• The MailBag
• More 2-Cent Tips
• The Answer Gang
• News Bytes
• Pl/Python and Cursors in Pl/Pgsql for PostgreSQL , by Mark Nielsen
• Red Hat and USB devices , by Mark Nielsen
• Configuring GDM 2.2 , by Mark Nielsen
• The Foolish Things We Do With Our Computers , by Mike Orr
• Qubism , by Jon "Sir Flakey" Harsem
• Wine: Raising a toast to your Windows Apps on Linux , by Gaurav Taneja
• My Guide To Linux Security , by Rob Tougher
• The Back Page

Linux Gazette Staff and The Answer Gang

Editor: Michael Orr
Technical Editor: Heather Stern
Senior Contributing Editor: Jim Dennis
Contributing Editors: Ben Okopnik, Dan Wilder, Don Marti

Linux Gazette, http://www.linuxgazette.com/
This page maintained by the Editor of Linux Gazette, gazette@ssc.com

Copyright © 1996-2002 Specialized Systems Consultants, Inc.

Send tech-support questions, Tips, answers and article ideas to The Answer Gang <linux-questions-only@ssc.com>. Other mail (including questions or comments about the Gazette itself) should go to <gazette@ssc.com>. All material sent to either of these addresses will be considered for publication in the next issue. Please send answers to the original querent too, so that s/he can get the answer without waiting for the next issue.

Unanswered questions might appear here. Questions with answers--or answers only--appear in The Answer Gang, 2-Cent Tips, or here, depending on their content. There is no guarantee that questions will ever be answered, especially if not related to Linux.

Before asking a question, please check the Linux Gazette FAQ (for questions about the Gazette) or The Answer Gang Knowledge Base (for questions about Linux) to see if it has been answered there.

• Kylix
• netbooting Linux from network by using pppoe
• Tools for altering installed kernel
• usb pegasus driver vend/prod id

Kylix

Hola amigos :

Alguien sabe como Podria ejecutar un programa compilado en Kylix, fuera del entorno de Kylix?

Hi, friends. Does anybody know how to run a program that's compiled in Kylix, but without having the Kylix environment around at runtime?

!ah! Un comentario demonio (daemon )siguifica Dinamic access memory, estoy equivocado?

Ah! A daemon commentary means dynamic access memory, or am I mistaken?

Gracias por su tiempo. Octavio.

Thanks for your time. -- Octavio

Octavio-- Sorry, I've never used Kylix. I just ran a demo once. I don't understand your second question. Memory is hardware; a daemon is software. And what's a "daemon commentary"?

Octavio-- Lo siento, no he usado Kylix. Ejecute' un demo de e'l una vez, no ma's. No entiendo la segunda pregunta. La memoria es hardware, un demonio es software. Que' significa un "comentario demonio"?

-- Mike Orr

Hola amigos :

Espero que puedan ayudarme, resulta que en mi maquina instalé firebird, y luego cree una base de daatos, luego desde Kylix me conecte sin problemas.

De la misma forma quise hacer en otras máquinas que tienen redhat 7.2, copie el instalador firebirdCC....rpm, pero resulta que me sale un error :

Hi friends, I hope you can help me. I installed Firebird on my computer, and then created a database, which I can connect to from Kylix without problem.

I wanted to install it the same way on other Red Hat 7.2 machines. I copied the installer FirebirdCC RPM, but got the following error:

--------------------------------------------------------------------------
warning : Expected size : 2676232=lead(96)+sigs(68)+pad(4)data(2676064)
warning : actual size : 2676232
warning : Expected size : 2676232=lead(96)+sigs(68)+pad(4)data(2676064)
warning : actual size : 2259998
error : unpacking of archive failed on file /opt/interbase/lib/libgds.so.0
;3 d021bc6 ; cpio:need
--------------------------------------------------------------------------

Intente bajar de internet el mismo paquete pero el resultado para instalarlo es el mismo error.

(Heather: oboy, my spanish is rustier than Mike's, but I'll try.)
I intend to go under the internet to packets (maybe: download the package?) but the result of installing is an error.

Como puedo solucionar este problema ?

Gracias por su tiempo Octavio

What's the solution to this problem? Thanks for your time. Octavio.

netbooting Linux from network by using pppoe

Hello,

Hi!

I want to install a Linux on my PC. This PC has a Ethernet connection towards a DSL modem. Across this Ethernet I had to use PPPOE. Is their an image available with PPPOE support? The standard netboot.img has no support for PPPOE. This image recognize my Ethernet but thus not allow me to activate PPPOE across it.

I want to install Redhat 7.3 across the net. Thanks in advance for your answer.

Grt

Wim

I'm pretty certain that most, if not all rather new images include that, exactly why you ask - to get dsl up and running.
Robos

So if someone knows the hint Wlm needs, let us know, and we'll publish your Two Cent Tip. -- Heather

Tools for altering installed kernel

I am helping someone out on another list I am on and a have a query.

Does anyone know of a reliable utility to alter parameters on an installed kernel other than rdev.

In particular whether there is anything that will get rid of debugging in the kernel?

I would just recompile personally but if anyone knows of a tool that would be useful.

usb pegasus driver vend/prod id

Hi,

I have a USB ethernet adapter based on the Pegasus/Pegasus-II chipset. I tried to use the pegasus driver with a 2.4.3 kernel but the driver is not claiming the vend/prod id of the USB adapter. Unfortunately I am unable to recompile a more recent kernel for my system. I would like to know if there is a way to spoof the vend/prod id somehow (without flashing the USB adapter) so the pegasus driver claims the device. I am thinking there is perhaps some kernel mechanism to increase the set of devices claimed by a particular driver at runtime. I think a more recent kernel version would solve my problems but a recompile is not an option for me because of my particular setup. Any ideas?

Ebo

Don't use quoted printable with no fancy characters to defend.

Maybe one of our readers can suggest the right tricks for compiling modules that are only major-version dependent instead of minor version dependent. I think some people who use linmodems might know a Tip or two.

Note you still have to have the right symbols present in your kernel. If they aren't, you won't be able to use the module safely, even with insmod -f. -- Heather

• Mistake in December 99 Linux Gazette
• Re: Making executables smaller
• Is it possible to have a latest issue link?
• Linux gazette 79 article improvement
• Kudos
• Difference between LJ and LG
• Etiquette Wars -- the final salvo.
• re : The Complaint Department: Typos and Grammatical Errors
• Did you see that Fabor gets famous?

Mistake in December 99 Linux Gazette

Hi,

I'm a French student and I tried to compile a little program I found in your magazine in your "Linuxthreads programming" section of December 99 magazine. The programm is not working and I didn't find why. Could you help me please to make it work.

Thanks

Your best bet is to contact the author, Matteo Dell'Omodarme. His address is at the top of the article. But you'll need to say something more specific than "not working". See http://www.linuxgazette.com/tag/ask-the-gang.html and scroll down to "doesn't work". How exactly did it fail? -- Mike

Re: Making executables smaller

John sent us a 2 Cent Tip, so the Answer Guy wondered...

Are you the same John Fisk that originally started the Linux Gazette? -- JimD

Hello Jim,

Thanks so much for the note. And yes, it's me. Since the last submission (now a couple years ago) I've been rather busy finishing up medical training (Pathology and Medical Informatics at Yale) and my wife and I have adopted a little girl from China (with another one on the way). Thomas Adam (the new maintainer of the Weekend Mechanic column) and I just started corresponding and I took the opportunity to look over the latest edition of the LG. The folks at SSC have done a commendable job of keeping the LG alive and very active. I really appreciate the great job they've done.

cheers,
John

Is it possible to have a latest issue link?

Can we have something like a www.linuxgazette.com/latest link?

http://www.linuxgazette.com/current -- Mike

linux gazette 79 article improvement

Hi Krishnakumar R.

Your series of articles "Writing Your Own Toy OS" on Linux Gazette is the best ever.

1. http://www.linuxgazette.com/issue77/krishnakumar.html Part I
2. http://www.linuxgazette.com/issue79/krishnakumar.html Part II

I am enjoying it very much and I'm eager to learn something about the 80x86 protected mode.

I also have a suggestion. In part II, I think the code for write.c can be improved a little.

When reading sect2 with

   file_desc = open("./sect2", O_RDONLY);
   read(file_desc, boot_buf, 510);
   close(file_desc);

instead of reading only 510, you can actually read 512 bytes with no harm.

It doesn't hurt to be 510, but I think it would be more easy to understand it's 512, so that learners won't be thinking that the second sector also have to finish with the magic numbers x55 xAA, just like the first sector does.

I'd like to know what do you think about it and if you agree.

See you and thanks a lot, man!

Silvio Luis Leite Santana

[The article was changed at the author's request. -Iron.]

Kudos

Guess what - I got a job . A guy I worked with a few years back was reading the Linux Gazette, and saw my article.

In May's issue, http://www.linuxgazette.com/issue78/tougher.html -- Heather

He emailed me, and one lunch and two meetings later I was employed again. You definitely get the assist on that one.

I also credit your magazine for keeping me sane the last 8 months - writing for you has been a great outlet. I'll keep the articles coming.

Difference between LJ and LG

At our company there has been some discussion as to the differences of the Journal and the Gazette. Would you please respond with SSC's definitive description of the difference and is both being published at this time. Thereby we can put to rest any further discussion at our work place.

Regards,
Wyman Griffin

Linux Journal is a commercial print magazine. Linux Gazette is a free e-zine that SSC donates some employee time to. LG was started by an individual, John Fisk, then SSC took over responsibility for it when John no longer could.

LJ pays authors for their work. It has a stricter standard for what it will accept, does professional proofreading and technical editing, won't publish anything that's been published elsewhere (with occasional exceptions), and doesn't allow republishing without permission. Each LJ issue has a theme, and articles are solicited for that theme (although any issue will have lots of non-theme articles too). Space issues determine the number of articles and their length. (Because the printer prints on 32-page sheets, you have to add pages in groups of eight.) LJ also publishes several series of publicly-accessible web articles at http://www.linuxjournal.com, for which we pay the authors just like magazine articles. (The magazine articles are also available to the public after three months.) For specific questions about LJ, contact the Linux Journal Editor, ljeditor@ssc.com.

LG does not pay authors, does less time-consuming proofreading, and publishes pretty much anything we have permission to publish that's about Linux, contains some significant content of a more or less permanent nature ("new information" or reference material), doesn't unfairly slam/slander anybody, and isn't a mindless advocacy rant. But we don't go looking for content, we let it come to us. LG issues do not have a theme, we just publish articles whenever they arrive. There is no particular number of pages to fill, so we don't pay attention to article count or article length, although we do try to keep the issue size to less than 2 MB compressed (occasionally 4 MB). LG does not accept advertisements, although we do have a sponsorship program that gets your logo on the home page. LG is published under the Open Publication License, so readers may copy and redistribute it (for free or profit) as much as they wish.

Etiquette Wars

the final salvo

Once upon issue 64 or so in the Mailbag, we pubbed a note from a cheeky fellow who decided to tell us off about our bad attitude. He also took the tack that we were a single person and that it was somehow our "fault" if a bunch of sensitive souls ever saw it. Obviously there are others in the world who think he was right, but Marko took umbrage with us for being cheeky enough ourselves, to post his message when we replied.

Scofflaws may enjoy reading the offending note and its reply together: http://www.linuxgazette.com/issue64/lg_mail64.html#mailbag/1 and a number of the Gang replying to Marko in issue 78: http://www.linuxgazette.com/issue78/lg_mail.html#mailbag/4 -- Heather

Dear Ben,

thanks for your reply, however I mastered etiquette course (sorry for misspellings) and know very well the difference between vulgar and etiquette.

Gazette is distributed under free licence and so when people reading it in a computer club (from 8 to 88) they should get a good viewpoint of gazette or bad view point?

And "vulgar" is a compliment when it comes to the Linux Gazette; we are indeed "of the common people".

re : The Complaint Department: Typos and Grammatical Errors

Boy, this topic is from the dusty shelf; http://www.linuxgazette.com/issue37/tag/23.html and http://www.linuxgazette.com/issue36/tag/67.html -- Heather

With all of that guy's complaints, I'm reminded of an old advertisement for learning shorthand:

u cn us unx.

Cheers, -- Jay R. Ashworth

Point being, that the informational difference between their/they're, or more appropriately, between "kernel core team has soundly reject suggestions that Linux adopt..." and "kernel core team has soundly rejected suggestions that Linux adopt..." as quoted the information content difference is nil, for a native English speaker. Don't let him get too down on you -- we just are pampered by having a MLA in the first place to standardize these issues.

Did you see that Fabor gets famous?

Famous? Obviously not famous enough for my name to get spelled correctly!

Hi all! Well, there is this "Is linux dead?" comment on /. and in the MSNBC article (http://www.msnbc.com/news/772215.asp) Fabor is quoted rather extensively. The /. news comment is really bad (doesn't fit at all) but Fabor comes along really nice.

Thansk for the compliment. When I first read the article, I was sounding like Chandler Bing's ex-girlfriend on "Friends"

<Janice>Oh - my - gawd!</Janice>

The article was, IMO, a back-handed compliment.

He should have mentioned TAG though

I think I did.

As a question to fabor: Why do you say (as the article quotes) "It's for geeks"? I mean, we're mostly geeks (ok, all) but those people who write us with questions are most certainly not geeks (most of them) since then they would probably figured out the thing themselves. These lusers might have some probs with "linux" but only because when they buy some win crap they wine to the support stuff of that firm, in GNU/linux they get all the tools at once and don't have such a technical support (suse and redhat for a short time at the beginning, ok..) to ask questions. So they come to us and thats what TAG is for. Bit I think that most luser get along with GNU/Linux pretty well given the fact that GNU/Linux is far more powerful and customizable. With most questions they come to either us or debianhelp for example, they wouldn't even ask those questions on win since there they wouldn't get the fix idea to run their own webserver just-for-fun since there it isn't that much free (as in beer) software to play with (that sentence is rather crap, granted, but I hope you get my general drift) -- Robos

I'll disagree to the extent that there is plenty of free-as-in-beer or shareware available for Windows, but to a certain degree you have to be geeky to know where to go looking for it. Been there, showing people cool stuff like virtual desktops and icon managers and replacement command shells. -- Heather

Well, you really need to hear the question I was asked! I was asked "Why is Linux popular in the enterprise and with upper management but not popular on home PCs?"

"It's for geeks and they thrive in the enterprise where the power of Linux is appreciated." Then I went on to mention about MS licensing practices taht forbid other OSes or changing of the boot sequesnce. I said THAT was why Linux isn't popular on the desktop. I even mention BeOS and Hitachi.

So, to make it clear: I think with a little help (about as much as you need in the beginning with windoze) and some distro like suse or mandrake a pretty normal user can now easily use linux and the accompained software (as long as they can and are willing to read).

I agree. And while my student/attendee, Dave Potter, did say those things, he came off alot different than the article sounds.

Anybody know of a course in "how to answer a journalists' questions without being misrepresented?"

-- Regards, Faber

What Robos had to say only works if you know what parts they are likely to misrepresent. Make them repeat it back. Squeeze 'em if they can't get it right. Unfortunately the cultural gap is likely to foster addiitonal assumptions based on whatever you say or do to try and keep matters straight.

If people don't want to understand, we can't make it happen. That's the real nature of freedom, folks. But we can say things our way in our own venue, and when they come looking for us, it'll still be here.

Remind me to ask Faber's question in the press room at LWE though... -- Heather

• ask-the-gang

ask-the-gang

Compliments to Ben for continuing to make tag/ask-the-gang.html better and better.

Thanks much! I treat it as a serious resource, and try my best.

Gosh, this is twice I've complimented Ben in one week. I promise it won't happen again.

"I guess his heart just couldn't stand the shock - we've got syncope and V-fib. All right, lets give him the whole 200j. ... Sync off... CLEAR!" <BZZZZT!>

"OK, got paced rhythm and pulse. He'll prob'ly pull through if he doesn't get any more of those compliments..."

-- Ben

Too many compliments? Just stack them over there next to the groceries, and I'll add them to the virtual beer and munchies in the Answer Gang fridge. (See tag/members-faq.html for more about the fridge.) No fuzzybears were harmed in the writing of this document -- Heather

This page edited and maintained by the Editors of Linux Gazette Copyright © 2002
Published in issue 80 of Linux Gazette July 2002

HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

• Fvwm Buttons
• Creating WAN "LAN' with one IP Address
• customized linux install cd?
• Ghostscript fails after printer driver install
• quick disaster recovery
• question
• checkinstall
• 'crypt' error !!
• demand dialing
• Modem speed and diald
• Exchange with Linux
• Grub vs LILO
• Email Linux To Windows - a simple solution for reference
• linuxconf setup
• Kernel Message: VM: Killing resource foo (bar)....
• Parsing Strings To Equations
• lpd/lpr problems with serial printer
• Getting files out of a .rpm file without installing it
• ramdisk and initrd fundamentals?
• Re: Making executables smaller
• Linux Journal's Weekly News Notes Tech Tips
  • sending Microsoft Word documents
  • Keeping a persistent session as you log in and out from different terminals
  • Outlook to Evolution
  • Mozilla port paranoia
  • Renaming a file that has a special character in the name
  • Have Ethernet cables, will travel
  • subscribe to LJWNN

Fvwm Buttons

This is in reply to Help Wanted #5, in issue 79. Thomas replied via the FVWM mailing list. More details about that can be found at: http://www.fvwm.org/mailinglist.html -- Heather

Hello,

In answer to your question as to why when you press a button on your panel, it stays depessed is to do with the way in which FVWM handles exec() a program via the $SHELL of the $USER.

If command is an fvwm Exec command, then the button will remain pushed in until a window whose name or class matches the quoted portion of the command is encountered. This is intended to provide visual feedback to the user that the action he has requested will be performed. If the quoted portion contains no characters, then the button will pop out immediately.

Note that users can continue pressing the button, and re-executing the command, even when it looks "pressed in."

There is a way around this, and I have found that if you append a "&" character at the end of your command that is bound to the button, then that sometimes solves your problem -- but not always.

Hope I have helped, Kind Regards,

Thomas Adam

-- "The Linux Weekend Mechanic" -- www.linuxgazette.com

Hi,

Thanks a lot for taking the time to answer my Q. I actually managed to sort it out some time ago, but you are right. I had some fiddle with the window name.

Best regards
Hans Borg.

Creating WAN "LAN' with one IP Address

If I'm given a network address 192.168.7 (Class C) and have to create a WAN with 5 routers, how do I do it?

I can Subnett but the 3rd router keep saying the Network address is already used, when i try to put the subnet address there.

How do i do it.
RURI!!

We have a very good piece on that sort of thing in the back issues. It's called "Routing and Subnetting 101" and is one of the longest postings ever written by Jim Dennis. Several professors have used it in their coursework and even though Linux was much younger then the principles are still valid.

It's in issue 36. A professor asked about it in issue 37's mailbag, and some followups appeared in issues 51 and 59. Of course you could have learned this by typing "Routing and Subnetting" into the Linux Gazette search page: http://www.linuxgazette.com/search.html

...and you can easily get to those articles by visiting the Answer Gang Knowledge Base: http://www.linuxgazette.com/tag/kb.html

customized linux install cd?

These are in reply to Help Wanted #1, Issue 79. -- Heather

Hi, I did about this on a RH7.2 but I don't think it'll change this drastically under RH 7.3. You can point your browser to

http://lcmpc10.epfl.ch:8080/Menu/Docus/RedHat%20CD

(sorry for the space in the url...) it even describes how you can add a kernel on your own and get it to run...

Ineiti

The easiest way to customise the install is with Kickstart. We have done it (and learned a few things on the way).

Haven't time for a comprehensive reply at present but, if kicksatart hasn't been covered, I could put something together.

Rgrds

Peter

Ghostscript fails after printer driver install

This is in reply to LG 79, Two Cent Tip #4. -- Heather

Hello,

Rich Price may want to try a Laserjet 4 driver instead of trying to figure out the Xt dependency.

I have the Samsung ML-1450 and it's quite happy pretending to be an HP Laserjet 4.

Matt Easton

quick disaster recovery

I may expand this into an article, but for the very common scenario of "no init found" "unable to open an initial console" (usually after hard crash) a couple of possible causes which I have not seen anywhere else

There is a fair chance that files on / have been corrupted wiped including /dev.

Solution (very Rpm specific)

So mount rescue media, check for files on /

if missing mount cdrom from install and do:

rpm -Uvh dev-<version>-rpm

to re-instate dev files

then progressively force re-install rpms until you can boot

Then when you have managed to boot do this:

rpm -Va|grep missing>filename

This will print to a file all the files that are missing from your system according to your rpm database.

Then for all the files given do rpm -qf <filename> which will give you the name of the rpm

Then re-install the rpms in turn.

This is best done manually so you can check whats missing.

Should only take around an hour in total at most.

Certainly preferable to doing a re-install.

I came across this on my own box a while ago after multiple power cuts in succession (I'm poor so no UPS)

The advantage is that your modifications are far less likely to be hosed as in a re-install.

question

To Whom It May Concern:

I went to ebay and found all these used laptops/notebook puters, but I

have no clue which one to select. For example, "Intel Pentium II=AE 366Mhz 290MB RAM 6.1GB HDD CDROM Sound Windows 98 Office 2000 ..."

What does all that mean? And how I go about finding a good, used laptop, like what trait(s) do I search for?

-Thanks ,
Desperately needing laptop

If your question aims at running linux on that thing, compare what http://www.linux-laptop.net has to say to the model you like.

checkinstall

I dont know whether anyone else on the list has used the utility checkinstall available at:

http://asic-linux.com.mx/~izto/checkinstall

What it is the solution to the problem of maintaining a rpm/deb based system with compiling programs from source

Basically what it does is runs make install and then makes a functional rpm and installs it.

It is not perfect but certainly works well enough to continued use

I regard it now as pretty much indispensible when | am following a project (eg: gnome2)

My feeling is that it ives the flexibilty of using source packages without losing package management

Hi!

Well, there was some coverage of that utility here in germany in the magazine (print) "linux-user". Seems to be quite nice, I've used it several times and it worked most of the times. Not always, but when it doesn't work you can still fall back to

./configure
make
make install

Robos

'crypt' error !!

I have got a peculiar problem in hand. Got this code compiled properly in red hat Linux 6.1(g++ compiler version 2.91.66 ) but giving error in red hat 7.1(g++ compiler version 2.96).But if compiled with red hat 7.1(gcc compiler version 2.96) , it is doing perfectly fine.

why this in-consistency ?

Source code: Server.c

Command:

g++ -lcrypt server.c

Error: 'crypt' undeclared

Since it seems nobody tried an answer yet I try to add some cents:

• first thing coming to my mind is probably a typo in the mail -- but server.c and Server.c might be different files....

You do the right includes (whatever they are, crypt.h problaby), do you? There might be a difference where crypt is stored for gcc and for g++ -- so gcc and g++ might behave differently. Also g++ might might have changed in default location or default behaviour of including C headers. Try to locate crypt.h (or wherever crypt is defined). Is there a g++ version of it? What happens if you put -I ad -L explicitly to the gcc crypt path?

Then crypt is probably compiled by gcc -- this has a different routine name mangling then g++, so you might have to call not crypt but '_crypt' or 'crypt_' or something like that (speaking from very little experience with how to use fortran subroutines in C -- and a peculiar problem lately: if I compiled a subroutine with gcc I got a "..." undefined from the linker. If I compile it with g++ everzthing works.

I would have expected "better" integration with gnu c and c++ -- but there you go.

K.-H.

demand dialing

This is in reply to LG 79, help wanted #7. -- Heather

There are two linux versions that I've used as a server for demand dialout for internet access, and both worked well. One is Coyote linux, which is a floppy disk boot version, and can be run on a 386 with numeric coprocessor or a true 486 (or 486sx with numeric coprocessor). I don't recall it's memory requirements. The other version that is good is the mitel (formerly e-smith) at www.e-smith.org. It requires a 586 class processor, but also setsup DNS, and other server functions.

Modem speed and diald

I've mainly been connecting to the internet using diald, but I've noticed that I'm only getting about 3.5 KBps , whereas on W98 I get about 5KBps. A little experimentation shows that dialling with kppp gives about 5KBps as well.

kppp seems to use an initialisation string of ATM1L1, but changing MODEM_INIT to "ATM1L1" in /etc/diald/connect, didn't improve the performance.

MODEM_INIT started out as "ATZ&C1&D2%C0". I changed "%C0" to "%C3" to ensure that compression was enabled, but this made no difference. I can't find an option in diald to log exactly what's sent to the modem and I can't see any conflicting options in the configuration for pppd.

Any suggestions for how to track down why kppp gets better performance than diald would be appreciated.

The modem is an MRI 56K internal modem.

I'm not sure how you would test this, but I suspect that it's not your PPP connection that's slowing you down - "diald" uses SLIP as a "fake interface" that's always up, which is why you don't get error messages from Netscape and such when you try to connect. It listens for requests, then makes the PPP connection "behind your back". It's been a long time since I've used it, and I'm rather fuzzy on the details, but ISTR that "diald" let you play around with SLIP settings... sorry I can't be of any more help, but that's pretty much the extent of what I remember. I also STR that "diald" had a good set of documents with it which I found very helpful in working around a problem that I had with it. Good luck.

Exchange with Linux

Hope this is the right address for answers as well as questions. Regarding Linux Exchange In my quest to use Linux without having to use Windows in our network I discovered a couple of simple solutions.

1. Most any email client will work with a default install of exchange if you enter your login as in the following:

domain/username

Of course all the group features will not work with this solution but simple email is no problem.

2. You can use a browser with a default install of exchange since it also installs IIS as a webserver. Various browsers will have different degrees of success since of course IE is the "prefered" browser. Type the following in your browser substituting your exchange servers correct IP address:

http://ipaddress/exchange

This will give you access to all the group features if your browser will render the Microsoft proprietary technology. Thanks for reading

John Helms

Grub vs LILO

Does anyone know what the Grub command is that replaces the LILO command append="hdb=ide-scsi"

It would be the same syntax, minus the "append=".

The append="" stuff is part of the Lilo syntax. Your grub.conf file should look something like:

default=1
timeout=10
splashimage=(hd0,0)/grub/splash.xpm.gz
title Example
        root (hd0,0)
        kernel /vmlinuz ro root=/dev/hda3 hdc=ide-scsi vga=1 console=/dev/tty2 CONSOLE=/dev/tty2
        initrd /initrd-2.4.18-whatever

where you are allowed to use "\" and the end of a line, to mean line continuation.

Email Linux To Windows - a simple solution for reference

1. Set up mail server on Linux with user ids as applicable (exim reads userids from linux box)
2. Set up fetchmail to poll server at isp
3. install and activate pop3 server on linux box
4. poll for mail by pop3 to the linux box account

a more detailed example is at:

http://www.linuxgazette.com/issue43/stumpel.html

linuxconf setup

This is in reply to LG 79 help wanted #3. Our reader wondered about setting up ACLs so he could access his Linuxconf remotely without letting everyone else in. -- Heather

http://www.redhat.com/docs/manuals/linux/RHL-7.1-Manual/custom-guide/linuxconf-lcinterfaces.html

That should help you Good luck

Kernel Message: VM: Killing resource foo (bar)....

Hi Gang,

Hello Thomas!

I am wondering if someone would be so kind as to expain to me why I get the following error messgae:

Grangedairy: kernel: VM: killing resource acroread

[ quickly find-grepping that in the kernel source tree ]

Ok, it's in "arch/i386/mm/fault.c". (Actually for 2.4.17 it's "killing process xxx" but I presume it doesn't really matter). It seems to occur whenever a process tries to read something from a memory page which is not accessible because of an out-of-memory condition.

The above message is brought about, by an abnormal exit of the adobe acrobat reader. For some reason, my computer will slow down to a snails pace when I am reading a pdf document,

That's when the system starts swapping out madly. Does it happen with any PDF document, or only really big ones? What if you disable "Use Page Cache" and/or "Allow Background Download of Entire File" in your Acroread preferences (assuming those are currently turned on)?

and then X will kill acroread without any warning.

X itself has nothing to do with the actual killing. It's the way the kernel preserves your system from a total crash. What amount of RAM do you have?

When I check my /var/log/messages file, I get the above message.

What causes this, and what does it mean???

OOM. Virtual memory exhausted.

I tried running "strace" on the "acroread" process, but the file ended up being 38.2MB, despite me telling it only to display a certain number of lines

So maybe your PDF document is a resource hog. Or there is some incompatibility between the version of Acrobat Reader installed on your system and your current set of libraries (even Acrobat 5 is dynamically linked, mmph... this has obvious advantages but since their reader is only distributed in binary form...) or some bug in Adobe's product turns it into an self-destructing madsoft.

It yeilded nothing useful, anyway from what I could see.

The kernel's divine intervention in such cases is kinda brutal. Expect the same sort of things as those resulting from a SIGKILL: perhaps the process was "innocently" trying to read() something then it just vaporized into limbo.

Thanks,

Well I hope this'll help you somewhat.

-- Didier Heyden.

Parsing Strings To Equations

Can anybody in this list suggest a method to parse a string into a mathematical equation and then compute the values according to the equation.I have to do this in C in Linux. Is there any such code available in C.

Is there a similar command in Linux.

I have tried myself a lot.It is getting more and more complex.This is going to the level of a compiler design.ie,the task is similar to "how a compiler reads the source code and manipulates it", i also have to do a somewhat similar task.

Yes. Maybe not as complex as a C parser, because the corresponding grammar will probably be much cleaner; but the part of the program which will be dedicated to the symbolic computation won't be easy to write. But if you try with tools such as `flex' and `yacc' (or `bison') the parser itself can be implemented in a relatively short time, at least once you have read enough documentation about those development tools, and thoroughly understood the basic concepts. The related `info' pages are most useful, and finding tutorials and likewise on the net is fairly trivial.

What's interesting with `bison'-based parsers is that one can more easily split such tasks into smaller, independent parts, i.e. separate completely the syntactical analysis of the source code from the rest. For the parser itself, the work essentially consists in writing a correct grammar for the corresponding developed language.

Then I tried google search but didn't find any useful links in C.I haven't done an extensive search.

For symbolic computation LISP-like languages may prove more adequate than C.

Then I contacted you,the answer gang. I thought if somebody here has previously done a similar task..it would be helpful for me if they share it.. or just give me some links to some resources on the net having information about this...

Well, what I'm saying here is nothing more than one possibility among others. Keep in mind that whatever solution you choose, you'll have to invest a good deal of time to fully work it out.

regards sree

HTH,

-- Didier Heyden.

lpd/lpr problems with serial printer

This is in reply to LG 79, help wanted #6. Rather than solving the problem as asked, Doug suggests a different approach. -- Heather

I have only used lpr over a TCP/IP network. Would it be possible to connect the printer to a serial to ethernet print server (Intel and HP work with Linux)? Then network to the Linux box. And then have all devices/terminals use lpr/lpd.

Doug

Raul knows which software he should be using instead. -- Heather

Hi,

From the article:

I am using Mandrake 8.2. I have recently installed a serial printer using a Digi Classic-8 ISA card. . . . I can print to this printer by using 'cat {filename} > /dev/ttyS11' and this works just fine, however I cannot get lpr to print to this printer.

AFAIK, lpd does not come with mdk 8.2. IIRC, no up to date distribution ships lpd anymore. At least they have replaced it with LPRng or CUPS.

I don't use MDK, but I do use Cups which is the default printer system in MDK.

I have two serial printers (ttyS0 and ttyS1) working fine.

All you got to do is to log into http://localhost:631 and add a new printer. Select the proper Serial Port and the most important, set the correct values to the printer. My printers works at 9600 and 4800 only.

This should be more than fine to live test.

If you will keep with lpd, the /etc/printcap is the place to go. The man page for printcap will provide the right arguments for setting the ttyS printer.

Regards,

Raul Dias

Getting files out of a .rpm file without installing it

Ever wanted to get those files out of a .rpm without installing it? Ofcourse it's easier to just install it, but sometimes it is not possible because a newer version may already be present on the system.

In such cases, the utility "rpm2cpio" can be used to extract the files of the RPM into a cpio archive.

For example:

$ rpm2cpio < xmms-2.4.rpm > xmms.cpio

The files can then be extracted from the cpio file using the "cpio" command. For example:

$ cpio -i -d < xmms.cpio

In this case the files will be extracted with absolute paths into the present directory.

A good 2-cents' worth, Ashwin. However, you can do this even easier by selecting the RPM and pressing 'Enter' in Midnight Commander; all the files are under "CONTENTS.cpio", and you can explore the rest of the RPM structure at will. There are also two executable metafiles called "INSTALL" and "UPGRADE"; if you're viewing the file as root, you can do either one simply by scrolling down to them and pressing 'Enter'. The above functionality also applies to Debian's DEB files.

ramdisk and initrd fundamentals?

This is in reply to Issue 79, Help Wanted #4 -- Heather

{1st time at this, let's see what happens}

What documentation is available for ramdisk and initrd fundamentals?

When "initrd" is specified in "lilo.conf" and the Linux kernel is configured for ramdisk support, and the system is booted, what ramdisk image is loaded first or at all?

You also need to configure the kernel to do an "initrd" as well as a "Ramdisk" if this is what you wish to do.

The "initrd" image or the kernel ramdisk image?

"initrd" loads first. It loads from the bootloader (LiLo, Loadlin, etc.) while still in "real" mode using the PC's BIOS calls, and loads before the kernel loads. The purpose of the initrd is to provide what ever support files that the kernel might need to find its root file system. Typically people wishing to have a generic boot disk that will boot just about anything can put disk controller drivers, various filesystems, PC-Card drivers and even networking drivers in here. Then the kernel is set to attempt to load all of them, but of course only the ones that match existing hardware will actually load. After this process, the initrd can be destroyed to free up the memory (but I don't know how this is done) as the boot process continues to find the real root filesystem to boot from. This file system can reside on anything the kernel has a driver loaded for, which includes a "ramdisk". I'm not sure if something in initrd may be needed to set up the "ramdisk". After the kernel finishes booting, control is handed off to /sbin/init which then begins executing things in /linuxrc.

In summery: the initrd is loaded using only BIOS and simply saves you from having to compile all these drivers into the kernel, only to have many of them unused and take up memory in the running system. It also the only way to deal with PC-Card devices that might be needed to boot that are only availble as external modules

What programs are responsible for the loading the "initrd" and "ramdisk" root images (kernel or LILO boot.d).

"initrd": is loaded with the bootloader. For the ramdisk, I'm not sure if the kernel can do this automatically, I expect you have to wait until an "init" "rc" script residing in initrd can do it

Note that the Linux kernel does not need LILO in order to load a ramdisk from a second floppy using the same drive.

Right, LiLo is completely out of the picture by this time.

When is "pivot_root" an implicit process, and when must it be explicitly invoked?

Don't know

What is the difference between a LILO specified "initrd" and the kernel specified ramdisk loaded as root?

See the above lengthy paragraph

I have a fundamental problem understanding the relationships between an "initrd" image, ramdisk root image, and the use of /initrd, /linuxrc, and swap-root.

Sorry that I don't have all the answers. In fact I may be wrong in places since I have never really done this sort of thing before.

James


Linux is the one competitor Microsoft
can't buy,
can't intimidate, and
can't stop.

If Linux were an "it" I might agree with the "one" part, but if you just take Linux as the kernel, it's like saying rotary engines (the natural competitor to the cylinder style) are competing with Ford, and if you don't, then there are a lot of different brand name Linuxen (Linuces?) out here to choose from. Plus I don't think it's fair to forget all the freely usable BSD variants out there. They could make something with a BSD core but they can't force "customer lock-in".

They could "buy" Linux - in the sense of using it commercially - but they wouldn't pay the price we charge (freedom for derivitive works).

Of course since we have so many varieties - and any effort to mention how un-glitzy and unready that Linux thingy is just causes more people to notice it as a possibility - plus the fact that many Linux distros are non commercial and all the parts and then some are easier to get than your average blinkylight at Radio Shack, I have to agree with the last. Look out for that "Palladium" chip trick they're trying to pull though. It sounds just like the Clipper Chip of yesteryear. -- Heather

Did you look at the files ramdisk.txt and initrd.txt in the Documentation subdirectory of the kernel source itself? It may be obvious, but many of your questions seem to be answered here.

The Slackware-style ramdisk at startup (see file ramdisk.txt) is loaded by the kernel from a device (almost always a floppy) where the offset can be specified. This ramdisk remains the root file system throughout the lifetime of the kernel. You can boot the kernel directly from a floppy (just dd the image to /dev/fd0) without a boot loader and still use this type of ramdisk.

The initrd ramdisk is much more flexible, but you do need a real boot loader to use it. It is loaded into memory by a boot loader. All Linux boot loaders (LILO, LOADLIN, SYSLINUX and GRUB) can use this type of ramdisk. The file /linuxrc inside the initrd ramdisk is the first program to be executed. Linuxrc can do any of the following things:

• load additional modules (e.g. scsi host adapter drivers that the kernel needs to access the hard disk)
• prepare other file systems, e.g. on the hard disk or on different ram disk devices (/dev/ram2)
• temporarily mount other file systems (e.g to retrieve extra programs).
• set the real root device (by writing to /proc/sys/kernel/real-root-dev)

When linuxrc terminates, the real root device (possibly a different ramdisk on /dev/ram2) will be mounted as root and /sbin/init will be run as usual. So the initrd ramdisk will be the root file system only as long as /linuxrc is running. Using this approach you can forget about the older type of startup ramdisk.

Because it is loaded by the boot loader, an initrd ramdisk can be loaded from the floppy image on an El Torrito bootable CD-ROM, while the older type of startup ramdisk can't.

The kernel command line is passed to the kernel by the boot loader (e.g. LILO) as a pointer to a string in memory. The kernel parses various arguments on that command line for itself and can pass the rest as a command line to init.

Just some minor additions:

• I don't think that the boot loader passes the pernel command line as a pointer (I think it's at a fixed location), but it _is_ a string in memory that the kernel processes later.
• The boot loader loads any initrd ramdisk into RAM using the BIOS disk device (therefore it works from a disk image on a bootable CD-ROM) and places the command line also in RAM. Next it jumps to the setup.S part of the kernel From then on the boot loader's work is over.
• The kernel moves itself, the command line and the initrd disk image around in RAM, decompresses itself and initializes a lot of things.
• If there is an initrd ramdisk, the kernel can gunzip it (if it is compressed) and copies it to /dev/ram (internal ramdisk device). Next it does the temporary root device trick that I have already explained.

Further reading: kernel source tree:

• arch/i386/boot/bootsect.S (is only used when the kernel is booted directly from a floppy, is skipped by LILO and other boot loaders).
• arch/i386/boot/setup.S (real-mode initialization, entry point for LILO).
• arch/i386/kernel/head.S (protected mode initialization).
• anything in arch/i386/boot/compressed (see how the kernel decompresses itself).
• init/main.c initialization flow, including temporary initrd mount and starting of init.
• drivers/block/rd.c (ramdisk driver, also for initrd ramdisk, decompression etc). Almost all code (including initialization) is shared between the initrd and normal ramdisk.

-- Lennart

Re: Making executables smaller

Hello Gals and Guys,

I enjoyed reading the latest batch of 2 cent tips and thought I'd pass along one more small bit of information:

Besides using "strip" to reduce the size of an executable, if you're into compiling from source you can use the "-Os" optimization, which will optimize for size (should work with an respectably recent version of GCC). For the ultimate in downsizing, you can also link your apps against any of a number of libc derivatives. Check http://freshmeat.net for the latest versions of:

• diet libc: http://www.fefe.de/dietlibc
  this site also provides documentation and helpful links
• uClibc: http://www.uclibc.org
  designed specifically for embedded systems; additional links avail.

Several other similar projects are out there; these were the first to come to mind. Thanks so much.

cheers,
John

Linux Journal Weekly News Notes Tech Tips

sending Microsoft Word documents

If someone asks you for a file in Microsoft Word format, don't panic or start to do something dumb, like dual-booting. Just convert the file to HTML, do a cp file.html file.doc and send the ".doc" file. Microsoft Word will automatically import it.

Keeping a persistent session as you log in and out from different terminals

Here's the Ten-Second Guide to screen for people who just don't want to type stuff over when they lose their net connection.

1. ssh to the server you need to work on.
2. Type screen.
3. Do what you need to do. It will be inside a screen session.

When your connection fails:

1. ssh to the server.
2. Run screen -r to resume your session where you got cut off.

Joy!

Outlook to Evolution

Upgrading from Microsoft Outlook to Evolution? To free your address book from Microsoft's proprietary format, just sync it to a Palm Pilot then sync it back into Evolution. (If you have to borrow a friend's Palm Pilot to do this, back it up first with pilot-xfer, then restore when you're done.)

Source: Ari Jort, New York Linux Users Group

Mozilla port paranoia

You've just hooked up a cool web-administered device; you type in the device's IP address and port number, and Mozilla says, "Access to the port number given has been disabled for security reasons." What?

Fix it in Mozilla's all.js configuration file, which probably lives in usr/lib/mozilla/defaults/pref/all.js or somewhere like that. If the banned devices are on ports 1080 and 31337, add the line:

pref("network.security.ports.banned.override", "1080,31337");

to the all.js file.

Renaming a file that has a special character in the name

Help! I can't rename a file with a special character in its name!

If you have a file called important?file, and the ? is really some character you can't figure out how to type, try this:

ls | grep important?file

Make the pattern after grep long enough that it matches only one file. Then, when the above command matches only one file, go back up and edit the command:

mv `ls | grep important?file` important-file

And you've renamed the file without ever figuring out its true name.

Have Ethernet cables, will travel

The lightest, most compact way to be prepared to hook up to whatever Ethernet connection you find is to carry one regular cable, one crossover adapter and one RJ45 coupler.

             Plug                           Jack

Straight     none, or coupler +             straight cable
             straight cable


Crossover    coupler +                      crossover adapter +
             crossover adapter +            straight cable
             straight cable

This page edited and maintained by the Editors of Linux Gazette Copyright © 2002
Published in issue 80 of Linux Gazette July 2002

HTML script maintained by Heather Stern of Starshine Technical Services, http://www.starshine.org/

Contents:

¶: Greetings From Heather Stern

How to Investigate a System Lockup

How to send email without a DNS server?

Starting many X sessions

Playing CD Music Digital Output

need help! Mandrake 8.1 fd0 unknown device, and downloading pictures from a digital camera

How to optimize space usage for multiple files on multiple CDs

/dev/scd0 is not a valid block device?

Greetings from Heather Stern

The Answer Gang's Editor Gal

Hello all, and welcome once more to the lively world of The Answer Gang. I'd like to hand out a special thank you to all the readers who definitely read the posting guidelines before asking the Gang their question. It's helped my work a lot.

Almost the only spam that escapes Dan's traps anymore are those dratted conman scams telling me about how their late uncle / business partner / revered general or whatever left them a quadzillion dollars / francs or whatever and they can't get at any of it unless you as a friend / distant relative / confidant / conveniently uninvolved sucker open your bank account to help them launder it. Whereupon I presume they take you to the laundry, raid all your spare quarters and leave you in the giant tumble-dryer with no socks. So that's the Peeve of the month. Although I suppose I should mention that the useless use of HTML attachments fought really hard for second place. The Klez worm gets a distant third since infected people usually figure it out and fix themselves.

So I didn't get very far in my personal project of the month this time. After a foray into the space for some clients of mine, I'm starting to seriously look into decent IMAP setups. UW IMAP is easy to install, sort of like saying a tricycle is easy to drive. Any kid can do it but it just was never designed for anything beyond being the reference implementation for IMAP as far as I can tell. Beyond that it's somewhat well known for being full of bugs, and their attitude towards client side problems in the FAQ feels laced with a lot of "your client sucks, use pine." Cyrus I glanced at and it seems sturdy enough, I guess. But the clients asked after Courier-IMAP. Hmmm, nice. About ten times as fast as UW. Just don't use the MTA that comes with it. Ugh, whatta mess. Maybe in a year or two you'll be reading this and it will have grown up. We went back to sendmail. If there are any other IMAP daemons to speak of I couldn't find them all that quickly. Courier is sufficiently cool and maildir delivery makes people happy for other reasons.

Did I remember to say IMAP is cool? As I've been going to more conferences with less techie features available, the idea of having my mail on the server be willing to deal with two workstations and a laptop carrying the same mail around doesn't sound half bad at all. And most people can't take the other tactic that I've done, which is to get involved with setting up internet lounge space

Which reminds me, the ConJosé, the 60th World Science Fiction Convention is coming up at the end of August, in my area, and I expect it'll be a lot of fun. I can tell you there's gonna be some Linux around.

Linux World Expo will be in my area in August also. WIth any luck I'll get to meet some more members of the Answer Gang. Group photo, anyone?

Anyways, back to IMAP. It turns out there are email specific appliances out there running Linux and Cyrus under the hood. Not only nice but I don't have to tinker? Oboy! To be fair I'm good friends with the folks over at IMAP Partners and the people who make the appliances they use host the system for my local sysadmins group, BayLISA.

But I didn't get as far along as I wanted in my great project to determine which IMAP clients suck the least. Our loyal readers may recall that I have a rather cynical opinion about the usefulness of so-called "productivity' software and and regardless of my deep dependence on email, MUAs (mail user agents) are no exception. So I will probably have an article about that next month. Contributions welcome, of course.

With that, it's time for a picnic. This year's Independence Day issue is full of some juicy bits just waiting for you to throw some CDs on the roaster. Me, I'll be spending a happy 4th of July hanging out with my family down in L.A.

How to Investigate a System Lockup

From Chris Gianakopoulos

Answered By Didier Heyden, Breen Mullins, Ben Okopnik, Jim Dennis, John Karns
with tidbits by Robos, Heather Stern

Hi Gang,

[Didier] Hello, Chris!

[Robos] Hi

I was running X tonight (with the ICEWM window manager), I had a couple of xterms running (one with kermit running), and I was using Acrobat Reader Version 4.0.

As I was making a mouse movement, the my console locked up.

[Robos] Don't you have to reboot when you make a mouse-movement? Oh, wait, that's that other thing that claims to be an os...

No less than 4 other gang members chimed in with some version of a sigblock fortune cookie about this. -- Heather

I could not even get a response, via the Ethernet, when trying to ping my crippled Linux system.

Which log files could I look at to try to determine what the impending disaster could have been? I have included the tail portion of /var/log/messages. I have included extra stuff, I suspect. I'm curious what those entries that say "MARK" mean. Could that be related to my lockup?

[Didier] Nope. From the `syslogd' man page:

      -m interval
              The syslogd logs a mark  timestamp  regularly.  The
              default interval between two -- MARK -- lines is 20
              minutes. This can be changed with this option.

(However it seems that this feature is disabled in some versions of the syslog daemon -- maybe through a compile-time option?).

Okay. I'll investigate other stuff.

[Breen] At least on Red Hat it's through a run-time argument.

The init script for syslogd reads /etc/sysconfig/syslog for its arguments:

# Options to syslogd
# -m 0 disables 'MARK' messages.
# -r enables logging from remote machines
# -x disables DNS lookups on messages recieved with -r
# See syslogd(8) for more details
SYSLOGD_OPTIONS="-m 0 -r -x"

"-m 0" is the default; I added "-r -x" on this machine.

[Didier] Fairly redhat-ish, indeed. My own system is based on an antediluvian RH 5.2 distro. I'm usually not too impatient to upgrade with a full new distro install (preferring recompiling packages from source -- RPM'ed or not -- iff I can't no longer avoid it). Believe it or not, I haven't drowned yet in the resulting mess

By that time they just had no such configuration file, and the syslog daemon was run without any argument by default. But somehow the `-- MARK --' feature was... erm, is still in my case... totally disabled: whatever -m xx option I try no timestamp appears in the logs.

[JimD] Actually I think this was a bug. I reported it to the upstream maintainer a few years ago (when I was running RH5.2) and he pointed me to the updated version that worked.

Naturally I'd advise that you simply fetch the latest version (in source form if you don't want to get trapped in RPM dependency upgrade hell) and build/install that.

[Didier] Thank you very much for your suggestions, Jim. Now I know what package to download next.

Regarding the RPM dependency hell, IIRC I once experienced core dumps from the `rpm' program itself after having fiddled with the `--nodeps' option (I was supposed to know what I was doing The problem was (hopefully) fixed with this simple command:

rpm --rebuilddb

I'm not sure it would have worked in all situations, though. And unfortunately I don't remember the exact version that was then installed on my system. In fact this has most probably been fixed ages ago...

[Didier] Note that I also have a couple of problems with the associated `klogd' daemon, as indicated by the last two lines of the following excerpt:

Jun  4 14:13:56 wallace kernel: klogd 1.3-3, log source = /proc/kmsg started.
Jun  4 14:13:57 wallace kernel: Loaded 15309 symbols from /boot/System.map.
Jun  4 14:13:57 wallace kernel: Symbols match kernel version 2.4.17.
Jun  4 14:13:57 wallace kernel: Error seeking in /dev/kmem
Jun  4 14:13:57 wallace kernel: Error adding kernel module table entry.

The other weird thing is that that ancient kernel log daemon cannot be stopped by anything but a plain SIGKILL. Doesn't prevent me from having nice dreams, however.

[Didier] Unfortunately, when one experiences such brutal lockups, the logs are often not of much use: the whole system freezes before the daemon is given a chance to write anything in them -- even if some kernel oops actually occurred. The only way to see this happening would be to have the kernel writing directly to the console (assuming you're currently viewing the console output, but it won't do in a X session unless, maybe, console output has been redirected to a serial port at boot time?)

Upgrading your kernel might help, provided the lockup was not caused by some hardware (RAM?) failure.

[Ben] That's pretty much what I would suspect - hardware. The only times I've seen Linux hang has been hardware-related stuff. In one very annoying case, my laptop would hang for a number of seconds, several times per day - and I had to live with it, because the PCMCIA card causing it was my wireless modem which was on 24x7. AFAICT, it took a huge chunk of CPU when it switched channels (sometimes the CPU load meter would actually catch the spike before everything froze); fortunately, it didn't do that very often.

[Didier] Another example is an IDE cd-rom or cd-writer device buggy enough to suck up every possible CPU clock cycle whenever it fails to read or burn the medium, the system thus becoming almost unusable -- especially in the case where the application which makes use of it is run with a static real-time priority (cf. `cdrecord'). Actually I've never figured out whether some ill-written code in the IDE / IDE-SCSI driver could be held responsible for such a misbehavior or if it was simply inevitable on this kind of architecture.

Real-time constraints in a multi-tasking operating system are often very difficult to deal with anyway.

[Ben] <sigh> Hardware stuff like PCMCIA has root-level access - has to, to access privileged ports, etc. - and unfortunately I know of no way to mitigate that. I wish there was a "nice" utility for hardware...

[Heather] ACPI might like to be that, someday.

[Didier] I once read that running a shell with a posix real-time scheduling policy could help in some situations. Unfortunately I've never heard either of a `nice'-like utility which could be used to launch `bash', `csh', etc. this way. I assume that in fact you must have a special version of your favorite shell, containing direct calls to sched_setscheduler(), in order to do that -- but I'm not sure.

RAM is always a possibility. The system seems awful reliable, though. Maybe it IS time to upgrade to a new distro just for the fun of it. I say distro rather than kernel so that I can use XFree86 version 4.x. My friends at work keep offering SuSE 8.0. I believe that the S3 Trio64v+ is supported, so nothing is really stopping me from going to the new distro.

I am guessing that it is related to whatever applications might have been running under X in combination with Acrobat (if not a hardware problem). Dynamic systems are always the most difficult to troubleshoot.

See attached chrisg.logfile.txt

[John] Continuing on the kernel side of the issue, a thread on a related subject just came up on a LUG list I'm on:

Thank you all (there are so many names to list!) for your quick responses to my question. I'm gonna do some detective work. My perception was that the system locked up. The only thing that I really know is that the console and the network did not respond. I got two serial ports on my system. I dedicate one to the modem, and I use the other for kermitting around. I think that I am going to use my nonmodem serial port for a login session. Would it not be funny if the system was still running and only my network stuff failed as a result of an X lockup?

That would seem odd, though. Since I was running X via my local console (you know -- with the keyboard and display), I would expect Unix domain sockets to be used, thus, bypassing TCP (the network stream stuff).

You all gave me lots of good ideas, and thanks much again. This email response is like a broadcast thanks to all of you!

How to send email without a DNS server?

From Faber Fedor

Answered By John Karns, Mike "Iron" Orr, Mike Martin, Heather Stern, Ben Okopnik

Hi Gang!

A client of mine is discovering the Joy That Is Linux. He recently asked me if it was possible to send email from applications (written in JBASE, a PICK shell that runs on Linux). I mentioned that you could do this:

mail -s "Some Subject" user1@domain.com < output_from_app

and he was impressed. He has since come up with several time- and resource-savings applications of that simple redirection to the mail command.

He has raised an intersting (to me) question and although I can come up with two solutions, I'm looking for a better one. Also, it's generated a question that I can't answer, hence this posting.

The question is: How do you send mail between two linux boxen if there is no DNS server from which to query for an MX record?

Now, I'm not asking for the impossible here. The two boxen run sendmail/postfix and they are on the same internal network (the sender is 10.10.10.1 and the receiver is 10.10.10.2).

I could tell him to install a DNS server, which he would do (I love clients that listen to me but there should be a simpler solution. We could send the email to a local user on .1 and fetchmail the mail from .2, but that's too much of a kludge for my taste.

I've been through the HOWTOS and google and didn't find anything applicable (at this point Ben walks in, donning his mirrored sunglasses, executes a google search using 1.5 words (without quotes), finds The Canonical Page that answers this question and makes me look like a fool (again ).

Ideas?

[Ben] <mock growl> Don't do that, Faber; the fact that one of us has the answer _does not make the rest of us look like fools. Unproductive attitude there, sir. Me, I'm always happy when I see one of the other Gangsters come up with an answer to something I don't know (how d'you think I got this smart? - I get to learn stuff, which is a very good thing indeed.

[Heather] half a search word?

[Iron] mail remote_user@10.10.10.2

Some mail transfer agents may balk at it, but I've found it to work most of the time.

[Heather] Works even better if the IP address is in brackets: remote_user@[10.10.10.2]

[Mike] Also, you can put the domain names in the poor man's DNS: /etc/hosts.

[Mike Martin] Assuming that the network is small with no server (as implied) /etc/hosts is the right idea

If you are sending externally you would need a DNS either locally or via the isp anyway

[Ben] "/etc/hosts", or maybe even a lightweight local (caching?) DNS. I'm coming to really like PDNSD; painless to install, dirt-simple to set up. Or, you could always use "ssmtp" to rotate the mail off the local hosts to the actual MTA, since it's local; that's my solution (farm the job out to the guy who's good at it.)

[Heather] Yes, you can use /etc/hosts, but it can't do an MX record lookup. Your smtp daemon will need some tweaking. In sendmail you can tell it that transport for a given named host will go through an explicit IP address host, and then it would never try an MX lookup, just use the burned in route.

[John] Depending on the distro, distro version, and MTA you're using ...

Using SuSE 7.x & sendmail, and a psuedodomain I use, I've come up with the following solution to that problem:

edit /etc/mailertable to add entries like the following

mylinuxServer1.my.domain		smtp:mylinuxServer1.my.domain
mylinuxServer2.my.domain		smtp:mylinuxServer2.my.domain

I don't have DNS setup, as the servers are isolated, but sendmail via dialup. I do have the hostnames listed in the hosts file on each server though. In this way, any mail addressed to user@mylinuxServer1.my.domain only goes out when connected to that server.

[Heather] The postfix equivalent to that is /etc/postfix/transports, although that's actually mentioned in /etc/postfix/main.cf to activate it, so you could name the file anything you like.

You can't use nullmailer for these even if it's only two boxes, because that won't accept mail for local deliveries. You could keep all the mail on only one of 'em and use nullmailer on the rest of the machines though. At that point the machine becomes a very tiny server and also needs to run a pop or IMAP daemon to allow mail pickup. At which point you may as well also set up a dummy local DNS, maybe for a silly domain name that won't be used by the outside world because there are no root servers for it, like fabershouse.local -- which would serve A and MX records for say... livingroom.fabershouse.local, den.fabershouse.local, etc.

You could skip mail and drop notes in a samba share at each other, or ssh in and share a kibitz session using your favorite text editor. Which means somebody becomes a server but may bother people a lot less.

Starting many X sessions

Answered By Ashwin M, Heather Stern, Jay R. Ashworth, Chris Gianakopoulos, Robos

[Ashwin] Hi,

I am a relative newbie to Linux, hence this tip may be common or may have been already published.

[Heather] This is a particularly clear description of it, and the thread turned a curious direction as well. Welcome to the pages of The Answer Gang, Ashwin.

[Ashwin] Yes! You can have more than one X session running at the _same time! By default your X session will be running on virtual terminal 7 (that is why you use Ctrl-Alt-F7 to get to it).

To start a new X session get to a terminal session (F1 - F6) and type "startx -- :1". A new X session starts up. Now to go back to the older X session, use Ctrl-Alt-F7. For the new X session use Ctrl-Alt-F8.

(For more X sessions use the next numbers after 1 like 2,3 and so on. They don't have to be in order like 1 next 2 and so on. No matter what number you use here the Ctrl-Alt-Fx number will be assigned the next available number after 7)

[Heather] Strictly speaking, the first available open terminal; on most major distro setups that's 7, after your 6th text console, but if you're an oddball like me who uses enough text consoles that you set up, say, 10, then it would show up on console 11 instead.

Of course if you do that too, make sure Xdm or its cousins don't think they are supposed to take a console used by a text getty.

[Ashwin] Now, why would anyone need another X session? Some reasons could be -

• To use different window managers at the same time like KDE and GNOME.
• To run X sessions logged in as different users.
• To try out things with the XFree86 system (like fonts and such)

You may think "why would I need this?". But, once you know this trick you will find new ways to use it for your work

Note: I found this not to work on some (very few) systems with old video cards and less VRAM.

thanks, Ashwin

[Chris G] That's a cool tip. I just tried it out, and it works with Version 3.3.6 of X.

[Heather] My favorite reasons to run multiple X sessions on the same machine are:

• Jim and I are both using the machine... so we want our own X session each. Since we're running in different chroot spaces (for an experiment) on one of the machines around here that's kind of handy.
• to have different native bit depths, also for experiment. Happily getting to be rare, but there is occasionally software that is happiest only at one color depth. If you're designing graphics for use at multiple color depths I swear this is the fastest way to tell which colors will wash out.

In the modern era you could also use Xnest to provide the reduced depth to the stupid app, but it's less confusing to do the above than to make sure you're running something inside the nest.

Not that it's hard mind you - make the nest's first client an xterm with a custom color fg and bg, then launch things you want in the nest from the special xterm.

[Robos] Well, I knew this, but did you know that the keyboard and mouse are bound to one? Meaning you can't - like I tried - run two xsessions on two grafik cards and share the keyboard between them. This is only possible if you use a ps/s keyboard and a usb version. I want to do that, but I haven't got the usb keyboard yet that I would like: small form factor (84 keys), english layout (I'm in germany) and usb...

[Jay] You are supposed to be able to do that.

And indeed, a quick test proves that you can.

RH 7.1, and whichever X ships with that. I did Ctrl-Alt-F1 to a text login, logged in as root, did 'startx -- :1', and was whisked over to tty8, where I got a KDE desktop as root (that being the default choice on my laptop).

Ctrl-Alt-F7 switched me right back to the desktop I'm on (KDE under KDM, logged in as myself), and I could C-A-F8 right back to the other one; mouse and keyboard worked in both.

And indeed they should: X switches to a free vt before it binds the hardware, and what it gets is the virtualization of the hardware that it produced by the multi-console driver. It doesn't bind directly to the hardware.

[Robos] Well, a misunderstanding and a wrong memory on my part (I run memtest86 later on me, it always takes so long...). I meant that I wanted to run two xservers on two graphic cards and have two keyboards too. My motivation: Running quake3 and chatting with my pals in irc at the same time. Thought about either two xservers or one xserver and one mga framebuffer. But, as said, no usb keyboard yet.

[Jay] What is hard is having two X servers on different physical display adapters with separate keyboards. Hard, but often very useful.

[Robos] And cheaper than two computers, not to mention the noise reduction.

[Heather] While I had interpreted this confusion as wanting both X servers to listen to the keystrokes in tandem, implying that they'd both be grabbing keyboard events from the same pool. Which given his usage example, of course is not what you'd want I could see some vague reason why somebody might want to do it (test the same app on two video cards at once) and I can certainly imagine the USB model of how devices work being much better for handling it.

X 4 lets you designate entirely different screen layouts to be stored in its config file, and you can specify which one to use by passing -layout "layout[1]" (or whatever ID you named it) down to X. Of course your usual methods of starting X may try to protect you from doing this, but with startx it's pretty easy: startx -- -layout secondhead :2

With X 3.3.x you probably have to just keep seperate config files.

If anyone out there is crazy enough to be trying this, let us know how it works out for you. And then there's always the Linux Terminal Server Project: http://www.ltsp.org

Or better yet the K-12 Linux Terminal Server Project: http://k12ltsp.org/contents.html

Playing CD Music Digital Output

Questions From Bill Parks, Mithra

Answered By Lennart Benschop, Dave Maxwell, Serkan Akdag, Rob McMeekin, Robos

This thread is in response to Issue 79, Help Wanted #2 and all of these people sent in Tips. I'm sure there's really at least a quarter's worth here; that'd be Two Bits instead of Two Cents -- Heather

I purchased an eMachine to run Linux on. It came with (sorry) XP which I used to check out the hardware. It plays music CD's fine but uses digital data over the IDE buss rather than a cable from the CD drive to the sound input.

Loaded Red Hat 7.3 and it plays sounds fine but it won't play music CD's...the player just runs and the CD spins along.

How do I configure the CD/sound system to pick up the digital sound data on the IDE buss to play the music?

Thank You, Bill Parks

[Lennart] Linux does have a program that can play a CD by reading the digital data over the IDE bus, it's cdda2wav, included with most Linux distributions.

cdda2wav -D /dev/cdrom -t 1 -eN

will play track 1, if the KDE sound system isn't keeping /dev/dsp busy. The -eN options cause cdda2wav to echo the data to the soundcard and not write the ridded data to a file.

You can also try to connect an analog cable between the CD-ROM drive and the sound cards, as most PC's have.

[Dave] There is a plugin for XMMS that does this. It can be downloaded at

ftp://mud.stack.nl/pub/OuterSpace/willem

The version numbers are a little confusing. The latest version is:

xmms-cdread-0.14a.tar.gz

The one labeled 0.9a is not it. XMMS comes with a cd player plugin that is supposed to do this but it didn't work for me. To play a cd, start XMMS and open your /dev/cdrom device or whichever device the cdrom drive actually is. I'm currently playing a cd from a drive without it's audio cable so I suppose that means it works . Come to think of it, if you add an audio cable from your cd to your sound card then the player you attempted to use will work as well. The audio from this plugin will be cleaner as it passes through one less analog stage.

Another nice thing about this plugin is it will pull title data from Freedb and let you play individual tracks by name. The stock plugin won't do this.

Oh yeah! Be sure to disable the stock plugin before attempting to use this one. Right click on XMMS and choose Preferences.

[Serkan] Hello, You will need to enable the "Enable Digital Sound" from one of the configuration options of KDE and/or Gnome. There is an option like that somewhere but I can't remember but I am sure its in X and not the console. Also, you could type audio:/ at the KDE browser to view your CD-Player contents and open a multimedia player like XMMS and drag the files in the browser to the playlist of XMMS and play that way. It should work.

[Rob] Hello. I have an IDE CD-ROM drive and an IDE CD-RW drive. The CD-ROM has audio cables attatched to the Sound Card while the CD-RW does not. The CD-RW would not play compact discs until I configured it for writing (enable SCSI-Emulation). I'm not an expert. I don't know if this is supposed to be the case. It could just be a fluke, but it worked for me.

And, since someone may also want to play other sorts of fun noise with their player, and most easily find this thread instead... -- Heather

[Mithra]

Hai,

I use Caldera Linux . i Would like to watch movies in my syatem. But There is no s/w that would play .Dat file for me. Can any one suggest me a downloadable s/w that can Play .dat files.

[Robos] Try mplayer, it can do this. Url is http://www.mplayerhq.hu.