Subject: Linux-Development Digest #731
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:     Wed, 18 May 94 14:13:10 EDT

Linux-Development Digest #731, Volume #1         Wed, 18 May 94 14:13:10 EDT

Contents:
  MCA ESDI driver problems ... (Arindam Banerji)
  Prob with static vs shared libX* (Aric D. Blumer)
  Re: workaround for SB16 on stupid motherboard? (David Monro)
  Re: Appletalk support? (David Hornsby)
  Re: Appletalk support? (Vince Adams)
  Re: Dual-Mon Patches (Andreas Buschmann US/END 60/1/25 Tel.71409)
  Re: Compressed filesystem??? (Georg Wiegand)
  New Beta of my MDM package now available for testing (Anthony Rumble)
  Re: Linux 1.1.12...compile problem (Helmut Lichtenberg(m))
  Re: 1.0.9 kernel bug (+ partial fix): when using non-loopback address to send to localhost, "from" address should also be non-loopback. (Rob Janssen)
  Re: Driver for smart card (Rob Janssen)
  Re: Help with GDB (Rob Janssen)
  Re: setlocale() bug or feature? (Nick Ruprecht)
  Re: setlocale() bug or feature? (Mitchum DSouza)
  Re: setlocale() bug or feature? (Mitchum DSouza)
  Re: rpc.lockd for Sun? (Jerry Eyers)

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

From: axb@defender.dcrl.nd.edu (Arindam Banerji)
Subject: MCA ESDI driver problems ...
Date: 18 May 1994 11:18:30 GMT


    Peter De Schrijver and I have managed to create a ps2 esdi driver and boot 
 up the Debian distribution. However, we are running into some problems. I 
 was wondering if any of you could give us a hand in solving some of them, based 
 upon your previous experiences. 
  Before I begin, a simple explanation of the structure of the driver ...
 1. The routine that actually processes the reads and the writes does the 
     following : 
     Fills up a command block (in memory) that is to be sent to the
     controller
     transfers the command block to controller. 
     waits for an interrupt from the controller indicating that it is 
      ready for data transfer (does an uninterruptible sleep_on)
    sets up dma and starts it.
    waits for interrupt to indicate that dma is complete. (similiar sleep on) 
    returns. 
    As you've probably guessed - the interrupt handler wakes up the 
    sleeping threads in both the (above) cases. 

  2. One more fact - when the controller is too busy to process a command
     it causes an attention interrupt. When this interrupt occurs the interrupt
     handler prints out "attention error". 

   Now here comes the strange part - on some machines, the driver gets a lot of
   attention errors. On others it does not. If this attention error printk 
   is removed from the interrupt handler - the machines invariably hang on dma. 
   What is it that printk does that causes these machines to work, when 
   the printk is included within the interrupt handler ?  Seeting & clearing
   interrupts do not help, timing out of the dma does not help. 
   
    I've tried alkinds of possibilities and to no avail. It's almost as if
   the driver goes into an endless loop handling interrupts. The code works
   - since most of the code is based upon a previous driver (for another 
    research OS) that worked just fine, on the very same machine. In this 
   case too, if the printks are left in - the driver otherwise works fine. 

  Any help will be appreciated.  


- thanx
=============================================================================
Arindam Banerji                              (219)-631-5273 (Voice)
384 FitzPatrick Hall                         (219)-631-5772 (Voice)
Dept. of Computer Science and Engineering    (219)-273-0862 (Voice)
University of Notre Dame                     (219)-631-9260 (FAX)
Notre Dame, IN 46556                         axb@cse.nd.edu (E-mail)
=============================================================================


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

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

From: ablumer@hubcap.clemson.edu (Aric D. Blumer)
Subject: Prob with static vs shared libX*
Date: 17 May 94 02:58:49 GMT

Could someone try something for me?  I ftped the xcalendar* (not sure
of the version) program from ftp.x.org in the /R5contrib directory.
When I compile the program normally (i.e., it uses shared libs), it
Segment Faults.  When I set the -static flag (or the -g flag) and
relink it, it works fine.  Could someone be so kind as to try this to
see if theirs does the same?  I have libX*.so.3.1.0 and GCC 4.5.8.
Linux kernel 1.0.  XFree86 2.0 (I checked: 2.1 has the same libs).

Thank you much,
Aric.
ablumer@eng.clemson.edu

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

From: davem@extro.ucc.su.OZ.AU (David Monro)
Subject: Re: workaround for SB16 on stupid motherboard?
Date: Tue, 17 May 1994 02:14:48 GMT

In response to a request I received, I am posting the details of the
motherboard I have which doesn't support 16 bit dma. It is a 486DX33,
"CH486 Mainboard", made in Taiwan, no brand name in the manual. It is
based on a 3 chip OPTI chipset consisting of an 82C495, 82C392 and
82C206, if that means anything to anybody. Apart from the dma problem
it is fine. Has an AMI bios. Runs the bus quite happily at 13.3 MHz if
tweaked with the shareware amisetup program (and yes, I did try
untweaking the bus to see if that was the problem). Is there anyone
keeping a list of defective/problem hardware I should send this to?

        David

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

From: djh@cs.mu.OZ.AU (David Hornsby)
Subject: Re: Appletalk support?
Date: Wed, 18 May 1994 12:11:12 GMT

Porting CAP in "Native EtherTalk" mode to Linux is not feasible unless

1. you can get packets delivered to individual processes based on the
packet content, specifically the AppleTalk destination socket number.
Normally this requires a packet filter based on the Stanford or Berkeley
models.  Or,

2. You can get all incoming AppleTalk packets delivered to each interested
process and have the process discard it if unwanted, and do so in a
relatively efficient manner.

On the other hand, porting UAR (UNIX AppleTalk Router, which supports CAP)
to Linux should be simple, provided that Linux can send and receive 802.3
format packets AND support ethernet multicast packets (09:00:07:ff:ff:ff &
09:00:07:00:00:XY). The last time I checked, these facilities were missing
from Linux. If someone can tell me otherwise, I will happily write the
relevant UAR code bits tomorrow ...

 - David.

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

From: adamsvm@Dunx1.OCS.Drexel.Edu (Vince Adams)
Subject: Re: Appletalk support?
Date: Wed, 18 May 1994 12:15:06 GMT

Peter Desnoyers (peterd@pjd.dev.cdx.mot.com) wrote:
: For a while now I've been toying with the idea of designing a
: Localtalk card for a PC; however, I don't know how much use it would
: be. It should be possible to make one for a hell of a lot less than
: $300... 

:                               Peter Desnoyers
: -- 

I would buy one providing it costs a hell of a lot less than $300.  

Vince


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

From: buschman@slsvirt (Andreas Buschmann US/END 60/1/25 Tel.71409)
Subject: Re: Dual-Mon Patches
Date: Wed, 18 May 94 12:17:29 GMT

Carlo James Calica (calica@cae.wisc.edu) wrote:
: I'm wondering modern dual-monitor patches exist or have they been added
: to the kernel?  The most recent ones are for 0.99pl13.  Thanks.

What is in these patches?


I am running a two monitor configuration VGA / et4000 and 19" b&w
Hyundai HGC-128 (with it's own graphic card). Linux Slackware 1.2.0 .

I can use the Hyundai with the XFree Mono Server, either alone, or
together with the VGA Monitor.

When using only the Hyundai for X11, the VGA Monitor goes blank, and I
can't see the console any more.


Tschuess
        Andreas

-- 
#include <stddisclaimer.h>

 /|)    Andreas Buschmann
/-|)    SEL Stuttgart US/END

        buschman@lts.sel.alcatel.de             # we have a new domain address
        buschman@us-es.sel.de                   # our old domain address

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

From: gw@gwcomp.e.open.de (Georg Wiegand)
Subject: Re: Compressed filesystem???
Date: Mon, 16 May 1994 13:45:08 GMT
Reply-To: gw@gwcomp.e.open.de

In article <2qpdkt$5jh@netnews.upenn.edu>,
Robert G. Smith <rob@bip.anatomy.upenn.edu> wrote:
>Andrew R. Tefft (teffta@cs690-3.erie.ge.com) wrote:
>: In article 16508@n5ial.mythical.com, jim@n5ial.mythical.com
>: (Jim Graham) writes:
>: >There is also a third option which, btw, is the one I use.  Some time ago,
>: >there appeared a util called gexe.  I think it was posted to one of the
>: >comp.sources.* groups, but I'm not sure.  Anyways, it allows you to
>
>: It is gzexe, and it comes with gzip.
>
>: I tried it for a while but it complained on every binary (they still
>: worked, just complained) so I gave up. It wasn't worth the effort.
>: I considered using it on seldom-used but big binaries but decided 
>: to just delete them instead ;-)
>
>Try "tcx", available on sunsite.  It "is a system designed for
>the transparent decompression, execution, and recompression of
>executibles under Unix. It allows ... timeouts between 
>recompressions, and emergency directories in case a decompression
>fails from shortage of disk space.  The system is designed with
>a reasonable amount of robustness..."
>
>It works, with less trouble than gzexe!
>
>Rob Smith

I don't know "tcx", but all the trouble that gzexe produces was
a message like: Unexpected EOF, looking for "`" (or something similar)

The reason was a missing "`" in the gzexe-script.

This one works with no trouble

======= cut =======
:
#!/bin/sh
# gzexe: compressor for Unix executables.
# Use this only for binaries that you do not use frequently.
#
# The compressed version is a shell script which decompresses itself after
# skipping $skip lines of shell commands.  We try invoking the compressed
# executable with the original name (for programs looking at their name).
# We also try to retain the original file permissions on the compressed file.
# For safety reasons, gzexe will not create setuid or setgid shell scripts.

# WARNING: the first line of this file must be either : or #!/bin/sh
# The : is required for some old versions of csh.
# On Ultrix, /bin/sh is too buggy, change the first line to: #!/bin/sh5

x=`basename $0`
if test $# = 0; then
  echo compress executables. original file foo is renamed to foo~
  echo usage: ${x} [-d] files...
  echo   "   -d  decompress the executables"
  exit 1
fi

tmp=gz$$
trap "rm -f $tmp; exit 1" 1 2 3 5 10 13 15

decomp=0
res=0
test "$x" = "ungzexe" && decomp=1
if test "x$1" = "x-d"; then
  decomp=1
  shift
fi

echo hi > zfoo1$$
echo hi > zfoo2$$
if test -z "`(${CPMOD-cpmod} zfoo1$$ zfoo2$$) 2>&1`"; then
  cpmod=${CPMOD-cpmod}
fi
rm -f zfoo[12]$$

tail=""
IFS="${IFS=     }"; saveifs="$IFS"; IFS="${IFS}:"
for dir in $PATH; do
  test -z "$dir" && dir=.
  if test -f $dir/tail; then
    tail="$dir/tail"
    break
  fi
done
IFS="$saveifs"
if test -z "$tail"; then
  echo cannot find tail
  exit 1
fi

for i do
  if test ! -f "$i" ; then
    echo ${x}: $i not a file
    res=1
    continue
  fi
  if test $decomp -eq 0; then
    if sed -e 1d -e 2q "$i" | grep "^skip=[0-9]*$" >/dev/null; then
      echo "${x}: $i is already gzexe'd"
      continue
    fi
  fi
  if ls -l "$i" | grep '^...[sS]' > /dev/null; then
    echo "${x}: $i has setuid permission, unchanged"
    continue
  fi
  if ls -l "$i" | grep '^......[sS]' > /dev/null; then
    echo "${x}: $i has setgid permission, unchanged"
    continue
  fi
  case "`basename $i`" in
  gzip | tail | chmod | ln | sleep | rm)
        echo "${x}: $i would depend on itself"; continue ;;
  esac
  if test -z "$cpmod"; then
    cp -p "$i" $tmp 2>/dev/null || cp "$i" $tmp
    if test -w $tmp 2>/dev/null; then
      writable=1
    else
      writable=0
      chmod u+w $tmp 2>/dev/null
    fi
  fi
  if test $decomp -eq 0; then
    sed 1q $0 > $tmp
    sed "s|^if tail|if $tail|" >> $tmp <<'EOF'
skip=18
if tail +$skip $0 | "/usr/bin"/gzip -cd > /tmp/gztmp$$; then
  /bin/chmod 700 /tmp/gztmp$$
  prog="`echo $0 | /bin/sed 's|^.*/||'`"
  if /bin/ln /tmp/gztmp$$ "/tmp/$prog" 2>/dev/null; then
    trap '/bin/rm -f /tmp/gztmp$$ "/tmp/$prog"; exit $res' 0
    (/bin/sleep 5; /bin/rm -f /tmp/gztmp$$ "/tmp/$prog") 2>/dev/null &
    /tmp/"$prog" ${1+"$@"}; res=$?
  else
    trap '/bin/rm -f /tmp/gztmp$$; exit $res' 0
    (/bin/sleep 5; /bin/rm -f /tmp/gztmp$$) 2>/dev/null &
    /tmp/gztmp$$ ${1+"$@"}; res=$?
  fi
else
  echo Cannot decompress $0; exit 1
fi; exit $res
EOF
    "/usr/bin"/gzip -cv9 "$i" >> $tmp || {
      /bin/rm -f $tmp
      echo ${x}: compression not possible for $i, file unchanged.
      res=1
      continue
    }

  else
    # decompression
    skip=18
    if sed -e 1d -e 2q "$i" | grep "^skip=[0-9]*$" >/dev/null; then
      eval `sed -e 1d -e 2q "$i"`
    fi
    if tail +$skip "$i" | "/usr/bin"/gzip -cd > $tmp; then
      :
    else
      echo ${x}: $i probably not in gzexe format, file unchanged.
      res=1
      continue
    fi
  fi
  rm -f "$i~"
  mv "$i" "$i~" || {
    echo ${x}: cannot backup $i as $i~
    rm -f $tmp
    res=1
    continue
  }
  mv $tmp "$i" || cp -p $tmp "$i" 2>/dev/null || cp $tmp "$i" || {
    echo ${x}: cannot create $i
    rm -f $tmp
    res=1
    continue
  }
  rm -f $tmp
  if test -n "$cpmod"; then
    $cpmod "$i~" "$i" 2>/dev/null
  elif test $writable -eq 0; then
    chmod u-w $i 2>/dev/null
  fi
done
exit $res
======= cut =======

yours,
george

-- 
  ***************************************************************************
  # Georg Wiegand                   | Email: gw@gwcomp.e.open.de            #
  # University Essen, Germany       | Phone/Fax: +49 201 495192             #
  ***************************************************************************

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

From: smilie@lsupoz.apana.org.au (Anthony Rumble)
Crossposted-To: aus.computers.linux,comp.sources.testers
Subject: New Beta of my MDM package now available for testing
Date: 19 May 1994 00:30:21 +1000

Begin2
Title        = Linux Modem Server
Version      = 2.11-beta1
Desc1        = A modem server for Linux
Desc2        = Replaces uugetty.
Desc3        = Is easier to configure, provides better debug
Desc4        = modes and error reporting.
Author       = Anthony Rumble
AuthorEmail  = root@lsupoz.apana.org.au
Maintainer   = Anthony Rumble
MaintEmail   = root@lsupoz.apana.org.au
Site1        = lsupoz.apana.org.au
Path1        = /pub/mdm
File1        = mdm-2.11-beta1.tar.gz
FileSize1    = 22K
Required1    = Linux 1.0.? or 1.1.?
Required2    = Intellegent Hayes compatable modems
Required3    = Serial port/s
CopyPolicy1  = GNU Copyleft
Keywords     = LINUX MODEM GETTY
Comment1     = Released to beta with known bugs
Commend2     = to fix problem with 1.0.? and 1.1.? kernels
Entered      = 19MAY94
EnteredBy    = Anthony Rumble
CheckedEmail = root@lsupoz.apana.org.au
End


-- 
root@lsupoz.apana.org.au   APANA Sydney UUCP regional hub (feeds available)
Anthony Rumble             Linux Support OZ (02) 418-8750 v.32bis - 5 lines
Voice (02) 418-8220        For information on APANA mail info@apana.org.au

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

From: heli@tzv.fal.d400.de (Helmut Lichtenberg(m))
Subject: Re: Linux 1.1.12...compile problem
Date: Mon, 16 May 1994 06:38:14 GMT

stimpson@panix.com (S. Joel Katz) writes:
:       Using the interim releases exposes you to risks that you may not
: be able to handle. If you are not doing serious work on your system, why
: does it matter if it takes you an extra week or two to get the latest
: features? If you are, why are you taking the risk of using dangerous code
: when you will likely be unable to handle the problems it may throw you?
: 
:       SJK
:       Stimpson@Panix.COM
: 

Ok, you are right with your arguments. It's a kind of impatience from my side.
I can do my work with the 'old' kernel version.
I don't know, if you can call it "serious work" what I am doing :-). Some weeks
ago I put my Sun Sparcstation SLC aside, with which I worked several years
and changed it to a Linux Workstation, as the Sun was too slow (diskless with
16 MB memory!) and Solaris 2.3 too buggy.
Linux works very stable and reliable, even interim releases work better than
expensive commercial production release. 
In our local network of Suns and Linux PC's we sometimes have problems with the
net, it seems to be slow between the Suns and Linux and especially between
the Linux workstations the connection hangs sometimes.
As Matthew Dillon pointed in c.o.l.d to some tcp patches in the new kernel I wanted to try
if they solve our netproblems.
The risk to get dangerous code is about 10% against working with the dangerous code
of Solaris :-) .
But I will hold myself back in future in misusing this channel.
Nevertheless thanks for the help, especially from Dave Goldsmith who additionally
pointed to changes in sock.c.
So my new kernel compiled successfully this morning and I can taste the smell
of danger. :-)
Bye
Helmut
-- 
========================================================================
|                      Helmut Lichtenberg                              |
|           Institut fuer Tierzucht und Tierverhalten (FAL)            |
|                     heli@tzv.fal.d400.de                             |

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

From: rob@pe1chl.ampr.org (Rob Janssen)
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.
Reply-To: pe1chl@rabo.nl
Date: Sun, 15 May 1994 22:29:36 GMT

In <2r3ith$q6f@brachio.zrz.TU-Berlin.DE> goldt@math.tu-berlin.de (Sven Goldt) writes:

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

>: Already fixed as of 1.1.8
>: Alan

>Well,you should learn that there is the 1.0.x series and
>the 1.1.x series and it's no upgrade to use 1.1.x,but a
>risk.

And you are trying to explain that to Alan?

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

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Driver for smart card
Reply-To: pe1chl@rabo.nl
Date: Sun, 15 May 1994 22:33:26 GMT

In <CpuJM0.2Gp@edus.oau.org> news@edus.oau.org (Bob Perkins) writes:

>I would like to have a driver for a smart card for my linux box.
>If someone will send me a roadmap for a device driver for the
>serial ports in the linux kernel, I will do what I can.

>I have all the board ports, registers, commands, etc, but am
>looking for a map to tie it all together. If this was assembler,
>I could have the driver working in a couple of hours, but just
>learning sanscrit ( some times known as 'C' ). :-)

Of course, nothing keeps you from coding Linux drivers in assembler...

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

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

Crossposted-To: comp.os.linux.help
From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Help with GDB
Reply-To: pe1chl@rabo.nl
Date: Wed, 18 May 1994 08:44:02 GMT

In <1994May18.020435.17729@lmpsbbs.comm.mot.com> lenz@ssd.comm.mot.com (Doug Lenz) writes:

>Linux Guru's:

>I just migrated to Linux (Slackware 1.2.0 Distribution) and
>I have one small problem...

>Simple example:

>$ gcc -g hello.c -o hello
>$ gdb hello
>GDB is free software and you are welcome to distribute copies of it
> under certain conditions; type "show copying" to see the conditions.
>There is absolutely no warranty for GDB; type "show warranty" for details.
>GDB 4.12 (i486-unknown-linux), Copyright 1994 Free Software Foundation, Inc...
>(gdb) list
>1       #include <stdio.h>
>2
>3       main()
>4       {
>5               printf("hello\n");
>6       }

>** works fine **

>$ make
>gcc -Wall  -g -c workbone.c
>gcc -Wall  -g -c hardware.c
>gcc -g -o workbone workbone.o hardware.o -s
                                          ^^
                                          why do you put this here??

>groff -man workbone.1 | compress >workbone.1.Z
>[doug]> gdb workbone
>GDB is free software and you are welcome to distribute copies of it
> under certain conditions; type "show copying" to see the conditions.
>There is absolutely no warranty for GDB; type "show warranty" for details.
>GDB 4.12 (i486-unknown-linux), Copyright 1994 Free Software Foundation, Inc...
>(no debugging symbols found)...
>(gdb) list
>No symbol table is loaded.  Use the "file" command.
>(gdb) 

>** this don't **

>I hope somebody else has seen this before...

I see two different examples with two different compiler flags, and
they give (the expected) different results.
So what are you getting at? :-)

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

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

From: ruprecht@euklid.informatik.uni-dortmund.de (Nick Ruprecht)
Subject: Re: setlocale() bug or feature?
Date: 18 May 1994 12:46:41 GMT

In article <9405171673@fangorn.moria>, Michael Haardt <(michael)u31b3hs@pool.informatik.rwth-aachen.de> writes:
>pd@kubism.ku.dk (Peter Dalgaard SFE) writes:
>> Would be a good idea if someone went over it with a Book of
>> Standards in hand. For instance, I spent quite a bit of time and
>> energy to persuade elm to print Latin-1. Turned out that
>> "iso_8859_1" (for LC_CTYPE) is misspelt as "ISO-8859-1" in the source.

Sad enough that elm hard-codes that in the source.

>So you do have some information about possible values for the
>categories? Neither the "POSIX Programmer's Guide" (has a section about
>X/OPEN message catalog functions) nor various other manuals and books I
>checked gave me any clue what possible values for LC_MESSAGES are.  The
>HP-UX manual pages I checked gave examples with abc or xyz :-/
>
>And: Am I right in thinking that LC_MESSAGES should allow any value to
>be specified?

Actually, the last statement is true for all LC categories. The ANSI
standard defines only two locales: the C locale (invoked by default or
by setlocale(LC_ALL,"C")) and the default locale (setlocale(LC_ALL,"")).
As far as ANSI C is concerned, all other locales are implementation
specific. For example, on Suns, the existing locales are controlled by
text files in /usr/share/lib/locale. This way, you can define your own
locales by putting your own locale files in the appropriate place. The
standard configuration under SunOS 4.1 is the following:

        /usr/share/lib/locale:
                LANGINFO/    LC_CTYPE/    LC_MONETARY/ LC_TIME/     
                LC_COLLATE/  LC_MESSAGES/ LC_NUMERIC/  
        /usr/share/lib/locale/LANGINFO:
                C        default
        /usr/share/lib/locale/LC_CTYPE:
                C        default     iso_8859_1  
        /usr/share/lib/locale/LC_COLLATE:
                C        default
        /usr/share/lib/locale/LC_MESSAGES:
                C/       default
        /usr/share/lib/locale/LC_MONETARY:
                C        En_GB    En_US    default  
        /usr/share/lib/locale/LC_NUMERIC:
                C        En_GB    En_US    default  
        /usr/share/lib/locale/LC_TIME:
                C        En_GB    En_US    default  

Maybe there is some other place where this is standardized. ANSI C at
least doesn't deal with it. Actually, I find it quite useful to allow
the users to define their own favorite locales the way it is done in
SunOS. 

-- 

Nick Ruprecht                   ruprecht@ls7.informatik.uni-dortmund.de
The Computer Graphics Group     phone:  +49-231-755 6134
Department of Computer Science  fax:    +49-231-755 6321
University of Dortmund          D-44221 Dortmund, Germany

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

From: Mitchum DSouza <m.dsouza@mrc-apu.cam.ac.uk>
Subject: Re: setlocale() bug or feature?
Date: 18 May 1994 12:07:17 -0400
Reply-To: m.dsouza@mrc-apu.cam.ac.uk

| pd@kubism.ku.dk (Peter Dalgaard SFE) writes:
| > Would be a good idea if someone went over it with a Book of
| > Standards in hand. For instance, I spent quite a bit of time and
| > energy to persuade elm to print Latin-1. Turned out that
| > "iso_8859_1" (for LC_CTYPE) is misspelt as "ISO-8859-1" in the source.
| 
| So you do have some information about possible values for the
| categories? Neither the "POSIX Programmer's Guide" (has a section about
| X/OPEN message catalog functions) nor various other manuals and books I
| checked gave me any clue what possible values for LC_MESSAGES are.  The
| HP-UX manual pages I checked gave examples with abc or xyz :-/

This is very true. All manuals/doccumentation I have read too give conflicting
or inconsistent definitions for their values and uses.
 
| And: Am I right in thinking that LC_MESSAGES should allow any value to
| be specified?

LC_MESSAGES and any LC_XXXX for that matter will accept "C", "POSIX" or
"ISO-8859-1" with respect with linux. The former two are synonymous.

In theory any setting apart from the above ones will be defaulted to the "C"
locale.

Mitch

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

From: Mitchum DSouza <m.dsouza@mrc-apu.cam.ac.uk>
Subject: Re: setlocale() bug or feature?
Date: 18 May 1994 12:07:20 -0400
Reply-To: m.dsouza@mrc-apu.cam.ac.uk

| Trying to use X/OPEN message catalogues, I found out that setlocale()
| does not accept anything for LC_MESSAGES.  As far as I understood the

What do you mean it doesn't accept anything ?? C/ISO-8859-1 arent accepted ??

| manuals, the string for LC_MESSAGES specifies what will be substituted
| for %L in NLSPATH.  When setlocale(LC_ALL,"") is called, further calls
| of setlocale(LC_MESSAGES,(char*)0) should return the value of the
| environment variable LC_MESSAGES (or "C"/"POSIX" if it is not set).
| I use library 4.5.21.
| 
| When I write a small setlocale() substitute which behaves as documented
| above, messages catalogues work fine.  I sent email to hlu, but got no

Huh ? Now you're saying it works ??

| response, and I don't quite understand the library code of setlocale
| other than the ctype[] initialisation for latin1.
| 
| It also does not make any sense to me that the libc message catalogues
| are installed under .../locale/C/libc.cat, with libc.cat being either an
| English or German version: That would make LC_MESSAGES obsolete.

or French or Polish currently. setting LC_MESSAGES to ISO-8859-1 gives you
extended diactric and accented characters.

| 
| I would also appreciate if getcat(1) would be included in the library
| image release.

gencat(1) is not actually part of the C library and shouldn't really be included.
The C library has been catalogued but it is for the user to provide the
localization. gencat(1) however in my opinion should be made available with Rik
Faith's util-1.x.tar.gz package. BTW you can get it from 

        sunsite.unc.edu:/pub/Linux/nls/

| Did I totally misunderstand the semantics of setlocale() or did nobody
| ever test that code?

No I think it works fine as per my limited testing.

Mitch

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

From: jeyers@eosc.osshe.edu (Jerry Eyers)
Subject: Re: rpc.lockd for Sun?
Date: Tue, 17 May 1994 15:55:29 GMT


> 
>    Is anyone working on a lock manager for Linux, ala Sun's rpc.lockd?
> We are using Sun's PC-NFS to mount filesystems from our Linux server to our
> PCs, and the lack of a lock manager is the only thing preventing us from 
> moving a lot of our MSDOS software to the Linux server.  Applications such as
> Excel, for instance, require locking.
> 

Mike, 

I was wondering the same thing about lockd and posted a message to the
net a while back.  I didn't get any responses, so I send some mail to
Rick, who wrote an NFS implementation package for Linux.  This was his
response:


> 
> From jrs@world.std.com Tue May  3 17:12:14 1994
> 
> > I have been working on setting up a linux server with PCNFS clients
> > and the nfs stuff works ok as far as it goes.  I was wondering if
> > you knew where I could find a "lockd" or "statd" implementation for
> > linux.  Some of the large database and application software that
> > we use requires file locking and using NFS, that means lockd (or
> > equivelent).
> 
> Sun was nice because they made the specification to RPC and NFS public
> domain.  Unfortunately, they did not do the same for the Network Lock
> Managment Protocol.  Therefore, unless someone reverse engineers it,
> it is unlikely that there will ever be a free version.  All of the
> other vendors just license the code from Sun.
> 
> The good news is that network locking sucks and barely works, even on
> SunOS.  So we aren't missing very much.  :-)
> 
> Sorry,
> 
> Rick
> 

As a result, we have had to move our proposed server from Linux to
Interactive unix.  The new Interactive product is much more stable than Solaris
and the new versions come with lockd (much faster too).  Hope this helps.

Jerry

                                   _\\|//_
                                 0(` 0-0 ')0
=================================ooO=(_)=Ooo==================================
 Jerry Eyers                                                    .->-.
 Programmer Analyst                                             `-<-'
 Eastern Oregon State College                             transmitted on 100%
 jeyers@eosc.osshe.edu                                    recycled electrons
______________________________________________________________________________
   |#|                   |#|     <__>   <__>        |#|                  |#|



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


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