Subject: Linux-Development Digest #776
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:     Mon, 30 May 94 00:13:04 EDT

Linux-Development Digest #776, Volume #1         Mon, 30 May 94 00:13:04 EDT

Contents:
  Re: 32-bit Novell desktop OS combines Unix, DOS 7 (Richard Hodson)
  Re: Why is my Emacs binary so big (2.2M) (Kevin Lentin)
  Re: Motif Development (dan@oea.hacktic.nl)
  Re: New Linux BBS under developement! NEED YOUR HELP! (dan@oea.hacktic.nl)
  Re: ISDN and Linux? (Florian La Roche)
  Re: Why is my Emacs binary so big (2.2M) (Robert Sanders)
  Re: Why is my Emacs binary so big (2.2M) (Robert Sanders)
  Re: Twain Scanner (CHRISTOPHER M MAY)
  Re: 1.1.15 breaks SCSI (Michael Schmidt)
  Re: Troubles with patches (Elaine Walton)
  Re: SIGHUP - Where do we go from here? (Leon Garde)
  Re: 1.1.15 breaks SCSI (Rob Janssen)
  kernel: Warning - bdflush not running (Scott McKinsey)
  Re: InterLink for Linux (NIIBE Yutaka)
  Re: SIGHUP: We tried it! (Colin Plumb)
  XFree SVGA and 1.1.16 (Richard W. Altheide)
  Brain dead name server, overcome ? (Barry Yip kam-wa)
  Dynamic linking with shared libraries. (owen taylor)

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

From: richard@radar.demon.co.uk (Richard Hodson)
Subject: Re: 32-bit Novell desktop OS combines Unix, DOS 7
Date: Sat, 28 May 1994 22:20:24 +0000

Mark Evans (evansmp@mb5194.aston.ac.uk) wrote:

: IPX/SPX is not the difficult bit.
: NCP is, however it would be quite possible to put a netware server as
: a userlevel programme, similar to the Linux NFS server.

Why would they have to use NCP? If they knock up a new version with stupid
cosmetic changes could they not release the code for a client version and
an NLM for the Netware server so that you can talk to it?

Or how about this for a conspiracy theory. Perhaps they don't know how
to get around the GPL yet, but they are reading this thread until one
of us comes up with a good idea :-)

--
Richard Hodson                          |  richard@radar.demon.co.uk
:1,$s/DOS/anal secretion from hell/g    |  rhodson@cix.compulink.co.uk

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

From: kevinl@NewsServer.cc.monash.edu.au (Kevin Lentin)
Crossposted-To: comp.os.linux.help
Subject: Re: Why is my Emacs binary so big (2.2M)
Date: 29 May 1994 15:48:35 GMT

On 27 May 1994 13:29:53 GMT, Michael MNUK wrote:
> I compiled the Emacs 19.24 using the standard distribution (no extra packages
> dumped). The resulting binary is about 2.2M big which is about twice as much
> as the binary I got with the Slackware distribution. I only changed the
> compilation flag from "-g" to "-O2".

Make sure you check the link flags as well. It's the linking where the big
executables are made (read static).

> I observed the same with e.g. fvwm. The ratio there is even more dramatic.
> E.g. FvwmBanner is now about 250K whereas the old (Slackware distrib.) was
> about 30K.

FvwmBanner is an unfair one to check. It has a 5x3 inch colour bitmap in
the binary. Mine is about 145k. The rest of the Fvwm modules are between 9
and 20k on my system (compiled today)

> Am I linking statically (-static option never specified!) or is
> something else wrong?

-g implies -static. And don't forget to check the CFLAGS and LFLAGS or
LDFLAGS.

-- 
[==================================================================]
[ Kevin Lentin                   |___/~\__/~\___/~~~~\__/~\__/~\_| ]
[ kevinl@bruce.cs.monash.edu.au  |___/~\/~\_____/~\______/~\/~\__| ]
[ Macintrash: 'Just say NO!'     |___/~\__/~\___/~~~~\____/~~\___| ]
[==================================================================]

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

From: dan@oea.hacktic.nl
Subject: Re: Motif Development
Date: Sun, 29 May 1994 09:14:41 GMT

tf49665@delphi.com wrote:
:    Where can I buy a low cost Motif Development library for Linux? Please
: reply via email.

contact:
        info@sequoia.com
        info@mtrolink.com

-- 
|< Dan Naas        dan@oea.hacktic.nl >|
+--------------------------------------+

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

From: dan@oea.hacktic.nl
Subject: Re: New Linux BBS under developement! NEED YOUR HELP!
Date: Sun, 29 May 1994 09:32:50 GMT

Dave Best (davidb@morgan.ucs.mun.ca) wrote:


:       Currently there is a NEW Linux BBS package under developement. Its in
: its primary stages and we need your ideas!  Our goal is to release the first
: version by September.  The BBS will be developed under the terms of the GNU 
: public license.

:       Please send any ideas/suggestions or options you would like to see in
: this BBS package to 

:       davidb@cs.mun.ca    OR    garyp@cs.mun.ca

Why not create a mailing list, if one doesn't already exist on, niksula.hut.fi.

-- 
|< Dan Naas        dan@oea.hacktic.nl >|
+--------------------------------------+

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

From: flla@stud.uni-sb.de (Florian La Roche)
Subject: Re: ISDN and Linux?
Date: 29 May 1994 19:23:25 +0200

: Ok I wasn't trying to complain about that I was just pointing out the diehl
: driver drops in more easily. Now the new tty drivers are done hopefully
: someone will drop the diehl driver into this and its all done. You can do
: IP over the diehl with the stuff provided but its a hack. The same ought

I have read through that Diehl ISDN driver and I also know what to change
for the new tty code. I would also like to change the existing Diehl driver
to not do any not-necessary copies of the data.
With the new ppp code, it should be possible to use that character device
driver for TCP/IP.

All I am waiting for is a free card from Diehl to get me starting...

Is there any other company, that would pay me for adjusting the existing
driver to the newest kernel? Don't wait to write me an email...

Florian

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

From: gt8134b@prism.gatech.edu (Robert Sanders)
Crossposted-To: comp.os.linux.help
Subject: Re: Why is my Emacs binary so big (2.2M)
Date: 29 May 1994 14:04:21 -0400

kevinl@NewsServer.cc.monash.edu.au (Kevin Lentin) writes:

>-g implies -static. And don't forget to check the CFLAGS and LFLAGS or
>LDFLAGS.

Not necessarily; it causes ld to link against libg.a instead of libc.sa.
If, like me, you made libg.a a symlink to libc.sa, it doesn't mean anything
at linktime.

Unless you have libc sources around, libg.a won't do you much good.

-- 
 _g,  '96 --->>>>>>>>>>   gt8134b@prism.gatech.edu  <<<<<<<<<---  CompSci  ,g_
W@@@W__        |-\      ^        | disclaimer:  <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~'  ro|-<ert s/_\ nders |   who am I???  ^  from Superman  '~~~**@@@@W
`*MV' hi,ocie! |-/ad! /   \ss!!  | ooga ooga!!    |    II (cool)!         `VW*'

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

From: gt8134b@prism.gatech.edu (Robert Sanders)
Crossposted-To: comp.os.linux.help
Subject: Re: Why is my Emacs binary so big (2.2M)
Date: 29 May 1994 15:40:03 -0400

gt8134b@prism.gatech.edu (Robert Sanders) writes:

>kevinl@NewsServer.cc.monash.edu.au (Kevin Lentin) writes:

>>-g implies -static. And don't forget to check the CFLAGS and LFLAGS or
>>LDFLAGS.

>Not necessarily; it causes ld to link against libg.a instead of libc.sa.
>If, like me, you made libg.a a symlink to libc.sa, it doesn't mean anything
>at linktime.

Er, of course, -g *does* imply -static.  What I meant to say was that -static
wasn't necessarily the only problem.

-- 
 _g,  '96 --->>>>>>>>>>   gt8134b@prism.gatech.edu  <<<<<<<<<---  CompSci  ,g_
W@@@W__        |-\      ^        | disclaimer:  <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~'  ro|-<ert s/_\ nders |   who am I???  ^  from Superman  '~~~**@@@@W
`*MV' hi,ocie! |-/ad! /   \ss!!  | ooga ooga!!    |    II (cool)!         `VW*'

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

From: cmay@titan.ucs.umass.edu (CHRISTOPHER M MAY)
Subject: Re: Twain Scanner
Date: 29 May 1994 19:47:37 GMT


There is a driver for the Logitech Scanman available on tsx-11.mit.edu:
/pub/linux/ALPHA/scanner/logiscan.tar.gz  

I haven't tried it yet, but I heard it works.
--

-Chris May, Computer Science, University of MA, Amherst
-       Technical Assistant, P.C. Maintenance Lab


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

From: michael@muc.de (Michael Schmidt)
Subject: Re: 1.1.15 breaks SCSI
Date: 29 May 1994 20:32:14 +0200

lellis@dmccorp.com (John Lellis) writes:

>Rene COUGNENC (rene@renux.frmug.fr.net) wrote:

>: The kernel 1.1.15 does not work at all with my Adaptec 1540B.
>: I get "Unable to reset SCSI host 0, probably a SCSI bus hang."

>: And I can't boot.

Same here, also with an 1540B.

>Funny, works fine with my Adaptec 1542CF.  Are you sure it patched correctly?

Could it be, that only 1540B's are hit?

        Michael

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

From: ewalton@magnus.acs.ohio-state.edu (Elaine Walton)
Subject: Re: Troubles with patches
Date: 29 May 1994 21:18:41 GMT

I think I did patch correctly; from another note, I may have gotten a
corrupted version.  So, I will download it from the source.
THANKS
-Sean

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

From: lgarde@scorch.hna.com.au (Leon Garde)
Subject: Re: SIGHUP - Where do we go from here?
Date: Sun, 29 May 1994 15:59:31 GMT

iiitac@uk.ac.swan.pyr (Alan Cox) writes:

>In article <1994May25.135713.31927@loreli.ftl.fl.us> nate@loreli.ftl.fl.us (Sean Puckett) writes:
>>Because of these facts:
>>
>>  1.  The kernel baseline will not change.  (The people who believe in
>>      100% Posix compliance are the people who control the baseline).
>>  2.  People will continue to have problems with non-compliant
>>      applications.

>All that is needed is to fix the shell programs that are managing the
>process groups and the SIGHUP - no other program is affected.

>Alan

So, all shells will have this largish section added especially for
linux. i wonder how many POSIX kernel's will handle SIGHUP'ing of
orphans the same way ...

leon

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: 1.1.15 breaks SCSI
Reply-To: pe1chl@rabo.nl
Date: Sun, 29 May 1994 21:38:22 GMT

In <2san3e$rm@colin.muc.de> michael@muc.de (Michael Schmidt) writes:

>lellis@dmccorp.com (John Lellis) writes:

>>Rene COUGNENC (rene@renux.frmug.fr.net) wrote:

>>: The kernel 1.1.15 does not work at all with my Adaptec 1540B.
>>: I get "Unable to reset SCSI host 0, probably a SCSI bus hang."

>>: And I can't boot.

>Same here, also with an 1540B.

>>Funny, works fine with my Adaptec 1542CF.  Are you sure it patched correctly?

>Could it be, that only 1540B's are hit?

I don't think so...  I have a 1542B and it works with 1.1.15 & 1.1.16.
The only difference between the 1540 and 1542 is supposed to be the
floppy controller.  And I have disabled it.

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

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

From: mckinsey@rainbow.sosi.com (Scott McKinsey)
Subject: kernel: Warning - bdflush not running
Date: 29 May 1994 17:52:01 -0700

The system is slackware 1.2 and ...
I have just installed the 1.1.16 kernel.  I am getting a message in
/var/adm/messages...

May 29 16:04:50 yggy kernel: Warning - bdflush not running
May 29 16:05:01 yggy last message repeated 2 times
May 29 16:05:47 yggy last message repeated 2 times
. 
. 

I also used the ide patch.  Everything went in without in rejects.
Anybody got a clue?
 Scott

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

From: gniibe@mri.co.jp (NIIBE Yutaka)
Subject: Re: InterLink for Linux
Date: Mon, 30 May 1994 10:00:54 GMT

In article <CqK41E.5oM@pe1chl.ampr.org> rob@pe1chl.ampr.org
        (Rob Janssen) writes:

   PLIP is a different thing.  At first, it was compatible with the Crynwr
   packet driver for parallel port.  Then, someone broke the compatability,
   probably not knowing about it.  Now, some people are again working on the
   PLIP driver to improve its efficiency, and I believe they also will restore
   compatability.

I'm working on PLIP, and it is compatible with the Crynwr.  It is not
as fast as old one, but it doesn't take much CPU power.

Since PLIP is busy-wait inherently, fast driver takes much CPU power.
For me, old one is too selfish.  Whenever there were tx packets, it
tries to transmit _all_ of them.  I changed the driver so that it
transmits only one packet at a time.

Well, current implementation of mine is not good for those who want
fast transmission.  I have an idea...

I'm thinking about implementing tx queue this weekend.  Changing tx
queue size, you can chose the balance between driver speed and cost of
CPU power.  To do this, I need some driver specific 'ioctl' scheme.

Thanks,
-- 
NIIBE Yutaka
Mitsubishi Research Institute, Inc.

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

From: colin@nyx10.cs.du.edu (Colin Plumb)
Subject: Re: SIGHUP: We tried it!
Date: 29 May 1994 19:09:43 -0600

In article <1994May21.164113.27409@loreli.ftl.fl.us>,
Rocco Caputo <troc@loreli.ftl.fl.us> wrote:
>Three questions and zero-content answers:
>
>  "Why is the sky blue?"
>      "Because it's supposed to be that way."
>  "Why is the earth round?"
>      "Because it's supposed to be that way."
>  "Why does SIGHUP only go to session leaders?"
>      "Because POSIX says so."
>
>Is that what folks around here consider a decent answer?

To expand on that last part "Because POSIX says so, therefore people writing
code assume t will behave that way."

I think that's a reasonable answer.

I don't have the POSIX standard, but the bits posted here say that the
leader gets SIGHUP, and when it exits (immediately unless it has trapped
SIGHUP somehow), the followers get SIGHUP.

That makes sense to me.
-- 
        -Colin

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

From: rwa@io (Richard W. Altheide)
Subject: XFree SVGA and 1.1.16
Date: Mon, 30 May 1994 01:02:57 GMT
Reply-To: rwa@umr.edu

I have been running XFree 2.1.1 SVGA without difficulty using v1.1.11.  
After moving up to v1.1.16 I find that the XVGA server hangs during 
startup via xdm.  The screen goes blank like the X server is trying to 
start, however, the root window is never displayed.  The keyboard and 
virtual console switching also fails to work so I can't switch to another 
virtual console.  The display may come up after pushing a few keys to 
switch virtual consoles and other time the keyboard locks up.  Everything 
works fine after rebooting with the v1.1.11 kernel.

I suspect this is related to the console and tty driver changes made at 
1.1.13, but, I have not had time to confirm this.  The strange thing is 
that my other system on which I first tested the 1.1.16 kernel which uses 
the Mach32 X server works without any difficulty (and before anyone asks 
it is the same kernel on both systems and I have retransfered it three 
times to eliminate the possibility of corrupted transfer).  Anyone else have 
any similar experiences??

--
Richard W. Altheide, Sr. Systems Programmer   |   E-mail:          rwa@umr.edu
UM-Rolla Computing Services                   |   URL: http://www.umr.edu/~rwa
114 Mathematics/Computer Science Building     |   Telephone:    1.314.341.4841
Rolla, MO   65401-0249                        |   Fax:          1.314.341.4216

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

From: g609296@win.or.jp (Barry Yip kam-wa)
Subject: Brain dead name server, overcome ?
Date: 30 May 94 01:19:35 GMT

First I must say this is not the problem of Linux but the Sun host which I
dial-in has a brain dead name server. Other than a few hosts in the
hosts file, it can't lookup other name server automatically. It makes me
great pain every time I use Mosaic, ftp or any other internet service.

I am currently using Term 1.15 and wonder is there any way to overcome
it. For example, if I setup my own name server, how to ask Mosaic or
ncftp to poll the local name server instead. Can this be done? I would
much appreciate any suggestions for work around.

--
Barry Yip
g609296@win.or.jp

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

From: otaylor@ellis.uchicago.edu (owen taylor)
Subject: Dynamic linking with shared libraries.
Reply-To: otaylor@midway.uchicago.edu
Date: Mon, 30 May 1994 03:19:52 GMT

Recently I've been doing some work on the question of dynamic linking
(By dynamic linking I mean here linking in modules at run time) and
I've gotten to a point where I'd like to my thoughts and code up for
comment:

Background
==========

My exploration was originally motivated largely by the impending
release of perl5, which will include dynamic linking as a standard
feature.  The ability to link dynamically is an important feature for
extensions to a interpreted language, as one does not (for reasons of
both convenience and disk space) want to make a separate executable
for each possible combination of extensions that one could desire.  

The only current solution for dynamic linking under Linux seems to be
dld.  dld is a nice package, being clean and portable.  However, it
has one major deficiency in its unmodified state that prevents it IMO
from being a really good solution for Linux: the lack of the ability
to link in objects linked to shared libraries.  For many of the
applications where one wants dynamic linking, i.e., interfacing to the
X libraries, big libraries that are prohibitive to static link to are
involved.  There are also other advantages to using shared libraries,
for example:

   Updates to libraries without recompilation.
   Easier inter-module coordination of symbols.

There are several possible solutions to this problem.  The cleanest,
most useful to have, and definitely the hardest-to-implement solution
would be to switch to a different executable format that supports the
easy creation and linking of shared objects.  (I understand that ELF
(or is it COFF?) fits this description.)  However, this would mean a
major departure from current practice, and, (I think), the support for
ELF shared libraries is not yet present in the the GNU tools.  For
these reasons this does not seem practical in the short term, although
it might be desirable in the long term.

What I've done
==============

In contrast to the above approach, the approach I have taken is very
elementary.  In essence, what I have done is modify the linker (the
standard Linux linker based on the GNU linker v1.9) to optionally
write out shared library information in combination with a relocatable
object, creating the extra necessary relocations within this
information.  I then modified dld to read this information and call
ld.so.  This is definitely a bit of a hack/kludge, which is especially
true of the modifications to GNU ld, which in the 1.9 form is a
somewhat of an ungainly monster. With the addition of the Linux shared
library additions + my patches, ughh!  The newer versions (current
2.4) seem much cleaner, but are, unfortunately without Linux shared
library support (a project for someone? It probably wouldn't offer,
however, too many tangible benefits other than beauty.)

Despite the hack nature of these patches, I think they may represent
something of use to people, and they seem work pretty well in their
domain.  For this reason I'd like to offer what I have currently to
any adventurous souls.  (This is definitely ALPHA code).  I've gotten
some fairly big things to work (primarily a version Malcolm Beatty's
tkperl5 modified for dynamic linking), but there are without doubt
tons of undiscovered bugs and possible improvements.  If you are
interested in a copy of the patches, email me and I'll send then to
you.

I would also be interested in any comments about the general approach
or information about any existing solutions.

                              Owen Taylor
                              o-taylor@uchicago.edu
-- 
*------------------------------------------------------------------*
* Owen Taylor              * SNOW *          o-taylor@uchicago.edu *

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


** 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
******************************
