Subject: Linux-Development Digest #40
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Thu, 18 Aug 94 17:13:13 EDT

Linux-Development Digest #40, Volume #2          Thu, 18 Aug 94 17:13:13 EDT

Contents:
  which company's motif library is better? (lam tao)
  Re: >Future of Linux (Byron A Jeff)
  Release Freeze Yet? (Jered Floyd)
  Re: MGA Matrox Ultima 64bit support? (Derek Seabury)
  Re: Intel syntax assembler available anywhere ? (Pete Chown)
  Re: tcl/tk XF application? (Neal Becker)
  Re: Suggest:SCSI Tape File System (Oleg Zabluda)
  Re: Novell interfacing experience?  Anyone using samba? (DahlinSmithWhite)
  Re: Which linux 1.1.x versions are stable? (Holger Mueller)
  Re: Report on SVNET: Bill Jolitz's Talk; Mind Overload. (Michael I Bushnell)
  Re: Which linux 1.1.x versions are stable? (James Smith)
  automatic connection closed in 1.1.4x (James Smith)
  Socket Timeouts in Linux 1.1 (Thomas Wallace Colthurst)
  Re: LOOPBACK Driver needs a little work. (Rene COUGNENC)
  tcl/tk XF application? (Simon Johnston)
  Howto write Device Drivers. (Mark Baker)
  1.1.45 Console Driver (Michael Faurot)
  On SIGCHLD signals (Marc Aurele La France)
  Re: Suggestion for Linux (Beverly J. Brown)
  KERNEL PANIC! after exiting XFree86 2.1.1 (Garrett D'Amore)

----------------------------------------------------------------------------

From: tlam2@ie.cuhk.hk (lam tao )
Subject: which company's motif library is better?
Date: Thu, 18 Aug 1994 09:40:09 GMT

Since recently I want to purchase the motif library to compile my Mosaic
program, I find that two companys provide the motif, but I don't know
which one is better? (SWiM motif or Metro link motif)?

Does anybody has installed any one of them. Please recommend me.
Thanks a lot!!!

 

------------------------------

From: byron@gemini.cc.gatech.edu (Byron A Jeff)
Subject: Re: >Future of Linux
Date: 18 Aug 1994 12:40:53 GMT

In article <32teqq$guj@bach.udel.edu>,
Jason Aaron Fager <jafager@bach.udel.edu> wrote:
--I would like to see a continuation of the UNIX "toolkit" approach
--where people build tools that do a fairly specialised job but that
--are easily combinable with other tools to do fairly complex jobs.
--For the most part, such tools have been "filters" that are
--connected together with pipes. 
-
-An excellent example of how to add filter/pipe functionality to a GUI
-framework is the Application Visualization System (AVS) from Advanced
-Visual Systems.  This program uses a "network editor" tool to form 
-complex visualisation systems.  The user simply drags the requisite 
-modules from a palette at the top of the screen, connects them together
-with pipe-fitting tools, and starts them running.  It's all done visually,
-with a "drag-and-drop" interface.

Now that's the beginning of a good idea. I'd only ask for two more features:

1) Have a mechanism that shows the user a script based equivalent. This is
   so that the user can still use these commands if for any reason they
   happen to be somewhere where the "network editor" is inaccessible.
2) Excellent documentation.

This AVS supports my point that powerful systems with a GUI basis can
be built. But I still have the caveat that the user must be trained to
use all the available tools. AVS sounds like a pretty good mix.

BAJ
-- 
Another random extraction from the mental bit stream of...
Byron A. Jeff - PhD student operating in parallel - And Using Linux!
Georgia Tech, Atlanta GA 30332   Internet: byron@cc.gatech.edu

------------------------------

From: jjfloyd@vela.acs.oakland.edu (Jered Floyd)
Subject: Release Freeze Yet?
Date: 18 Aug 1994 02:01:04 GMT

Has there been a release freeze yet? I'm just about to install 1.1.18,
and I don't want to upgrade to an unstable release (i.e. 1.1.4x), so I
am urgently awaiting 1.2 for bug fixes.

Any word of a 1.2 release?

How about XFree86 3.1? I want my Diamond Stealth 64 driver!


--
Jered Floyd - jjfloyd@vela.acs.oakland.edu
Geek Code 2.1 - GAT d? H- s-:- g- p? !au a-- w+ v+ C++++ UL++++ P+ L++
N+++ K+++ W++ M-- V-- -po+ Y++ tv+ 5+++ j++ R v++ b+++ D+++ B--- e* u**
h++ f? r? n- !y+ (Finger for PGP key, picture, humor anOUT OF SPACE

------------------------------

From: sseabury@galileo.cs.jhu.edu (Derek Seabury)
Subject: Re: MGA Matrox Ultima 64bit support?
Date: 18 Aug 1994 13:21:42 GMT

In article <94229.014203U62323@uicvm.uic.edu> <U62323@uicvm.uic.edu> writes:
>Does any of you know, if there is ever going to be support for MGA video cards
>I am thinking about geting another video card, that is supported now.
>Any suggestions?
>My E-mail rgebis@ernie.eecs.uic.edu
>Thanx!!!!

I just installed XInside Inc. X server on one of my machines with a MGA II+
and it flies!  Much easier to setup though.  (you still need the rest of
X to use it - the programs and libraries etc...)

Derek

(If you want more info email info@xinside.com)


------------------------------

From: pc@dale.dircon.co.uk (Pete Chown)
Subject: Re: Intel syntax assembler available anywhere ?
Date: Wed, 17 Aug 1994 07:55:09 GMT

In article <CuMAs5.3LG@pe1chl.ampr.org> rob@pe1chl.ampr.org (Rob Janssen) writes:

> In <32oquf$f0i$1@heifetz.msen.com> jsturm@garnet.msen.com (Jeffrey Sturm) writes:
> 
> >             MOV     [BX], 1         ; move 1 to address bx, MASM-style
> >and
> >             movl    $1,(%ebx)       /* move 1 to address ebx, gas-style */
> >
> > The MASM example moves 2 bytes and the gas example moves 4 bytes.  Doesn't
> > this mean your entire data segment would have to be reorganized to do a
> > conversion?

> Hey, but here you are comparing two different instructions, aren't you??
> One is a word move, the other is a longword move (to EBX instead of BX).
> No wonder they yield different results...

These are not different instructions, unfortunately.  There is a
descriptor table bit which makes instructions default to 16 or 32
bits.  In real mode and in Crashdows 386 enhanced mode it is 16 bit,
but in Linux it is 32 bit.  Thus both the example instructions will
generate the same code!

Actually this is not too big a problem, because as long as you ask the
assembler for the correct word size, the output should work.  I
suspect a bigger problem will be the memory management.  Windoze
handles memory in segments of 64k.  If you need a lot of memory, well,
you need a lot of segments as well.  This makes programming rather
interesting.  It also makes it very difficult to convert Windows code
to an architecture like Linux, where each process has essentially a
flat memory model.

--

==========================================================================
Pete.Chown@dale.dircon.co.uk          "The Pen is mightier than the Quill"
                                      -- anonymous

------------------------------

From: neal@ctd.comsat.com (Neal Becker)
Subject: Re: tcl/tk XF application?
Date: 18 Aug 1994 12:51:10 GMT

I have tried it.  I don't recall having any difficulty building from
source.  If you are having trouble, I'll take a closer look.

------------------------------

From: zabluda@math.psu.edu (Oleg Zabluda)
Crossposted-To: comp.os.linux,comp.os.linux.misc
Subject: Re: Suggest:SCSI Tape File System
Date: 18 Aug 1994 02:47:11 GMT

Great Idea. I'd really like to see it implemented in such a way
that It can be used with QIC-80 tape drives using existent ftape
driver. I think there are more people owing these kind of drives
then SCSI tape drives.

Oleg.


------------------------------

From: dswi@xmission.com (DahlinSmithWhite)
Crossposted-To: comp.os.linux.misc,comp.os.linux.help
Subject: Re: Novell interfacing experience?  Anyone using samba?
Date: 17 Aug 1994 20:32:23 -0600

Zaphod Beeblebrox (zaphod@teleport.com) wrote:
: If anyone has any good or bad experiences attempting to network up with
: novell, can you please give me any words of advice?  I have this linux
: package called samba that does interfacing to novell.  Does anyone have
: this installed in their system?

Could you explain how samba allows you to interface with Novell? Is Samba
a Netware requester, or what does it let you do with Novell (access 
Netware volumes, print queues, etc)?

Pete.

------------------------------

From: holgi@rbg.informatik.th-darmstadt.de (Holger Mueller)
Subject: Re: Which linux 1.1.x versions are stable?
Date: 18 Aug 1994 07:15:05 GMT

pageone@netcom.com wrote:

:       That's the question... I'm trying to figure out which versions are
: safe to use... or should I just stick with 1.0.9 for a while?

I could just post my experience:
1.1.3x --> 1.1.44 were more or less instable for me, having panics and more.

1.1.45 now works again since a few days without any panic.
Before 1.1.45 i was using 1.1.9 all the time which never crashed on my machine.

Remember, this is only MY experience, but you may go and try it out.
Keep a copy of your old kernel handy!

Maybe you should just cp your old vmlinuz to vmlinuz.rescue and add an
rescue entry into your lilo.conf.


Type ya,
--
/======================================\
| Olav "Mac" Woelfelschneider          |
| wosch@rbg.informatik.th-darmstadt.de |
+--------------------------------------+
| I refuse to grow up,                 |
| I don't want to lose my humor...     |
\======================================/

------------------------------

From: mib@geech.gnu.ai.mit.edu (Michael I Bushnell)
Crossposted-To: comp.unix.bsd,comp.os.386bsd.development,comp.os.386bsd.bugs,comp.os.386bsd.apps,comp.os.386bsd.questions,comp.os.386bsd.misc,comp.os.os2.programmer.misc,comp.os.minix,comp.os.mach,comp.unix.pc-clone.32bit
Subject: Re: Report on SVNET: Bill Jolitz's Talk; Mind Overload.
Date: 18 Aug 1994 14:27:35 GMT

In article <jmonroyCuq5J1.Fy8@netcom.com> jmonroy@netcom.com (Jesus Monroy Jr) writes:

               Second: The issue of code reuse....

                       Bill and Lynn plainly stated that the code is
                       freely re-distributable and freely re-modifiably,
                       *AS ALWAYS*.

               Fourth: The CD-ROM

                       The CD-ROM and the annotations may *NOT* be placed
                       on any FTP'able site for redistribution.

Don't these seem just a little contradictory?

        -mib
--
+1 617 623 3248 (H)    |   The soul of Jonathan was bound to the soul of David,
+1 617 253 8568 (W)   -+-   and Jonathan loved him as his own soul.
1105 Broadway          |  Then Jonathan made a covenant with David
Somerville, MA 02144   |    because he loved him as his own soul.

------------------------------

From: smith@meishan.animal.uiuc.edu (James Smith)
Subject: Re: Which linux 1.1.x versions are stable?
Date: 18 Aug 1994 13:46:08 GMT

debruijn@cs.utwente.nl (S.G. de Bruijn) writes:

>Holger Mueller (holgi@rbg.informatik.th-darmstadt.de) wrote:
>: pageone@netcom.com wrote:

>: :    That's the question... I'm trying to figure out which versions are
>: : safe to use... or should I just stick with 1.0.9 for a while?

>: I could just post my experience:
>: 1.1.3x --> 1.1.44 were more or less instable for me, having panics and more.

>: 1.1.45 now works again since a few days without any panic.
>: Before 1.1.45 i was using 1.1.9 all the time which never crashed on my machine.

>: Remember, this is only MY experience, but you may go and try it out.
>: Keep a copy of your old kernel handy!

>: Maybe you should just cp your old vmlinuz to vmlinuz.rescue and add an
>: rescue entry into your lilo.conf.

>I have had great stability with 1.1.17 and 1.1.36. Stick with it until I see no more
>problems with the new ones... (ftape/1.1.45 for example)

>Bye,
>Steef
>--------------
>S.G. de Bruijn              E-Mail: debruijn@cs.utwente.nl
>Twente University of Technology, Dept. of Computer Science 
>Enschede                                   The Netherlands
>Phone: Work: +31 53 894191             Home: +31 53 334812
>-----------------------==== @@ ====-----------------------
>signature: file not found

in my experiance i see that 1.1.11 or 1.1.23 is stable

=======================================================

------------------------------

From: smith@meishan.animal.uiuc.edu (James Smith)
Subject: automatic connection closed in 1.1.4x
Date: 18 Aug 1994 13:55:32 GMT

hi,

i see that in 1.1.4x  telnet will close automatically for some
idle time.   How can I make it connect forever like in old kernel.

THANKS

------------------------------

From: thomasc@ATHENA.MIT.EDU (Thomas Wallace Colthurst)
Subject: Socket Timeouts in Linux 1.1
Date: 18 Aug 1994 16:15:09 GMT

I know of two ways of making a socket timeout while 
attempting to read data, neither of which seems to work
with Linux.

Method #1:  setsockopt( skt, SOL_SOCKET, SO_RCVTIMEO,
                        &tv, sizeof(tv) );
Problem:  SO_RCVTIMEO isn't defined in sys/socket.h or linux/socket.h

METHOD #2:  fd_set readfds;  FD_ZERO(readfds);  FD_SET( skt, &readfds );
            select( 1, &readfds, NULL, NULL, NULL, &tv );
Problem:  select always waits till the timeout, and then returns 0.


Am I using either of these methods incorrectly?  Is there another
method I should be using?

Thanks for your help,

-Thomas C



------------------------------

From: rene@renux.frmug.fr.net (Rene COUGNENC)
Subject: Re: LOOPBACK Driver needs a little work.
Date: 18 Aug 1994 04:04:55 GMT
Reply-To: cougnenc@blaise.ibp.fr (Rene COUGNENC)

Ce brave Chip Edwards ecrit:

> running v1.1.45.
>    If I do a:
>
>       ping -f loopback
>
> It has a packet loss of about 90% or more! I've noticed every once in a 

--- 127.0.0.1 ping statistics ---
5726 packets transmitted, 203 packets received, 96% packet loss
round-trip min/avg/max = 5.5/266.7/542.2 ms        

Linux plux 1.1.45 #1 Mon Aug 15 15:37:40 MET DST 1994 i486  

Interesting :-)

--
 linux linux linux linux -[ cougnenc@renux.frmug.fr.net ]- linux linux linux 

------------------------------

From: skj@oasis.icl.co.uk (Simon Johnston)
Subject: tcl/tk XF application?
Date: Thu, 18 Aug 1994 07:52:03 GMT

Has anyone had any joy finding/using the XF interface builder for tcl/tk
under Linux. If so could you let me know where it can be found and how to 
install.

Thanks.


MODULE Sig;
FROM ICL IMPORT StdDisclaimer;
FROM Interests IMPORT Modula2, Modula3, Linux, OS2;

BEGIN
(* ------------------------------------------------------------------------.
|Simon K. Johnston - Development Engineer              |ICL Retail Systems |
|------------------------------------------------------|3/4 Willoughby Road|
|Unix Mail : S.K.Johnston.bra0801@oasis.icl.co.uk      |Bracknell, Berks   |
|Telephone : +44 (0)344 476320   Fax: +44 (0)344 476084|United Kingdom     |
|Internal  : 7261 6320    OP Mail: S.K.Johnston@BRA0801|RG12 8TJ           |
`------------------------------------------------------------------------ *)
END Sig.

------------------------------

From: mbaker@phal.chilli (Mark Baker)
Subject: Howto write Device Drivers.
Date: 18 Aug 1994 08:54:25 GMT
Reply-To: mbaker@phal.chilli

I have a frame grabber card and suspect that the simplest way of getting a device
driver for it is to write one myself.

Problem:- Is there an authoritive text on how to write Linux/Unix device
drivers. I have designed sbc's before so it should be a managable task. Its
just that I do not know where to start...


Thanks in advance.


------------------------------

From: mfaurot@phzzzt.atww.org (Michael Faurot)
Subject: 1.1.45 Console Driver
Date: Thu, 18 Aug 1994 14:08:52 GMT

I installed the v1.1.45 kernel yesterday without any difficulties with
a v1.2 Slackware system that has been running kernel 1.0.8.  However I
noticed a quirk/bug, which might be the console driver.  When I ran
the "workbone" program (for playing audio CDs), the display was messed
up.  The same binary under 1.0.8, display just fine.
-- 
+--------------------+----------------------------+--------------------------+
|   Michael Faurot   | mfaurot@phzzzt.atww.org    |      I don't like        |
|   ------- ------   | ...!netcomsv!phzzzt!mfaurot|      lima beans!!        |
+--------------------+--------------------+-------+--------------------------+

------------------------------

From: tsi@gpu.srv.ualberta.ca (Marc Aurele La France)
Subject: On SIGCHLD signals
Date: 18 Aug 1994 17:52:23 GMT

For some time I've been having a problem for which I now need guidance to
solve.  The problem relates to the handling of SIGCHLD signals.

When a child task is created via a (v)fork or clone system call, it inherits
its parent's signal handlers.  Then, when the child executes one of the exec
system calls to run another program, all of its signal handlers, except those
set to SIG_IGN, are reset to SIG_DFL.  This is documented behaviour in many
(if not all) Unix implementations, not only in Linux.

When the child executes an exit system call, it sends a SIGCHLD signal to its
parent and becomes a "zombie" task.  In Linux, this zombie can only be
discarded when its parent executes one of the wait system calls.  If the
parent does not wait before it exits, its children will be made children of
the init task and init, being the new parent, will discard the zombies.

The parent can respond to SIGCHLD signals in several different ways:

1) If the parent is already (interruptibly) sleeping in a wait system call,
   the parent will wake up and its wait call will return the child's process
   id.

2) If the parent's SIGCHLD handler is SIG_IGN, the kernel will make the wait
   system on behalf of the parent and the parent will be none the wiser.

3) If the parent's SIGCHLD handler is SIG_DFL, the kernel ignores the signal
   thereby making the assumption that the parent will call wait directly.

4) If the parent's SIGCHLD handler is anything other than SIG_IGN or SIG_DFL,
   the indicated function is called and, presumably it will do the wait call.

There is a design hole here that occurs when the parent makes a wait system
call while its SIGCHLD handler is set to SIG_IGN, either intentionally or
because it did not bother to reset the SIGCHLD handler it inherited.  If the
child exits while the parent is waiting, then the child's process number will
be returned to the parent (as per 1 above).  If the child exits at any other
time, the parent will be told it has no children, which causes problems when
the parent is attempting to keep track of its children.

Unfortunately, this affects many widely available programmes, such as GNU find
and GNU xargs, to mention only a few.  My take on this is that programmes that
assume they are not inheriting a SIGCHLD handler of SIG_IGN should be fixed
and programmes that intentionally set their SIGCHLD handler to SIG_IGN and
then make a wait system call (such as fvwm) should be looked at to determine
if such behaviour is justified.  A daunting task, to be sure...

On the other hand, there are several ways a kernel could be changed to remove
the time dependency of a wait call's behaviour while the caller's SIGCHLD
handler is set to SIG_IGN.  But there are compatibility issues to be
considered before any kernel solution is adopted.  What does POSIX, or any
other Unix "standard" have to say about this?

Any comments would be appreciated (before I potentially waste my time spewing
out diffs :-)).

+----------------------------------+-----------------------------------+
|  Marc Aurele La France           |  work:   1-403-492-9310           |
|  Computing and Network Services  |  fax:    1-403-492-1729           |
|  352 General Services Building   |  email:  tsi@gpu.srv.ualberta.ca  |
|  University of Alberta           +-----------------------------------+
|  Edmonton, Alberta               |                                   |
|  T6G 2H1                         |     Standard disclaimers apply    |
|  CANADA                          |                                   |
+----------------------------------+-----------------------------------+

------------------------------

From: bjb@shore.net (Beverly J. Brown)
Subject: Re: Suggestion for Linux
Date: 18 Aug 1994 15:34:25 -0400
Reply-To: bjb@shore.net

In article <DRAGAN.94Aug16094322@norma.gmd.de>,
dragan@norma.gmd.de (Dragan Cvetkovic) wrote:
> In article <32olv5$2r9@charnel.ecst.CSUChico.EDU> dranch@ecst.csuchico.edu (David A. Ranch) writes:
> > 
> > Oh Man....  can you say file-name completion?  I believe its been in
> > Unix since its inception.
> > 
> > ksh - hit ESC twice
> > sh, csh, tcsh, bash - hit TAB
> > 
>       What? I don't think that "normal" sh and csh have this
> feature. Yours are most probably links to tcsh and bash. 
>                       Dragan
> 

csh uses ESC. (at least it does on SunOS 4.1.3)
tcsh will use ESC twice or TAB
sh doesn't do it at all.


Beverly J. Brown
bjb@shore.net
beverly@datacube.com

------------------------------

From: garrett@athena.sdsu.edu (Garrett D'Amore)
Crossposted-To: comp.os.linux.help,comp.windows.x.i386unix
Subject: KERNEL PANIC! after exiting XFree86 2.1.1
Date: 18 Aug 1994 18:07:43 GMT

Hi all!  Sorry about the cross post, but nobody responded to my original
query on comp.os.linux.help three or four days ago, so I'm trying a wider
distribution.

Basically, after exiting XFree86 2.1.1 on a box running Slackware 2.0
(using XF86_S3 on an STB-VL24), I encounter *very* strange behavior.
Sometimes I get kernel panics, other times various normal process (e.g.
"ls" and "login") just start seg. faulting on me.  If I don't exit X,
everything works fine.  (Also, if I don't start X everything is Ok.)

I am completely unsure about how to go about fixing this problem, or
even how to track it down. If anyone out there has any experience with
similiar problems, or has suggestions about how to proceed, I would be
very grateful.

Incidentally, this same system used to run MCC-1.0+ and XF_S3 2.1 just
fine.  The only thing new in terms of hardware is that I have replaced
an older ESDI drive with an IDE drive.  I am even using the same Xconfig.

Thanks for any help you can render!

=================================================================
Garrett D'Amore                 |     garrett@sciences.sdsu.edu
Student Programmer/Analyst      |     SDSU College of Sciences
SDSU Computer Science Major     |     Physics Bldg., Room 238
=================================================================

------------------------------


** 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: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.development) via:

    Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Development Digest
******************************
