NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/unix/question/13766

Internet Message Format | 1992-11-22 | 85.6 KB

Path: sparky!uunet!stanford.edu!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 16:46:27 GMT
Sender: [e-mail]
Lines: 2569


 ----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: 21 Nov 92 14:05:00 EST
From: [e-mail]
Subject: INFO-UNIX Digest  V17#003
To: "rubel" <[e-mail]>
cc: "slosser" <[e-mail]>

Return-Path: <[e-mail]>
Received: from SEM.BRL.MIL by ntsc-rd.navy.mil with SMTP ;
          Sat, 21 Nov 92 13:59:48 EST
Received: from SEM.BRL.MIL by SEM.BRL.MIL id ab23909; 21 Nov 92 6:05 EST
Received: from sem.brl.mil by SEM.BRL.MIL id aa23746; 21 Nov 92 5:50 EST
Date:       Sat, 21 Nov 92 10:50:00 EST
From:       The Moderator (Mike Muuss) <[e-mail]>
To:         [e-mail]
Reply-To:   [e-mail]
Subject:    INFO-UNIX Digest  V17#003
Message-ID:  <[e-mail]>

INFO-UNIX Digest          Sat, 21 Nov 1992              V17#003

Today's Topics:
               Fastest way to transfer whole filesystems?
                           Re: IS UNIX DEAD?
                             IS UNIX DEAD?
                           Re: IS UNIX DEAD?
                        Re: IS UNIX DEAD? (long)
                 THE SOURCE TO NAMEDP, A FINGER PROGRAM
                     Serious question - please read
                        cron not finding script
                      Re: cron not finding script
                            PRINTCAP needed
               Change Modification Date of Symbolic Link
                       Reliable signals in Unix ?
                       Running "dd" from .profile
                     Re: Running "dd" from .profile
           Proper way to read/write a struct from/to a socket
         Re: Proper way to read/write a struct from/to a socket
                     Need help with 'tar' command.
                     2nd time:How to remove files?
                   Re: 2nd time:How to remove files?
                          Ride-Share Software
                       Need help with "MAKEFILE"
                    MAXOptics Earasble optical drive
                      creeping featurism reference
                 How do I make dynamic link libraries?
                        Re: IS UNIX DEAD (long)
                        Mail - signature files.
       Re: Does anyone have any good insulting acronyms for VMS?
            Re: Bourne Shell Programming Considered Harmful
           Newmbox-length of mbox  != spool mailbox length!!
          Re: Newmbox-length of mbox != spool mailbox length!!
                  Help with generating random numbers
                            Re: RFCs via FTP
                       HELP: tech terms reference
                                  grep
                                Re: grep
                            Re: Vi-question
   mmap() with shared access fails over NFS using AIX 3.2 on RS 6000
                               PS command
         (Q) SLIP, Xylogics Annex boxes, MacX, MacTCP, MacSLIP
                             UUENCODE help!
                  Printing within Uniplex spreadsheet
-----------------------------------------------------------------

From: Susan Cole <[e-mail]>
Subject: Fastest way to transfer whole filesystems?
Date: 8 Nov 92 12:30:16 GMT
Followup-To: comp.unix.admin
To:       [e-mail]

What's the fastest way to transfer an entire filesystem from one
disk partition to another (different type disk, larger partition)?
I thought that a tar would be much faster than a dump and restore
from tape, but when I tried it, using a command like:

    tar cf - . | ( cd /otherfs; tar xpf - )

I got a surprising slow rate, something like 2 minutes per megabyte.  Am I
using the wrong command, or using the right command the wrong way, or is this
just an inevitably slow procedure?  What I'm moving is the news partition,
and I've already learned that restoring it from dump tape is horribly slow.
The system is a Sun SPARC machine.

Thanks for any assistance.

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

From: Budi Rahardjo <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 10 Nov 92 02:26:45 GMT

 .hacktic.nl>
Sender: [e-mail]
Nntp-Posting-Host: antares.cc.umanitoba.ca
To:       [e-mail]

[e-mail] (Peter Busser) writes:

>[e-mail] (Budi Rahardjo) writes:

>>>in a similar way. Tell me, how do I do that with UNIX?
>>Well just buy a pre-configured Unix machine.

>That'll cost extra because it takes a few hours to do that.

Nope. Looak at UnixWorld (and other magazine) they sell 486s with

 ...
>You know how people learn using computers. But for some it's 'logical' and
>therefore easy and for others it's black magic and difficult. A good system
>should support people with all levels of knowledge. The Mac is a bad example,
>it only supports 'stupid' users.

I have different opinion. If somebody ask me what kind of computer to buy
I always ask what they want to do with it and how 'deep' they want to get
involve with the computer. Most of the time I suggested them to buy a Mac,
because it's a better computer for them (most of them just want to use it
for word processing or simple spreadsheet). The Mac will survive.
(Note: I wouldn't buy a Mac, I am a power cmdline user :-)

 ...
>>The problem with F3 approach is that F3 in one package is different with
>>other package. F1 may be help in one program, but it's Exit in WordPerfect.

>F1 is a de facto standard key under MS-DOS. I've used many programs and 90%
>has help (if any) under the F1 key.

Well, F1 in Word Perfect is different and most of the people I know here
use Word Perfect.

>>Try CRISP.

>I might, if I get it to compile...

No, I mean the newer (comercial) version. I heard it's much better.
I have used the old version and was impressed.

 ...
>>Ever run .EXE file compiled for 386 on 8086 ?

>You mean that BSD/386 is to System V.4 what the 386 is to the 8086??? Or the
>other way around??? It's obvious that a 386 executable does not run on a 8086,
>but it really isn't that obvious that BSD/386 executables don't run under
>Sys V.4. UNIX is UNIX right? And a 8086 is not a 386 right?

Ofcourse executables for BSD/386 would be different from Sys V.4 .
They are different versions (one is BSD one is sys V). Just like running
Windows 3.1 programs under Windows 2.

Look, if Joe User doesn't want to learn a little, then stay away from UNIX
and buy a Mac instead. He'll be happier. I mean why would he need a UNIX
if all he wants to do is type a letter.

-- budi
PS: I heard Sun is going to announce "something" tomorrow ?
    What is it going to be ?

--
Budi Rahardjo <[e-mail]>
Unix Support - Computer Services - University of Manitoba

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

From: "Hahn;Harley;;;;HAHN;personal;930630;" <[e-mail]>
Subject: IS UNIX DEAD?
Date: 10 Nov 92 06:01:43 GMT
Sender: [e-mail]
To:       [e-mail]


This is number 10 in a series of 22 responses to the question:

What do you think about the Byte magazine cover that asked:
IS UNIX DEAD?

(moderated by Harley Hahn)

 ----------
From: [e-mail] (Dan Busarow)

I sincerely doubt it.  Microsoft and IBM both have a huge stake
in seeing their respective OS take over the market, and I think
that for the average DOS user Microsoft has already won the
battle with Windows, these users do not need NT nor is it
targeted at them. >From my brief experience with OS/2, I feel
that the only people who will use it are big IBM shops who like
the warm fuzzys they get by keeping the shop true blue.

The hype that this battle between IBM and Microsoft has generated
would lead one to believe that nothing currently exists to supply
the role of a true multi-tasking OS on Intel machines.  This is
of course, untrue.  There are many Unix ports running on this
hardware already and there have been for several years.  The
marketing people at both companies have had to invent a need in
order to sell their product.

Unix already satisfies need which these two OS's purport to
solve, multi-tasking, networking and GUI's are all available and
Unix also throws in multi-user which these two seem to have
forgotten about.

We sell vertical market software for Unix systems and have seem a
continual growth in interest in Unix over the past several years.
This growth in interest is not slowing down, in fact it seems to
be increasing as Open Systems awareness comes to more people.

This (Open Systems) is where Unix has always shined and always
will. Sure, Microsoft may port to MIPS and Alpha chips as well as
Intel but three processors out of the hundreds out there hardly
qualifies. I haven't heard about IBM's plans for OS/2 in this
regard but I don't see the Closed Systems giant licensing there
product to other vendors at a reasonable price as a very likely
event.  Have you seen what the are asking for the APPC license?
100,00K and up to license a networking protocol!  Amazing.

While USL and the Unix vendors most definitely want and need to
make money from their products I do not think that they have the
megalomania of either Microsoft or IBM.  Therefore we will
continue to see advances in Unix coming out of small and not so
small companies who are dedicated to continuing the growth of the
Unix market.

I've rambled a bit, again the key to me is that Microsoft and IBM
are trying to convince the world that they have THE solution for
all our computing problems when in fact, they have just given us
a couple more OS's that do what Unix ALREADY does.

==========


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

From: "Marc G. Frank" <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 10 Nov 92 21:20:13 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Paul Prescod) writes:
>Why is it, in VI if I try to save a file without a name it doesn't prompt
>me for a name?  Instead it kicks me back to vi.  Or if I try to quit without
>saving.  Why doesn't it ask me "Save y/n."  There is no reason VI couldn't
>have an "expert" mode that got rid of messages for those that don't want
>them.

Because that's not how vi works.  But if you don't like it, get elvis
and *change* it.  That's what the source is for.  How come I can't quit
WordPerfect without it asking me if I'm really really really really sure
I want to quit, even though I clearly typed the quit key?  How come I
can't supress that behavior?  Where can I get the WP source so I can fix
it?

>Why is it, in VI that there is no way (I know of) to get help, or exit,
>without knowing the exit key?  What doesn't "h" in command mode present
>ANY kind of help?

Why can't I quit QEdit or MS-DOS 5 edit or WordPerfect or Windows
without knowing the exit key?  How come I can't get the computer to work
if I don't turn it on?  What the hell am I supposed to do if I can't
find the switch?  I could call tech support, but I don't know how to
dial a phone!

>Why is it, in RN that when I hit 'h' to ask about how to find a command,
>it doesn't give me a menu "message read commands","post commands" "misc
>commands."  Instead it makes me read two pages of (mostly useless) commands.

Because that way you don't have to waste keystrokes going through
useless menus when all the available options can be displayed on a
couple of pages, which you can abort as soon as you find the key you're
looking for.

>Why don't Unix "tools" have a convention about help and exit keys AT THE
>VERY LEAST.

UNIX "tools" aren't even supposed to be interactive.

>Where am I supposd to get help about VI?  The man pages tell me nothing
>except how to load it.

Well, if you're the bookish sort you could get Linda Lamb's excellent
book *Learning the vi Editor*, a Nutshell Handbook published by O'Reilly
& Associates.

If you're not, you could always ask someone.

>Why do I have 12 function keys on my keyboard that seem to do absolutely
>nothing?

Because you haven't bothered to program them yet?

>Why doesn't help do anything in SH (and others).

Answer 1: Because that's what "man" is for.
Answer 2: In bash:

11:51:32 mgfrank [1/2048] ~ $help if
if: if COMMANDS; then COMMANDS; [else COMMANDS;] fi
    `if' executes the `then' COMMANDS only if the final command in the
    `if' COMMANDS has an exit status of zero.
11:51:39 mgfrank [2/2049] ~ $

>Why doesn't help do anything in VI?

For the same reason that :q! doesn't exit the file without saving in
Microsoft Word.

>Why does VI default to beeping at you when you try to type as opposed to
>editing?

For the same reason that WordPerfect inserts characters when I try to
move the cursor back with h.

HINT: Because they aren't the same program.


>Why is it that if I want to WordProcess in VI I have to remember to hit return at the end of the line otherwise bizarre things happen.  Why not just wrap, beep, or slide over (i.e. let the line go long).

Because you didn't turn wordwrap on?

>Is there a wordwrap mode in VI?  And if so, why doesn't it kick in when I
>use VI from RN?

Because you didn't bother putting "set wrapmargin=8" in your .exrc?

--
Marc G. Frank                          Gee, Canada's sure looking good right
[e-mail]                    about now, isn't it?

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

From: Jeff Wasilko <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 11 Nov 92 04:02:06 GMT
X-Mailer: uAccess - Macintosh Release: 1.6v1
To:       [e-mail]


In article <[e-mail]> (comp.unix.questions), [e-mail] (Ade Barkah) writes:
> [e-mail] (Paul Prescod) writes:
> : [discussion about people should have CD-ROMs deleted]
>
> : BUT THEY HAVEN'T.  SO WHO CARES!  Sure the world would be a better place
> : if everyone had CD-ROMs.  But if they *DON'T* then selling an OS only
> : on CD-ROM is sorta silly, isn't it!
>
> Well, that's a weak argument. Just as you wouldn't expect the Grolier
> Encyclopaedia to ship on floppies, one shouldn't expect that all versions
> of Unix are available on floppies. It may not be practical. To give
> you an example, (if I remember correctly) the complete info-explorer
> manual pages for the IBM RS/6000 & the AIX operating system _alone_ is
> over 200 megabytes, and is shipped on CD-ROMs. And we haven't touched the
> operating system code. Who wants *that* on floppies ?
>

Sun now ships all of their servers with CD-ROM standard. You can't
buy the OS on anything other than CD-ROM.

Sun has also put their full documentation on CD-ROM--it's called Answerbook.
It includes Postscript versions of all of the standard man pages,
plus the full paper documentation set (normally 8 or so volumes weighing
in at 80 or so pounds!). I also have hardware and configuration manuals
for most of the desktop systems. It completely fills a 600 meg CD.

All of these things are on-line, fully-indexed, cross-referenced and
hypertext-linked, and just a command away. Just today, I was looking
for info on how the sticky bit affects directories. I just fired up
the searcher and typed in 'sticky'. I had 3 hits, and found my answer
in just a couple of seconds. I could have never done this with paper
documentation.

Unix is a bit cryptic, and the sheer volume of documentation is overwhelming
at times. Having it all shrunk down to a fully indexed CD where I
have a prayer of finding what I want makes things a lot easier!

CD-ROMs could be the saviour of UNIX.....

Jeff

--
Jeff's Oasis at Home.                   Jeff can also be reached at work at:
                                                        [e-mail]
"A system admin's life is a sorry one.  The only advantage he has over
Emergency Room doctors is that malpractice suits are rare.  On the other
hand, ER doctors never have to deal with patients installing new versions
of their own innards!"  -Michael O'Brien

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 11 Nov 92 21:27:33 GMT
Sender: [e-mail]
To:       [e-mail]

>Yes, but when will it be available ? Solaris 2.0 will be ready by then.

I hope so, and I hope it's everything I dream it will be.  But if it is,
it won't be the unix that some people here know and love.  It may have some
ease-of-use enhancements.  It will probably be graphical.  It may "pander
to the lowest common denominator."  I'm crossing my fingers.

>If Joe User prefers a comercial program why not go with SCO or NeXTstep.
>
Because a NeXT costs >$5000.00 (for greyscale) and a PC with scos almost
as much.  SCO costs more then many systems for sale today.

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

From: Martin McCormick <[e-mail]>
Subject: Re: IS UNIX DEAD?
Date: 11 Nov 92 21:54:09 GMT
Sender: USENET News System <[e-mail]>
Nntp-Posting-Host: datacomm.ucc.okstate.edu
To:       [e-mail]

In article <[e-mail]> [e-mail] writes:
>Sun has also put their full documentation on CD-ROM--it's called Answerbook.
>It includes Postscript versions of all of the standard man pages,
>plus the full paper documentation set (normally 8 or so volumes weighing
>in at 80 or so pounds!). I also have hardware and configuration manuals
>for most of the desktop systems. It completely fills a 600 meg CD.

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.  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.

>
>All of these things are on-line, fully-indexed, cross-referenced and
>hypertext-linked, and just a command away. Just today, I was looking
>for info on how the sticky bit affects directories. I just fired up
>the searcher and typed in 'sticky'. I had 3 hits, and found my answer
>in just a couple of seconds. I could have never done this with paper
>documentation.

To do that with the system I use, you look through the directory until
you find a likely-looking title and then strip the Postscript from it.
Next, use "less's" text finding capability to look for the string.  It works,
but it is as slow as molasses in January, and you run the risk of missing
important information which might be in files whose names don't betray
the knowledge, within.
    As I have also said, before, I realize that not all formatting effects
could be duplicated on non-graphical terminals, but a good Postscript-to-ASCII
converter would simply do what it could with what it had.  Then, we could
all use the hypertext capabilities for indexing.

>
>Unix is a bit cryptic, and the sheer volume of documentation is overwhelming
>at times. Having it all shrunk down to a fully indexed CD where I
>have a prayer of finding what I want makes things a lot easier!
>
>CD-ROMs could be the saviour of UNIX.....

If handled right, they could be the answer to all kinds of problems.  So far,
for anybody needing non-standard output, they are just another headache.

This doesn't mean I disagree with anything said, here, or I hate Windows
or anything more than that we need to look long and hard at the slash-and-burn
approach which seems to have taken over the data I/O business.  We are creating
a Tower of Babyl which seems to be creating more and more specialized forms
of I/O rather than trying to create universal formats that have unlimited
utility.
    There are software solutions which are beginning to make Window formats
available to the blind, but my problem is that in order to use the "AnswerBook"
or any similarly formatted output, any user must have a graphics terminal
such as another Sun work station to access the data.  This is nice if you are
selling Suns, but quite limiting if you have a perfectly good PC/XT running
"Kermit" at home.

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

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

From: Budi Rahardjo <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 10 Nov 92 02:35:56 GMT
Sender: [e-mail]
Nntp-Posting-Host: antares.cc.umanitoba.ca
To:       [e-mail]

[e-mail] (Peter Busser) writes:

>[e-mail] (Budi Rahardjo) writes:

>>Why is that more intuitive than 'ls' ? I have to learn CAT ,D2 ...
>>Why is that important to have DIR. You can setup 'dir' as an alias for 'ls'.
>>ls -l is not as difficult as  DIR /W

>Not *because* it is ls, but because most people know DIR and don't know ls.

I am not so sure about that. If s/he knows about DIR then it's easier
to teach her/him to remember ls. I taught some people with DOS
background to use UNIX, in around 3 hours. They can edit a file (with 'pico',
don't bother teaching vi :-), use elm, print a file, do ftp, hytelnet, gopher,
(and some even try nn :-) and use X Window (Open Windows).

-- budi
--
Budi Rahardjo <[e-mail]>
Unix Support - Computer Services - University of Manitoba

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

From: Leslie Mikesell <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 11 Nov 92 05:31:21 GMT
To:       [e-mail]

In article <[e-mail]> [e-mail] (Paul Prescod) writes:
>Why is it, in VI if I try to save a file without a name it doesn't prompt
>me for a name?
[and many more similar questions]

The simple answer is that these programs were written in an era when
computers were expensive and impressive enough that people were expected
to read the manuals before touching them.  While I happen to like vi
(and the other user-unfriendly programs you mentioned), I agree that
times have changed.  No one starting to use computers now is going to
put up with the terse interfaces designed years ago and no one cares
anymore that it is more efficient to be terse.

>Why don't Unix "tools" have a convention about help and exit keys AT THE
>VERY LEAST.

This is a little different - the tool functionality kind of depends on
processing to end-of-file (so exit is normally ^D if you are typing
to them from a terminal). And since they are typically used in pipelines
an interactive style "help" wouldn't be available anyway.

Les Mikesell
  [e-mail]

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 11 Nov 92 21:22:03 GMT
Sender: [e-mail]
To:       [e-mail]

>Hey, why doesn't my toaster ask me for confirmation if I press the little
>bar without inserting any bread?

Try it, it will usually pop up after a few seconds.

>>Why is it, in VI that there is no way (I know of) to get help, or exit,
>>without knowing the exit key?  What doesn't "h" in command mode present
>>ANY kind of help?
>
>Why doesn't my hard drive realize that I didn't really want to delete
>that data?  Why don't I have a team of ergonomics specialists sitting
>at my desk doing all my typing by dictation?

Two different things.  Asking for a help key is really not as
"pampered" as you would make it out.  As far as the hard drive:
OS/2 and DOS both have reasonably good undelete procedures.  OS/2 just
copies the file to a temporary directory in the background.

>Again, vi is not a drop-in replacement for Microsoft Write.  It's
>an extremely powerful editor of text files.  If you want to whine,
>consider asking Microsoft why they don't have a command line in
>their editors.

Microsoft has a  *PROGRAMMING LANGUAGE* built into word.  I can't
imagine why someone would prefer a command line to keystrokes
and requester windows.
>
>>Why don't Unix "tools" have a convention about help and exit keys AT THE
>>VERY LEAST.
>
>Control-C works for me.  That, or Control-Z kill %1.

And assembly is a high level language..right.  And your programming language
of choice is binary.

>Your lack of attentiveness is simply not our problem.  The bottom of
>every man page I've seen about 'vi' contains a pointer to the actual
>vi manual.

Would you like me to mail you our man pages or will you take my word for it
that it says nothing about other man pages.

>>Why do I have 12 function keys on my keyboard that seem to do absolutely
>>nothing?
>
>dumb question is this?  Are you blaming your keyboard manufacturer's
>ills on Unix, or are you incoherently expecting that Unix assign some
>sort of meaning to every key assignable?

No it's not a dumb question.  There is no standard for help in the Unix
world.  The f keys are seldom used in unix programs.  Most terminals
have f keys.  It would seem this would be a good time to start convincing
people to use F1 for help, n'est ce pas?

>>Why doesn't VI use the jkl; keys for cursor movement, since your fingers
>>are already on them.  Or why not use keys somewhere far from them.  A
>>person can easily get confused.
>
>Whose fingers are already on them?  Not mine.  If you're having a hard
>time moving your fingers one key away, might I recommend not using
>a computer?  These things are dangerous.

It's got nothing to do with difficulty.  It has to do with intelligence.
With 95% of all keyboards having little bumps on the "home" key it makes
absolutely NO SENSE to use keys one character right of the home keys
as cursor keys.  None whatsoever.  Espeicially when h could be a help
key, instead of wasted on the right arrow, when J should be the right
arrow.

>>Why does VI default to beeping at you when you try to type as opposed to
>>editing?
>
>Because you're Making a Mistake, and vi appropriately tells you so.

Why is typing text into a text editor "a mistake."  VI should default
to typing mode.  It would 100% or more user friendly, as well as compatible
with the other 1000 programs out there that claim to be editors.

>Again, you're confused.  My vi, and every vi I have used, has dealt with
>too-long lines in a rational manner.  All my vis also have word-wrap,
>which you would discover if you took the time to read the man page and
>read the manual listed at the end of the man page.

Sigh.  See above.

>>Is there a wordwrap mode in VI?  And if so, why doesn't it kick in when I
>>use VI from RN?
>
>a) yes.  b) because your .exrc file is not set up correctly.

How often does someone want to edit news without word wrap?  Wouldn't that
make it logical to put word wrap in the standard .exrc file?

>Hey, I'm getting into the rhythm of this.  Let me try:
>
>Why is it that people who have no idea how to use Unix and who think
>that it should operate at the lowest common denominator keep asking
>stupid questions?

Because they are not stupid questions.  They are good questions.  Obviously
VI has flaws.  Obviously unix has flaws.  Everything has flaws.  It is
frightening the way many Unix users refuse to admit that Unix has flaws.

If you launch an editor from VI, that editor should be a wordprocessor.  If
it is vi, vi should be in wordprocess mode.  If it doesn't, that is a flaw.

If help does nothing at the command prompt.  That is a flaw.

If the cursor keys are illogical that is a flaw.

Anything can be improved if we discuss the flaws and discuss ways to improve
them.  If we treat those that notice flaws as blasphemers, we all end up
using 1970s text editors in 1992, and unix dies.

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 11 Nov 92 21:36:08 GMT
Sender: [e-mail]
To:       [e-mail]

>>Why doesn't VI use the jkl; keys for cursor movement, since your fingers
>>are already on them.  Or why not use keys somewhere far from them.  A
>>person can easily get confused.
>
>That is one of the nice features of vi.   The hjkl keys are where you
>normally place your fingers and in command mode h moves you left, j
>down, k up and l right.   I guess you just didn't try them.   And if
>your terminal is configured properly you can use the arrow keys too.

What touch typing course did you take?  asdfjkl;.  I practiced them
for hours.  That's why most keyboards have little bumps on j and f,
so you can find them without looking.  Does it make much sense to
"find them without looking" and then shift my fingers to move around?
Besides, why waste the h key.  It is very valuable...it should be help!

>Yes - I noticed that you let the words continue on a very long line.
>But since vi needs to be a program editor too, there are options of
>continuing long lines, or to have them wrap just like vi is doing for
>me right now.   set wrapmargin=5 will do nicely.

Does VI need to be program editor when I'm in RN?  Surely RN could set
VI to word wrap mode, as I am certain they work together often.

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD? (long)
Date: 11 Nov 92 21:46:05 GMT
Sender: [e-mail]
To:       [e-mail]

> BTW, I do think you all are overestimating the power of the "average
>user" when it comes to deciding the course of the PC world. The "average
>user" couldn't afford to buy a machine powerful enough to run GUIs when
>they first came out, but look where we went... Yeah, Joe Stupid may be a
>large part of it, but he ain't the only thing determining what the next
>operating system will be. (We all gotta keep in mind that not every
>"average user" is stupid to the core, as several of the "common user
>advocates" have suggested.)

Overestimating the power of the "average user?"  Underestimating his/her
intelligence?  Not me!

Look at the most popular operating system today: DOS.  A 1970s cpm rehash
with a 640k memory limit.  Guis took off when Joe Average could afford
to buy a gui based computer.  It's not that they are stupid, just poorly
informed.  Many, many people don't know there are opearting systems
other then DOS.  Those that do, usually crinkle their noses if you tell
them about Unix.  "I heard that has a brutal learning curve."

In other words, while "average users" are smart, they are not computer
literate, they don't know a good OS from a bad one, and they will not
use something with a reputation for user-unfriendliness.

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

From: Steve Franklin <[e-mail]>
Subject: THE SOURCE TO NAMEDP, A FINGER PROGRAM
Date: 10 Nov 92 14:07:09 GMT
Sender: USENET News <[e-mail]>
Nntp-Posting-Host: ug2.cs.dal.ca
To:       [e-mail]

Here's the source to NAMEDP. I'm not sure if it made it to the group already.
This is the 4th time I have posted it. I think I mucked up the last time.
 Please do not mail me asking for help, unless
a) i know you
b) You can make me laugh.
   Otherwise I have no time to help you install it.
steve

# This is a shell archive.  Remove anything before this line,
# then unpack it by saving it in a file and typing "sh file".
#
# Wrapped by Scratchy <napalm@molest> on Wed Jan 15 12:32:17 1992
#
# This archive contains:
#    README        MANIFEST    Makefile    logfinger
#    plan.c        plan.nr
#
# Existing files will not be overwritten.
# Error checking via wc(1) will be performed.
# Error checking via sum(1) will be performed.

LANG=""; export LANG
PATH=/bin:/usr/bin:$PATH; export PATH

if sum -r </dev/null >/dev/null 2>&1
then
    sumopt='-r'
else
    sumopt=''
fi

if test -f README
then
    echo Ok to overwrite existing file README\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f README
    if test -f README
    then
        echo Error: could not remove README, aborting
        exit 1
    fi
fi
echo x - README
cat >README <<'@EOF'
All I did was to put in a function that keeps track of the PID for
plan by writing it to a file called ".planpid" in your home directory.
Now you run plan in your .login, and, to kill the process when you log
out, put something like this in your .logout:

    kill `cat /home/mine/.planpid`
    rm -f /home/mine/.planpid

You have to remove the .planpid (or modify the code), because the
program will not run if a plan is already running, which it assumes
when it finds a .planpid file already there.

If someone wanted to run multiple plans, maybe one for a .signature
and another for a .plan, the save name for the PID could be changed,
and the program compiled twice (the binary is not that big, ~40K on
the HP 425's I used), or, if someone is really motivated, they could
add a command line argument for the save name.

Sure, my modification isn't very complex, but it does what I needed
it to do, and maybe someone else can use it.

Karen    ([e-mail])

****  Additions by Geoff Loker ****

I have modified Tony Rems' code for plan a bit to get rid of the
necessity for hard-coding in the file name to be set up as a FIFO.
The program will now accept an optional argument that specifies the
name of the FIFO to be used.  If that argument is not set, the default
FIFO is $HOME/.plan.

In order to use this program, your O/S needs to support named pipes.
You also need to link in getopt for the changes to work.

Any executable program can be set up to run when the specified FIFO is
opened, but don't forget that the program you specify to run is being
run under your userid.

****  Original README ****

Date: Thu, 11 Apr 91 14:09:10 MST
From: Jim Armstrong <[e-mail]>
Subject: RE: fingeree ...

> >About a month ago there was a sample program posted to this newsgroup that
> >set up a FIFO named pipe as your .plan file.  I modified the code to set up
> >a simple (perhaps naive) finger monitor for users on my machine.  The process
> do you still have the source for it? would you send it to me?

Here is the article which appeared in comp.unix.questions a while back.  It
contains generic code that will run any program you want whenever a certain
file is accessed (i.e. the .plan file in this case).  All you have to do is
compile it (it creates an executable called 'plan') and then to get it
running say 'plan a.out &' where a.out is some program you have written.
I don't have my program any more, but basically what I did was a ps au
within that program.  It used egrep to search for someone currently fingering
me and appended the output to a file.  To get really fancy you could then
read from that file to find out exactly who it is (all in the same program)
and print out a nice personal message to whoever is fingering you as part
of what looks to be your .plan file.  A couple of things to watch out for:
If the finger is remote, the ps au won't find anything.  Also, if two people
finger you at the same time you may run into trouble, but I didn't try
experimenting with this too much.  I also found it helpful to timestamp a
date to the file, too (I used localtime() for efficiency).  This makes it
easier to look back later at the file and see who's been fingering you and
when while you were not logged on.  It also helpful in debugging your program.

So set up this code and experiment with different programs.  Just be creative
and see what else you can do with it.  One idea I used for a while is making
it print a different quote each time.  The possibilites are endless.  Enjoy.

Jim


Article 31270 of comp.unix.questions:
From: [e-mail] (Tony Rems)
Newsgroups: comp.unix.questions
Subject: Re: Finger
Date: 22 Feb 91 02:44:17 GMT
Reply-To: [e-mail] (Tony Rems)
Organization: UniSoft Corporation -- UNIX R Us.

In article <[e-mail]> [e-mail] (Min-Zhi Shao) writes:
>
>    When I fingered our system administrator, I got the following result:
>
>_________________________________________________________________________
>Login name: gardella              In real life: Ed Gardella [CETS]
>Directory: /home/cets/gardella          Shell: /usr/local/bin/bash
>On since Feb 15 19:49:04 on ttyp1 from TSTEST.SEAS.UPEN
>14 minutes Idle Time
>No unread mail
>Project: System Administrator eniac.seas.upenn.edu
>Plan:
>         Meander about until something interesting comes along.
>
>Office: 154 Moore Building           Work Phone: [phone]
>                                     Home Phone: [phone]
>
>I have been fingered 3 times today
>_________________________________________________________________________
>
>the .plan file in his home directory looks like:
>
>prw-r--r--  1 gardella        0 Feb 15 23:48 /home/cets/gardella/.plan
>^

As you have found out by now, I'm sure, the p means that this is
a named pipe aka a FIFO.  If you'd like to do this yourself, here
is a little program I wrote to do it (see the comments at the
the beginning of the plan.c file for usage info):

Here's the shar of my plan program, just cut up until it says
"cut here", and then type 'sh filename' using whatever filename
you save it as.  If you use 'plan' it will get overwritten.

The code here should compile w/o any problems on any BSD machine,
I have tried it on a Sun, Vax 750, and Pyramid 90x.  It should
also work properly on any SVR4.0 machine.

The code is pretty heavily commented so it should be self
explanatory.

Note that you should put a -DFILENAME="your_home_dir/.plan"
to get it to put your path in, or you can just edit the
source and change the value of FILENAME permanently.

If you have any problems getting it compiled, just send me mail.

Enjoy.

-Tony

@EOF
set `sum $sumopt <README`; if test $1 -ne 65223
then
    echo ERROR: README checksum is $1 should be 65223
fi
set `wc -lwc <README`
if test $1$2$3 != 1339725668
then
    echo ERROR: wc results of README are $* should be 133 972 5668
fi

chmod 644 README

if test -f MANIFEST
then
    echo Ok to overwrite existing file MANIFEST\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f MANIFEST
    if test -f MANIFEST
    then
        echo Error: could not remove MANIFEST, aborting
        exit 1
    fi
fi
echo x - MANIFEST
cat >MANIFEST <<'@EOF'
MANIFEST - This file
Makefile - The makefile
README   - Original notes, notes for first set of changes, and notes
       for my changes
plan.c   - The program
plan.nr  - A man page
logfinger - a perl script included with the shar I got
@EOF
set `sum $sumopt <MANIFEST`; if test $1 -ne 19972
then
    echo ERROR: MANIFEST checksum is $1 should be 19972
fi
set `wc -lwc <MANIFEST`
if test $1$2$3 != 743233
then
    echo ERROR: wc results of MANIFEST are $* should be 7 43 233
fi

chmod 644 MANIFEST

if test -f Makefile
then
    echo Ok to overwrite existing file Makefile\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f Makefile
    if test -f Makefile
    then
        echo Error: could not remove Makefile, aborting
        exit 1
    fi
fi
echo x - Makefile
cat >Makefile <<'@EOF'
DEST          = /home/napalm/bin/fifo

EXTHDRS          = /usr/include/fcntl.h \
        /usr/include/signal.h \
        /usr/include/stdio.h \
        /usr/include/sys/fcntl.h \
        /usr/include/sys/file.h \
        /usr/include/sys/stat.h \
        /usr/include/sys/sysmacros.h \
        /usr/include/sys/sysmacros.h \
        /usr/include/sys/types.h \
        /usr/include/sys/types.h

HDRS          =

LDFLAGS          =

LIBS          =

LINKER          = cc

MAKEFILE      = Makefile

OBJS          = plan.o

PRINT          = pr

PROGRAM          = plan

SRCS          = plan.c

all:        $(PROGRAM)

$(PROGRAM):     $(OBJS) $(LIBS)
        @echo -n "Loading $(PROGRAM) ... "
        @$(LINKER) $(LDFLAGS) $(OBJS) $(LIBS) -o $(PROGRAM)
        @echo "done"

clean:;        @rm -f $(OBJS)

depend:;    @mkmf -f $(MAKEFILE) PROGRAM=$(PROGRAM) DEST=$(DEST)

index:;        @ctags -wx $(HDRS) $(SRCS)

install:    $(PROGRAM)
        @echo Installing $(PROGRAM) in $(DEST)
        @install -s $(PROGRAM) $(DEST)

print:;        @$(PRINT) $(HDRS) $(SRCS)

program:        $(PROGRAM)

tags:           $(HDRS) $(SRCS); @ctags $(HDRS) $(SRCS)

update:        $(DEST)/$(PROGRAM)

$(DEST)/$(PROGRAM): $(SRCS) $(LIBS) $(HDRS) $(EXTHDRS)
        @make -f $(MAKEFILE) DEST=$(DEST) install
###
plan.o: /usr/include/sys/types.h /usr/include/sys/sysmacros.h \
    /usr/include/sys/sysmacros.h /usr/include/sys/file.h \
    /usr/include/sys/fcntl.h /usr/include/sys/types.h \
    /usr/include/fcntl.h /usr/include/stdio.h /usr/include/sys/stat.h \
    /usr/include/signal.h
@EOF
set `sum $sumopt <Makefile`; if test $1 -ne 45642
then
    echo ERROR: Makefile checksum is $1 should be 45642
fi
set `wc -lwc <Makefile`
if test $1$2$3 != 641331401
then
    echo ERROR: wc results of Makefile are $* should be 64 133 1401
fi

chmod 644 Makefile

if test -f logfinger
then
    echo Ok to overwrite existing file logfinger\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f logfinger
    if test -f logfinger
    then
        echo Error: could not remove logfinger, aborting
        exit 1
    fi
fi
echo x - logfinger
cat >logfinger <<'@EOF'
#!/usr/bin/perl -- # -*-Perl-*-

$me = "geoff";
$logfile = "/prod_10/geoff/.fingerees";
$pscommand = "ps -auw";

open(PS, "$pscommand |");
@fingers = grep(/(f\s+$me)|(finger\s+$me)/, <PS>);
close(PS);

($num) = unpack("A9", `wc -l $logfile`); # =~ s/^\s*(\d+).*\n/\1/;
printf "I have been fingered %d time", $num;
print "s" if ($num - 1);
print " today\n";
close(STDOUT);

open(LOG, ">>$logfile");
foreach $line (@fingers) {
    ($user) = $line =~ /^\s*(\S+)/;
    print LOG "$user, at ", `date`;
}

print(LOG "somebody, at ", `date`) if (! ($#fingers + 1));
close(LOG);
@EOF
set `sum $sumopt <logfinger`; if test $1 -ne 41255
then
    echo ERROR: logfinger checksum is $1 should be 41255
fi
set `wc -lwc <logfinger`
if test $1$2$3 != 2478571
then
    echo ERROR: wc results of logfinger are $* should be 24 78 571
fi

chmod 755 logfinger

if test -f plan.c
then
    echo Ok to overwrite existing file plan.c\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f plan.c
    if test -f plan.c
    then
        echo Error: could not remove plan.c, aborting
        exit 1
    fi
fi
echo x - plan.c
cat >plan.c <<'@EOF'

/*    THIS IS THE UNPUBLISHED SOURCE CODE OF REMBO        */
/*    The copyright notice above does not evidence any       */
/*    actual or intended publication of such source code.    */
/*    So, use it if you like, but give me credit.        */


/*     Usage: plan [-f file_name] program_name            */


/*    Description:                    */

/*     This program takes the full pathname of an    */
/*     executable and runs it on a fifo in the     */
/*    user's home directory named .plan.  This    */
/*     way, when finger is executed, the output    */
/*     of the program goes to the fifo.        */

/*    Written by:  Tony Rems                 */

/*     Send bugs and flames to /dev/null or         */
/*     [e-mail]                 */

/*    Modifications:                    */

/*    September 1991                    */
/*           (by Geoff Loker [e-mail])    */
/*    Modified the program so that the path to the    */
/*    .plan file is not hardcoded in.  Now any number    */
/*    of users can use the program at the same time.    */
/*    I also modified the program to use an optional    */
/*    argument to specify which file to use.  The    */
/*    default file used is still the user's .plan,    */
/*    but this can now also be used to set up        */
/*    .signatures or any other file the user wants.    */

/*    Even more modifications                */

/*    January 1992                    */
/*          (by Karen Bruner [e-mail]) */
/*    Added pid_deal function, so people can stick    */
/*     the program in their .login, and then have it    */
/*    killed by their .logout.  Program will not    */
/*    run if a .planpid file, the file with the PID    */
/*    for plan, already exists in the user's home    */
/*    directory.                    */

#include <sys/types.h>
#include <sys/file.h>
#include <stdio.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <signal.h>
#include <string.h>

/* Defines */
#define PERMS 0666
#define USAGE "%s [-f file_name] program_name\n"

/* Function prototypes */
void sig_handler();
int pid_deal();

main (argc, argv)
int argc;
char *argv[];
{
        int c, fflg;
    char *file;
        extern char *optarg;
    extern int optind;
    int fd;
    int pid;
    int status;
    char *getenv(), *home, plan[256], *strcat(), *strcpy();
    int pid_check;

    fflg = c = 0;
    while ((c = getopt(argc, argv, "f:")) != EOF) {
      file = optarg;
      fflg++;
    }

/* Comment out the next line if you don't want to have PID recorded
   to file .planpid */

    pid_check = pid_deal(); /* check for .planpid, if none, write
                   .planpid */

/* Uncomment next line if you commented out previous line */

/*  pid_check = 1; */

  if (pid_check == 1)  /* execute remaining part of program if .planpid
              does not exist, i.e., no other plan process is
              running */
    {
    if (fflg)
      strcpy(plan, file);
    else {
      home = getenv("HOME");
      strcpy(plan, home);
      strcat(plan, "/.plan");
    }
/*    setenv("PLAN", plan, 1); */

    if ( argc != optind + 1 ) {
        fprintf (stderr, USAGE, argv[0]);
        exit(1);
    }  /* if */

/* Catch interrupts for cleanup */
    signal(SIGTERM, sig_handler);
    signal(SIGINT, sig_handler);
    signal(SIGHUP, sig_handler);

    unlink (plan);

/* Make the fifo */
    if ((mknod(plan, S_IFIFO | PERMS, 0)) < 0 ) {
        perror("mknod");
        exit(2);
    }  /* if */

    while (1) {
        if ((fd = open(plan, O_WRONLY)) < 0 ) {
            perror("open");
            exit(3);
        } /* if */

/* Once our open completes we know that someone else has
 * opened the FIFO for reading, so we can know run our
 * program on it.  So, we fork, exec our program and
 * wait for the child to complete.
 */
        switch (pid = fork()) {
            case -1:
                perror("fork");
                exit(4);
                break;
            case 0:
/* If we're in the child, we copy our fifo to stdout */
/* and exec the program given */
                dup2(fd, 1);
                execlp(argv[optind],argv[optind],(void *)NULL);
                perror("child returned");
                exit(5);
                break;
            default:
/* If we're in the parent, we close the pipe and wait */
                close(fd);
                while (wait(&status) != pid)
                    ;
                break;
        } /* switch */
        sleep(2);
        close(fd);
    } /* while */
    }  /* end of my if (pid_check... */

  else
    printf("plan already running\n");

} /* main */

void sig_handler()  /* cleanup */
{
    char *plan, *getenv();

    plan = getenv("PLAN");

    unlink(plan);
    exit(0);
}

int pid_deal()    /* function for recording pid and making sure process
           isn't already running */
{
  char savepid[100];    /* string for file name */
  FILE *sp;
  int checker;        /* return value:  0 if .planpid exists, and
               program shouldn't be run, 1 if not */
  char *home;

  home = getenv("HOME");        /* put save name for file */
  strcpy(savepid, home);        /* in savepid */
  strcat(savepid, "/.planpid");

  if ((sp = fopen(savepid, "r")) != NULL) /* test for existence of
                         .planpid by trying to open
                         the file for reading */
    checker = 0;     /* return a zero if read was successful, i.e.,
               file already exists */

  else
    checker = 1;    /* file doesn't exist, return a 1 to execute
               the rest of the program */
  fclose(sp);

  if (checker == 1)
    {
      sp = fopen(savepid, "w");
      fprintf(sp, "%d", getpid());     /* puts PID for plan into file */
      fclose(sp);
    }

  return checker;
}
@EOF
set `sum $sumopt <plan.c`; if test $1 -ne 51212
then
    echo ERROR: plan.c checksum is $1 should be 51212
fi
set `wc -lwc <plan.c`
if test $1$2$3 != 2088235034
then
    echo ERROR: wc results of plan.c are $* should be 208 823 5034
fi

chmod 644 plan.c

if test -f plan.nr
then
    echo Ok to overwrite existing file plan.nr\?
    read answer
    case "$answer" in
    [yY]*)    echo Proceeding;;
    *)    echo Aborting; exit 1;;
    esac
    rm -f plan.nr
    if test -f plan.nr
    then
        echo Error: could not remove plan.nr, aborting
        exit 1
    fi
fi
echo x - plan.nr
cat >plan.nr <<'@EOF'
 .TH PLAN L "\*(V)" "4BSD"
 .SH NAME
plan - run an executable when a specified FIFO is opened
 .SH SYNOPSIS
 .B plan
[
 .BR -f file
]
 .B executable
 .SH DESCRIPTION
 .I plan
takes the name of an executable program and runs it on a FIFO that
is specified on the command line or on a FIFO in the user's home directory
named .plan.  Whenever the FIFO is accessed, the output of the executable
program goes to the FIFO.  It also writes the PID of plan to a file, so
the process can be killed during logout, and it checks for the existence
of this file to make sure plan isn't already running.
 .SH OPTIONS
 .TP
 .B \-f file_name
The \fB\-f flag causes \fIplan\fR to set up the specified \fBfile_name\fR
as a FIFO rather than the default \fB$HOME/.plan\fR.
 .PP
Some sample uses of \fIplan\fR would be:
 .TP
 .B plan /usr/games/fortune &
This will display a randomly selected fortune as the contents of
your .plan file whenever you are fingered.
 .TP
 .B plan -f ~/.signature ~/bin/gensig &
This will run the program gensig whenever your .signature file is accessed,
allowing you to change your .signature whenever you post.
 .TP
 .B plan ~/bin/logfinger &
This will run the program logfinger whenever your .plan file is accessed,
and you can log all instances of people fingering you.
 .SH AUTHOR
Tony Rems ([e-mail])
 .PP
Modifications by Geoff Loker ([e-mail])
More modifications by Karen Bruner ([e-mail])
 .SH BUGS
Your system must support named pipes in order for this to work.
@EOF
set `sum $sumopt <plan.nr`; if test $1 -ne 57204
then
    echo ERROR: plan.nr checksum is $1 should be 57204
fi
set `wc -lwc <plan.nr`
if test $1$2$3 != 432551495
then
    echo ERROR: wc results of plan.nr are $* should be 43 255 1495
fi

chmod 644 plan.nr

exit 0
--

Wed Jan 15 12:50:58 PST 1992

Thirty-two footsteps leading to the room where the paint doesn't wanna dry.


--
aasdSteveFranklin-Subliminal Psychology Major.ks;dlasBlueJaysRULEkasdfeahsdbfl
sd;lfaswoq[eBuyMeAQuadra!!!mbnZMXCNdfsba;KdSPAMiuroqiyetIBMSuxiweuryth'ewr;mxn
qpuepriuPartyOneqtuj;,n.,xnc,kjasFlameMeNot!;lkj;lkgkjd;askElvisLivesjhfquweru
[e-mail],sk;t;lrut[[e-mail],Physics!eq3rwkh;oHA

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

From: MIGUEL BRAXTON FARAH <[e-mail]>
Subject: Serious question - please read
Date: 10 Nov 92 15:27:03 GMT
Sender: News Manager <[e-mail]>
X-Newsreader: Tin 1.1 PL5
To:       [e-mail]

Hi! I'd like to make a serious question:

some days ago, in my account appeared a directory called "??????" - and I
wasn't the owner. It was protected so I couldn't even erase it. Using
the quota command I determined that it had about 300Kb of things that were
occupying my disk space - I could not get rid of them. Three days after
that, the directory disappeared (obviously, the owner erased it). My
question is: how can I access to such a directory? I know now that one
should protect his home directory - but if it happens again (I suspect
of an administrator), how will I be able to take measures about it?

BTW, I'm using ULTRIX-32.
The "cd ??????" responded "ambiguous command".
I'm using Cshell.

Thank you in advance.

--
MIGUEL FARAH F.         |          (__)   "I like EBCDIC."     | Yjod od sm
[e-mail]   |          (oo)  /                     | rmvtuqyrf
                        |   /-------\/      "I do too."        | zrddshr.
#include <disclaimer.h> |  / |     ||                  \  (__) | Vtsvl oy smf
#include <flames.h>     | *  ||W---||                     oo ) | upi eoaa eom
#include <signature.h>  |    ||    ||                     |_/\ | noh qtovrd.

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

From: Thom Anthony McCarty <[e-mail]>
Subject: cron not finding script
Date: 10 Nov 92 17:28:34 GMT
To:       [e-mail]


I am executing a Bourne shell script, newsrunning, from within cron.
I have given the absolute filename (including the path) but cron says
it cannot find the script:


helpdesk% crontab -l
5,10,20,25,35,40,50,55 * 1-31 * 1-6 '/usr/lib/newsbin/input/newsrun'
30 8    1-31 *  1-5     '/usr/lib/newsbin/input/newsrunning off'
00 17   1-31 *  1-5     '/usr/lib/newsbin/input/newsrunning on'
#40 *    1-31 *  0-6     '/usr/lib/newsbin/batch/sendbatches'
59 0    1-31 *  0-6     '/usr/lib/newsbin/expire/doexpire'
10 8    1-31 *  0-6     '/usr/lib/newsbin/maint/newsdaily'
00 5,13,21      1-31 *  1-5     '/usr/lib/newsbin/maint/newswatch
| mail page'
1,31 *   1-31 *  1-5     '/usr/lib/newsbin/utils/create-desc'


Although the script works from when I'm logged on the "news" account,
I get the following mail when cron attempts to run the script:


From daemon Tue Nov 10 08:30:02 1992
Return-Path: <root>
Received: by helpdesk.inre.asu.edu (4.1/SMI-4.1)
        id AA09380; Tue, 10 Nov 92 08:30:02 MST
Date: Tue, 10 Nov 92 08:30:02 MST
From: root (Operator)
Message-Id: <[e-mail]>
To: news
Subject: Output from "cron" command
Status: R

Your "cron" job

    '/usr/lib/newsbin/input/newsrunning off'

produced the following output:

sh: /usr/lib/newsbin/input/newsrunning off: not found


And here's the kicker; I have given an absolute filename and cron
cannot find it:


helpdesk% ls -al /usr/lib/newsbin/input
total 144
drwxrwsr-x  2 news          512 Oct 29 10:53 .
drwxrwsr-x 10 news          512 Nov  5 10:08 ..
-rwxrwxr-x  1 news        24576 Sep 23 08:55 bdecode
-rwxrwxr-x  1 news        24576 Sep 23 08:55 c7decode
-rwx--x--x  1 news         1499 Sep 23 08:55 cunbatch
-rw-------  1 root        51740 Oct  7 10:32 mail.man
-rwxr-xr-x  1 news         3693 Nov  2 13:03 newsrun
-rwxr-xr-x  1 news          510 Sep 23 08:55 newsrunning
-rwsrwsr-x  1 news        32768 Sep 23 08:55 newsspool
-rwx--x--x  1 news          239 Sep 23 08:55 recenews
-rwx--x--x  1 news          279 Sep 23 08:55 recpnews
-rwx--x--x  1 news         1499 Sep 23 08:55 rnews

Any words of wisdom (or even plain and simple guesses) would be
greatly appreciated.

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

From: Henry Spencer <[e-mail]>
Subject: Re: cron not finding script
Date: 12 Nov 92 17:29:17 GMT
To:       [e-mail]

In article <[e-mail]> Thom Anthony McCarty <[e-mail]> writes:
>30 8    1-31 *  1-5     '/usr/lib/newsbin/input/newsrunning off'
> ...
>Your "cron" job
>
>    '/usr/lib/newsbin/input/newsrunning off'
>
>produced the following output:
>
>sh: /usr/lib/newsbin/input/newsrunning off: not found

Get rid of the quotes.  The shell is looking for a command named
"/usr/lib/newsbin/input/newsrunning off", and there is no such command.
There is a command "/usr/lib/newsbin/input/newsrunning", but that's not
what your cron is looking for, because those quotes make the command all
one word.
--
MS-DOS is the OS/360 of the 1980s.      | Henry Spencer @ U of Toronto Zoology
              -Hal W. Hardenbergh (1985)|  [e-mail]  utzoo!henry

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

From: Spencer Greene <[e-mail]>
Subject: PRINTCAP needed
Date: 11 Nov 92 07:05:48 GMT
Sender: Spencer Greene <[e-mail]>
To:       [e-mail]

OK, I've checked the comp.unix.questions faq, and searched through
comp.unix.admin, but I can't find any reference to what I thought
would be an obvious question: where can I get a "monster" printcap
file that will include entries for any printer my users may think
to connect and ask me for help with?

I'm *sure* this is in an faq somewhere, & I'm just looking in the
wrong place, so will someone please enlighten me?

thx,
--
- Spencer Greene                                          [phone] -
- Nimbus Technology Inc.                               [e-mail] -
- Santa Clara, California USA            {sgiblab,ub-gate}!c2tech!sgreene -
- "Patriotism is the last refuge of a scoundrel."    --Dr. Samuel Johnson -

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

From: Martin Berli <[e-mail]>
Subject: Change Modification Date of Symbolic Link
Date: 11 Nov 92 08:36:28 GMT
To:       [e-mail]

Is there a way to change the modification date of a symbolic
link? I would like to set it to the same date as the file it's
pointing to.

 -- Martin Berli

 ------------------------------------------------------------------------
Internet: [e-mail]  | SWITCH, Swiss Academic and Research Network
Phone   : [phone]   | Limmatquai 138, CH-8001 Zurich

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

From: Brian Day <[e-mail]>
Subject: Reliable signals in Unix ?
Keywords: signal
Date: 11 Nov 92 16:36:24 GMT
To:       [e-mail]

OK, time for a question from the brain-dead.

I have two Unix processes: one sends a burst of signals to the
other.  The receiving process misses several of these signals -
i.e.  they don't get 'queued up'.

I know I could set up a kluge with a semaphore to act as an up-down
counter, but I was hoping for a more elegant solution.  This is
on a Sparcstation 4 running SunOS 4.1.1.

Any suggestions ?

Many thanks,

bd


--
Brian Day                                       [e-mail]
New Technology, Inc.                            [phone]
Mission Operations Support Systems              Opinions are my own -
Marshall Space Flight Center, Huntspatch, AL    May be fatal if swallowed

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

From: steve scrivano <[e-mail]>
Subject: Running "dd" from .profile
Keywords: dd .profile
Date: 11 Nov 92 17:29:27 GMT
Sender: netnews admin account <[e-mail]>
To:       [e-mail]

Can anyone tell me how to write a script that could be placed in the user's
 .profile file that would check which /dev/tty port I am logged into and pass
that variable to a "dd" command line that is placed in the .profile file which
would then execute upon login.  Example .profile entries:

script to check for login terminal variable
dd if=file of=/dev/ttyvariable

Steve Scrivano
[e-mail]

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

From: Martin McCormick <[e-mail]>
Subject: Re: Running "dd" from .profile
Keywords: dd .profile
Date: 11 Nov 92 22:10:45 GMT
Sender: USENET News System <[e-mail]>
Nntp-Posting-Host: datacomm.ucc.okstate.edu
To:       [e-mail]

In article <[e-mail]> [e-mail] (steve scrivano) writes:
>Can anyone tell me how to write a script that could be placed in the user's
>.profile file that would check which /dev/tty port I am logged into and pass
>that variable to a "dd" command line that is placed in the .profile file which
>would then execute upon login.

If you want to keep things simple, you might pipe the output of the "who"
command through awk or nawk and locate the user by matching all of the
output lines with the "whoami" variable.  I am leaving this as an exercise
for the user, because there are a couple of nasty little problems which will
need attention before the script is bullet-proof.  You will need to resolve
what to do with a user who is already logged in to 15 or 16 other sessions
and has just logged in, again.
    Also, it would be helpful, later, to have this script as a separate
shell script so that it can be better understood, two years later, when
it needs modification.  Good luck.

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

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

From: Robert Earl <[e-mail]>
Subject: Re: Running "dd" from .profile
Keywords: dd .profile
Date: 12 Nov 92 09:38:00 GMT
To:       [e-mail]

[e-mail] (steve scrivano) writes:

| Can anyone tell me how to write a script that could be placed in the user's
| .profile file that would check which /dev/tty port I am logged into and pass
| that variable to a "dd" command line that is placed in the .profile file which
| would then execute upon login.  Example .profile entries:

| script to check for login terminal variable
| dd if=file of=/dev/ttyvariable

what about

    dd if=file of=`tty`

or the ever-popular:

    cat file

--
robert earl          |
[e-mail]       | "I got a bottle of tequila, baby, who needs friends?"
[e-mail] |   --Johnette Napolitano, Concrete Blonde

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

From: Ron Cecchini <[e-mail]>
Subject: Proper way to read/write a struct from/to a socket
Date: 11 Nov 92 17:39:55 GMT
Nntp-Posting-Host: gehrig.cs.rpi.edu
To:       [e-mail]

Hi all,

o read()/write() expect you to read/write a "char *buf"
o I have a variable "foo" which is of type "struct foo_type"
o I've been reading/writing "(char *)&foo"

It seems to be working.  Am I just lucky, or is this correct?
I've never tried reading into a structure like this before...

Ron

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

From: Kartik Subbarao <[e-mail]>
Subject: Re: Proper way to read/write a struct from/to a socket
Date: 12 Nov 92 15:11:14 GMT
Sender: USENET News System <[e-mail]>
Nntp-Posting-Host: tex.princeton.edu
To:       [e-mail]

In article <[e-mail]> [e-mail] (Ron Cecchini) writes:
>Hi all,
>
>o read()/write() expect you to read/write a "char *buf"
>o I have a variable "foo" which is of type "struct foo_type"
>o I've been reading/writing "(char *)&foo"
>
>It seems to be working.  Am I just lucky, or is this correct?
>I've never tried reading into a structure like this before...

Sure, it's fine. Think about it. The read() and write() calls transfer
some bytes from where you tell it to start. C stores the elements
in a struct contiguously in memory, so when you give the address of the struct
as the starting address, then say sizeof struct for the number of bytes to
be sent, and assuming the other end has the same byte-ordering scheme, then
you can read it right into another struct.

    -Kartik

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

From: hong vuong <[e-mail]>
Subject: Need help with 'tar' command.
Date: 11 Nov 92 18:19:50 GMT
Sender: [e-mail]
Nntp-Posting-Host: ucrengr
To:       [e-mail]

I am having a problem with the 'tar' command.  We archived some
directories,  tar cvf /dev/rst0 dir1 dir2.  But now that we are
trying to read it back,  tar xvf /dev/rst0, we got the following
error message:

tar: read error: I/O error
st0:   Error for command 'read', Error Level: 'Fatal'
       Block: 33620      File Number: 0
       Sense key:  Media Error

Is there a way to skip over this error and continuing on reading?
Please help.  My e-mail is [e-mail].  Thank in advance.

Mike.

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

From: michael <[e-mail]>
Subject: 2nd time:How to remove files?
Keywords: remove, unix, corefiles
Date: 11 Nov 92 18:41:21 GMT
Followup-To: comp.unix.questions
To:       [e-mail]


    I posted this article in comp.unix.questions with a narrow
distribution, so I am reposting with a wider range. If you see it
twice, sorry.

I keep getting files added to my working directory such as 'xerrors'
and '.nfs***'. If I put a line in my .login that says 'rm xerrors', a
'xerrors:no such file' message appears every time I log in. If I leave
the 'rm xerrors' line out of my .login, the 'xerrors' file starts to
show up in my root directory.  The '.nfs***' files appear whenever I
am working in the text editor, emacs in this case. I have uncommented
the line 'limit coredumpsize 0' in my .login.  Any suggestions? If it is
more convenient to email, that's fine.

        Thanks,
            Mike Puchalek

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

From: Josef Moellers <[e-mail]>
Subject: Re: 2nd time:How to remove files?
Keywords: remove, unix, corefiles
Date: 12 Nov 92 15:46:48 GMT
Sender: [e-mail]
NNTP-Posting-Host: uranium.sto.pdb.sni.de
To:       [e-mail]

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


>    I posted this article in comp.unix.questions with a narrow
>distribution, so I am reposting with a wider range. If you see it
>twice, sorry.

>I keep getting files added to my working directory such as 'xerrors'
>and '.nfs***'. If I put a line in my .login that says 'rm xerrors', a
>'xerrors:no such file' message appears every time I log in. If I leave
>the 'rm xerrors' line out of my .login, the 'xerrors' file starts to
>show up in my root directory.  The '.nfs***' files appear whenever I
>am working in the text editor, emacs in this case. I have uncommented
>the line 'limit coredumpsize 0' in my .login.  Any suggestions? If it is
>more convenient to email, that's fine.

I can't say anything about "xerrors", but as far as ".nfs***" is
concerned:
NFS is a stateless system. That means that neither the server, nor the
client can assume anything about the state of the other side.

One of the things the server is not allowed to know is the idea of "open
files" on the client's side. Therefore, whenever the client opens a file
on an NFS-mounted file system, and subsequently unlinks the file
_before_ closing it, the client cannot completely remove the entry from
the directory, as this might be the last reference to the file and the
server (who doesn't know the file is still open on the client's side)
will remove the file, freeing all disk blocks that are/were allocated to
the file. UNIX's notion is, however, that the file remains on disk
_until the file is closed for the last time_, i.e. until either the
process closes the file or the process dies (closing the file
implicitly).
In order to allow the file to remain open and still "free" the directory
entry (or at least allow the file's name to be re-used), the NFS client
silently renames the file to ".nfs<number>", where <number> is generated
automagically. When the process closes the file, the NFS client software
will detect this and send an "unlink" command to the server which will
then _really_ remove the file and free up all disk space.

This will also pop up when You "open" the last file in a directory, then
"unlink" it and then try to "rmdir" the directory it is in.
Under S5 or UFS file systems, this works OK: the directory is gone.
Under NFS, You'll get a "directory not empty" error, and when You look:
voila! It's empty! B-{)

Does this make sense to You?

I assume that "xerrors" is created by some X-windows product and removed
whenever the product terminates. So, why worry?
--
| Josef Moellers        | c/o Siemens Nixdorf Informationssysteme AG  |
|  USA: [e-mail]    | Abt. STO-XS 113       | Riemekestrasse   |
| !USA: [e-mail]    | Phone: [phone] | D-4790 Paderborn |

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

From: MARK GUM <[e-mail]>
Subject: Ride-Share Software
Date: 11 Nov 92 19:18:34 GMT
NNTP-Posting-Host: oavax.csuchico.edu
To:       [e-mail]

Has anyone heard of a Ride-Share software package that runs on xenix?  If
you have any information to share on this please E-Mail to
[e-mail]
Thanks.

**********************************************************
MARK GUM, CALIFORNIA STATE UNIVERSITY, CHICO
[e-mail]

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

From: Frank Pinto <[e-mail]>
Subject: Need help with "MAKEFILE"
Date: 11 Nov 92 19:54:27 GMT
Sender: USENET News System <[e-mail]>
To:       [e-mail]

Hello all

I'm new in the subject, therefore here it goes:
I've just 'ftp' some files from "digital", uncompressed
them, unarchived them,
and found out that there are about 20 files all together,
now i want to compile them bu i'm sopposed to
??????  Makefile ?????? them.
Is this some sort of arrengement I must performe in
a certain order to have the program compiled?
I have no manuals (I'm home)
and one person which could help me is in Jury Duty.

thanks in  advance

Frank Pinto

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

From: HOSSEIN <[e-mail]>
Subject: MAXOptics Earasble optical drive
Keywords: optical drive sco
Date: 11 Nov 92 20:32:05 GMT
Expires: Dec. 1, 1992
Sender: [e-mail]
NNTP-Posting-Host: ida.ces.cwru.edu
To:       [e-mail]

Hi there,
    We have a 486/33 running under SCO unix 3.2v4.0.  We are thinking of
adding an earasble optical drive to our system.  Does anybody have any
experience with Maxtor's MAXOptics earasable optical drive?
Is it supported under SCO unix 3.2v4.0?
Any other brand which is tested?

Any help is greatly appreciated( Please respond through Email),
Seyed.

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

From: Jon Beck <[e-mail]>
Subject: creeping featurism reference
Date: 11 Nov 92 20:37:32 GMT
Sender: [e-mail]
To:       [e-mail]

Can anyone give me a reference for the first use of this term to
describe unix programs? ("I've already forgotten more than you ever
knew..." Good grief, now I've forgotten who wrote *that*. Dylan?)

Thanks.
--
Jon Beck, SoRReL Project GRA
West Virginia University, Stat & CS
[e-mail]

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

From: Eric Tse <[e-mail]>
Subject: How do I make dynamic link libraries?
Date: 11 Nov 92 20:52:17 GMT
Sender: Mr News <[e-mail]>
X-Newsreader: Tin 1.1 PL5
To:       [e-mail]

Hi,
    I am new to dynamic-link libraries and wonder how I can
create one. For example, if I want to make f1.o f2.o into a
libtest.so.0.1, do I do:
    ld -Bdynamic f1.o f2.o -o libtest.so.0.1
?
    Then if I want to compile a programm a.out originally
using f1.o and f2.o, do I do:
    cc -Bdynamic -o a.out other1.o other2.o libtest.so.0.1
?
    If I want to execute a.out, do I have to
    setenv LD_LIBRARY_PATH /path_of_the_so
before I run a.out?

But it does not work! Any help? Please?

--
========================================================================
Eric Tse
[e-mail]
========================================================================

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

From: Paul Prescod <[e-mail]>
Subject: Re: IS UNIX DEAD (long)
Date: 11 Nov 92 20:56:17 GMT
Sender: [e-mail]
To:       [e-mail]

>>As far as OS/2 (and AmigaDOS, BTW) goes, only one file matching the
>>pattern (unix parlance) [Mm][Yy][Ff][Ii][Ll][Ee] can exist in a
>>directory.
>
>Geez, I can't count how often I saved a version of a file by typing
>    mv Driver.o Driver.O
>    mv source.c source.C

And what's so evil about Driver.bak, or driver.1,driver.2.  There are
100 ways to differentiate two files with close names, case sensitivity
is only one.

>My 2ct:
>UNIX doesn't have these "user friendlyness" of DOS, because it isn't DOS
>that's user friendly, it's the applications that were written for it.
>One OS is no more "user friendly" than the other.
>
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.

If people like you say: Let's see how we can 'layer' this ugly operating
system with nice applications, shells and help functions.

BTW, while I maintain that comparing unix to DOS is useless, DOS does
come with a quite user friendly (though poor for power-users) shell.

I have set up that shell many, many times, and people who couldn't
copy a file suddenly can back up their systems, copy, move, delete
and edit.  I, for one, hate the thing.  But I recognize it's value to
new users.  Unix could do the same thing.  Probably someone has written
it.

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

From: [e-mail]
Subject: Mail - signature files.
Date: 11 Nov 92 21:16:10 GMT
To:       [e-mail]

Is there a way that you can have your signature file automatically appended
to all your mail messages.


Thanks
Nitin Seth
[e-mail]

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

From: Brett Lymn <[e-mail]>
Subject: Re: Does anyone have any good insulting acronyms for VMS?
Date: 11 Nov 92 22:30:08 GMT
NNTP-Posting-Host: coolibah.awadi.com.au
To:       [e-mail]

>>>>> On 7 Nov 92 06:49:22 GMT, [e-mail] (Mark) said:
Mark> NNTP-Posting-Host: 150.203.76.2
Mark> Organisation: Wassat?!

Mark> [e-mail] (Elizabeth Haley) writes:

>Verifiablely Maniacal System
>Viciously Mangled System
>Vogon Made System
>Vacuous; moreover, stupid.
>Vaxes mostly Suck.

Mark> Said with a european accent:
Mark> Ve Make Shit

Mark> Vulgar Madman Society
Mark> Virgin Mind Screwer
Mark> Valium Mandatory System
Mark> Viscuous Menstral Slop
Mark> Violent Miscarriage Starter
Mark> Viable Monday Symbol
Mark> Very Moronic Setup

My favourite (from someone in GNU I think):

Vomit Making System


--
Brett Lymn

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

From: "John N. Stewart" <[e-mail]>
Subject: Re: Bourne Shell Programming Considered Harmful
Date: 11 Nov 92 23:20:31 GMT
Sender: News Administrator <[e-mail]>
To:       [e-mail]

In article <1992Nov5.225431*[e-mail]> [e-mail] writes:
>While I cannot say that I entirely enjoy the freehand Perl style, I
>certainly agree that it takes something more than basic *sh languages
>as we've gotten to know them to do "real" programming.  But since you
>mention Makefiles, some of the toughest programming tasks I've ever been
>through has actually been planting and maintaining pieces of Bourne shell
>language into ravel Make configs.
>
>Which really suggests that if someone could come up with a Perl based
>alternative to Make which could understand Perl language itself, the
>programming world should do little but rejoice.


This is a plug :) At LISA VI down in Long Beach, a group of us got
together (thanx to Paul Anderson) and started discussing software
configuration and installation issues. An email list formed, and was
announced over the "net" a couple of weeks back. If you are
interested, I attached information on how to subscribe.


--Ace







  This is to announce a new email list focusing on software
installation and maintenance issues, following the workshop talk of
LISA VI. Those who have received announcements from me already know
that they are on the list, those that haven't should contact me
directly please :)

  TO SUBSCRIBE:

        please send mail to: [e-mail]


  TO SEND TO THE LIST

        please send mail to: [e-mail]


Questions may be sent directly to me, and I will begin an FAQ. For
those of you who have not received the notes from the meeting (thanx
to Rob Slater for typing them in!!) please let me know and I will be
sure to send them directly. They have been posted accordingly to
comp.org.usenix

Cheers!

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

From: Ronda Hauben <[e-mail]>
Subject: Newmbox-length of mbox  != spool mailbox length!!
Date: 12 Nov 92 07:23:57 GMT
Sender: Usenet News Admin <[e-mail]>
Nntp-Posting-Host: yfn.ysu.edu
To:       [e-mail]


I have just found it impossible to get into my mailbox and
wonder if someone can explain what the message I am getting
means and what to do about it.

when I type mail I get a message that says
newmbox -length of mbox.! = spool mailbox length!!
exciting leaving mailbox in tact

Thanks.

Ronda
--
Ronda Hauben                    email address:[e-mail]
Amateur Computerist Newsletter              or
P.O. Box 4344                   [e-mail]
Dearborn, MI 48126

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

From: Jan-Piet Mens <[e-mail]>
Subject: Re: Newmbox-length of mbox != spool mailbox length!!
Date: 12 Nov 92 16:21:36 GMT
Followup-To: comp.mail.elm
To:       [e-mail]

Note  Followup-To!

In <[e-mail]> [e-mail] (Ronda Hauben) writes:

>when I type mail I get a message that says
>newmbox -length of mbox.! = spool mailbox length!!
>exciting leaving mailbox in tact

I'm looking for the problem myself. As a temporary solution, specify
the full pathname of your mailbox

        $ elm -f /usr/spool/mail/username

and if that doesn't work

        $ elm -f /usr/../usr/spool/mail/username
             ^^^^^^^^
This seems to be the only way to fool ELM... :-).

[ BTW, I have only noticed this problem since installing 2.4PL8 last night]

    -JP
--
    __  _____   __  __
   |  ||  _  \ |  \/  |    Logix GmbH                             [e-mail]
 __|  ||  ___/ |      |    Moritzstrasse 50, [phone]   [e-mail]
|_____||__|    |__||__|    D-6200 Wiesbaden  ...!uunet!mcsun!unido!logixwi!jpm

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

From: [e-mail]
Subject: Help with generating random numbers
Date: 12 Nov 92 09:12:43 GMT
To:       [e-mail]

I need help with generating random numbers. If I use the normail C routines
on UNIX for generating random numbers (random, rand, etc.) everytime the
program is executed, the sequence of random numers remains the same.

How can I fix this so that each time I execute that program, the initial (and
thus the following sequence of) random numbers is different? On Turbo C the
subroutine randomize() takes care of this. Is there such a routine available on
UNIX/ULTRIX?

Any help will be greatly appreciated. Thanks a lot.

Nitin Seth
[e-mail]

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

From: Samuel Ko <[e-mail]>
Subject: Re: RFCs via FTP
Date: 12 Nov 92 10:03:54 GMT
Sender: [e-mail]
To:       [e-mail]

[e-mail] (Steven J. Sobol) writes:

>I know how to receive the text of Internet RFC's via e-mail, but is there
>an FTP site which holds them?

ftp site: nic.ddn.mil (IP: 192.112.36.5)
directory: rfc (this one is sooooo obvious) ...

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

From: [e-mail]
Subject: HELP: tech terms reference
Date: 12 Nov 92 13:36:24 GMT
Sender: USENET News System <[e-mail]>
To:       [e-mail]

In article <[e-mail]>, [e-mail] (Michael Bradley) writes:
>
> Novice question: I want to flush a stream socket after
> a send call. Can do? Email answers please.
>
>  Michael Bradley
>

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

From: [e-mail]
Subject: grep
Date: 12 Nov 92 13:41:49 GMT
Sender: [e-mail]
To:       [e-mail]

A quick question I was asked this morning. "Why is grep so called?"

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

From: Michael Salmon <[e-mail]>
Subject: Re: grep
Date: 12 Nov 92 14:43:22 GMT
Sender: [e-mail]
Nntp-Posting-Host: eos6c02.ericsson.se
To:       [e-mail]

In article <[e-mail]>,
[e-mail] writes:
|> A quick question I was asked this morning. "Why is grep so called?"

It comes from an ed command:

g/re/p

i.e. globally search for a regular expression and print those lines
that match.

--

Michael Salmon

#include    <standard.disclaimer>
#include    <witty.saying>
#include    <fancy.pseudo.graphics>

Ericsson Telecom AB
Stockholm

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

From: Tom Parker <[e-mail]>
Subject: Re: Vi-question
Date: 12 Nov 92 18:05:34 GMT
Sender: USENET Maintenance <[e-mail]>
To:       [e-mail]

In article <[e-mail]> [e-mail] (S. Vemula) writes:
>Hi,
>
>     Sometime back, somebody mailed on how to use arrow keys in insert mode
>     also, to move around in the document.  I think, it is by mapping the
>     arrow keys.  Could somebody mail me as how i have to set up my keys.
>
>     thanks,
>     Shank Vemula


I do it with these lines in my .exrc file:

" Allow arrow keys to work in VI "text insert" mode
:map! ^[OA ^[ka
:map! ^[OB ^[ja
:map! ^[OC ^[la
:map! ^[OD ^[ha
--
+--------------------------------------------------------------------+
| Tom Parker             |  National Center for Atmospheric Research |
| [e-mail]  |  [phone]                           |
+--------------------------------------------------------------------+

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

From: 25235-gokhman <[e-mail]>
Subject: mmap() with shared access fails over NFS using AIX 3.2 on RS 6000
Keywords: memory mapped files over NFS - is there a problem for write ?
Date: 12 Nov 92 15:27:10 GMT
Sender: USENET System Software <[e-mail]>
To:       [e-mail]

Hello netters,

I am trying to do mmap-ing to a file accross NFS using the following
arguments:

 caddr_t pointer = mmap(0, file_length, PROT_READ | PROT_WRITE,
                MAP_FILE | MAP_VARIABLE | MAP_SHARED, fd, 0);

It appears to be successful. However, when I am trying to write
to the pointed region I get segmentation faults. Segmentation faults
do not necessarily occur right at the start of the region but also
in the middle and almost at the end at what seems to be page breaks,
e.g. location 4096 with 0-4095 bytes written successfully. The problem
does not occur when files are not NFS mounted. I am aware of the fact
that the earlier versions of mmap() under AIX 3.1.X were not allowing
writes over NFS, but my understanding is that AIX 3.2 and up do not
have this limitations - manuals at least do not mention it. Could it
be a router-related problem ? Help !!!

Following is the program that I use to examine this situation. Again,
it always failed over NFS and always worked locally on the same set
of files:

=========================== CUT HERE ==============================

                Thanx in advance,
                Ed Gokhman

#    include        <stdio.h>
#    include        <stdlib.h>

#    include        <sys/stat.h>
#    include        <sys/signal.h>
#    include        <fcntl.h>
#       include     <unistd.h>

#    include        <errno.h>

#    include        <sys/mman.h>
#       include         <sys/stat.h>
#       include         <sys/mode.h>
#       include         <sys/file.h>
#    include        <unistd.h>

char *ptr;

caddr_t pointer;

void bus_err_or_segment (int sig)
{
   fprintf (stderr, "Can not finish this test at byte=%ld", ptr - (char *)pointer);
}

main (int argc, char **argv)
{
    size_t    file_length;
    struct stat  statbuf;
    int    fd;

    if (argc < 2) {
      fprintf (stderr, "Usage: check <filename>\n");
      exit (1);
    }
    (void) signal ((int) SIGBUS, (void (*)(int)) bus_err_or_segment);
    (void) signal ((int) SIGSEGV, (void (*)(int)) bus_err_or_segment);

    if (!stat(argv[1], &statbuf))
    {
       if (!(file_length = statbuf . st_size))
       {
          fprintf (stderr, "%s - empty file\n", argv[1]);
          return;
       }
    }
    else
    {
          fprintf (stderr, "%s - does not exists\n", argv[1]);
          return;
    }

    fprintf (stderr, "Mapping %ld bytes in %s\n", file_length, argv[1]);

    if ((fd = open ( argv[1], O_RDWR, 0660 )) == -1)
    {
       fprintf(stderr, "Can not open %s with O_RDWR - %s\n", argv[1], strerror(errno));
       exit (1);
    }

    errno = 0;

    pointer = mmap(0, file_length, PROT_READ | PROT_WRITE,
        MAP_FILE | MAP_VARIABLE | MAP_SHARED, fd, 0);

    if (pointer == (caddr_t) -1 || errno)
    {
       fprintf(stderr, "Can not map %s - %s\n", argv[1], strerror(errno));
       return;
    }

    for (ptr = (char *)pointer; ptr < (char *)pointer + (int) file_length; ptr++) {
       char a = *ptr;
       *ptr = a;
    }
    munmap (pointer, file_length);
    fprintf(stderr, "Success for %s\n", argv[1]);
}

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

From: Karl Larson <[e-mail]>
Subject: PS command
Keywords: PS command
Date: 12 Nov 92 15:46:25 GMT
Sender: [e-mail]
Nntp-Posting-Host: mirage.rss.dl.nec.com
To:       [e-mail]

This could possibly be SUN specific, but does anyone know what it means when
you have a process name "COMMAND" surrounded by "()" ie. "(biod)".  I thought
that possibly it implied a child of init, but I ruled this out when I saw
"-ksh (ksh)".  This is not a problem, it just came up in conversation and I
did'nt know. Further, the manuals don't describe such a beast.

 -----------------------------------------------------------------------
Karl Larson                                 (email) [e-mail]
Software Engineer                           (home)  214-328-2563
NEC America - Richardson, Texas             (work)  214-907-4463
 -----------------------------------------------------------------------

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

From: Erik Hoel <[e-mail]>
Subject: (Q) SLIP, Xylogics Annex boxes, MacX, MacTCP, MacSLIP
Date: 12 Nov 92 16:08:31 GMT
Sender: [e-mail]
To:       [e-mail]

Basically, I would like to be able to display xterms on my Mac at home
while dialed into a Xylogics Annex box via a high speed modem and run SLIP,
etc.

On the Mac side, I have MacX 1.7, MacTCP 1.1.1, MacSLIP (a SLIP driver), and
a v.32bis modem. Various postings to the comp.sys.mac.comm newsgroup have
confirmed that this is everything that is necessary on the Mac end.

On the unix side, I am dialing into a Xylogics Annex box and rloging into
a DecStation. The Annex boxes have not been configured yet to run SLIP (it
is still undecided whether they _ever_ will). One of our system gurus has
built a SLIP kernel (under Ultrix 4.2) and provided me with a special login
that runs slattach.

At this point, I am able to login to my slip account via MacSLIP, but after
the standard login messages scroll by, I am promptly logged out. I have not
been able to determine whether the Annex boxes are blocking SLIP (do they
have this functionality?) or whether there is a problem elsewhere.

Does anyone have any experience running SLIP through an Annex box when the
Annex boxes have not been configured to do so (in the underhanded manner
that I have described ;-)?

Any pointers, etc. would be most appreciated.

Erik Hoel
[e-mail]

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

From: "jeffrey.n.jones" <[e-mail]>
Subject: UUENCODE help!
Date: 12 Nov 92 17:32:29 GMT
To:       [e-mail]

I am having a heck of time with uuencode. If I do the format that it calls
for, uuencode file remotefile, I get a screen full of ascii characters.
Rather then into remotefile. Then is seems after this happens uuencode
dies and won't do anything other then "begin 666 qlist.z" and that's it.
This happens on all files I try it on. I have had the same problem on
this maching, on my SUN, on a UNIX box and even on my PC using a this
uuencode. HELP!

Jeff
--
 Jeff Jones  AB6MB         |      If we do not succeed,
 [e-mail]   |      then we run the risk of failure.
 Infolinc BBS 415-778-5929 |      Vice Pres. Dan Quayle

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

From: Seamus Conlon <[e-mail]>
Subject: Printing within Uniplex spreadsheet
Date: 12 Nov 92 18:15:41 GMT
To:       [e-mail]


I have been asked by a novice Unix user to post a quesiton about Uniplex
to you Unix experts.

The question concerns the Uniplex spreadsheet.  The requirement is to
print from within a macro.  Apparently this does not work and when
a key tape is used, the print works during the recording of the macro
but not when it is being executed.

All sounds a bit strange to me but anyone got a solution?

Please e-mail ([e-mail]).

Tks.

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


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