss of data (I said ANY).
I thank very much Svend Olaf for the help. He gave me a very precious
support in using the software he produced (findpart, cyldir etc...) in order to
fully recreate MBR, partitiona tables and fats.
--
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net
--
Posted from mx0.gmx.de [213.165.64.100]
via Mailgate.ORG Server - http://www.Mailgate.ORG
------------------------------
From: "Rich Rudnick" <[EMAIL PROTECTED]>
Subject: Re: Voodoo (usual) problems
Crossposted-To: comp.os.linux.x
Date: Mon, 23 Apr 2001 18:00:47 GMT
In article <[EMAIL PROTECTED]>, "Steve Martin"
<[EMAIL PROTECTED]> wrote:
>
> What's your monitor vertical refresh rate? If it's 60 fps, then you probably
> have DRI running right. DRI can't display more fps than your monitor can
> show anyway.
export FX_GLIDE_SWAPINTERVAL=0
this will de-link the fps from the vertical refresh. I get ~ 450fps with
4.0.1.
------------------------------
Date: Mon, 23 Apr 2001 20:05:09 +0200
From: Michael Heiming <[EMAIL PROTECTED]>
Subject: Re: How about those kernel-panics :(
ief wrote:
>
> Hi
>
> I've been experiencing several kernel-panics. With as result an uptime never
> longer then 21days, usually shorter even.
>
> A friend of mine suggested a mem-test, which I ran for a couple of
> hours...no result. Will be running it an all-nighter somewhere this week,
> 'cause it is getting VERY irritating.
Don't rely on some mem-test tools, just compile a kernel, there is an
option
to start a new gcc for everything that will be compiled, I can't
remember at the
moment (check the kernel-howto) but it may be enough even without this
option.
If make stops with:
gcc caught fatal Signal 11
It should be your RAM, change some, to test which is broken. I had
similar
problems with a machine and this "tss.cr3" error, but could never really
work out
where the problems was (RAM/Mobo).
Good luck
Michael Heiming
> The panic:
> Unable to handle kernel paging request at virtual address 000a0050
> current -> tss.cr3 = 02def000, %cr3 = 0257a000
------------------------------
From: "James Moore" <[EMAIL PROTECTED]>
Crossposted-To: alt.windows98
Subject: Re: RAID question.
Date: Mon, 23 Apr 2001 19:56:33 +0100
Steve Wolfe wrote in message <2cm1c9.339.ln@helix>...
>> >My guess is that I cant but.. I am confused at what RAID does to
>> >the partitions and at what layer it operates.
>>
>> 1. RAID doesn't improve performance, it degrades it.
>
> What? Common sense, practical experience, and benchmarks all prove that
>wrong.
>
>> Some forms of RAID provide redundancy at the cost of performance, some
>> increase size by combining disk drives, which has no effect on
>> performance.
>
> I think that you're totally confused. You're thinking of the hacks that
>simply start writing at the first of one disk, fill it up, then start on
the
>second disk. That is not how most RAID arrays work. In a RAID 0 array,
>when you write your data, it is split up and parts are written across all
of
>the disks. When you go to read that, since the drives are working in
>parallel, there is a higher throughput than a single drive.
Mirroring, or RAID0, is simple disk mirroring. This provides complete
redundancy, but there is a downside: absolutely no capacity advantage Speed
is affected somewhat, and to use RAID0 you need 2 seperate controllers
holding identical pairs of disks. For all other RAID levels, data is striped
between the disks, sector by sector, and in the higher levels of RAID, a
portion of the space (25%) is given over to parity checking. Depending on
which level of RAID you choose to implement, parity data could be stored on
one disk, in which case you need a minimum of 4 disks in a multiple of 4
(easy example: 4x10GB disks gives you 30GB total storage space available
with 10GB left over to parity checking), or it could be striped across the
disks as with the actual data. Either way provides redundancy. If one disk
fails, parity data could be used to reconstruct the 'lost' segments on a
replacement. In case the parity drive fails, the data itself can be used to
reconstruct the parity disk on a replacement drive.
>
> Now if you're talking about an IDE RAID setup, things get a little
>trickier, because IDE wasn't designed for multiple devices, and has a
higher
>CPU usage than SCSI drives. However, with two drives, one on each IDE
>controller, overall performance is definitely increased.
>
>
> The one area where performance can suffer is in writing data under
>certain RAID levels. Under RAID 1, there is a small performance hit in
>writing. In RAID 5, there is a bigger hit, as checksums must be computed.
>A lot of people will say "write performance sux under raid 5, dude!", but
in
>my experience, that's not true. Even a relatively small SCSI RAID array
can
>*write* significantly faster than a good IDE drive can *read*. While it's
>not as fast as it would be under RAID 0, that's still pretty good write
>performance.
>
Being as how SCSI drives are generally faster anyway, a small performance
hit when utilised in a RAID won't make a hit of difference even if you're
heavily into video capture/editing or live studio work.
>steve
>
>
>
HTH
======================================================================
I defragged my brain the other day, emptied the recycle bin and found I
could back up my entire knowledge onto one 3.5" disk! One wonders at the
possibilities that disk compression can bring...
[EMAIL PROTECTED] (correspondence only)
[EMAIL PROTECTED] (advertising only)
=======================================================================
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Crossposted-To: comp.os.linux.development.system,comp.os.linux.hardware
Subject: Re: A Linux emulator for Linux, does this exist?
Date: 23 Apr 2001 19:03:41 GMT
In article <[EMAIL PROTECTED]>,
Jonadab the Unsightly One <[EMAIL PROTECTED]> wrote:
| At some kind of signal from the user, the kernel writes
| the entire contents of RAM, plus the CPU registers and
| any other relevant data, to a big file on disk, saves
| a pointer to this so that it'll see it at boot time,
| and then just powers down.
|
| Next boot time, it sees that flag file, loads the big
| file off the disk to RAM, restores everything, and
| picks up right where it left off, in the middle of
| whatever it was doing.
Like a laptop...
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"I am lost. I am out looking for myself. If I should come back before I
return, please ask me to wait." -seen in a doctor's office
------------------------------
** 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 by posting to comp.os.linux.misc.
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
******************************
