NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/comp/unix/question/10847

Internet Message Format | 1992-09-09 | 66.9 KB

Path: sparky!uunet!sun-barr!rutgers!cmcl2!adm!news
From: [e-mail] (SMTP MAILER)
Newsgroups: comp.unix.questions
Subject: Mail Delivery Problem
Message-ID: <[e-mail]>
Date: 8 Sep 92 23:48:10 GMT
Sender: [e-mail]
Lines: 1720


 ----Reason for mail failure follows----
Sending mail to recipient(s) woodfordm :
  Couldn't make final delivery.

 ----Transcript of message follows----
Received: from gw1.hanscom.af.mil by vd1.hanscom.af.mil with SMTP ;
          Tue,  8 Sep 92 08:42:20 EST
Date: 8 Sep 92 08:41:00 EST
From: [e-mail]
Subject: INFO-UNIX Digest  V16#002
To: "woodfordm" <[e-mail]>

Return-Path: <[e-mail]>
Received: from SEM.BRL.MIL by gw1.hanscom.af.mil with SMTP ;
          Tue,  8 Sep 92 08:36:15 EST
Received: from SEM.BRL.MIL by SEM.BRL.MIL id aa03664; 7 Sep 92 23:41 EDT
Received: from wharf.brl.mil by SEM.BRL.MIL id aa03619; 7 Sep 92 23:32 EDT
Received: from b62150.STUDENT.CWRU.Edu by WHARF.BRL.MIL id aa06893;
          7 Sep 92 23:10 EDT
Received:  by b62150.STUDENT.CWRU.Edu (5.65b+ida+/CWRU-1.4-client)
    id AA00609; Mon, 7 Sep 92 23:10:55 -0400 (from alex for [e-mail])
Return-Path: <[e-mail]>
Received: from sem.brl.mil by b62150.STUDENT.CWRU.Edu with SMTP (5.65b+ida+/CWRU-1.4-client)
    id AA03307; Mon, 7 Sep 92 18:38:52 -0400 (from [e-mail] for alex)
Received: from SEM.BRL.MIL by SEM.BRL.MIL id ab29929; 7 Sep 92 15:32 EDT
Received: from sem.brl.mil by SEM.BRL.MIL id aa29905; 7 Sep 92 15:15 EDT
Date:       Mon, 07 Sep 92 15:15:46 EST
From: The Moderator (Mike Muuss) <[e-mail]>
To: [e-mail]
Reply-To: [e-mail]
Subject:    INFO-UNIX Digest  V16#002
Message-Id:  <[e-mail]>

INFO-UNIX Digest          Mon, 07 Sep 1992              V16#002

Today's Topics:
      Re: I want a program that prints out base gid's of any user
    Re: how come : ! out of vi always uses sh instead of $SHELL ???
                                Inittab?
                                  FAQ
        Init: process table full?  What do you tune to fix this?
                           Re: Dear vi guru,
                  Re: Awk and getting the current date
                       Awk: Beginner needs help!
                     Re: Awk: Beginner needs help!
                  Re: missing /etc/resolv.conf - help
      Question: which is faster 'find -exec' or 'find | xargs' ??
                       Re: Extended chars in Unix
                  Problem building new ISC 3.0 kernel
                Re: Problem building new ISC 3.0 kernel
              Re: What does * mean in csh history subst ?
                    Re: Shell Scripts vs. C programs
                How to store data files in binary format
              Re: How to store data files in binary format
             Question: Stallion Technologies email address
        Have QIC-24(?) tape-drive - Need info on jumper settings
                      Re: Does free() really free?
              net.views -- which vendor is the most open?
                  Re: signal when window is destroyed
         Re: SUMMARY: What Unix HW/SW I will buy with my $8000
                More problems with find (the stderr bug)
             Re: need script to rename uppercase filenames
                      SIGIO - when does it occur ?
         Re: /usr/mail/glenr.lock not creatable after 10 tries
                        Berkely Fast File System
                      DRS/NX documentation anomaly
          Re: Problem with "umount" on SCO-ODT (UNIX v. 5.3.2)
            termcap entry for a Tandem TS530 Series Terminal
          Playing CD in workstation A, hearing it in B, Can I?
                        Re: Ytalk source code...
-----------------------------------------------------------------

From: "Daniel B. Suthers" <[e-mail]>
Subject: Re: I want a program that prints out base gid's of any user
Date: 27 Aug 92 23:17:03 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Tom Milner) writes:
>In comp.unix.questions, [e-mail] (Benjamin Z. Goldsteen) writes:
>
>
>|       I have been trying to write a program that prints out the
>|    base gid of a user (given on the command line).  I have tried
>|    using awk and perl, but I can not get anything decent.

The following shell script should work in a resonably portable and robust way:
GRPNO=`grep "^$1:" /etc/passwd | cut -f4 -d:` &&
sed -n "s/^\(.*\):.*:$GRPNO:.*/\1/"p /etc/group

Daniel B. Suthers,  CCP
Technology Consultant, PCS & IN Development Lab,        Pac*Bell
Voice: [phone]               UUCP:  [e-mail]
================================================================================
=  Commit unexpected kindnesses and senseless acts of beauty.                  =
================================================================================

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

From: Erick Rudiak <[e-mail]>
Subject: Re: I want a program that prints out base gid's of any user
Date: 28 Aug 92 14:22:43 GMT
Sender: "Usenet on news.acns" <[e-mail]>
To:       [e-mail]

|> >In comp.unix.questions, [e-mail] (Benjamin Z. Goldsteen) writes:
>|       I have been trying to write a program that prints out the
>|    base gid of a user (given on the command line).  I have tried
>|    using awk and perl, but I can not get anything decent.

    Assuming you give the user name on the command line, such as

    % getgroup username

you could do something as easy as:

#! /bin/sh
#
#  Script to generate group id for user from /etc/password file

GROUPID=`grep $1 /etc/passwd | cut -d: -f4`
echo "The group id for $1 is: " $GROUPID


    What this does is look for the user's entry in /etc/passwd.  You might
want to put something in there to look for only that user specifically, because
their password entry might be a string that is contained in someone *else*'s
entry as well.  Just save the above script as getgroup and you're set.

-- Erick
_____________________________________________________________________
                                                |
I tell myself, hey, only fools rush in          | Erick Rudiak
Only time will tell if we stand the test of time| [e-mail]
And all I know, you've got to run to win        |____________________
And I'll be damned if I get hung up on the line | "I'm not worthy..."
                                                |      \o/      \o/
    -- Van Halen, "Why Can't This Be Love"      | O__   \\  O__  \\
________________________________________________|_/__|__||__/__|_||__

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

From: "Daniel B. Suthers" <[e-mail]>
Subject: Re: how come : ! out of vi always uses sh instead of $SHELL ???
Date: 27 Aug 92 23:27:56 GMT
To:       [e-mail]

[e-mail] writes:
>: whenever i :! out of vi it uses /bin/sh instead of $SHELL
>: so all of my aliases are lost, etc and makes it very limited


Checking the source for VI on our SVR3 machine shows that VI uses the
execl(2) call to run sub-shells.  It incorrectly uses the hardcoded
string "sh" as the second argument to that call.  This means that the
shell you use THINKS it was invoked as "sh", as this is what's in ARGV[0].

It actually uses whatever is in your (vi) shell variable.  This variable is
loaded from a getenv() call and possibly overwritten by your EXINIT string.

Now, to address the complaint: Aliases are lost because aliases are not
passed to child processes.  This applies to Bourne and KSH.  I don't
know about CSH.

My solution is to have all of my aliased commands backed up as shell
commands.  This allows a reasonable environment whether in a subshell or
not.

Daniel B. Suthers,  CCP
Technology Consultant, PCS & IN Development Lab,        Pac*Bell
Voice: [phone]               UUCP:  [e-mail]
================================================================================
=  Commit unexpected kindnesses and senseless acts of beauty.                  =
================================================================================

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

From: Benjamin B Thomas <[e-mail]>
Subject: Inittab?
Date: 28 Aug 92 00:03:05 GMT
Sender: News system <[e-mail]>
Nntp-Posting-Host: e40-008-4.mit.edu
To:       [e-mail]




I was wondering if someone could help me clear something up.
The documentation I have leads me to believe that if I want to run a script from
/etc/inittab at run level 3 and insure that it remains alive, I need to add this
to /etc/inittab:

foo :3:respawn:/bin/sh -c 'exec /usr/local/bin/my_script'

The script gets executed (ps gives:  ...../usr/local/bin/my_script /usr/local/)

The script writes to a tmp file, sleeps, checks some stuff, and then writes over
the tmp file (ls -ct > /tmp/foo) This works fine when
I execute from the command line, but won't write anythingwhen started from initt
ab.
However, it will write a blank /tmp/foo if I place anything in the file while th
e process is sleeping.

Any ideas?   BTW It's a csh script.

                                 Thanx,
                    Benjy

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

From: Kai Zhu <[e-mail]>
Subject: FAQ
Date: 28 Aug 92 01:44:08 GMT
NNTP-Posting-Host: panda.cs.utexas.edu
To:       [e-mail]

I would to like to have a copy of FAQ of this news group.
Could anyone send me a copy or tell me the FTP site ?
Thanks in advance.

-Kai Zhu
[e-mail]

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

From: Rob Yampolsky <[e-mail]>
Subject: Init: process table full?  What do you tune to fix this?
Date: 3 Sep 92 15:49:07 GMT
Sender: Rob Yampolsky <[e-mail]>
To:       [e-mail]

I have been attaching terminals to an Interactive 3.0 system using Digiboard
concentrators.  After adding the 7th group of 16 async ports, I started getting
"Init: process table full" (not an exact quote) errors.  At this point, I could
not shut the system down cleanly.  Changing the getty lines in /etc/inittab
to "off" for the last 16 terminals clears up the error condition, but of course
I can't use the terminals.

Is there a tuneable parameter "init" uses to size it's process table?  I don't
think it's NPROC because the remaining terminals can run lots of processes
with no problem.
--
================================================================
=     Mail replies appreciated if possible (who has time to    =
=     read this stuff?)               Thanks, [e-mail]    =
================================================================

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

From: Bryan Curnutt <[e-mail]>
Subject: Re: Dear vi guru,
Date: 3 Sep 92 16:30:59 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Tom Parker) writes:
>In article <[e-mail]> [e-mail] (Brian D. Schieber) writes:
>>Dear vi guru,
>>
>> I like to use tabs alot in just about every session of vi I
>>work with, It would be great if I had a macro or whatever that
>>could replace the tabs with the appropriate number of spaces
>>in a file WHILE in vi. Do you have such a macro?
>
>Check if your system has the 'expand' command.  This command converts
>tabs to spaces, and vice versa.

 ... but is not what Brian asked for.

Assuming your tabs are for indentation at the beginning of the line
only, i.e. that you don't have text characters followed by tabs,
something simple like

    :map g :%s/^I/    /g^M
[tab character]^^       ^^[carriage return]

will do what you want.

If your system does have the 'expand' command, then something like

    :map g :%!expand -4^M

will do the trick, and do it correctly (unlike the %s above, which
only works if the tabs are all at the beginning or coincidentally
aligned so they start at a tabstop).

There may be a simple way to correctly expand tabs from within vi
without resorting to external programs, but I'm not really a guru,
just a Joe Random User.
--
Any opinions above are mine, and do not necessarily reflect the views of SAI.
Bryan Curnutt
[e-mail]

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

From: Phil Hughes <[e-mail]>
Subject: Re: Awk and getting the current date
Date: 3 Sep 92 17:12:01 GMT
X-Newsreader: Tin 1.1 PL3
To:       [e-mail]

mbrodesky@sc9 writes:
: How can I get the current date in a AWK program

This is the most awk-ish way to do it:

    awk 'BEGIN  {print dt}' dt="`date`"  file

The "program" in this example just consists of the "print dt" statement
performed at begin time.  Following the program, dt="`date`" executes
the date command (because of the backquotes) and then assigns the result
to the awk variable dt.  The double quotes are needed as executing date
expands to multiple words.

--
Phil Hughes, SSC, Inc. P.O. Box 55549, Seattle, WA 98155  (206)FOR-UNIX
>>> Publishers of pocket references for UNIX, C, VI, Emacs, Ksh, MS-DOS, ... <<<
     uunet!nwnexus!ssc!fyl or [e-mail]            [phone]

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

From: "Steven C. Velasco" <[e-mail]>
Subject: Awk: Beginner needs help!
Keywords: awk unix
Date: 3 Sep 92 19:32:05 GMT
Sender: [e-mail]
To:       [e-mail]


HI,
I am new to unix, and really new to awk, so hopefully somebody can
answer this question real easily.  I have hundreds of SAS program
files written for MVS, that we now need to run on UNIX.
The files from MVS look like this :

COMMAND? list unnumbered
//MVS JCL statements
//MVS JCL statements
//MVS JCL statements
DATA _NULL_ ;
   SET SASLIB.DATASET ;
MORE SAS STATEMENTS
RUN ;

I can get awk to print out the lines that don't contain the string  "//"
or "COMMAND?" in $1, but I would like to have
something that replaces the entire line  "COMMAND? list unnumbered"
(this is always the first line, and occurs just once)
with the string: "libname saslib /usr2/username/datasets"  or, to
somehow put the "libname" string in the first line of each file.
So the resulting file would look like this:

libname saslib '/usr2/username/datasets' ;
DATA _NULL_ ;
   SET SASLIB.DATASET ;
MORE SAS STATEMENTS ;
RUN ;

below is my feeble attempt to do this.  When I run this program, I
get a message saying something like 'awk: cant set $0 at record 1':

$1 !~ /\/\// { if ( $0 ~ /COMMAND? list unnumbered/ )
$0 = "libname saslib /usr2/username/datasets ;"
print >FILENAME }

What am I doing wrong?  I would appreciate any help or suggestions.

Steven C. Velasco
[e-mail]  -OR-
[e-mail]

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

From: Greg Hunt <[e-mail]>
Subject: Re: Awk: Beginner needs help!
Keywords: awk unix
Date: 6 Sep 92 19:41:39 GMT
Sender: Greg Hunt <hunt@robin>
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Steven C. Velasco) writes:
>
> HI,
> I am new to unix, and really new to awk, so hopefully somebody can
> answer this question real easily.  I have hundreds of SAS program
> files written for MVS, that we now need to run on UNIX.
> The files from MVS look like this :
>
> COMMAND? list unnumbered
> //MVS JCL statements
> //MVS JCL statements
> //MVS JCL statements
> DATA _NULL_ ;
>    SET SASLIB.DATASET ;
> MORE SAS STATEMENTS
> RUN ;
>
> I can get awk to print out the lines that don't contain the string  "//"
> or "COMMAND?" in $1, but I would like to have
> something that replaces the entire line  "COMMAND? list unnumbered"
> (this is always the first line, and occurs just once)
> with the string: "libname saslib /usr2/username/datasets"  or, to
> somehow put the "libname" string in the first line of each file.
> So the resulting file would look like this:
>
> libname saslib '/usr2/username/datasets' ;
> DATA _NULL_ ;
>    SET SASLIB.DATASET ;
> MORE SAS STATEMENTS ;
> RUN ;
>
> below is my feeble attempt to do this.  When I run this program, I
> get a message saying something like 'awk: cant set $0 at record 1':
>
> $1 !~ /\/\// { if ( $0 ~ /COMMAND? list unnumbered/ )
> $0 = "libname saslib /usr2/username/datasets ;"
> print >FILENAME }
>
> What am I doing wrong?  I would appreciate any help or suggestions.

I think this awk script might be easier:

    BEGIN {
        getline;
        print "libname saslib '/usr2/username/datasets' ;";
          }
    {if (substr ($1,1,2) != "//") print $0}

Then run it like this for each mvs file you have:

    awk -f awk_script mvs_file > unix_file

The BEGIN pattern runs once when the awk script is started.  It does
a getline to read the next line of the input file, thus ignoring the
first line of the file (the COMMAND? line).  Then it prints the line
that you want instead.

The next pattern runs for each line of the input file.  It looks at
first two bytes of the first field, and if they aren't "//", it prints
the whole line.  This will ignore the JCL statements.

Enjoy!

--
Greg Hunt                            email: [e-mail]
DG/UX Core Development
Data General Corporation
Research Triangle Park, NC, USA      These opinions are mine, not DG's.

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

From: Sabina Wolfson <[e-mail]>
Subject: Re: missing /etc/resolv.conf - help
Date: 4 Sep 92 02:34:43 GMT
Sender: netnews admin account <[e-mail]>
X-Disclaimer: Nyx is a public access Unix system run by the University
    of Denver for the Denver community.  The University has neither
    control over nor responsibility for the opinions of users.
To:       [e-mail]

In article <[e-mail]> [e-mail] (Simon Langley) writes:
>Sabina Wolfson ([e-mail]) wrote:
>: Or is there some way I can send a request to another name server
>: (I know which namer server my site should be using)?
>:
>my resolv.conf just says:
>nameserver 158.152.1.65
>
>and it uses this host as the nameserver.  However, there are other changes
>that may need to be made and these are system dependent.

One of the problems is that I don't have the correct access to be changing
around these files (and I probably don't have anough UNIX knowledge either
though it would be a lot of fun to try :)  However, as many helpful people
pointed out in e-mail, all I really needed to do was:

    nslookup <site-to-look-up> <name-server>

All done quickly and beautifully on the command line.

Sabina :)- [e-mail]

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

From: John Navarra <[e-mail]>
Subject: Question: which is faster 'find -exec' or 'find | xargs' ??
Date: 4 Sep 92 04:10:33 GMT
Sender: "Usenet on news.acns" <[e-mail]>
To:       [e-mail]


I was experimenting with find and xargs. I issued the command
  find . -exec ls -ld {} \;
and the command
  find . -print | xargs ls -ld
and found they gave the same results. However, I noticed the execution
times for the two were MUCH different. If I time the two commands
from my home directory, which contains many files, here is what I get:

[casbah:47] ~ -> time find . -exec ls -ld {} \; > /dev/null
      287.4 real        39.1 user       219.9 sys
[casbah:48] ~ -> time find . -print | xargs ls -ld > /dev/null
       47.6 real         0.0 user        17.3 sys


Question: why is the xargs command MUCH faster?

While running the following commands, I ran ps -j many times. Here are some
of those results:

using exec:
PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21398   SOE   453  0:03 -bash (bash)
 9120 21383 21383  9120 r6 21398    SE   453  0:00 find . -exec ls -ld {} ;
 9120 21398 21398  9120 r6 21398    RE   453  0:00 ps -j
21383 21401 21383  9120 r6 21398    RE   453  0:00 ls -ld ./.elm/lsig.bak
 PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21409   SOE   453  0:03 -bash (bash)
 9120 21383 21383  9120 r6 21409    SE   453  0:00 find . -exec ls -ld {} ;
 9120 21409 21409  9120 r6 21409    RE   453  0:00 ps -j
21383 21412 21383  9120 r6 21409    RE   453  0:00 ls -ld ./bin/file_clean
 PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21422   SOE   453  0:03 -bash (bash)
 9120 21383 21383  9120 r6 21422    SE   453  0:00 find . -exec ls -ld {} ;
 9120 21422 21422  9120 r6 21422    RE   453  0:00 ps -j
21383 21426 21383  9120 r6 21422    RE   453  0:00 ls -ld ./bin/lastarg

In this case, bash is the parent to the find command, and find is the
parent to the ls -ld filename command. Thus, there is one find process,
plus X ls -ld filename processes where X is the number of files to be
listed. (obviously, I was not fast enough to capture all of them)


using xargs:
 PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21857   SOE   453  0:06 -bash (bash)
 9120 21817 21817  9120 r6 21857    SE   453  0:02 find . -print
 9120 21818 21817  9120 r6 21857    SE   453  0:03 xargs ls -ld
 9120 21857 21857  9120 r6 21857    RE   453  0:00 ps -j
 PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21867   SOE   453  0:06 -bash (bash)
 9120 21817 21817  9120 r6 21867    SE   453  0:02 find . -print
 9120 21818 21817  9120 r6 21867    SE   453  0:03 xargs ls -ld
 9120 21867 21867  9120 r6 21867    RE   453  0:00 ps -j
21818 21868 21817  9120 r6 21867    RE   453  0:00 ls -ld ./awk/shellreport.naw
 PPID   PID  PGID   SID TT TPGID  STAT   UID  TIME COMMAND
 9119  9120  9120  9120 r6 21900   SOE   453  0:07 -bash (bash)
 9120 21817 21817  9120 r6 21900    RE   453  0:04 find . -print
 9120 21818 21817  9120 r6 21900    RE   453  0:07 xargs ls -ld
 9120 21900 21900  9120 r6 21900    RE   453  0:00 ps -j
21818 21908 21817  9120 r6 21905    RE   453  0:00 ls -ld ./perl/perl_info/docs

In this case, bash is the parent of both the find and xargs commands. And
xargs is the parent of the ls -ld filename commands. Thus, there is
one find process, one xargs process, and X ls -ld processes (again, where
X is the number of files to be listed.)

The way I count it, the xargs command has one extra process, the overhead
of setting up a pipe, and the fact that two commands are being used vs
one. But, for some reason the xargs command was MUCH faster than the
find -exec command.

What's the deal? From what I can gather, find -exec will write the
stdout X times in this case where xargs writes in chunks. Is that
correct?


-tms
--
You can get further with a kind word | You can get further with a kind word
and a gun than a kind word alone.    | and a phaser than a kind word and a gun.
          --al capone                |           -- John Navarra
=======From the Lab of the MaD [e-mail]========

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

From: "S.Riehm" <[e-mail]>
Subject: Re: Extended chars in Unix
Date: 4 Sep 92 12:55:30 GMT
Sender: [e-mail]
To:       [e-mail]

[e-mail] (Scott Barnhart) writes:

>I am using Unix SystemVR4, and want to find a way to generate the
>extended ASCII chars from the keyboard.  The reason for this is
>to generate the appropriate foreign characters (such as the vowels
>with the accents, etc) needed in the various different languages
>(specifically French, German and other European languages).

>I first thought it could be done somehow with the LOCALE stuff
>in Unix, setlocale(), chrtbl(), etc.  But the man pages don't go
>into how to generate or translate these characters.

The only IMPORTANT thing that I found about this stuff is that you
MUST have 8-bit clean programs. For example, I use Xterms, so I had to
set the eightBitInput and eightBitOutput resources ( although I think
they are default ), then its a matter of using Meta or Alt keys to set
the eighth bit, which in turn gives you these weird characters. In my
set up, an alt-k gives an e with umlauts ( two little dots ), the
combinations are pretty cryptic, but I got around that by setting up
translation tables for xterms, ( read the manual for them.. or ask for
a listing... I'll post it if I get > 3 requests ). So I now use a
compose sequence, compose + " + e = ALT-K for example. What you are
supposed to do at a dumb terminal without Alt keys, I have no idea..
get your boss to buy you a new terminal! :-)

This gets the input side of thing working, you then need to make sure
that your output device can handle 8bit characters, for XWindows you
just set the default font to an ISO font or equivalent ( I am using
iso1.16 but I have no idea what it maps to, cus I can't find it in any
font.aliases files ). I also had to set the environment variable
LESSCHARSET=latin1 to get less to understand that an 'international'
character ( foreign sounds so..... US v's THEM'ish ) is not a control
character to be barfed at.

( fx: sarcasm on :xf )
see its really quite simple isn't it??
( fx: sarcasm off :xf )

( fx: flame thrower set to: Light Toast :xf )
Of course in a perfect world vendors would realise that
internationalisation is not a bad thing, and that the default
settings should be 'international character sets' friendly, then those
that only use the simple a-z characters got no problems, and everyone
in not-America, and all those corresponding with not-America would also have
no problems. ( I say not-America because its pretty difficult to type
a real pound sign, not a # but a # <- what goes after british
currency, using the 'standard' character set )

Am I really asking too much??
It only took me several days to set up international characters right,
and I still have a couple of bugs to get around, and I have had experience
at setting up Xwindows etc.. it wasn't difficult to do, just so
incomprehensibly interconnected and confusing, setting an environment
variable does nothing!

 -----------------------------------------------------------------
Stephen Riehm        Configuration Management       _-_|\
[e-mail]    Philips Kommunikations Industrie  /     \
Work: [phone]    N|rnberg, Germany          \_.-.*/
Fax:  [phone]    "I was there, now I am here!"           v
"My company speaks another language, I CAN'T speak on it's behalf"
PS: if you can't read the name of my city, you are not being
international enough! and it's YOUR problem.. not mine!

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

From: A Wizard of Earth C <[e-mail]>
Subject: Re: Extended chars in Unix
Date: 7 Sep 92 09:51:57 GMT
Sender: [e-mail]
To:       [e-mail]

In article <smr.715611330@hitkw14> [e-mail] (S.Riehm) writes:
>[e-mail] (Scott Barnhart) writes:
>
>>I am using Unix SystemVR4, and want to find a way to generate the
>>extended ASCII chars from the keyboard.  The reason for this is
>>to generate the appropriate foreign characters (such as the vowels
>>with the accents, etc) needed in the various different languages
>>(specifically French, German and other European languages).
>
>>I first thought it could be done somehow with the LOCALE stuff
>>in Unix, setlocale(), chrtbl(), etc.  But the man pages don't go
>>into how to generate or translate these characters.
>
>The only IMPORTANT thing that I found about this stuff is that you
>MUST have 8-bit clean programs. For example, I use Xterms, so I had to
>set the eightBitInput and eightBitOutput resources ( although I think
>they are default ), then its a matter of using Meta or Alt keys to set
>the eighth bit, which in turn gives you these weird characters. In my
>set up, an alt-k gives an e with umlauts ( two little dots ), the
>combinations are pretty cryptic, but I got around that by setting up
>translation tables for xterms, ( read the manual for them.. or ask for
>a listing... I'll post it if I get > 3 requests ). So I now use a
>compose sequence, compose + " + e = ALT-K for example. What you are
>supposed to do at a dumb terminal without Alt keys, I have no idea..
>get your boss to buy you a new terminal! :-)
>
>This gets the input side of thing working, you then need to make sure
>that your output device can handle 8bit characters, for XWindows you
>just set the default font to an ISO font or equivalent ( I am using
>iso1.16 but I have no idea what it maps to, cus I can't find it in any
>font.aliases files ). I also had to set the environment variable
>LESSCHARSET=latin1 to get less to understand that an 'international'
>character ( foreign sounds so..... US v's THEM'ish ) is not a control
>character to be barfed at.
>
>( fx: sarcasm on :xf )
>see its really quite simple isn't it??
>( fx: sarcasm off :xf )
>
>( fx: flame thrower set to: Light Toast :xf )
>Of course in a perfect world vendors would realise that
>internationalisation is not a bad thing, and that the default
>settings should be 'international character sets' friendly, then those
>that only use the simple a-z characters got no problems, and everyone
>in not-America, and all those corresponding with not-America would also have
>no problems. ( I say not-America because its pretty difficult to type
>a real pound sign, not a # but a # <- what goes after british
>currency, using the 'standard' character set )
>
>Am I really asking too much??
>It only took me several days to set up international characters right,
>and I still have a couple of bugs to get around, and I have had experience
>at setting up Xwindows etc.. it wasn't difficult to do, just so
>incomprehensibly interconnected and confusing, setting an environment
>variable does nothing!

    You are Assuming Scott was even talking about an X application.

    If he was not talking about an X application, then he was asking
a question about his keyboard driver, which would probably be better
answered by his UNIX vendor (the wonderful thing about standards is that
there are so many to choose from).

    If he *was* asking about X (and thus a high probability of
meaning the "xterm" application instead), you have almost given the
correct advice.

    The real question is what is meant by international characters.
In the first instance, if he's on the console, he can only be talking
about the PC "extended" ASCII character set.  This supports only a subset
of the characters needed for full internationalization.  While it is
true that one can provide an alternate character set, to my knowledge,
only SCO supports this feature, and SCO doesn't do SVR4, the indicated
environment.

    Generally, as you say, one of the most important pieces is an *
bit clean OS and utilities.  This is generally unavailable in UNIX
systems in the US, so if he bought a US UNIX shipped abroad (there are
several disreputable companies that transship US software for sale in
Europe and Australia to avoid the international surcharges), he's
probably SOL.

    Second is to determine what character set(s) his user interface
supports.  On the console, this is generally the PC "extended" ASCII
character set.  It may be impossible to do what he wants with this set,
since it generally supplies only lowercase "international" characters.
In addition, unless the file system supports storage in directory
entries (and, in particular, the UNIX utilities support the idea) of
8 bit names, he will have to do input/output translation of the characters
so that they may be stored as 7 bit values.  This is generally done by
making the scan codes generated by the keys report a 7 bit value, thus
replacing American ASCII with a local version.  This is what DEC does
with what they call NRCS (National Replacement Character Set) in a
VT200/VT300 series terminal.  The old way to do this in UNIX is with
"mapchan".  The new way is to modify the keyboard mapping.  Again, SCO
does this best, from what I've seen, although I am sure that if Scott
owns an appropriately "internationalized" version of UNIX, some support
for access to the scan code table and console driver is provided that
will do this.  Again, he will need to contact his vendor.

    The "setlocale()" and "chrtbl()" functions are part of the XPG3
standard for internationalization of *applications* in a nominally non
international environment.  Since Scott appears to have posted from
Germany, the Bundespost would be the appropriate place to get these
standards.  Dell, among other SVR4 vendors, supports XPG3; however,
there are still significant bugs in the 4.0.2 UNIX release of these
libraries.

    For X windows, the "iso1.16" or "latin1" font you refer to is
a 16 point font from the ISO 8859 standard called "Latin 1".  If you are
familiar with the DEC Multinational set, then it can be best described as
the DEC Multinationl set with the column 3 characters rearranged, plus
support for the "eth" and "thorn" characters used in Icelandic.  The
promary benefit of this is the ability to represent most non-Slavic
European langues in their entirety.  The drawback is that there is still
the requirement that your UNIX be 8-bit clean.

    If Scott was talking about using a VT200/VT300 series terminal,
he has it made.  If his UNIX isn't 8-bit clean, he can use the NRCS and
an appropriate keyboard.  If it is, he can use the DEC Multinational set,
or he can even turn on the ISO Latin-1 set (through escape sequences).


    The advantage of using the NRCS in these circumstances is that he
can apply the techniques to supply the same representation as the NRCS to
the console, and thus the file names will look the same on his terminals
as they do on the console.  X everywhere (when using the same fonts)
automatically has the advantage of being the same representation everywhere.

    Another possibility is the use of an X font which is ASCII except
for the NRCS characters within it.  This has the advantage of being usable
with an internationalized keyboard when "xmodmap".  If Scott wishes to
do "international computing" in an X environment, he shoud contact his X
terminal vendor.  NCD, probably the foremost manufacturer of X terminals,
located in San Jose, California, USA, has done significant work in
providing an internationalized product (keyborads, etc.).

    It's unlikely that if mixed I/O devices (ie: X on terminals and
not on the console, or different vendors terminals, etc.) are used
that you will be able to reconcile character set differences without a
great deal of effort.

    The real answer for internationalization is a combination of XPG3,
for processing of currency values (your complaint), and Unicode (draft
standard available through ANSI?), a 16 bit character set standard which
contains most symbols humans use for writing.  The wonderful thing about
a 16 bit Unicode set (aside from the fact that X supports it and DOS and
current DOS ROM character sets do not, and it will thus probably kill
them -- I can hope, can't I? 8-)) Is that not only does it solve all the
problems of internationalization for a *long* time, it allows memory
manufacturers to sell twice as much memory to character terminal vendors
and halves the size of hard drives so that more and bigger ones can be
sold!  I know, I know, no soloution is perfect, but this one is mighty
expensive to stomach, given that most computer work is done in 8-bit
ASCII anyway... someone has to pay the piper, I guess.

    Anyway, I hope this helps make things a bit clearer and gives you
some avenues of exploration to solve your problem.


                    Terry Lambert
                    [e-mail]
                    [e-mail]
---
Any opinions in this posting are my own and not those of my present
or previous employers.
--
 -------------------------------------------------------------------------------
                                                       [e-mail]
 "I have an 8 user poetic license" - me
 -------------------------------------------------------------------------------

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

From: Dean Carpenter <[e-mail]>
Subject: Problem building new ISC 3.0 kernel
Date: 4 Sep 92 19:52:41 GMT
Followup-To: comp.unix.questions
To:       [e-mail]


  We just installed ISC 3.0 on a generic 386DX machine here, and we
already have it running fine on a couple of other (different) 386s and
486s.  This new install won't let us rebuild a kernel though ...

  The system is pretty standard, 386DX, SVGA, WD 1007-SE2 esdi.  Almost
immediately after starting the build from with sysadm, it halts with the
following errors :

    LINE: swap    dsk0t    2
    ERROR: Unknown device 'dsk0t'
    LINE: dump    dsk0t    2
    ERROR: Unknown device 'dsk0t'
    LINE: root    dsk0t    1
    ERROR: Unknown device 'dsk0t'
    LINE: pipe    dsk0t    1
    ERROR: Unknown device 'dsk0t'
    ERROR: Missing required device 'root'
    ERROR: Missing required device 'swap'
    ERROR: Missing required device 'pipe'
    ERROR: Missing required device 'dump'
    FATAL ERROR: Errors encountered. Configuration temrinated.

  This above here shows up in a window offering to let you escape.  The
install went no differently from the others, and the default controller
is the athd on irq 14, i/o 1f0.  There doesn't appear to be a dsk0t
device on the other machines, though they all are scsi, not esdi.

  Anyone have any ideas ?  We're stumped.

--
Dean Carpenter        uunet!areyes!deano        [phone]
Areyes, Inc.        [e-mail]

"No matter where you go, there you are"  sayeth Buckaroo                                  across the Eighth Dimension

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

From: "Conor P. Cahill" <[e-mail]>
Subject: Re: Problem building new ISC 3.0 kernel
Date: 5 Sep 92 12:31:22 GMT
To:       [e-mail]

[e-mail] (Dean Carpenter) writes:

>  This above here shows up in a window offering to let you escape.  The
>install went no differently from the others, and the default controller
>is the athd on irq 14, i/o 1f0.  There doesn't appear to be a dsk0t
>device on the other machines, though they all are scsi, not esdi.

Even when you use a default athd, you need to add a disk on the next
sysadm screen after configuring the controller (select BASIC and then
use the space key to say that the first device is a disk).

This should solve your problem.

Conor.
--
*** SENTINEL(tm) The ultimate Debugging Environment - email for more info ***

Conor P. Cahill              [phone]            [e-mail]
Virtual Technologies, Inc.  46030 Manekin Plaza          Dulles, VA 21066

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

From: Dean Carpenter <[e-mail]>
Subject: Re: Problem building new ISC 3.0 kernel
Date: 5 Sep 92 19:07:35 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Dean Carpenter) writes:
>  We just installed ISC 3.0 on a generic 386DX machine here, and we
>already have it running fine on a couple of other (different) 386s and
>486s.  This new install won't let us rebuild a kernel though ...
>
>  The system is pretty standard, 386DX, SVGA, WD 1007-SE2 esdi.  Almost
>immediately after starting the build from with sysadm, it halts with the
>following errors :
>
>    LINE: swap    dsk0t    2
>    ERROR: Unknown device 'dsk0t'
        :    :
  Got it.  We had to actually add a disk in the BASIC screen that follows
when configuring the boot controller.  We had gone to advanced and simply
exited.  Thanks to Conor P. Cahill for the pointer !

--
Dean Carpenter        uunet!areyes!deano        [phone]
Areyes, Inc.        [e-mail]

"No matter where you go, there you are"  sayeth Buckaroo                                  across the Eighth Dimension

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

From: Robert Garvey <[e-mail]>
Subject: Re: What does \!* mean in csh history subst ?
Date: 5 Sep 92 00:10:13 GMT
Sender: [e-mail]
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Zhuo Zang[~{j0WA~}]) writes:
|> I am confused by the meaning of the
|> csh history substitution \!* .

Ah, csh history is inherently confusing to many.

|> the man says it means: take word 1-$ from previou cmd.
|> my question is: if the previous cmd have pipes, like
|> > cat foo.1 | lpr
|> what is the 1-$ ? if you use:
|> >more !*
|> ?: 1-$ is foo.1 or foo.1 |lpr

You might want to have everything following a csh alias as parameters to
the first command in a pipe, like:
    alias mtoday 'from \!* | /usr/bin/fgrep "`date | colrm 11`"'

The words of a csh command shell include pipes, semicolons, and
redirect symbols.  If you would want to more the file in your example
command line, use:
    more !:1
--
Robert Garvey [e-mail] {sun,lll-tis,pyramid,pacbell}!sybase!robert
          Sybase, Inc   6475 Christie Ave   Emeryville, CA   94608-1010

If Sybase were to pay employees for their opinions, they couldn't afford me.

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

From: Martin McCormick <[e-mail]>
Subject: Re: Shell Scripts vs. C programs
Keywords: shell script, C
Date: 5 Sep 92 02:25:23 GMT
Sender: USENET News System <[e-mail]>
To:       [e-mail]


I have really been enjoying this thread.  There are some
really interesting philosophical points raised here that have
crossed my mind many times.  My background is not Computer
Science, but is a mixture of Journalism and Broadcasting plus
electronics.  I got into computers because they have such
potential to emulate discrete logic circuitry without as many
of the hassles one finds when trying to build logic-oriented
hardware.  Having been blind all my life, I also see the
potential of computers and programmable controllers for
providing a way to design all kinds of nifty devices that
solve various logistical problems associated with functioning
in a world where information and instrumentation are so
important.
     I am also lazy.  On PC's, when I want to do something in
a hurry, I try it in Basic.  If it works, I am apt to leave
it that way.  If it runs too slowly, I might be inclined to
try Pascal or C since a good compiler will assemble code that
runs much faster than the Basic interpreter can ever hope to
run.  In unix, I feel the same way about shell scripts.  They
are quick to write but slow to run.  Most of the time, who
cares?  I know I must continue to learn C because of its
sophistication and because there are C compilers which
generate machine code for programmable controllers like the
Motorola 68hc11.
     I am just glad to know that there are others who like to
keep it simple, if possible.

Martin McCormick WB5AGZ   Stillwater, OK
O.S.U. Computer Center Data Communications Group

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

From: Mike Stefanik <[e-mail]>
Subject: Re: Shell Scripts vs. C programs
Keywords: shell script, C
Date: 5 Sep 92 22:38:29 GMT
Followup-To: comp.unix.shell
To:       [e-mail]

In an article, [e-mail] (Stephen Harris) writes:
>[...]
>Heck!  On this machine the C compiler would barf on its own include files!
>#include <machine.h>
>That one line program through up tonnes of cpp errors.  Blugh.

What, pray tell, are you talking about?  Of *course* this is going to
choke up blood ... for one, a machine dependent include file probably
requires that you include <sys/types.h> among others.  What sort of
errors was cpp returning?

Regardless, dismissing C in general because one vendor's compiler may (or
may not) have broken header files is poor logic indeed.
--
Mike Stefanik  [e-mail]  ...!uunet!pacsoft!mike
Pacific Software Group, Riverside, CA

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

From: Ramesh Balasubramanian <[e-mail]>
Subject: How to store data files in binary format
Date: 5 Sep 92 03:10:51 GMT
Sender: [e-mail]
To:       [e-mail]


We have data files (ascii now) which are huge. I'd like to know how
one goes about storing these in binary (or some other compressed)
format and have programs read these files to build the necessary
data structures?

Thanks in advance,

 -Ramesh ([e-mail])

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

From: Bill Miller <[e-mail]>
Subject: Re: How to store data files in binary format
Date: 7 Sep 92 05:43:52 GMT
Followup-To: comp.unix.questions
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Ramesh Balasubramanian) writes:
>
> We have data files (ascii now) which are huge. I'd like to know how
> one goes about storing these in binary (or some other compressed)
> format and have programs read these files to build the necessary
> data structures?

Depends on the format.  If large portions of the file are numeric, you
might be able to write a program to read in the file, storing each field
in a separate part of a structure variable, but when you write the
file out, write it in binary mode, using fwrite() or write(), not
fprintf().  The numeric stuff will be smaller in the binary file.  For
example, it would take up to 10 digits to describe an int in ascii, but
only 4 for a binary integer.

Your textual parts of the file won't be affected.  The other way to
save space would be to use compress - but that might cause some problems
when you go try and read it.  (Can 'awk' read compressed files - or
can you pipe the compressed file, and pipe it to awk?)

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

From: Nickolay Saukh <[e-mail]>
Subject: Question: Stallion Technologies email address
Date: 5 Sep 92 07:53:35 GMT
Sender: Usenet News Administrator <[e-mail]>
X-Newsreader: Tin 1.1 PL5
To:       [e-mail]

Do they still have it? If so, what is it?

Thanks
--
Nickolay Saukh

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

From: Morten Bonde <[e-mail]>
Subject: Have QIC-24(?) tape-drive - Need info on jumper settings
Date: 5 Sep 92 08:59:23 GMT
To:       [e-mail]

One of my friends gave me a tape-streamer and now I'm trying to make it
work on my 386-based UNIX-box.

My problem is this: There is no documentation. All I've got is the streamer
(and my friend claims that it works) and the adapter-board (ISA-bus).

On the adapter-board there is three different jumpers:

 ----------------------------------------------------------
:                                                     :
:                                                     :
:                                                     :
:                                                     :
:            9     4                                  :
:            A     A                   J8 J10         :
:            :::::::         IRQ      ::: :::         :
:              ADDR      ... ... ...    DMA           :
 -----------------------  2 3 4 5 6 7          --------:
                      :                       :
                      :::::::::::::::::::::::::

I've figured out the IRQ-jumbers (I'me not THAT stupid :-), and I think
the DMA-jumper is pre-set to DMA channel 1 (But I don't know for shure).

My problem is the ADDR-jumpers: I would like to change the I/O address range
to 0x300 - 0x301 (the default for my driver on my UNIX-box), but I have
no idea how to set the two ADDR-jumpers.

The adapter-board is labeled:

    SMS MODEL SIGEN 4044-___  ASSY 000 _____  REV B


The tape-streamer is a TDC 3319 (from Tandberg Data in Norway)

I don't know if this is enough information for you out there, but if
you could help I would really apreciate it.

Please E-mail because I don't have time to read news at work :-(

Thanks,
Morten
--
Morten Bonde                 Voice:  Int. [phone]
Dansk Data Elektronik A/S    Fax:    Int. [phone]
Central Government           Email:  [e-mail]
DK-2730 Herlev, Denmark      uucp:   ...!uunet!mcsun!sunic!dkuug!dde!mbo

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

From: Brian Fitzgerald <[e-mail]>
Subject: Re: Does free() really free?
Date: 5 Sep 92 19:48:55 GMT
Nntp-Posting-Host: mml0.meche.rpi.edu
To:       [e-mail]

Simon Marshall writes:
>    So, please could anyone tell me what is going on here?  Does free()
>    actually free, and is the SZ column of ps reliable in any way?  Or is

malloc() uses the sbrk() system call to get more space when it needs
it.  free() performs no system calls, and merely frees the space for
subsequent reuse by malloc.

I base my comments on perusal of the malloc sources in wuarchive.wustl.edu:
/archive/systems/unix/4.3bsd-reno/lib/libc/stdlib

Brian

void
free(cp)
    void *cp;
{
      register int size;
    register union overhead *op;

      if (cp == NULL)
          return;
    op = (union overhead *)((caddr_t)cp - sizeof (union overhead));
#ifdef DEBUG
      ASSERT(op->ov_magic == MAGIC);        /* make sure it was in use */
#else
    if (op->ov_magic != MAGIC)
        return;                /* sanity */
#endif
#ifdef RCHECK
      ASSERT(op->ov_rmagic == RMAGIC);
    ASSERT(*(u_short *)((caddr_t)(op + 1) + op->ov_size) == RMAGIC);
#endif
      size = op->ov_index;
      ASSERT(size < NBUCKETS);
    op->ov_next = nextf[size];    /* also clobbers ov_magic */
      nextf[size] = op;
#ifdef MSTATS
      nmalloc[size]--;
#endif
}

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

From: Mitch Wagner <[e-mail]>
Subject: net.views -- which vendor is the most open?
Date: 6 Sep 92 03:32:04 GMT
To:       [e-mail]


       Which computer and/or operating-vendor has the greatest
       commitment to open systems? Which has the least? Why?


     This question is being posted to gather responses for a regular
 opinion column in OPEN SYSTEMS TODAY called "net.views."

     By posting a response to this topic, you are granting
 permission for OPEN SYSTEMS TODAY us to publish your response. A
 compilation of the responses to this post will be posted here about
 two weeks from today.

     Please include in your response your name, your employer or
 university, your job title or class standing (that is to say, if
 you identify yourself as a student, are you a freshman, sophomore,
 junior, senior, graduate student?), and a telephone number where
 you can be reached during the daytime (overseas readers note:
 that's daytime in North America).

     Please be brief; try to limit your responses to two screens
 (24x80) of text, not including headers and other administrivia.

     Thanks!

                         -- mitch w.
--
Mitch Wagner, senior editor, Open Systems Today
2353 Massachusetts Ave. Suite 47, Cambridge, MA 02140
[e-mail]  [phone]  CIS:70212,51  GEnie:MITCH.WAGNER
For subscription information, please call 516/562-5882

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

From: der Mouse <[e-mail]>
Subject: Re: signal when window is destroyed
Date: 6 Sep 92 06:29:43 GMT
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Paul J. Wickman) writes:

> What (if any) signal type is sent when a window is destroyed to the
> program running in it[?]

Depends on what you mean by "the program running in it".  It would also
help if you were more precise about what "destroyed" means, though in
this case I think I can probably guess.  "running in it" is similarly
vague.

> The program is an xterm exec-ing another program

You are probably thinking of "a window" from a user's perspective.
>From that point of view, destroying a window amounts to making a
request of your window manager.  When you do this, no signal in the
technical UNIX sense is sent to the X client responsible for the
window.  Depending on the window manager and various other things, the
client may see the connection go away, or it may receive a request to
make its window disappear.  In neither case, though, does the event
take the form of a UNIX signal.

However, your comment about xterm makes it sound as though "running in
it" here does not refer to the X client, which would be xterm, but
rather to some process running connected to the pseudo-terminal xterm
has set up.  In this case, what *should* happen is that the virtual
terminal loses carrier, which usually generates a SIGHUP.  Whether this
actually happens, and if it does whether your program will see it,
depends on many factors.  Some systems don't seem to close down ptys
correctly to begin with, at least not always.  Even if the pty is shut
down correctly, your program may be running in the background and hence
not get the signal.

                    der Mouse

                [e-mail]

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

From: Gene Kochanowsky <[e-mail]>
Subject: Re: SUMMARY: What Unix HW/SW I will buy with my $8000
Date: 7 Sep 92 00:35:37 GMT
To:       [e-mail]

[e-mail] (Republic of Taiwan) writes:

>Get a Sparc II clone for $4843(16M memory,17"color).
>Add 420M HD $1197.
>Upgrade to 19" Trinitron $1197
>Each Extra 16MB $785, (Add up to 64MB)
>Add Sun GX Accelerator $585

>So a decent SparcII clone with 420MB HD, 64MB Memory will be $8395.

>Or 840MB HD, 32MB Memory will be $8022.

>For more info call 800-mobius-1.

>PS. I have no connection with this mobius.

Does this include the operating system and RTU.
How about the warranty?


--
 ------------------------------------------------------------------------------
Gene Kochanowsky                           | "And remember ....
Associated Electronic Engineers, Inc.      |       The better you look ...
2864 Cercy Trace                           |

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

From: John Navarra <[e-mail]>
Subject: More problems with find (the stderr bug)
Date: 7 Sep 92 03:00:46 GMT
Sender: "Usenet on news.acns" <[e-mail]>
To:       [e-mail]


Here is an interesting little problem with the find command:

I want to remove certain files with my userid in the /tmp directory.
Here is the command I tried:

[casbah:219] ~/tmp -> find /tmp -user navarra -ok /bin/rm {} \; 2>/dev/null

This finds all the files with my userid in /tmp and prompts me in turn
if I want to remove the file. Now, the problem is the redirect to
/dev/null. The rationale for doing this is so I won't see things like:

cannot change directory to blah: Permission denied.

However, when I run this command, it just hangs. At first, I thought it
had something to do with the way I constructed the command. I ruled that
out. Then, I thought it had something to do with my shell (bash) but I
ruled that out by trying other sh-shells (remembering that redirecting
stderr in csh/tcsh is real *fun* -- read stupid). So, then I figured
it had something to do with my OS (SunOS) but no, that wasn't it
either!

Then it was time to look at the man page again:

    -ok command    Like -exec except that the generated  command
                    is  written  on the standard output, then the
                    standard input is read and the  command  exe-
                    cuted only upon response y.

Now read that. It *says* it writes to standard output right?

Well:
[casbah:226] ~ -> find /tmp -user navarra -ok /bin/rm {} \; 2>yeah_right
n
n
n
n
screw you find!
yeah whatever
^C
[casbah:227] ~ -> cat yeah_right
find: cannot chdir to /tmp/screens/S-jeff: Permission denied
find: cannot chdir to /tmp/screens/S-triona: Permission denied
< /bin/rm ... /tmp/screens/S-navarra >?   find: cannot chdir to
/tmp/screens/S-j
find: cannot chdir to /tmp/hpa: Permission denied
find: cannot chdir to /tmp/lunde: Permission denied
find: cannot chdir to /tmp/jmaltzen: Permission denied
find: cannot chdir to /tmp/triona: Permission denied
< /bin/rm ... /tmp/navarra >?   < /bin/rm ... /tmp/navarra/chapters >?   <
/bin/
rm ... /tmp/navarra/nutshell >?   < /bin/rm ... /tmp/navarra/vi >?   <
/bin/rm .

So, they lied! Those bastards!

Well, I know in this circumstance I can use:

find /tmp -user navarra -exec /bin/rm -i 2>/dev/null

but, really, this is not always going to help me and I was wondering
if this problem with find is well known and if anyone has done anything
about it? I tested this on a number of machines (Sun, Next, Sequent
Risc) and they all didn't work.

Ideas, suggestions, comments?
-tms
--
You can get further with a kind word | You can get further with a kind word
and a gun than a kind word alone.    | and a phaser than a kind word and a gun.
          --al capone                |           -- John Navarra
=======From the Lab of the MaD [e-mail]========

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

From: Republic of Taiwan <[e-mail]>
Subject: Re: need script to rename uppercase filenames
Keywords: script,tar,msdos
Date: 7 Sep 92 07:00:22 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Bill Miller) writes:
=Hi, everyone.
=
=I'm fairly new to unix, and I need a script or procedure to do the following:
=
=I have some source code in DOS (many separate files) that I tarred under
=DOS and untarred under 386BSD.  The big problem is that all the files
=came through in UPPERCASE and I need a script to mv (rename) them all
=to lowercase quickly.
=
=Since they're in DOS text format, I realize I also need to strip the
=extra carriage returns on each line.  I've been successful in doing this
=with:
=
=  cat (file) | tr -d '\015' > (newfile)
=
=It would be nice to combine both of these so that I could rename the
=files to uppercase and strip the extra newlines all in one fell swoop
=instead of doing it one file at a time.
=

#!/bin/csh -f
# ^M is contrl M, ^Z is control Z
#
foreach i ($*)
sed -e 's/^M//g' -e 's/^Z//g' $i  > `echo $i | tr A-Z a-z`;
rm $i;
end

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

From: SIMON LAI <[e-mail]>
Subject: SIGIO - when does it occur ?
Date: 7 Sep 92 07:28:07 GMT
Sender: Usenet system <[e-mail]>
To:       [e-mail]


There seems to have been some discussion in this group about signals and
sockets, so hopefully those same people can help me.

I am writing an application that uses SIGIO to service a number of non-
blocking sockets with incoming data.  Everything seems to work fine, though
as the subject line says I need to know exactly when SIGIO occurs.

It doesn't seem to occur until a certain number of characters are in the
socket waiting to be read (buffer full?).  Is this correct?  If yes, then
how many characters cause SIGIO to occur, and can I vary this number?

I've looked at sys/socket.h, and though some of the options seem related
none seem to be exactly what I want.

Any help appreciated.

Simon

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

From: Heikki Karhunen <[e-mail]>
Subject: Re: /usr/mail/glenr.lock not creatable after 10 tries
Date: 7 Sep 92 08:22:37 GMT
Sender: Uutis Ankka <[e-mail]>
To:       [e-mail]

In article <[e-mail]> [e-mail] (Glen Reesor) wrote:
> I'm using the ftpmail server at decwrl.  When I request large files it usually
> takes a few tries to get all parts.  Up until now I just thought that some
> site along the way was dropping them.  But the other day I noticed a number
> of bounced messages going *back* to decwrl, so it appears that it is
> *my* site dropping parts.  My system configuration is:
>
> Network of HP9000/730's running HPUX 8.07.  All machines share a single
> /usr/mail directory using NFS.  The machine that /usr/mail actually resides
> on is our UUCP machine.  My account is not on the UUCP machine, thus the mail
> will be routed from the UUCP machine to my machine using SMTP.
>
> I've configured sendmail to mail a copy of all bounced headers to postmaster,
> and this is what I get:
>
> --------------------------------------cut--------------------------------------
> From MAILER-DAEMON@onyx Sat Sep  5 21:27 EDT 1992
> Received: from ruby by onyx with SMTP
>         (16.8/16.2) id AB20350; Sat, 5 Sep 92 21:26:49 -0400
> Date: Sat, 5 Sep 92 13:32:58 -0700
> From: Mail Delivery Subsystem <MAILER-DAEMON@onyx>
> Return-Path: <MAILER-DAEMON@onyx>
> Subject: Returned mail: Can't create output
> To: Postmaster@onyx
> Status: RO
>
>    ----- Transcript of session follows -----
> mail: /usr/mail/glenr.lock not creatable after 10 tries
> 550 <glenr@onyx>... Can't create output
>
>   ----- Message header follows -----
> --------------------------------------cut--------------------------------------
>
> onyx is the machine that my account is on and ruby is the UUCP machine.  I've
> RTFM'ed and can't find any reference to "Can't create output".  Since I don't
> understand the in's and out's of mail lock files all I can postulate is that
> more than one piece of mail is trying to be delivered at once.  This
> particular mail bounced on Saturday in the evening, so there was no mail
> other than from decwrl coming to me.
>
> So.....is it a problem with NFS mounting /usr/mail?  Or is sendmail tripping
> over itself because of the volume of mail coming to me all at once?

This sounds very similar to one problem I had this summer. And no, I
didn't find any solution to it, but here is what I discovered.

We have at work several Apollos that co-exist very nicely with each
other. Now, we got a HP 9000/710 on loan from HP (they obviously
wanted us to buy it, but that is another story altogether).

My home directory is in Apollo, but I mainly used the HP, because I
was supposed to try it out and to see that it too co-existed with the
rest of the boat anchors hanging around the net.

As the only way to share HP-UXen and Apollo's filesystem is via NFS,
that is exactly what we did (by the way, we run Domain/OS 10.4 on the
Apollos with NFS 2.3 and the HP-UX was around 8.07).

I configured my own environment so that my home directory (among other
things) was mounted via NFS from the Apollo cluster. I also mounted
HP's /usr/mail to our mail handling Apollo's /usr/spool/mail.

Now begins the fun part. I could use Elm (2.3 PL11, not the one HP
supplies as I use Elm 2.3 in Apollos too and HP's Elm didn't
understand all the elmrc switchery) in HP to read and send my mail.
But.  Whenever I logged in as root (or even a normal user), I had to
wait about 20 seconds after the copyright texts while mail was
checking whether I had any mail. It invariably failed with: 'mail:
can't lock /usr/mail/<username> after 10 tries'.

I digged around this for a while, and it seems that mail (when invoked
with -e or otherwise) creates a lock file, /usr/mail/<username>.lock
in the mail directory. For some reason it couldn't create one when the
/usr/mail was NFS mounted from Apollo.

Actually, it _did_ create something... Namely a file called
/usr/mail/root.lock, the length of which was 0 and all the protection
bits were clear. I tested Apollo's mail (the sys5.3 variety) and its'
lockfile was some bytes long and its' protection was r+w for owner (I
think, not sure about this anymore :) ).

I had more or less synchronized the GIDs and UIDs in both machines (I
didn't do 'em all, the HP was on loan, after all...) and the NFS
mounting was done with default options.

I _know_ that Apollo's NFS 2.3 is not fully functional (lacks lockd
among other things) and HP will ship version 3.0 anytime-real-soon-now,
but by the previous problem the glitch might be in HP's NFS. Comments?

>
> If I can get this worked out, FTP mail may actually be semi-painless!  Any
> suggestions or solutions would be greatly appreciated.
>
> Please e-mail.  I will summarize if there is interest.

Hope someone can unravel this... I'd sure like to know how...

>
> --
> Glen Reesor                       |Internet Style: aruba![e-mail]
> Systems Administrator, Project Zed|UUCP     : ...!uunet!uu2.psi.com!aruba!glenr

    Heikki Karhunen

/-----------------------------+-----------------------------------------\
| [e-mail]  | There is always a job for a theoretical |
| [e-mail] | physicist -- at least in theory.        |
\-----------------------------+-----------------------------------------/

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

From: Stephen Wynne <[e-mail]>
Subject: Berkely Fast File System
Date: 7 Sep 92 08:53:36 GMT
Sender: [e-mail]
X-Newsreader: Tin 1.1 PL5
To:       [e-mail]

I'm having difficulties understanding the Berkeley Fast File System.
I can see how fragmentation would be a problem under the old method,
but I'm getting lost in the details of cylinder groups, bitmaps, etc...
Would anyone care to start a discussion on this topic with me?

Thanks,

STEVE

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

From: Simon Tyrrell <[e-mail]>
Subject: DRS/NX documentation anomaly
Date: 7 Sep 92 09:01:07 GMT
NNTP-Posting-Host: daisy.ee.und.ac.za
To:       [e-mail]

We've recently acquired a DRS3000 running DRS/NX and I'm trying to
implement mail as per the instructions in Appendix E of the System
Administrator's Guide.  Quoting from page E-8 (no, that's not a smiley)

>    2.  The list of machines that will accept SMTP mail is specified
>        by the netdird service.  See netdird(1M) in the System
>        Administrator's Reference Manual to see how to add services
>        to this database.

There is no such netdird man page in my SAR manual :-(  Have other
DRS/NX users noticed this problem?  We want to send mail over the
Internet via TCP/IP.  Unless I've missed something, the SAR manual
seems to have failed me, and I'm now stuck.

How does one get over this step?  Any help will be greatly appreciated.

Simon Tyrrell
 -------------
Academic Computer Services
University of Durban-Westville

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

From: Staale Deraas <[e-mail]>
Subject: Re: Problem with "umount" on SCO-ODT (UNIX v. 5.3.2)
Date: 7 Sep 92 12:59:32 GMT
Sender: NetNews Administrator <[e-mail]>
To:       [e-mail]


On SCO 3.2 you have a command that is called "fuser", that will tell you who are using
the file system.

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

From: Geert Lindemulder <[e-mail]>
Subject: termcap entry for a Tandem TS530 Series Terminal
Date: 7 Sep 92 13:14:02 GMT
Sender: Geert Lindemulder <[e-mail]>
To:       [e-mail]

Dear network readers ,

This morning we received a Tandem TS530 Series Terminal.
We tried to connect this to our gould powernode pn6040 with
a utx 2.1b operating system.
Unfortunately our system does not have an entry, for this kind
of terminal in the termcap file.
Did anyone of you ever had to connect one of these terminal to his
unix system.
I am very interested in any number of termcap lines , before i
start searching in the terminal manual to edit a termcap entry myself.

regards , Geert Lindemulder
Encore Computer Nederland B.V.
Maarssenbroek , Nederland
Email : [e-mail]

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

From: Eduardo Rodriguez <[e-mail]>
Subject: Playing CD in workstation A, hearing it in B, Can I?
Date: 7 Sep 92 13:37:45 GMT
Sender: Network News <[e-mail]>
Originator: erodrigu@tolten
To:       [e-mail]


I want to play a CD in a device connected to some workstation A,
but i want to hear it in other workstation B.
  Is there an utility that do that?. Both workstation are Sun Sparc
connected via Ethernet.

Thanks for any hint!

edo...

Eduardo Rodriguez S        - --- - --- - --- -    [e-mail]
Dpto de Ciencias de la Computacion  - --- -    ...!uunet!uchdcc!erodrigu
Universidad de Chile   --- - --- - ---   [e-mail]
- --- - --- - --- - --- - --- - --- - --- - --- - --- - --- - --- - ---

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

From: Eduardo Rodriguez <[e-mail]>
Subject: Re: Ytalk source code...
Keywords: ytalk
Date: 7 Sep 92 15:11:05 GMT
Sender: Network News <[e-mail]>
Originator: erodrigu@tolten
To:       [e-mail]


In article <[e-mail]>, [e-mail] (amy) writes:
> I'm looking for the source code for the most recent version of
> ytalk. if anyone knows where i can get it, let me know.
>

>From archie:

Host uhunix2.uhcc.hawaii.edu

    Location: /incoming/blaine
           FILE -rw-r--r--      65536  Mar 30 19:48  ytalk
    Location: /incoming/unix
           FILE -rw-r--r--      33223  Jul 26 07:42  ytalk-2.0.tar.Z

Host bode.ee.ualberta.ca

    Location: /pub/newstuff
           FILE -rw-r--r--      27215  May 27 15:23  ytalk.tar.Z

edo...

Eduardo Rodriguez S        - --- - --- - --- -    [e-mail]
Dpto de Ciencias de la Computacion  - --- -    ...!uunet!uchdcc!erodrigu
Universidad de Chile   --- - --- - ---   [e-mail]
- --- - --- - --- - --- - --- - --- - --- - --- - --- - --- - --- - ---

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


End of INFO-UNIX Digest
***********************