Subject: Linux-Development Digest #766
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:     Fri, 27 May 94 18:13:09 EDT

Linux-Development Digest #766, Volume #1         Fri, 27 May 94 18:13:09 EDT

Contents:
  Re: Why is my Emacs binary so big (2.2M) (Robert Sanders)
  Handy Scanners (Re: Twain Scanner) (Karl Eichwalder)
  SLIP problems with kernel 1.1.15 (Andreas Kabel)
  lynx and select timeouts (lots of em) (Scott McKinsey)
  Re: 8k nfs performance (Michael Griffith)
  Re: SIGHUP: We tried it! (Kevin Brown)
  Re: 1.0.9 kernel bug (+ partial fix): when using non-loopback address to send to localhost, "from" address should also be non-loopback. (Kevin Brown)
  Re: NI5010 net driver ? (R.C.Van-Den-Bergh)
  Re: Q: Should recompiling the kernel update /etc/issue? (Charles E Meier)
  Re: 32-bit Novell desktop OS combines Unix, DOS 7 (Mark Evans)
  Re: Using emx/gcc for OS/2 and Linux (Tony Coates)
  Re: 32-bit Novell desktop OS combines Unix, DOS 7 (Russell Nelson)
  iBCS-2 support (Mike Jagdis)
  Re: SIGHUP - Deep Kernal Guts question! (Ken Pizzini)

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

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: 27 May 1994 12:46:40 -0400

petterr@stud.cs.uit.no (Petter Reinholdtsen) writes:

>Michael MNUK (mmnuk@risc.uni-linz.ac.at) wrote:
>> Am I linking statically (-static option never specified!) or is
>> something else wrong?

>You could try to strip the files. 

Furthermore, you should be aware that most GNU programs use "-g" in the
default CFLAGS, so that even if you strip the resulting executable, it'll
still be larger than normal.

My emacs 19.24, with X11 and toolkit options enabled:

$ ls -l =emacs
-rwxr-xr-t   2 root     bin       1490948 May 24 15:00 /usr/bin/emacs
$ file =emacs
/usr/bin/emacs: sticky Linux/i386 QMAGIC demand paged executable
$ ldd =emacs
        libXaw.so.3 (DLL Jump 3.1) => /usr/X386/lib/libXaw.so.3.1.0
        libXt.so.3 (DLL Jump 3.1) => /usr/X386/lib/libXt.so.3.1.0
        libX11.so.3 (DLL Jump 3.1) => /usr/X386/lib/libX11.so.3.1.0
        libc.so.4 (DLL Jump 4.5pl21) => /lib/libc.so.4.5.21
        libm.so.4 (DLL Jump 4.5pl21) => /lib/libm.so.4.5.21

(Yes, I know the sticky bit is useless, but the install script put
 there and I see no reason to take it off.  It's a sentimental link
 with emacssen past).

-- 
 _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*'

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

Crossposted-To: alt.sources.wanted
From: karl@pertron.central.de (Karl Eichwalder)
Subject: Handy Scanners (Re: Twain Scanner)
Reply-To: keichwa@gwdg.de (Karl Eichwalder)
Date: Fri, 27 May 1994 08:51:19 GMT

Byron A Jeff (byron@cc.gatech.edu) wrote:

In general : what is about handy scanners?

> In article <769949835.MC3516@mclink.it>,
> Francesco Duranti  <MC3516@mclink.it> wrote:

> >Has anyone written driver or program for the TWAIN scanner interface? I've
> >found two scanner driver on Sunsite but none of those is for Logitech
> >ScanMan.

"Andreas Beck" <becka@hp817s.rz.uni-duesseldorf.de> has start writing a
Logitech driver; I think he will mail you the recent version.

But I have still problems to get it work with the ScanMan256; I think you have
to hack /linux/kernel/ksyms.c -- I am not able to do this.

Another question: Does any free available scanner software (OCR / image
processing) exist, which could be ported to Linux?

--
                        | keichwa@gwdg.de
                        | karl@pertron.central.de
Karl Eichwalder         | 2:2437/209

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

From: kabel@marvin.tphys.uni-heidelberg.de (Andreas Kabel)
Crossposted-To: comp.os.linux.help
Subject: SLIP problems with kernel 1.1.15
Date: 27 May 1994 15:18:11 GMT

I used to set up  a SLIP connection to my DOS box via a null
modem cable by doing

        setserial /dev/cua0     
        slattach -s 38400 -p slip /dev/cua0 &

under kernel version 1.1.0. Today, I upgraded to v1.1.15 and 
couldn't connect to the remote machine after.

When I killed slattach, I got this error message: 

        tty_set_state: invalid argument
        tty_hang_up(DROP): invalid argument

Any clues?


 
==================================================================
Andreas Kabel                                   Bahnhofstr. 9-13
                                                D-69115 Heidelberg
kabel@marvin.tphys.uni-heidelberg.de            Tel.: 06221/181552

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

From: mckinsey@rainbow.sosi.com (Scott McKinsey)
Subject: lynx and select timeouts (lots of em)
Date: 27 May 1994 10:41:41 -0700

I have a performance problem with lynx 2.3.  I Have slackware 1.2 with a
1.0 kernel, also ncsa http 1.3.  All http files are local to the
machine, which is a 486dlc-33 (cyrix) with 8Meg ram.  I have done a
strace on lynx and mosaic.  Mosaic seems to work fine (most of the time!)
But lynx get a shitload of timeouts on the select call.  Here is the lynx
strace...
. 
. 
. 
write(1, "\1b[23;1H\1b[7mHTTP request sent; wa".., 59) = 59
sigaction(SIGTSTP, {0x48bd0, [SIGCONT SIGTSTP], 0}, (struct sigaction *)0) = 0
select(256, fdset0:[ 0], 0, 0, tv:[0,100]) = 0 (Timeout)
select(256, fdset0:[ 4], 0, 0, tv:[0,100000]) = 0 (Timeout)
select(256, fdset0:[ 0], 0, 0, tv:[0,100]) = 0 (Timeout)
select(256, fdset0:[ 4], 0, 0, tv:[0,0]) = 0 (Timeout)
select(256, fdset0:[ 0], 0, 0, tv:[0,100]) = 0 (Timeout)
select(256, fdset0:[ 4], 0, 0, tv:[0,0]) = 0 (Timeout)
select(256, fdset0:[ 0], 0, 0, tv:[0,100]) = 0 (Timeout)
. 
. 
. 
 
And on and on, until finally...

select(256, fdset0:[ 0], 0, 0, tv:[0,100]) = 0 (Timeout)
select(256, fdset0:[ 4], 0, 0, tv:[0,0]) = 1 (fdset0:[4])
read(4, "HTTP/1.0 200 OK\nDate: Friday, 27".., 1023) = 178
sigaction(SIGTSTP, SIG_IGN, {0x48bd0, [SIGCONT SIGTSTP], 0}) = 0

I have recompiled lynx several times to include the ncurses modifications.
I have also noticed that the binaries are static.

Has anybody observed this behavior before?  If so is there a fix?

Any help would be appreciated.

Thanks in advance.

Scott 

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

From: grif@tempest.ucr.edu (Michael Griffith)
Subject: Re: 8k nfs performance
Date: 27 May 1994 16:54:59 GMT

In article <KIRK.94May27085523@speech.braille.uwo.ca>,
Kirk Reiser <kirk@speech.braille.uwo.ca> wrote:
>A question I have is, is there benefit to be gained by using larger
>transfer blocks on linux to linux boxes?

Yes.  Early reports from Alan and others suggested that the 8k nfs
client would only be useful for improving performance between Linux
and Sun.  Sun uses a braindead algorithm for dealing with linux's 1k
blocks.  Mr. Varner just posted some statistics that show that it
doubles read performance linux <--> linux.  Similar speedups on writes
can be expected.

To use the 8k NFS client, you need a kernel >= 1.1.13, and you have to
add rsize and wsize options to your fstab (or AMD maps :-)).

>Running /usr as a nfs mounted drive is considerably slower than a
>standard mount fs.

We still have a way to go until Linux NFS performance can match a
SPARCstation.  Even on a good NFS implementation you can expect a 20%
slowdown over local filesystems.












-- 
Michael A. Griffith (grif@cs.ucr.edu)
Department of Computer Science
University of California, Riverside


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

From: kevin@frobozz.sccsi.com (Kevin Brown)
Subject: Re: SIGHUP: We tried it!
Date: Tue, 24 May 1994 01:20:29 GMT

In article <2rj7n2$rv6@charm.magnus.acs.ohio-state.edu> cemeier@magnus.acs.ohio-state.edu (Charles E Meier) writes:

>Why not a program that is put in inittab in place of getty which cruises the 
>/proc filesystem and kills anything that is being run by a user who isn't 
>logged in?  When it has cleaned things out, have it call getty.  Put a 
>config file together that lists legitimate stuff - you don't want to kill 
>update or lpd.  Might be easier than hacking the kernel.

...well, you'd have to be able to specify the set of users whose processes
you'd want it to kill and the set whose processes you'd want it to ignore.
Reason: you'd want people to be able to have crontabs, and not have the
program kill off those people's crontab entries.

I would presume the program would send SIGHUP...

>BTW - how were your kernel mods going to handle programs started by users 
>with a "nohup" option?  This is at least 1 reason why the kernel doesn't
>automatically SIGHUP all of a user's processes just because they hung
>up the phone line.

If a user explicitly wants a program to ignore SIGHUP, he can do so by
using the "nohup" program.  The kernel knows not to send SIGHUP to
processes which explicitly ignore it, so the resulting process can stick
around even though SIGHUP is sent to the process group (via whatever
means.  Apparently, POSIX decided that this task would be the responsibility
of the process group leader).

Admittedly, one would have to insure that any kernel mods retain this
behavior...


-- 
Kevin Brown                                     kevin@frobozz.sccsi.com
This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end >
            This is your .signature virus on drugs: <>
                        Any questions?

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

From: kevin@frobozz.sccsi.com (Kevin Brown)
Subject: Re: 1.0.9 kernel bug (+ partial fix): when using non-loopback address to send to localhost, "from" address should also be non-loopback.
Date: Thu, 26 May 1994 22:52:11 GMT

In article <Cpxv7J.91J@pe1chl.ampr.org> pe1chl@rabo.nl writes:
>In <1994May17.003827.14965@emc.rvt.com> remco@emc.rvt.com (Remco Treffkorn) writes:
>
>>Rob Janssen (rob@pe1chl.ampr.org) wrote:
>>: I think the original issue was only the existance of a bug and how it
>>: could be fixed.  When the bug is fixed in an alpha version and you don't
>>: want to upgrade to it, then so be it.  You still have your stable version
>>: with the bug.
>
>>: Rob
>
>>Do your emotions run away from you? 
>>What the heck is a "stable version with a bug"? An oxymoron?
>
>How do you define "stable"?
>
>1. it does not change
>
>2. it does not crash

I think most people, particularly here, use the latter definition.  But
see below.

>I think it is perfectly possible to have a stable version, under both
>definitions, of some piece of software.  Even if it includes known bugs
>that are nonfatal.

True enough.  However, the kernel has quite a few pieces.  It's possible to
have a stable kernel *and* unstable networking at the same time.

What people want is really stability *and* reliability.  That is, people
want the software to not only work continuously, but work *properly* as
well.

I think many embody all these concepts into the term "stable".  For
instance, stable networking often means that the networking code won't
do anything to hinder the performance of network-oriented tasks, and
certainly won't do anything to cause other network entities to crash.

Similarly, a "stable" operating system means, to many, that the operating
system will not only remain operational for extreme periods of time, but
that it will not prevent programs from doing the same (provided, of
course, that those programs are properly written).  Furthermore, all of
the services the operating system provides need to be similarly stable,
for the OS as a whole to be considered "stable" (at least by the above
"definition").


-- 
Kevin Brown                                     kevin@frobozz.sccsi.com
This is your .signature virus: < begin 644 .signature (9V]T8VAA(0K0z end >
            This is your .signature virus on drugs: <>
                        Any questions?

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

From: rcv@ukc.ac.uk (R.C.Van-Den-Bergh)
Subject: Re: NI5010 net driver ?
Date: Fri, 27 May 94 15:28:27 GMT

Hi,

has anyone done any work on the NI5010 ether card ?

I have acquired some and wonder if there is any support for these 8-bit
INTERLAN cards.

Thanks,

Cedric

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

From: cemeier@magnus.acs.ohio-state.edu (Charles E Meier)
Subject: Re: Q: Should recompiling the kernel update /etc/issue?
Date: 26 May 1994 22:30:26 GMT

In article <n9044144.769976582@rowlf>,
Khan M. Klatt <n9044144@rowlf.cc.wwu.edu> wrote:
>Recently I upgraded my kernel from version 0.99.15 to 1.1.0. I sync'd and 
>rebooted, and was greeted by
>
>Welcome to Linux 0.99.15 (or whatever)...
>
>What's wrong? Did I miscompile? So I recompile a couple more times. 
>Finally, I see 'version 1.1.0' on the bootup, but the login greeting is 
>still 0.99.15. Being a Linux/UNIX newbie, I decided it wasn't the kernel.
>

/etc/issue is just a simple text file that is displayed as part of the
login sequence before you type your login name.  It has nothing to do 
with the kernel.  Use your favorite editor and modify it to read whatever
you want.

A standard book that aspiring sysadmins should read is _Unix System
Administration Handbook_, by Nemeth, Synder, and Seebass, Prentice Hall.
It is a bit out of date and geared more toward BSD than SysV, but still 
a usefull.

cem

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

From: evansmp@mb5194.aston.ac.uk (Mark Evans)
Subject: Re: 32-bit Novell desktop OS combines Unix, DOS 7
Date: Fri, 27 May 1994 17:08:15 GMT

chris ulrich (insom@galaxy.ucr.edu) wrote:

: Except they would have to include IPX/SPX to sell it in an existing
: novell network, and the only way they could do that is if they put
: it in the kernel.  Bugger the rest, just give me netware interoperability
: and I can burn novell and toss netware.
: chris

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.

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

From: coates@kelvin.physics.uq.oz.au (Tony Coates)
Crossposted-To: comp.os.os2.programmer.porting
Subject: Re: Using emx/gcc for OS/2 and Linux
Date: 27 May 94 16:22:42
Reply-To: coates@physics.uq.oz.au

In article <2rv2ka$qsm@omega.gmd.de> veit@borneo.gmd.de (Holger Veit) writes:

   Another problem is that you have to rewrite all the DLLs that exist
   under OS/2 for Linux. Only some of them are specified in detail in the
   documentation. You would be surprised how many of them could be involved directly
   or indirectly even for a small OS/2 program. The Wine developers already have a
   hard job with porting about a dozen files containing the (public) API for Windows,
   the OS/2 API is much more complex, provided you do not want to restrict yourself
   to text mode apps (emulating OS/2 1.X).

In truth, my perspective here in one of wanting to port Linux apps to
OS/2, rather than vice-versa.  My OS/2 apps are usually text-mode and
written using the UNIX-compatible libraries that come with emx/gcc.
So, my own thoughts were towards having libraries, etc., for OS/2
which make porting of UNIX-style apps to Linux as easily as possible.
The thing is, GNU and free software support for Linux is much better
than for OS/2, and this is the software I would like to get at more
easily.  There would be no need in this case to support 16:16 code (I
hope), and once XFree86 is ported to OS/2, a compatible shell might be
possible too (if it seems worth the effort), though for myself I am
more interested in having X running under the Presentation Manager.

My assumption then is that, since all the Linux sources are available,
there wouldn't be a need to try to support undocumented or badly-
documented functions.
        Anyway, thanks for some valuable technical insights.
                                        Cheers,
                                                        Tony.
--
_____________________________________________________________________________
A.B.Coates, Dept. of Physics,
The University of Queensland  QLD  4072  Australia.
Email: coates@physics.uq.oz.au
Phone: (07/+617) 365-3424  Fax: (07/+617) 365-1242
Disclaimer: The University is ignorant of my
            opinions, let alone guilty ...
_____________________________________________________________________________

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

From: nelson@crynwr.crynwr.com (Russell Nelson)
Subject: Re: 32-bit Novell desktop OS combines Unix, DOS 7
Date: 27 May 1994 17:20:12 GMT

In article <1994May26.124452.11394@uk.ac.swan.pyr> iiitac@uk.ac.swan.pyr (Alan Cox) writes:
   In article <NELSON.94May25141928@crynwr.crynwr.com> nelson@crynwr.crynwr.com (Russell Nelson) writes:
   >You don't understand -- Corsair is a desktop system, not a server
   >system.  It's IPX support is going to be as a client, not a server.
   >You'll still need your netware server.

   Russ - NCP the layer the client would need to talk is a Novell trade secret
   and fairly well guarded. If they give out a client source it'll only be a
   month before someone knocks out a clone server (at least for file services)
   as it is it might take 6 months to deduce from documents and ethernet traces

It's already been done -- Puzzle Systems has NetWare server software
that runs on various Unix machines.

--
-russ <nelson@crynwr.com>      ftp.msen.com:pub/vendor/crynwr/crynwr.wav
Crynwr Software   | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St.      | +1 315 268 1925 (9201 FAX)    | Quakers do it in the light
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.

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

From: jaggy@purplet.demon.co.uk (Mike Jagdis)
Subject: iBCS-2 support
Date: Wed, 25 May 1994 22:16:00 +0000

* In message <MICHEL.94May24193313@tornado.seas.ucla.edu>,
  Scott Michel said:

SM> Sorry I came to the "Linux Distributions thread a wee bit late, but
SM> hey, that's what I get for working full time and holding three courses
SM> down. But the question is: Will there be iBCS-2 binary support?

Not the only thing you've missed by the sound of it. Take a look in
tsx-11.mit.edu:/pub/linux/ALPHA/ibcs2.

                                Mike  
 

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

From: ken@chinook.halcyon.com (Ken Pizzini)
Subject: Re: SIGHUP - Deep Kernal Guts question!
Date: 26 May 1994 22:28:34 GMT

In article <2s33un$r5m@paperboy.osf.org>,
Dan Swartzendruber <dswartz@pugsley.osf.org> wrote:
>Sigh.  That's precisely my point.  The original poster was
>talking about checking for something (SIGHUP), which doesn't
>apparently need to be checked for on a POSIX system, but does
>on many non-compliant Unix systems.  A following poster (if I
>remember correctly) took him to task for advocating coding not
>compliantly to POSIX, which is a different issue entirely, as
>the SIGHUP case demonstrates.

Sorry, I was unclear on what I was insulting -- I was insulting
the idea of "not checking for an error condition [...] because
it can never happen on a POSIX system".  I realized that it was
not you (Dan Swartzendruber) who was advocating this position.
Rereading my post, this was not obvious; my apologies.  My intent
was to agree with you and amplify with some additional justification
for this position.

                --Ken Pizzini

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


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