NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/unix/question/13774

Internet Message Format | 1992-11-22 | 68.4 KB

Path: sparky!uunet!europa.asd.contel.com!gatech!psuvax1!rutgers!cmcl2!adm!news
From: [e-mail] (SMTP MAILER)
Newsgroups: comp.unix.questions
Subject: Mail not delivered yet, still trying
Message-ID: <[e-mail]>
Date: 22 Nov 92 17:58:09 GMT
Sender: [e-mail]
Lines: 1863


 ----Mail status follows----
Have been unable to send your mail to <[e-mail]>,
will keep trying for a total of three days.
At that time your mail will be returned.

 ----Transcript of message follows----
Date: 22 Nov 92 04:19:00 EST
From: [e-mail]
Subject: INFO-UNIX Digest  V17#008
To: "rubel" <[e-mail]>
cc: "slosser" <[e-mail]>

Return-Path: <[e-mail]>
Received: from SEM.BRL.MIL by ntsc-rd.navy.mil with SMTP ;
          Sun, 22 Nov 92 04:09:13 EST
Received: from SEM.BRL.MIL by SEM.BRL.MIL id aa25264; 21 Nov 92 7:23 EST
Received: from sem.brl.mil by SEM.BRL.MIL id aa25030; 21 Nov 92 7:05 EST
Date:       Sat, 21 Nov 92 07:04:42 EST
From:       The Moderator (Mike Muuss) <[e-mail]>
To:         [e-mail]
Reply-To:   [e-mail]
Subject:    INFO-UNIX Digest  V17#008
Message-ID:  <[e-mail]>

INFO-UNIX Digest          Sat, 21 Nov 1992              V17#008

Today's Topics:
                             UNIX Training
              where no vi has gone before...? (CORRECTION)
  Tutorials - nine topics (12/7, San Jose)  Sun User Group Conference
                          tcsh-like subroutine
                           Re: IS UNIX DEAD?
                        Re: IS UNIX DEAD (long)
                      Adoptation of Unix in Europe
             Re: Looking for Time Series Analysis software
             Re: What full-screen file managers are there?
                     Re: SCO TCPIP >9 LOGINS AGAIN
                         Unlocking named pipes
                                Re: grep
                              FSCK command
                       pwd: getwd: can't open ..?
             Re: What full-screen file managers are there?
                lib77 "flush" call. Where is it in UNIX?
                           Re: Wait for child
                      Re: name of file descriptor
                                Re: talk
                  Re: Changing the owner of a process
                            file .fingerees
         Re: Whence Unix? (was Re: IS UNIX DEAD?) (New Thread?)
                       Curses support under DOS?
                            MPE, NS, TCP/IP
              Re: lib77 "flush" call. Where is it in UNIX?
                    Re: Searching for E-mail package
                        C++ Debugger with Motif
             Re: Looking for Time Series Analysis software
                          Re: file .fingerees
            Re: BIG IMPORTANT QUESTION PLEEEEZE HELP ME!!!!!
                Shortened repost: mmap over NFS problem
       Sun Printserver? IPX-stack available ? NCP documentation ?
                Tool for email, and snail mail addresses
                  Re: Changing the owner of a process
                       I need some pic-like thing
                 Re: C Source Code for Pattern Matching
                        Re: IS UNIX DEAD? (long)
            Re: BIG IMPORTANT QUESTION PLEEEEZE HELP ME!!!!!
                        Re: IS UNIX DEAD? (long)
                        Re: tcsh-like subroutine
                  Re: Changing the owner of a process
                        Re: IS UNIX DEAD (long)
                rexec does not search for ~/.netrc file
                           Re: IS UNIX DEAD?
                     Re: IS UNIX DEAD? (very long)
                   Re: Overflow warnings (SCO SV3.2)
                 CURSES - No delay option, SunOS 4.1.1
              Re: lib77 "flush" call. Where is it in UNIX?
                        Re: IS UNIX DEAD? (long)
             Re: What full-screen file managers are there?
-----------------------------------------------------------------

From: "Mr. Themos Pentakalos; ADMIN-COMP" <[e-mail]>
Subject: UNIX Training
Date: 17 Nov 92 02:28:08 GMT
Sender: News posting account <[e-mail]>
To:       [e-mail]

Hi everyone,

I am not very familiar with the UNIX market and would like to ask
a couple of questions.

1) are there any companies which teach UNIX topics such as system
   management, networking, etc? addresses? prices?

2) what are the most popular magazines in the UNIX area?

    Thanks a lot. Please reply directly to:   [e-mail]
and i'll be glad to summarize if comments are good.

Themos.

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

From: Vinay Kashyap <[e-mail]>
Subject: where no vi has gone before...? (CORRECTION)
Date: 17 Nov 92 02:35:07 GMT
Sender: News System <[e-mail]>
To:       [e-mail]

I had posted a list of questions on vi, and Tuomas Lukka pointed out
that question 3 didn't make sense:

>3. Is there a way to prepend or append characters to macros without
>   (a) introducing newlines between the additions, &/or
>   (b) writing to the file and deleting to a named buffer
It should have read as follows (I apologize for the error):

3. Is there a way to prepend or append characters to BUFFERS without
   (a) introducing newlines between additions (yanking to an uppercase
   buffer appends to text already in the buffer, but on a new line) &/or
   (b) writing to the file and deleting to the named buffer (in other
   words, are there ways to put characters in named buffers other than
   yanking/deleting text)?

Vinay ([e-mail])

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

From: Nancy Frishberg <[e-mail]>
Subject: Tutorials - nine topics (12/7, San Jose)  Sun User Group Conference
Keywords: tutorials seminar Unix security programming tools network debugging
Date: 17 Nov 92 00:19:49 GMT
Sender: Mr USENET himself <[e-mail]>
Nntp-Posting-Host: sug.org
To:       [e-mail]


If you're concerned about advanced Unix security, Posix-conforming
systems, or moving from Unix programming to Unix system
administration, plan to be at the Sun User Group Conference (San Jose
Convention Center) on Monday, December 7, 1992 for one of the full day
tutorials.  The Conference and Exhibition extends through Thursday.

All-day tutorials will focus on these topics and others of interest to
novice and experienced Unix users as well as Unix programmers who want
to expand their theoretical and practical knowledge.

Special offer: 5 full conference registrations (each includes a day of
tutorial) for the price of 4 when preregistering with a single payment.

Full-day tutorials currently scheduled include:
- Advanced Unix Security (Matt Bishop, Dartmouth College)
- Preparing for Disaster (a.m. - Brent Chapman, Great Circle Associates),
  plus, Why have Computer Security? (p.m. Bob Baldwin, Tandem Computers)
- Sun Network Debugging (Smoot Carl-Mitchell, Texas Internet Consulting)
- Topics in Perl (Tom Christiansen, Convex Computer Corporation)
- Programming in POSIX (Jeffrey S. Haemer, Canary Software)
- UNIX Programming Tools (Kenneth Ingham, consultant)
- The Internet and its Protocols (William LeFebvre, Northwestern University)
- Introduction to UNIX System Administration (Dinah McNutt, Tivoli Systems)
- Integrating C Code and Xt Widgets (Craig Rudlin, MD, Medical Software and Computer Systems)

If you just want to go to the exhibits, ask for a free show-only pass.

To get more information by email about these tutorials, the technical
program, or exhibits at the Sun User Group conference, send requests
to [e-mail] .

You will receive the full tutorials and program description with
registration information.  Or call 1-800/727-EXPO.  (Outside the U.S.,
use 512/331-7761 (voice) or 512/331-3950 (FAX).)

--
Nancy Frishberg, Sun User Group.

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

From: Karl Glazebrook <[e-mail]>
Subject: tcsh-like subroutine
Date: 16 Nov 92 16:07:56 GMT
Nntp-Posting-Host: dur.dust6
To:       [e-mail]

Hello,

I'm looking for a tcsh like input routine which would be callable from a C or
FORTRAN program. I just want to be able to read a line of text from the user
with tcsh-like (or emacs-like!) command line editing/recall features.

I'm working on a SPARC with SUNOS 4.1.3 but obviously the more general and
portable the solution the better!

As a bonus there would also be another routine which could modify the recall
list, but this is not essential as I could probably hack this myself given
the first subroutine.

Thanks for any help,

Karl Glazebrook.

---
 ------------------------------------------------------------------------------
| Karl Glazebrook,      | INTERNET:   [e-mail]           |
| Dept. of Physics,     |    JANET:   [e-mail]           |
| Univ. of Durham,      |     SPAN:   19463::DUVAD::KGB                      |
| United Kingdom.       |      FAX:   [phone]                           |
 ------------------------------------------------------------------------------

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

From: Leslie Mikesell <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 16 Nov 92 18:17:47 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Martin McCormick) writes:

>First, I completely agree with Jeff's enthusiasm for this system.  It is good
>and probably is the wave of the future.  I do have one gripe, however.
>As I have said, in other postings, Postscript is a half-finished job until
>somebody comes up with a method for converting it back into standard ASCII
>text like the lines you are reading, now.

I think GNU Ghostscript has an ASCII output option that will generate
text positioned approximately where it is supposed to be on a page
with anything else omitted.

> The windows-oriented approach is
>totally useless for those of us who are blind and use speech synthesizers.
>We were able to cobble together a small C program to strip out all that
>Postscript language and display a completely deformatted stream of data.
>It beats nothing, though not by much.  I would love to help fix this problem,
>but I need to find a document explaining Postscript in electronic form so
>I can read it and figure out how to tackle this beast.

It's non-trivial in the general case because you have to be able to handle
the entire postscript language to interpret the positioning and scaling
effects.  Plus you'll miss things where fonts have been converted to
curves and manipulated for artistic but still readable effects.  A
quick and dirty approach is to just extract the strings in parens
from the postscript and try to assemble them into text.  A more ambitious
project would be to modify the Ghostscript ascii output to maintain
some font size and type info that could be used by the speech synthesizer
to vary the output effects.

Les Mikesell
  [e-mail]

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

From: Andy Newman <[e-mail]>
Subject: Re: IS UNIX DEAD (long)
Date: 16 Nov 92 23:44:31 GMT
Sender: [e-mail]
To:       [e-mail]

[e-mail] (Paul Prescod) writes:
>
>True, the problem is there are people, in this very newsgroup, who see
>no reason to try to make user friendly applications for unix.  To them,
>if you can't use VI right off the bat, or enjoy learning obscure,
>nonsensical, illogical keystrokes, you should go back to the mac.
>THIS will kill Unix.
>

Define "user friendly". Aren't the people who like vi users too?

--
Andy Newman ([e-mail])

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

From: Nick Gavrielatos <[e-mail]>
Subject: Adoptation of Unix in Europe
Keywords: Unix, Europe
Date: 17 Nov 92 04:51:55 GMT
To:       [e-mail]

I am preparing some work on adoptation of Unix in Europe.

WOuld anyone happen to have any old articles, or know
of any other magazine and/or book articles that can
recommend? Any info via personal mail would be greatly
aprreciated.

Thank you much,

Nick Gavrielatos
[e-mail]

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

From: s jonathan silverman <[e-mail]>
Subject: Re: Looking for Time Series Analysis software
Date: 17 Nov 92 05:16:27 GMT
Sender: News System <[e-mail]>
To:       [e-mail]


The leading program for time series analysis is probably RATS, which runs
under DOS, the Macintosh, Unix and perhaps some other operating systems
as well.

The company that writes this is located in Evanston, IL.  Their name used
to be VAR econometrics, but they might well have changed it about a year ago
(perhaps to "Estima", but I could have them confused with another company).

RATS stands for something like "regression analysis for time series" and
contains a broad array of time series analysis functions including specral
analysis.

If you cant track this down yourself, contact me by email and I will send
you the phone number.

--
Jonathan Silverman

[e-mail]
[e-mail]

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

From: Pete Holsberg <[e-mail]>
Subject: Re: What full-screen file managers are there?
Keywords: utree, maint
Date: 16 Nov 92 22:33:40 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Paul Kramer) writes:
=Hello,
=
=I found while reading articles in this electronic conference, a
=reference to a public-domain program called utree.  It is full-screen
=file manager which allows you to perform file functions with single
=keystrokes.  For example, after I start the program I see a screen
=that lists the files on my account.  If I want to delete one of those
=files I move the cursor on top of it, press a certain key.  Boom it
=gone!
=
=Well I am wondering:  Is there anything better than utree?  I am aware
=of a program called 'maint' but it doesn't have the full functionality
=of 'utree'.

I prefer HDSCAN.

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

From: Bill Campbell <[e-mail]>
Subject: Re: SCO TCPIP >9 LOGINS AGAIN
Date: 17 Nov 92 02:00:52 GMT
To:       [e-mail]

In <[e-mail]> [e-mail] (Doug Pavey) writes:

>Well, I read Aris' and Barry's messages about what is involved in
>getting more than 9 tty's to login.  ttyp00 -ttyp08 work fine, ttyp09
>will login, but hang TCP.

Xenix has a max of 32 ptys and SCO UNIX a max around 254.

>How does one debug streams limits problems - determine how many are in
>use, or how many one should set up if the user may be using as many as
>10 processes per login (Progress Database is wonderful).  All users are
>using TCP via 3 8 port Terminal servers. I can get the login messages on
>all ports, until the user actually signs in, we don't seem to have
>problems.  Seems like streams may be the culprit.  How to fix???

You don't say in this post whether you're running Xenix or UNIX
and there are different tools to determine problems with streams
in each.  Xenix comes with a program 'sw', UNIX comes with crash, and
there are some freely distributed monitor programs such as u386mon
available from various ftp sites.

Xenix sw displays:

Pause=1   Calls=74                                Host=camco1

   Resource  Cnt       Use   Total     Max    Fail
     stream: 128        40    4020      53       0
      queue: 256        83    8074     111       0
     mblock:1152        7818508524     636       0
dblk totals:1152        7813966687     543    4469

Mem Size Cnt Med Low   Use   Total     Max    Fail
  1    4 256 230 204     0  893844     155       0
  2   16 128 115 102     2  449445     102     889
  8   32 256 230 204     6 2644776     207    3339
 16   64 256 230 204     5 8602774     147       0
 16  128 128 115 102    50  386387     103      18
 12  256  48  43  38    15  220139      37       0
  4  512   8   7   6     0  102435       7     223
 32 1024  32  28  25     0   50083       8       0
 64 2048  32  28  25     0  584718      13       0
 32 4096   8   7   6     0   32086       5       0
  0 8192   0   0   0     0       0       0       0

Buffers (used/total) = 11/187 Kbytes

This is from a system that has been running for a while now.  Uptime gives:
camco1:ttyp02 /usr/local/bin # uptime
  5:53pm  up 20 days, 55 mins,  12 users,  load average: 1.19, 1.04, 1.02

You would probably want to use crash and it's strstat command on a UNIX
system which gives similar information.  A typical crash display is:
beeson1:ttyp8:/onager/etc > crash
dumpfile = /dev/mem, namelist = /unix, outfile = stdout
> strstat
ITEM                  CONFIG   ALLOC    FREE         TOTAL     MAX    FAIL
streams                  256      158      98           821     171       0
queues                  1280      704     576          2306     786       0
message blocks          2825      309    2516      11357083     705       0
data block totals       2260      309    1950       9675047     669      63
data block size    4     384       51     333         66265     364      63
data block size   16     384        8     376         39861      65       0
data block size   64     512       24     488       4950469      68       0
data block size  128     448      157     290       2974970     392       0
data block size  256     220       33     187        632454     107       0
data block size  512     136       36     100        495575      46       0
data block size 1024      52        0      52         84628      44       0
data block size 2048     104        0     104        430474      25       0
data block size 4096      20        0      20           351      12       0

Count of scheduled queues:   0
>

Bill
--
INTERNET:  [e-mail]   Bill Campbell; Celestial Software
UUCP:   ...!thebes!camco!bill   6641 East Mercer Way
             uunet!camco!bill   Mercer Island, WA 98040; [phone]
SPEED COSTS MONEY -- HOW FAST DO YOU WANT TO GO?

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

From: "frederick.d.true" <[e-mail]>
Subject: Unlocking named pipes
Keywords: named pipe
Date: 17 Nov 92 06:11:16 GMT
Followup-To: [e-mail]
To:       [e-mail]

I'm having a bit of trouble figuring out what's wrong with the
following usage of named pipes. I have 12 large compressed files which
I would like to merge/sort (they are sorted) without having to
decompress all of them before merging. So, I've written a simple script
to create 12 named pipes with mknod, say pipe1..pipe12. I begin
background processes to zcat each of the 12 files to one of the pipes
in sequence. I then start 'sort <switches> -m pipe1 .. pipe12 | <stuff> >
outputfile'.

Things go fine for a while, then the entire process freezes. The
output stream stops, and all processes go into IW wait, as if waiting
for something to happen. While in deadlock, if I try to read any of
the pipes, they each have data waiting, so all seems to be fine. So
who stopped everything and why?

Sort is able to merge at most 16 files at once, so I don't think
there's anything wrong there. Is there another anomoly in sort that
might cause this failure?

As best I can tell, sort found a depleted pipe that zcat couldn't keep
up with (perhaps sort is reading large buffers?), but no EOF, and
started sleeping. However, this isn't logical since the sort process
is followed by 3 other heavyweight processes in the pipe before the
output is written, so I would think that zcat could feed the pipes
fast enough. Besides, I thought processes would handle FIFO's normally
and sleep for at most a few seconds if it found and empty one before
hitting EOF.

By the way, in case it proves significant: I'm using SunOS 4.1.2 on a
SPARC. Standard sort, zcat, etc. Named pipes are created with mknod
-p. I've also tried sleeping for a few seconds before starting things,
with no results.

Can someone clear up the confusion, or at least add postulations to
the heap?

Please e-mail to [e-mail] or post here.

--
Fred True
AT&T

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

From: "David L. Parker" <[e-mail]>
Subject: Re: grep
Date: 17 Nov 92 03:36:38 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Michael Salmon) writes:
>In article <[e-mail]>,
>[e-mail] (David L. Parker) writes:
>|>
>|> Actually it's Global Regular Expression Parser.
>
>Unfortunately Brian Kernigan and Rob Pike disagree with you, see
>page 18 of:
>
>The Unix Programming Environment
>Prentice Hall
>ISBN 0-13-937699-2
>

WOOPS!  Okay, so I got three out of four.
--
Dave Parker
Automated Data Management Services, Pleasant Hill, MO 64080-1331
[phone]/5218 voice/fax          -          dlpinc00!dlparker

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

From: [e-mail]
Subject: FSCK command
Date: 17 Nov 92 04:50:34 GMT
To:       [e-mail]

When using the fsck (file system check), when it comes upon part 4 of
the check, which is a reference check, it says, "UNREF DEVICE".  What
does this mean?

Thanks.. please send all replies via mail..

suicide
[e-mail]

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

From: Jonas Furrer <[e-mail]>
Subject: pwd: getwd: can't open ..?
Date: 16 Nov 92 11:10:22 GMT
To:       [e-mail]

Hi,

why I have this error -> pwd: getwd: can't open ..


OS          : OS/MP 4.1A.1 (Solbourne -> SUN 4.1)
fstab entry : /dev/sd2d /files/disk2/imagery  4.2 rw 1 9

belfast:/% df /dev/sd2d
Filesystem            kbytes    used   avail capacity  Mounted on
/dev/sd2d             564349  311204  196710    61%    /files/disk2/imagery

belfast:/% ls -agld files
drwxrwxrwx  6 root          512 Oct 21 15:48 files

belfast:/% cd /files
belfast:/files% ls -agld files
drwxrwxrwx  4 root          512 Oct  2 14:07 disk2
belfast:/files% pwd
/files

belfast:/files% cd disk2
belfast:/files/disk2% ls -agld files
drwxrwxrwx  5 root          512 Nov 13 11:16 imagery
belfast:/files/disk2% pwd
/files/disk2

belfast:/files/disk2% cd imagery
belfast:/files/disk2/imagery% pwd
pwd: getwd: can't open ..
belfast:/files/disk2/imagery% df .
Filesystem            kbytes    used   avail capacity  Mounted on
Could not find mount point for .

Can anyone help?

Thanks.

          +---------------------+---------------------------+
          |Jonas Furrer         |E-Mail : [e-mail]     |
          |Emch + Berger AG     |Tel    : [phone]   |
          |Gartenstrasse 1      |Fax    : [phone]   |
          |3001 BERN (CH)       |                           |
          +---------------------+---------------------------+

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

From: Arnaldo MANDEL <[e-mail]>
Subject: Re: What full-screen file managers are there?
Date: 17 Nov 92 09:01:45 GMT
Sender: Le Facteur <[e-mail]>
Nntp-Posting-Host: litp4.ibp.fr
Full-Name: Arnaldo MANDEL
To:       [e-mail]

In article <[e-mail]> [e-mail] (Pete Holsberg) writes:

> In article <[e-mail]> [e-mail] (Paul Kramer) writes:
> =Hello,
> =
> =I found while reading articles in this electronic conference, a
> =reference to a public-domain program called utree.  It is full-screen
> =file manager which allows you to perform file functions with single
> =keystrokes.  For example, after I start the program I see a screen
> =that lists the files on my account.  If I want to delete one of those
> =files I move the cursor on top of it, press a certain key.  Boom it
> =gone!
> =
> =Well I am wondering:  Is there anything better than utree?  I am aware
> =of a program called 'maint' but it doesn't have the full functionality
> =of 'utree'.

> I prefer HDSCAN.

Better go all the way and get emacs, with tree-dired.  You not only get all
funtions most file managers execute, but you can add some of your own.  And
then, this would be just the beginning of using emacs for almost everything,
instead of hundreds of little frozen utilities.
--
 ............................................................
Arnaldo Mandel   ---  [e-mail]
Temporarily: [e-mail]

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

From: Bruce Sams <[e-mail]>
Subject: lib77 "flush" call. Where is it in UNIX?
Date: 17 Nov 92 12:23:03 GMT
To:       [e-mail]

Dear fellow UNIX and HP-UX folks,

   I am trying to bring up the graphics package MONGO on our HP-720 machine
   and I have encountered a problem.  It seems that MONGO uses a "flush"
   call deep in its IO guts.  Looking through all the HP-UX libraries I
   find no such routine.  Rather, I find "pflushli", "flushinp", "intrflush",
   "__cflush", "fflush", and "fflush_p."  Some of these are clearly not
   what I want, but others are less certainly wrong.  I find no documentation
   for any of these readily available.  How do I proceed?  Has anybody out
   there installed MONGO under HP-UX?

Cheers,

Lowell Tacconi-Garman

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

From: der Mouse <[e-mail]>
Subject: Re: Wait for child
Date: 17 Nov 92 12:54:28 GMT
Followup-To: comp.unix.questions
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Kurt George Gjerde) writes:

> On Unix, how can I have a process wait for it's child process to do
> something (not terminate).

This is a UNIX question, not a C question.  Consequently, I'm
cross-posting to comp.unix.questions and directing followups there.

There are multiple ways.  You can have the child send some signal to
the parent.  You can create a pipe (or other stream connection) and
have the child write and the parent read.  You can use semaphores, if
your system provides them.  There are probably other ways that don't
come to mind at the moment.

Which one is best?  That depends on many factors, such as the required
response time, how much information needs to be passed (besides the
bare fact of the wakeup), how widely portable the code needs to be,
what if anything the parent is doing while waiting....

                    der Mouse

                [e-mail]

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

From: der Mouse <[e-mail]>
Subject: Re: name of file descriptor
Date: 17 Nov 92 12:59:31 GMT
Followup-To: comp.unix.questions
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Kurt George Gjerde) writes:

> How can I get the filename from a filedescriptor, on Unix??

This is a UNIX question, not a C question, and I'm therefore
crossposting to comp.unix.questions and redirecting followups there.

In general, you can't.  The file descriptor may not be connected to a
file (it may be a pipe or (other) socket or TLI stream or a device or
maybe something else).  Even if it is a file, the file may have no
names any longer, or it may have multiple names.  Even if it has at
least one name, there may be no name by which it is accessible to your
process or possibly even other processes.  And even if you assume it
has a name that you could use to refer to it by, there is no way to
find any such name short of a full scan of all filesystems, which can
take a *long* time.

You're usually best off writing wrappers for open/close (or
fopen/fclose if you're using stdio) that remember filenames.  The major
case where this doesn't work is when you're dealing with redirection to
or from files, in which case you're pretty much out of luck.

                    der Mouse

                [e-mail]

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

From: der Mouse <[e-mail]>
Subject: Re: talk
Date: 17 Nov 92 13:06:42 GMT
To:       [e-mail]

In article <[e-mail]>, [e-mail] (The Beast) writes:

> I tried to use 'talk' to converse with a friend in another college.
> I got the following message:
> 'Target machine does not recognize us'.

This means simply that talkd on the remote machine did a
gethostbyaddr() on your machine's address and got a failure indication.

Either the remote machine is broken somehow or whatever nameserver is
responsible for the relevant .in-addr.arpa zone is misbehaving, or
(worst case) hasn't been set up.

> If so, how can this situation be circumvented?

The right way to fix it is to fix whatever's wrong.  A workable
substitute is to fix talkd so that if gethostbyaddr() fails, it prints
the numeric address instead (and in parallel, fix talk to accept
numeric addresses).  This is what I've done, and sometimes when the net
is flaking out (and thereby breaking the DNS) it's proven useful.

                    der Mouse

                [e-mail]

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

From: Ed Posnak <[e-mail]>
Subject: Re: Changing the owner of a process
Keywords: process owner
Date: 17 Nov 92 14:28:37 GMT
To:       [e-mail]


In article <[e-mail]> I wrote:
"Is there an un-general way of changing the euid of a process from another
"process?  We have an unusal requirement on our project for a 'shift-change'
"where a user logs in and 'inherits' ownership processes that were running
"under another euid.  Thanks in advance.

Thanks to all who responded.  I was surprised at the number of different
responses I received, (most involving setuid()) which led me to realize my
wording must've been real vague.

What I was looking for was something along the lines of how to change the
effective user id of a process who's source I may not be able to modify, by
some other means, e.g. from another process.  Many suggested writing a device
driver or system call to do this.  Here is one answer along those lines.
---

This is one of those dirty tricks I've always wanted to get around to
figuring out a way to do... (without kernel source, that is)

I believe it could be done by writing a device driver. Open the device
driver and write commands to it, and it does the dirty work. For
instance, send it 8 bytes containing the process to change and the uid
to change it to.

The uid, in every version of Unix I've seen, is stored in the proc
structure in the kernel. You should be able to fiddle with this at
will in a device driver, and since it's in the proc structure and not
the u area, it'll always be there even if the process is swapped out.
It *should* be just a matter of searching for the entry and changing
it.

This technique should also serve to implement something like renice()
under Xenix, which doesn't have it. (The nice value also being stored
in the proc structure.)
--
Mark Buda

I get my monkeys for nothing and my chimps for free.
---

I would enjoy hearing from anyone who might have done this already.  Thanks.



--
--
Ed Posnak
Harris Space Systems Corporation
[e-mail]

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

From: Luke Higgins <[e-mail]>
Subject: file .fingerees
Date: 17 Nov 92 13:56:41 GMT
To:       [e-mail]


I have found a program which attempts to execute a file when you are
fingered and I am having problems getting it to work.  My problem is
that on my system I can't seem to find the .fingerees file.  Is this
a standard unix file and if so how do I go about finding it.
    Thanks
    Luke  (a computer novice  to say the least)

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

From: Lamar Owen <[e-mail]>
Subject: Re: Whence Unix? (was Re: IS UNIX DEAD?) (New Thread?)
Date: 16 Nov 92 15:57:21 GMT
To:       [e-mail]

In <[e-mail]> [e-mail] (Steve VanDevender) writes:
>In article <[e-mail]> [e-mail] (John 'Fritz' Lowrey) writes:


>       My seed:
>       Microsoft DOS    ->    Intended as a stepping stone while DR
>                   wrapped up CP/M-86, and now the program
>                   loader of choice for countless millions.

>You really need to study up on your computing history.  You seem
>to imply that Microsoft got MS-DOS from Digital Research.
>Microsoft got MS-DOS from a small firm called Seattle Computer,
>which had written a quick-and-dirty CP/M clone called SC-DOS.
>Then Microsoft hacked it up and marketed the hell out of it.

The original marketing niche for MS-DOS was as a stepping stone from the
8-bit CP/M world to the multiuser 16-bit world of Xenix.  Microsoft, in
the early 80's, fully intended to make Xenix their high-end OS.  However,
the market chose otherwise.


>Steve VanDevender

--
--

Lamar Owen, Systems Consultant, GE Lighting Systems, Hendersonville, NC
***********************************************************************
 Opinions expressed herein are the author's and do not reflect policy
   or opinions of the General Electric Company or its subsidiaries.
***********************************************************************

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

From: Jeffrey Reilly <[e-mail]>
Subject: Curses support under DOS?
Date: 17 Nov 92 15:56:17 GMT
Sender: USENET News System <[e-mail]>
Nntp-Posting-Host: mipos2
To:       [e-mail]

Is anyone aware of a curses package that works/can be ported to a DOS
environment?

Any pointers would be appreciated!

Jeff

Jeff Reilly            | "Do not route the (Ethernet)
Intel Corporation        | cable near a cyclotron" -
[e-mail]    | LAN/586 Board User's Guide
(408) 765 - 5909        |

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

From: Steven Lon George <[e-mail]>
Subject: MPE, NS, TCP/IP
Date: 17 Nov 92 17:53:29 GMT
To:       [e-mail]

Hello there.  What I am wondering relates to the HP UNIX environment.

I am applying for a job under this environment, in a networking
application. The job has some criteria that I am not sure what it means.
I am hoping I can find someone who does. Anyway, the line reads as
follows.?

Must have knowledge of MPE, NS, TCP/IP & Network Components and
Architecture.

What is MPE (Multiplatorm Environment?), and NS.  Has anyone ever heard
these acronyms before.  The interview is on Thursday, so you can answer
by mail at '[e-mail]'.  Please notice the plural
georges with an s.  otherwise, someone else will get the message.

Thanks in advance.

Steve....

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

From: "System Admin (Mike Peterson" <[e-mail]>
MMDF-Warning:  Parse error in original version of preceding line at BRL.MIL
Subject: Re: lib77 "flush" call. Where is it in UNIX?
Date: 17 Nov 92 17:29:55 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Bruce Sams) writes:
>Dear fellow UNIX and HP-UX folks,
>
>   I am trying to bring up the graphics package MONGO on our HP-720 machine
>   and I have encountered a problem.  It seems that MONGO uses a "flush"
>   call deep in its IO guts.  Looking through all the HP-UX libraries I
>   find no such routine.  Rather, I find "pflushli", "flushinp", "intrflush",
>   "__cflush", "fflush", and "fflush_p."  Some of these are clearly not
>   what I want, but others are less certainly wrong.  I find no documentation

You can write a "flush" that calls "fflush" (which is a standard C
routine, and you should have a man page for it). You will need to use the
HP-UX 'fstream' routine to map the FORTRAN unit number to a C stream number.

I can't send you source for flush since it is licensed, but you can
have these routines I used to make it work:

      function igetfstream (lunit)
c
c     Purpose:
c        Return the FILE pointer corresponding to unit 'lunit',
c        or 0 if unit 'lunit' is not open.
c
      logical opened
c
c
c     write (6,*) 'igetfstream called for unit ', lunit
      inquire (unit=lunit, opened=opened)
      if (opened) then
         igetfstream = fstream (lunit)
         if (igetfstream .eq. 0) then
            if (lunit .eq. 5) then
               igetfstream = igetstdin ()
            else if (lunit .eq. 6) then
               igetfstream = igetstdout ()
            else if (lunit .eq. 7) then
               igetfstream = igetstderr ()
            end if
         end if
      else
         igetfstream = 0
      end if
c     write (6,'(1x,a,z9.8)') 'igetfstream returning ', igetfstream
      return
      end

To get the stream number for stdin/stdout/stderr, try:

/*
 * Return the FILE pointer to stdin.
 */

#include <stdio.h>

    FILE *
igetstdin_ ()
{
    return (stdin);
}


/*
 * Return the FILE pointer to stdout.
 */

#include <stdio.h>

    FILE *
igetstdout_ ()
{
    return (stdout);
}


/*
 * Return the FILE pointer to stderr.
 */

#include <stdio.h>

    FILE *
igetstderr_ ()
{
    return (stderr);
}


Note that always use "+ppu" for f77 compilations; you will need
to remove the trailing '_' on the C routines if you don't.

I have (cross-)posted this to comp.sys.hp.
--
What are the chances that any HP computer system will ever "work" properly?
 ... and Slim just left town. -*- Mike Peterson, SysAdmin, U/Toronto Chemistry

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

From: Tim Evans <[e-mail]>
Subject: Re: Searching for E-mail package
Date: 17 Nov 92 17:00:53 GMT
To:       [e-mail]

[e-mail] (Karl Nicholas) writes:

>In article <[e-mail]> [e-mail] (ANDY KUMEDA) writes:
>>
>>
>>
>>We are currently searching for E-mail/Office Automation packages that will
>>serve several thousand users based on the following criteria:
>>
>>  1) Must support X-Window (OpenLook or Motif) either through an X-terminal
>>     or a PC running an X-server.
>>  2) Must also support PCs running DOS, with a TCP/IP network connection.
>>  3) In addition, a non-X version (character-based -- ASCII terminals) is
>>     preferable, but not required.
>>  4) Must be able to 'customize' -- ie original text to be replied or
>>     forwarded should not be modifiable.
>>  5) Must support SMTP.
>>  6) Vendor must have good technical support.
>>

>    I don;t know all the answers to the questions here,
>but you should call the cc-mail people. no I do not know the
>phone number either, sorry, but cc-mail is the name of the
>product. What do I know? well ...

cc:Mail is only made by Lotus.  You know, that unknown 1-2-3 company.
--
Tim Evans                     |    E.I. du Pont de Nemours & Co.
[e-mail]  |    Experimental Station
[phone]/7395           |    P.O. Box 80357
EVANSTK AT A1 AT ESVAX        |    Wilmington, Delaware 19880-0357

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

From: Ioi Kim Lam <[e-mail]>
Subject: C++ Debugger with Motif
Date: 17 Nov 92 14:16:28 GMT
Sender: Ford WDL <[e-mail]>
To:       [e-mail]


I am using X11R4 on the SunOS. I am writing Motif programs in C++, using the
GNU compiler g++. Does anyone know how to debug these programs. I tried
gdb-3.2 and gdb-3.5 but they don't seem to understand the C++
function-name convention. Gdb-4.4 gives an segmentation fault with the
Motif program, although it can debug normal C++ programs under Unix. I
tried UPS but it always shows the header file instead of the source file.

Are there any public-domain X-Debugger available that can support C++?

Please reply to [e-mail]

Thank you in advance.

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

From: Stephen Daedalus <[e-mail]>
Subject: Re: Looking for Time Series Analysis software
Date: 17 Nov 92 16:56:04 GMT
Sender: Stephen Daedalus <farrar@firewall>
Nntp-Posting-Host: 192.153.52.178
To:       [e-mail]

You might take a look at TSP, which runs on mainframes to PC's.  We used the
student version in one of my econometrics classes and, although limited, did
a very good job on the project, and kept me out of the comp. lab.
---
Richard Farrar        These views are my own, not my company's or country's.
ADAC Laboratories    "A man without hand is not a man."
[e-mail]        - George Costanza, Seinfeld
        Spam is a registered trademark of Hormel

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

From: Fiona Wong <[e-mail]>
Subject: Re: file .fingerees
Date: 17 Nov 92 17:51:59 GMT
Sender: [e-mail]
Nntp-Posting-Host: eddie.cdf
To:       [e-mail]

In article <[e-mail]> [e-mail] (Luke Higgins) writes:
>
>I have found a program which attempts to execute a file when you are
>fingered and I am having problems getting it to work.  My problem is
>that on my system I can't seem to find the .fingerees file.  Is this
>a standard unix file and if so how do I go about finding it.

make it by yourself. :-)

>    Thanks
>    Luke  (a computer novice  to say the least)


--
|       ____                    | You know you've been spending too much time |
|      /    . __   __   __      | on the computer when your friend misdatess  |
|     /--- / /  ) /  ) ___)     | a check, and you suggest adding a "++" to   |
|____/    / (__/ /  / (__(______|_fix it._____________________________________|

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

From: Pete Hardie <[e-mail]>
Subject: Re: BIG IMPORTANT QUESTION PLEEEEZE HELP ME!!!!!
Date: 17 Nov 92 14:45:36 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Rouben Rostamian) writes:
>In article <[e-mail]> [e-mail] (Tim Pierce) writes:
>>My car's making a funny noise somewhere under the hood.  It's one of
>>these cheapo models, though, and it doesn't have pull-down help from
>>under the visor available!!!  Can anyone tell me what it's doing
>>wrong!!!!
>
>Switch to emacs and go into the overdrive mode.

Shouldn't that be 'drive-over' mode?


--
Pete Hardie:  phardie@nastar  (voice) [phone]
Digital Transmission Systems, Inc., Duluth GA
Member, DTS Dart Team           |  cat * | egrep -v "signature virus|infection"
Position:  Goalie               |

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

From: 25235-gokhman <[e-mail]>
Subject: Shortened repost: mmap over NFS problem
Keywords: Does mmap with MAP_SHARED work for write over NFS ?
Date: 17 Nov 92 18:54:22 GMT
Sender: USENET System Software <[e-mail]>
To:       [e-mail]

Hi:

This is a (shortened) repost of an article for which I got no responses. Writing
to a memory region allocated with mmap call using MAP_FILE and
MAP_SHARED causes a segmentation fault (RS6000 under AIX 3.2) when
it is done over NFS. Mapping to the same file locally works fine.
Documentation, as far as I can tell, does not mention
NFS as a limitation. Is mmap supposed to work over NFS and this is an AIX
bug, or is it not supposed to work and it is a documentation problem ?

            Thanx in advance, Ed Gokhman

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

From: Ruurd van der Meer <[e-mail]>
Subject: Sun Printserver? IPX-stack available ? NCP documentation ?
Date: 16 Nov 92 11:28:19 GMT
Sender: USENET News System <[e-mail]>
Originator: rvdm@st33-sys4
To:       [e-mail]

I want to have a printserver on a Sun SParc (SunOS 4.1.x, in future Solaris)
for a Novell environment.  So I have some questions:

Does anybody know if such a product exists ? If not:

Is there an IPX protocol stack on a Sun SParc available, commercial or public
domain ? If this is in the form of a Netware shell as in DOS, what about
a C-interface for it (to write a printserver)?

And what about the NCP protocol of Novell? Is this available somewhere ?

Please direct answers to

[e-mail]

If I have collected the answers I will put them back into the newsgroups.

      ###########################################################
      #  This note does not necessarily represent the position  #
      #     of Oce-Nederland B.V. Therefore no liability or     #
      #      responsibility for whatever will be accepted.      #
      ###########################################################

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

From: Matz Engstrom <[e-mail]>
Subject: Tool for email, and snail mail addresses
Date: 17 Nov 92 19:11:00 GMT
Sender: Matz Engstrom <[e-mail]>
To:       [e-mail]

I am interested in a UNIX tool for keeping my
addressbook, both electronic and snailmail. I
store them as VM NAMES files today, which also
gives me the possibility to store phone numbers,
and free text notes about the addressee.

Do you know of any UNIX tools for that, including
a fullscreen interface to maintain the files,
and some extracting tools to make the email
addresses available to my mail tools?
--
H{lsningar (Sincerely) /
          Matz Engstrom, Gothenburg Universities' Computing Centre;
          Address Box 19070; Kapellg}ngen 5; S-400 12 Gothenburg; Sweden
          Phone [phone]; Fax [phone];

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

From: "Michael S. McLean" <[e-mail]>
Subject: Re: Changing the owner of a process
Keywords: process owner
Date: 17 Nov 92 20:08:13 GMT
Sender: "Michael S. McLean" <[e-mail]>
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Ed Posnak) writes:
|>
|> What I was looking for was something along the lines of how to change the
|> effective user id of a process who's source I may not be able to modify, by
|> some other means, e.g. from another process.  Many suggested writing a device
|> driver or system call to do this.  Here is one answer along those lines.
|> ---
|>
|> This is one of those dirty tricks I've always wanted to get around to
|> figuring out a way to do... (without kernel source, that is)
|>
|> I believe it could be done by writing a device driver. Open the device
|> driver and write commands to it, and it does the dirty work. For
|> instance, send it 8 bytes containing the process to change and the uid
|> to change it to.

There is no need to write another device driver for this.  /dev/kmem will
suffice.

--
Michael S. McLean ([e-mail])
Engineering Computing Systems             "Imagination is the one weapon
College of Engineering                     in the war against reality."
University of Florida                         -- Jules de Gaultier

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

From: Artur Romao <[e-mail]>
Subject: I need some pic-like thing
Keywords: pic
Date: 17 Nov 92 15:19:19 GMT
Sender: USENET News System <[e-mail]>
Followup-To: comp.unix.questions
To:       [e-mail]


Hello, net!

I need to build some docs, and some of the stuff needs pic. I know
it isn't public domain (or am I wrong?), so what I would like was
to find something that does the work, but available in the public domain.

Is there any such thing.

Thanks in advance for any info.

BTW, are theese the right groups to ask this kind of questions. Please
adress me to the apropriate ones if not.

Best regards,

 .........................................................................
Artur M. P. Romao : PUUG - Grupo Portugues de Utilizadores do Sistema UNIX
[e-mail]     : Av. 24 de Julho, 134 - 7.
[phone]  : P-1300 LISBOA - PORTUGAL
 .................:.......................................................

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

From: Eduardo Rodriguez <[e-mail]>
Subject: Re: C Source Code for Pattern Matching
Keywords: text pattern matching source
Date: 17 Nov 92 22:11:25 GMT
Sender: Network News <[e-mail]>
Originator: erodrigu@tortel
To:       [e-mail]


In article <leb.721935245@edgar>, [e-mail] (Larry Bright) writes:
> Could anyone tell me where I could find C source code for some
> simple text pattern matching s/w, preferably with a syntax and
> semantics based on "regular expressions" as in grep, awk, sed, etc.
> Anything that implements a non-trivial subset of these functions
> would be useful.

The agrep package (approximate pattern matching) has many of the grep functions
and other stuff besides.  You can pick it up from cs.arizona.edu using
anonymous ftp.
Regards,
    Edo...

 ////                                ```\                             .
 c-OO  Is there something that       OO-c     [e-mail]  .
    \  hurt you?                     /      Dpto Cs de la Computacion .
   -                    Ya... Pain    -        Universidad de Chile   .

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Keywords: n
Date: 18 Nov 92 01:32:12 GMT
To:       [e-mail]

>and will probably miss the end:
>
>I disagree.  Not having used too many OS's, I can only recall 2
>that the help command did something.  Vax VMS, and Microsoft Dos 5.
>I take that back, i just thought of Wylbur/MVSX, and VM/CMS.  I believe
>help does something on those systems.

Add in OS/2 and (I believe) AmigaDOS.  Plus many, many, many, command
line oriented programs that are not operating systems like symbolic
math packages and graphing packages.  Plus every text mode adventure
game.
>
>Why should help do something at the command prompt?  I would much, much
>rather have the hardcopy manual in my hand.  While online help might
>be nice, I really just do not see its absence as an inherent flaw.

You think the average university should issue everyone textual manuals?

Especially for the huge mass of files that come with Unix???

And what about when those files change?

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

From: Mike Leibensperger <[e-mail]>
Subject: Re: BIG IMPORTANT QUESTION PLEEEEZE HELP ME!!!!!
Date: 16 Nov 92 22:12:51 GMT
Sender: Netnews <[e-mail]>
To:       [e-mail]


   In article <[e-mail]> [e-mail]
   (Tim Pierce) writes:

   >My car's making a funny noise somewhere under the hood.  It's one of
   >these cheapo models, though, and it doesn't have pull-down help from
   >under the visor available!!!  Can anyone tell me what it's doing
   >wrong!!!!

Hey Tim!  I think I know what you're (*smirk*) DRIVING at!!!  ;-)  ;-)  ;-)

Listen up, gang!  Tim thinks maybe if you Read The Fine Manual you
might fair better with some of these questions!!!  What a Fine idea!
Or, as we net.czars say, WAFI!

In article <[e-mail]> [e-mail] (Steve
Franklin) writes:

   Wrong group Timbo... This is comp.unix.questions, not
   comp.eunuchs.questions. Try again, and this time read your damn warning
   when you post (are you REALLY REALLY sure you wanna do this?)
      steve

Hey Steve-a-rino!  Maybe your fine manual is floating up in the air,
near the AC duct in the ceiling!  You'll have to (*smirk*) LIGHTEN UP
if you want to grab it and read it!

    Too wise for his own good,
    mjl
--
Michael J. Leibensperger      "Rats and roaches live by competition under the
Locus Computing/Boston         laws of supply and demand; it is the privilege
25 Mall Road; Burlington MA         of human beings to live under the laws of
01803 [phone] [e-mail]       justice and mercy."  -- Wendell Berry
Member of the League for Programming Freedom --- write [e-mail]

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

From: fred j mccall 575-3539 <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 17 Nov 92 21:17:47 GMT

hacktic.nl>
To:       [e-mail]

In <[e-mail]> [e-mail] (Peter Busser) writes:

>[e-mail] (Chris Sherman) writes:

>>In <[e-mail]> [e-mail] (Peter Busser) writes:


>>>user interfaces. With NT or OS/2 you only need to learn only *1* user
>>>interface.

>>Suppose you don't like it...

>Then you're stuck with it. But then at least you're stuck consistently... :-)

Yes, and all cars should look like an EDSEL.  ;-)

--
"Insisting on perfect safety is for people who don't have the balls to live
 in the real world."   -- Mary Shafer, NASA Ames Dryden
 ------------------------------------------------------------------------------
[e-mail] - I don't speak for others and they don't speak for me.

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

From: Mike O'Connor <[e-mail]>
Subject: Re: tcsh-like subroutine
Date: 18 Nov 92 00:14:47 GMT
NNTP-Posting-Host: slee01.srl.ford.com
To:       [e-mail]

In article <[e-mail]> [e-mail] writes:

:I'm looking for a tcsh like input routine which would be callable from a C or
:FORTRAN program. I just want to be able to read a line of text from the user
:with tcsh-like (or emacs-like!) command line editing/recall features.

You need the GNU readline library or a clone of the GNU readline
library.  The GNU stuff is FTPable at prep.ai.mit.edu:/pub/gnu...

                        ...Mike

--
 Michael J. O'Connor           |  Internet:  [e-mail]
 Ford Motor Company, OPEO      |  UUCP:      ...!{backbone}!fmsrl7!mjo
 20000 Rotunda, Bldg. 1-3001   |  Phone:     [phone]
 Dearborn, MI  48121           |  Fax:       [phone]

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

From: Tim Ramsey <[e-mail]>
Subject: Re: Changing the owner of a process
Keywords: process owner
Date: 18 Nov 92 01:40:58 GMT
Followup-To: comp.unix.internals
NNTP-Posting-Host: hilbert.math.ksu.edu
To:       [e-mail]

[ Followups set to comp.unix.internals ]

[e-mail] (Michael S. McLean) writes:

>There is no need to write another device driver for this.  /dev/kmem will
>suffice.

You can do it by poking about in /dev/kmem, but it's not safe.  Consider:

Process A                     Process B               Process C
Open /dev/kmem                <runs>
Find B proc table entry       <runs>
Update proc table entry       <exits>                 <starts up>

Between the time when Process A locates the proc table entry for Process B
and the time that it updates the entry, Process B exits and Process C is
started -- and given Process B's old table entry.  You end up updating
the wrong process.

--
    Tim Ramsey, 913.532.6750  |  I'm a sysadmin and I'm okay;
    Department of Mathematics |  I work all night and I sleep all day.
    Kansas State University   |

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD (long)
Date: 18 Nov 92 01:50:24 GMT
To:       [e-mail]

>
> It isn't a question of preservation of case sensetivity; it is a
>question of teaching the kernel about case insensetivity. Right now,
>the kernel doesn't know anything about the character set you (or
>anyone else on the system) is using; all it knows about are two
>special 8-bit blobs, one for 'end of filepath' and another for
>'filepath sepperator'. If you teach it about case insensetivity,
>it is now working in whatever character set you chose, and people
>who wish to use it with other ones are out of luck.

Am I to understand that unix file management is in the "kernel" and can
not be intercepted before and after kernel operations?  Even DOS seperates
the file system and kernel enough that you can use stacker or something
between the files system and the kernel.

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

From: "Sheetal V. Kakkad" <[e-mail]>
Subject: rexec does not search for ~/.netrc file
Keywords: rexec, netrc
Date: 17 Nov 92 22:29:42 GMT
NNTP-Posting-Host: boogie.cs.utexas.edu
To:       [e-mail]

I am trying to use rexec(3N) on a Sun SparcStation to start a server
on a remote host (also a Sun SparcStation). The synopsis of the call
(as given in the man page) is:

SYNOPSIS
     rem = rexec(ahost, inport, user, passwd, cmd, fd2p);
     char **ahost;
     u_short inport;
     char *user, *passwd, *cmd;
     int *fd2p;

Here's the relevant portion of the description from the man page:

DESCRIPTION
    ...
     If  a  username  and password are both specified, then these
     are used to authenticate to the foreign host; otherwise  the
     environment  and  then  the  user's  .netrc file in his home
     directory are searched for appropriate information.  If  all
     this fails, the user is prompted for the information.
    ...

Now, I have a .netrc file in my home directory which has lines of the
following form, one line for each foreign host:

machine machinename login mylogin password mypassword

Here's how I call rexec in my program (I have also shown some other
relevant variables used in the call):

int port;    /* set using getservent ("exec", "tcp") call */
char *rhost;    /* set to the foreign host name */
char *srvr;    /* set to full path name of my server */

s = rexec (&rhost, port, (char *) NULL, (char *) NULL,
       srvr, (char *) NULL);

Even though I have the .netrc file set up in my home directory, the
call still prompts me for the username and password.

My question is: Am I doing something wrong here, or is this a commonly
known bug? Any pointers will be appreciated.

My machine is a SparcStation ELC running SunOS 4.1.2. All the remote
hosts (where I am trying to start up my server) are SparcStation 1's
also running SunOS 4.1.2.

Thanks in advance,

- Sheetal

Note: Follow-up set to comp.sys.sun.misc.

 ----------------------------------------------------------------------
Sheetal V. Kakkad    [e-mail]    Office: [phone]
Dept of Computer Sciences, UT-Austin        Home:   [phone]
======================================================================

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 18 Nov 92 02:35:41 GMT
To:       [e-mail]

>>Because that often saves (a lot of) money.
>
>I doubt it. If you insist to install the OS yourself then you better
>start learning how manage it.

No, no, no, no, no....OS/2, Windows, DOS etc. have PROVED that you don't
have to be a system administrator to install an OS.  Why do you insist on
this dogma "if you aren't computer literate enough to install Unix, you
should pay someone to do it."  Why not make unix easy enough to install
that you don't have to do either?  Why force the problem onto the user
instead of just FIXING it?  Why defend a flaw?

I won't defend Dos, or windows' or OS/2s' flaws.  I will highlight their
strong points.  Highlight unixes' strong points, but FIX IT'S FLAWS.

Overly complex installation is a flaw.

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD? (very long)
Date: 18 Nov 92 02:30:43 GMT
To:       [e-mail]

>Mine popped up after 45 seconds (cheap Korean brand).  In the same
>vein, my microwave doesn't ask for confirmation, my light switches
>don't ask for confirmation, and my toilet definitely does not ask
>"are you sure? (y/n)".

All of these are considerably easier to use then a text editor.  Many cars
will (rightly) make a sound if you leave your keys in.  They will also
(rightly) make a sound if you leave your lights on when the engine is
off.  It would be (and perhaps is, in some cars) nice if you could turn
these off.  But for my, girlfriend, who is very intelligent, but occasionally
forgets her lights and keys, the buzzers are a Godsend.  And for me,
who can forget and type qw instead of wq, a prompt "haven't saved yet"
would be nice.  BTW, this is not asking for superflurous time wasting
prompting.  VI will NOT LET YOU QUIT with a Q.  It must be wq or Q!.  So
if I type Q, why not ask me if I meant WQ?  Easier for the user, easeir
for the power user that forgets.

>But then, we're arguing apples and oranges, as I said before.  You can't
>do anything useful with a Microsoft Write file except print it, as far
>as I can tell, and I haven't printed anything out in about a month.

Or compile it?  write can write ascii as well as VI!

>No, assembly is a low level language, C is my favorite programming language,
>and control-c is the de facto standard for process interruption on Unix
>systems.  The bit about kill %1 was facetious, but everyone who intends to
>use Unix seriously should have control-c pretty firmly imbedded in their
>minds.

process interruption is different from exit.  Exit implies "update setup
files etc., and perhaps prompt me if I want to save"  Process interruption
implies "quit unconditionally."

>Definitely not.  My keyboard doesn't have an F1 key.  If it did, there
>would be no guarantee that it would produce what your F1 key does.
>There are good reasons why nobody uses F-keys -- their original purpose
>was to be bindable to whatever the user felt appropriate.

That's fine...your keyboard doesn't have an F1 key and you don't want help.
My keyboard DOES have an F1 key, and unix doesn't have a standard F1 key.
If I want to bind F1 on my computer, having help on it doesn't matter.  Think
about it...it would be intercepted before it gets sent, and translated,
wouldn't it?  And for those of us that don't know how to bind keys yet,
it would go through to help.

>You mean one character left, but you're still terminally unconscious.

Yes...I keep saying right because I'm thinking about what hand I use to
type it.

>Little bumps on the keyboard don't mandate any sort of cursor control
>arrangement whatsoever.  It makes perfect sense to not use ; for a
>control key -- it's not right of l on all keyboards, for one thing.

Non standard QWERTY-incompliant keyboards are certainly not my problem.
99.9% of qwerty keyboards (and every one I have seen) has a semicolon beside
the l.  The other .1% are so nonstandard that perhaps their hjkl aren't
in a row either.

>It does NOT make perfect sense to have an unshifted help key in the
>middle of the goddamn keyboard in an editor people are supposed to use.
>Why don't you take some courses on human-computer interaction and
>ergonomics and try to get back to Unix later?

Thus the beauty of the (till now) unused F1 key.  Besides, one would
assume that VI would be configurable enough that you could change the
key mappings of the keys!  Or at least turn off help if you wanted to.

>I assume that you meant "if you launch an editor from RN."  Pray tell,
>how does vi figure out from where it was launched, and whether or not
>that program would necessarily need word-wrap on?  Please go on to
>inform me how you've determined that I want word-wrap when I write news --
>I don't -- and how you've managed to instantiate artificial intelligence
>on any other competing system that you're now comparing with Unix?

Well, in the PC world (blaspheme!!!!) you tell programs what editor you
want to use and what command line options to send to them.  And guess what
if you want to change that information, you don't have to edit some
invisible file that the computer has (annoyingly) placed in your work
directory.  The program updates it's OWN config file.  (although you could
do it manually if you're into that kind of thing)

>For you, it's a flaw.  For enough other people, it's not.  Your systems
>administrator has decided not to add a file that does something at the
>prompt.  Contact him for further information.

I see.  For other people it is a benefit?  Can you explain how it is to
their benefit to have help do nothing?

>If they were, then it would be.

They are.

>Programs do not decay.  vi is unchallenged as a small, powerful text editor;
>that it was originally written years ago is just not important.

B
A
A
A
yes programs DO decay.  See that crap above this line?  That's what
happens when I try to use my arrow keys.  Let's see what happens when
I try to use my f1 key?  Beep.  Where is built in help?  It's not there.

VI is decaying.  It is not keeping up with the hardware.  It is not keeping
up with competing software.  Anyone in their right mind given a choice
would not use vi unless they are already familiar with it (or the only
choice is emacs).

Let's not let this decay happen to unix.

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

From: Kevin Smith <[e-mail]>
Subject: Re: Overflow warnings (SCO SV3.2)
Date: 17 Nov 92 15:49:54 GMT
Followup-To: comp.unix.sysv386
To:       [e-mail]

In article <[e-mail]> [e-mail] (Anthony P Lawrence) writes:
:>[e-mail] (The Golden Gryphon) writes:
:>:
:>:cd /etc/conf/cf.d
:>:./configure
:>:
:>:find NREGION and NFILE, and increase their size quite a bit.  I would recommend
:>:doubling them.  Then relink the kernel.  DO THIS IN SINGLE USER MODE!  Back the
:>:system up first.
:>
:>While this is certainly not bad advice, isn't backing up the system somewhat
:>like cautioning the use of seatbelts while vacuuming the car?  All that's
:>really going to be affected is /unix.  I could see making a /unix.good, and
:>he certainly should have backups anyway, but this makes it all sound very
:>scary and system-threatening!
:>
:>What is the point of single user mode? Increasing these parameters doesn't
:>affect anything but the space in the new kernel, so how is it going to
:>cause any problems for the running system?
:>
:>Just wondering why we're being so cautious here.
:>

Relinking the kernal and installing the new unix automatically creates
/unix.old.  This is fine unless you screw up the kernel twice in a row
and have nothing to boot.  This is where the /unix.good can save the day.
I always create a /unix.good from a known, working, kernel.  It, however,
does not have to always be the latest one, just one that will boot.

Single user mode is important because 'ps' may not run with the new
kernel (without '-n /unix.old'), which may, in turn, cause shutdown to
hang or not shut everything down.  Many applications (ingres...) use
ps to find their daemons and servers and so may not terminate properly.
--
        | Email - !shady!kevin uunet!shady!kevin
Kevin Smith | Voice - [phone]
        | Mail  - ShadeTree Software, Inc., 192 Capricorn Dr. #10,
        |         Somerville, NJ  08876

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

From: Narendra Bhandari <[e-mail]>
Subject: CURSES - No delay option, SunOS 4.1.1
Date: 18 Nov 92 03:30:30 GMT
Sender: USENET News System <[e-mail]>
To:       [e-mail]


Is there a nodelay function in curses.

I am using SunOS 4.1.1.( I have used one on SVR4).

If not , is there any other technique of

Non-blocking character input from Keyboard.


--
      ...........................................................
      .  Narendra Bhandari             [e-mail]    .
      .  950 S Terrace Road, #D366     602.921.3401(Home)    .
      .  Tempe, AZ 85281        602.965.1780(Work)    .

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

From: Paul_Dineen <[e-mail]>
Subject: Re: lib77 "flush" call. Where is it in UNIX?
Date: 17 Nov 92 23:31:11 GMT
Sender: Notes Administrator <[e-mail]>
X-Newsreader: TIN [version 1.1.4 PL6]
To:       [e-mail]

: You can write a "flush" that calls "fflush" (which is a standard C
: routine, and you should have a man page for it). You will need to use the
: HP-UX 'fstream' routine to map the FORTRAN unit number to a C stream number.

  Note that 'fstream' is a FORTRAN routine, rather than an HP-UX routine.

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

From: Tim Pierce <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Keywords: n
Date: 18 Nov 92 03:57:30 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Paul Prescod) writes:

>>I take that back, i just thought of Wylbur/MVSX, and VM/CMS.  I believe
>>help does something on those systems.
>
>Add in OS/2 and (I believe) AmigaDOS.  Plus many, many, many, command
>line oriented programs that are not operating systems like symbolic
>math packages and graphing packages.  Plus every text mode adventure
>game.

Which, as we all know, collectively constitute the zenith of software
development.  It is doubtful that any modern operating system could
succeed that was not modeled after a text adventure.  I can see it now
 ...

amhux3% hello sailor

--
____ Tim Pierce                /
\  / [e-mail] /               Rocks say goodbye.
 \/ (BITnet: TWPIERCE@AMHERST) /

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

From: Thomas Richard Stevenson <[e-mail]>
Subject: Re: What full-screen file managers are there?
Date: 17 Nov 92 23:35:49 GMT
To:       [e-mail]

[e-mail] (Arnaldo MANDEL) writes:

>In article <[e-mail]> [e-mail] (Pete Holsberg) writes:

>> In article <[e-mail]> [e-mail] (Paul Kramer) writes:
>> =Hello,
>> =
>> =I found while reading articles in this electronic conference, a
>> =reference to a public-domain program called utree.  It is full-screen
>> =file manager which allows you to perform file functions with single
>> =keystrokes.  For example, after I start the program I see a screen
>> =that lists the files on my account.  If I want to delete one of those
>> =files I move the cursor on top of it, press a certain key.  Boom it
>> =gone!
>> =
>> =Well I am wondering:  Is there anything better than utree?  I am aware
>> =of a program called 'maint' but it doesn't have the full functionality
>> =of 'utree'.

>> I prefer HDSCAN.

>Better go all the way and get emacs, with tree-dired.  You not only get all
>funtions most file managers execute, but you can add some of your own.  And
>then, this would be just the beginning of using emacs for almost everything,
>instead of hundreds of little frozen utilities.
>--
>............................................................
>Arnaldo Mandel   ---  [e-mail]
>Temporarily: [e-mail]

If you are going to install emacs and you have X-terminals, also think
about epoch. I just installed it today, and it's great!
--
     ____  __   __       [e-mail]
      /   /_/  /_        [e-mail]
     /.  /\ . __/.       [e-mail]
                         [e-mail]

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


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