Subject: Linux-Misc Digest #851
From: Digestifier <Linux-Misc-Request@senator-bedfellow.MIT.EDU>
To: Linux-Misc@senator-bedfellow.MIT.EDU
Reply-To: Linux-Misc@senator-bedfellow.MIT.EDU
Date:     Tue, 22 Mar 94 14:13:15 EST

Linux-Misc Digest #851, Volume #1                Tue, 22 Mar 94 14:13:15 EST

Contents:
  Re: STRAW POLL RESULT: Linux groups automonitoring (dan@oea.hacktic.nl)
  Re: Opinions wanted about SCO-unix (vs AIX/Linux). (merlin)
  Re: Effective rights in Linux/Unix (Paul Stravers)
  Re: STRAW POLL RESULT: Linux groups automonitoring (Mark Allender)
  Re: Mosaic 2.2 with FORMS? (Mitchum DSouza)
  Re: NEW PRODUCT : 3 Linux CD's and a T-Shi (Dan Tran)
  Re: Impressions: FreeBSD vs Linux (J Wunsch)
  SCSI Magneto-Optical Drive Supported? (Duncan C THOMSON)
  floating point benchmarks (486 vs. Pentium) (Klaus Liedl)
  Xconfig: truecolor/pseudocolor/staticcolor what does it mean ? (Enrico Scotoni)
  Re: STRAW POLL RESULT: Linux groups automonitoring (Mark Lord)
  Incorrect IRQ detect of mouse on boot (Douglas Donahue)
  Re: Impressions: FreeBSD vs Linux (Michael L. VanLoon)

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

Crossposted-To: news.groups
From: dan@oea.hacktic.nl
Subject: Re: STRAW POLL RESULT: Linux groups automonitoring
Date: Mon, 21 Mar 1994 21:42:03 GMT

Ian Jackson (iwj@cam-orl.co.uk) wrote:

: I think that this gives me a fairly clear mandate to go ahead; I'm not
: going to apply the overmajority rules that apply to group creation.

Gives you a mandate to do what? People don't take straw polls seriously
enough for them to count. If the status of the linux groups is to be
changed, it must done the same way other changes (splitting) are. And
aren't you being presumptous about giving yourself the right to apply
rules or not! 

This whole damn business about trying to control the linux groups is just
plain silly. Probably every c.o.l.* reader feels that the groups are over-
crowded but most people find personal solutions (a good newsreader comes
to mind) instead of trying to impose a solution on everybody.

Nobody has to read everything in c.o.l.*. Choose subjects you want to follow
and let others make their choices.

: Note that discussion of this probably belongs in news.groups;
: comp.os.linux.misc is crowded enough already.

No, it is not too crowded for this unless you want to pull a fast one.

-- 
|< Dan Naas        dan@oea.hacktic.nl >|
+--------------------------------------+

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

From: merlin@neuro.usc.edu (merlin)
Crossposted-To: comp.unix.advocacy,biz.sco.general,biz.sco.opendesktop
Subject: Re: Opinions wanted about SCO-unix (vs AIX/Linux).
Date: 22 Mar 1994 06:52:09 -0800

>The icc compiler seems a little flaky.  It doesn't seem to produce
>optimized code for the Pentium that is much of a win over cc.

Based on my own personal experience compiling hundreds of thousands 
of lines of locally developed and network distributed image processing,
numerical analysis, ray tracing, and scientific visualization source
code using rcc, cc, gcc, and icc under SCO ODT-3.0 I can assure you
that the icc compiler is not in the least little bit flakey.  

The icc compiler generates highly optimized code with a minimum of 
source code changes for typical network distributed sources (BRLCAD, 
IRIT/RayShade, NetPBM, KHOROS, etc).

The icc compiler is rock solid and, if used by someone with even the 
smallest clue about setting source code compilation flags and compiler 
optimizer options, the icc compiler generates code which typically 
ranges from 35 to 65 percent faster (on 386, 486 & Pentium processors)
than equivalent nonoptimized code generated by the current generation 
of SCO cc and GNU gcc compilers.

I invite anyone who has any question about the relative performance
of icc vs. cc or gcc code to compile your own copy of the extensive
US Army BRLCAD ray tracing benchmark code -- collect code execution
statistics for code compiled by rcc, cc, gcc, and icc -- collect a
set of statistics for each compiler type on each architecture (386,
486 & Pentium) -- then judge for yourself whether icc generates the
best code across the entire series of Intel processor architectures.
The BRLCAD benchmark is particularly appropriate because it permits
comparison of Intel processor performance with other parallel and
nonparallel desktop and minisupercomputer systems.

==============================================================================
Alexander-James Annala
Principal Investigator
Neuroscience Image Analysis Network
HEDCO Neuroscience Building, Fifth Floor
University of Southern California
University Park
Los Angeles, CA 90089-2520

Telephone:  (213)740-3406
FAX:        (213)740-5687
==============================================================================
-- 
AJ Annala, Principal Investigator, USC Neuroscience Image Analysis Network

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

From: stravers@cuba.gmd.de (Paul Stravers)
Subject: Re: Effective rights in Linux/Unix
Reply-To: Paul Stravers <paul.stravers@gmd.de>
Date: Tue, 22 Mar 1994 12:32:06 GMT

In article <1994Mar17.002518.3035@compiler.tdcnet.nl> jvoosten@compiler.tdcnet.nl (J.S. van Oosten) writes:

   What you're looking for is the (in)famous 's'-bit. Its an attribute just
   like the 'r', 'w' and 'x' bits on files. It's called the 'sticky' bit,
   because it "sticks" the EUID (effective userID) to the program while

No, it is called set-uid-bit or set-gid-bit, hence the ``s''. The ``sticky
bit'' hints the kernel that an executable better remain in the swap space,
*even* if no process is currently executing it. The idea is that this speeds up
the fork() system call for often used executables, because loading a
``sticky'' executable does not require access to the file system.

On HPUX, the sticky bit is shown as ``t'' with ls. I am not sure if Linux has a
sticky bit. With todays disk performance, I don't think there is much benefit
in implementing the ``sticky'' attribute.

Paul Stravers
-- 
Gesellschaft fuer Mathematik        |
und Datenverarbeitung i5-SET        |
                                    |
Schloss Birlinghoven                |  email paul.stravers@gmd.de
53754 Sankt Augustin                |  fax   +49 2241 142035
Deutschland                         |  phone +49 2241 142586
====================================|==============================
--
-- 
Gesellschaft fuer Mathematik        |
und Datenverarbeitung i5-SET        |
                                    |
Schloss Birlinghoven                |  email paul.stravers@gmd.de
53754 Sankt Augustin                |  fax   +49 2241 142035
Deutschland                         |  phone +49 2241 142586
====================================|==============================

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

Crossposted-To: news.groups
From: allender@caffeine.tucson.ibm.com (Mark Allender)
Subject: Re: STRAW POLL RESULT: Linux groups automonitoring
Date: Tue, 22 Mar 1994 15:13:09 GMT
Reply-To: allender@vnet.ibm.com

In article <Cn1Ay3.LII@oea.hacktic.nl>, dan@oea.hacktic.nl writes:
|> This whole damn business about trying to control the linux groups is just
|> plain silly. Probably every c.o.l.* reader feels that the groups are over-
|> crowded but most people find personal solutions (a good newsreader comes
|> to mind) instead of trying to impose a solution on everybody.

Yup.  A good newsreader does wonders.  Also, KILL files are an
excellent way to trim down those pesky, unwanted articles.  With a
good KILL file, I can remove probably 1/2 the stuff that I don't want
to read in the high traffic groups.

I personally don't understand why the c.o.l.* groups are any more
special that some of the other high traffic groups.  To my knowledge,
they aren't trying to automate their groups.  Why should it happen
here?

-- 
-Mark Allender                            < insert standard disclaimer here > 
-IBM -- SSD Tucson
-Device Driver Development
-allender@vnet.ibm.com

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

From: Mitchum DSouza <m.dsouza@mrc-apu.cam.ac.uk>
Subject: Re: Mosaic 2.2 with FORMS?
Date: 22 Mar 1994 11:11:49 -0500
Reply-To: m.dsouza@mrc-apu.cam.ac.uk

| Mitchum DSouza (m.dsouza@mrc-apu.cam.ac.uk) wrote:
| >You need to run libc.so.4.5.8 at the least to stop this messages. Also you
| >should upgrade your ld.so to version 1.4.3. If you do the latter you are able
| >to stop warning messages by setting the environment variable LD_NOWARN. See
| >the GCC-FAQ on sunsite.unc.edu /pub/linux/docs/faqs for more info.
| 
| However, this will not fix the forms problem.  I have libc 4.5.21,
| XFree 2.1, and kernel 1.0.2 on my system.  Forms are still broken.

I never said it would fix the problem - and in fact I do not use Mosaic on my
linux machine at all. I was merely answering the query as to how one could
stop the warning messages when using a version of Mosaic with shared libraries
less than what it expects. The original question you have convinently deleted
from the message.
 
| The only way to type into a form is by holding down all 3 mouse buttons
| while you do so.  :(  Somebody said there's a bug in the current
| Motif libs for linux.

I take your word for it.

Mitch

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

From: dant@minerva.rolm.com (Dan Tran)
Subject: Re: NEW PRODUCT : 3 Linux CD's and a T-Shi
Date: 22 Mar 1994 02:31:23 GMT
Reply-To: dant@minerva.rolm.com

> Can anyone explain how it costs $14.95 to ship 3 cd's?  It's not like there
> is any heavy documentation.  Maybe it's a *really really* thick t-shirt?

> --bob

I agree, the 3 cd + t-shirt + box < 2 pounds.  2-day priority mail costs only
$2.90.

-Dan





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

From: j@uriah.sax.de (J Wunsch)
Crossposted-To: comp.os.386bsd.misc
Subject: Re: Impressions: FreeBSD vs Linux
Date: 22 Mar 1994 12:41:45 +0100

elandal@tower.nullnet.fi (Ismo Peltonen) writes:

>       Brant Katkansky (brantk@atlas.com) wrote:

>> * FreeBSD has a more polished look and feel.  Linux definately looks and
>> feels like a beta product.  FreeBSD seems to have consistancy where
>> Linux does not.

Q:
>What do people mean with this (`looks and feels like a beta/not finished')?
>What in Linux makes that unfinished look'n'feel?

...
A:
> (I have hard time trying to keep
>up with updates - last time I got route-binary I noticed I'd better
>update my libs, which lead to downloading about 7 megs, some installing,
>some compiling, and cursing for not to having yet changed my system to
>conform to FSSTND)...

What you're describing there *is* the ``beta look'n feel''. Inacceptable
for a release. Not that FreeBSD doesn't need beta's or development -
but people getting a release are not suspected to run into those upgrade-
by-the-patch-of-the-day troubles.

With {Free,Net}BSD, you can easily live with one distribution on a fairly
stable basis. Unless you really need the new features of an upcoming
release, you might stay with your old one until the new stuff is out
of beta. (Though i'm running a FreeBSD-almost-current at home for
development purposes, i'm just sitting on a box that ran the 1.0-GAMMA(!)
version for quarter of a year, it was quite stable with average uptimes
of 14 days or more.)
-- 
cheers, J"org                             work:    joerg_wunsch@tcd-dresden.de
                                          private:   joerg_wunsch@uriah.sax.de
Steinbach's Guideline for Systems Programming:
        Never test for an error condition you don't know how to handle.

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

From: duncan@sunart.spd.eee.strathclyde.ac.uk (Duncan C THOMSON)
Subject: SCSI Magneto-Optical Drive Supported?
Date: 22 Mar 1994 16:06:44 GMT


We're looking to get an SCSI MO drive (probably about 1GB) on our
system.  Before I order it / install it, does anybody know if these
are supported by the Linux kernel, by patches, or would I have to
write a devide driver for it.  I know SCSI card drivers exist, but
which (if any) of the SCSI-CDROM, SCSI-tape and SCSI-generic drivers
work for MO drives.

I'd like to know before I install it on the Linux NFS server, or the
other PC which occasionally runs DOS...

Great job Linus, by the way!

        -Duncan

--
Duncan C Thomson, Signal Processing Division  | .__ _.._ .__._  _ , .___ _  
University of Strathclyde, Glasgow G1 1XW, UK | |_ (_ |_)|_ |_)/_\|\| | / \ 
+44 41 552 4400 X2205    Fax: +44 41 552 2487 | |__._)|  |__| \| || | | \_/ 
     <duncan@spd.eee.strathclyde.ac.uk>       | -==- lingvo internacia -==- 

--
Duncan C Thomson, Signal Processing Division  | .__ _.._ .__._  _ , .___ _  
University of Strathclyde, Glasgow G1 1XW, UK | |_ (_ |_)|_ |_)/_\|\| | / \ 
+44 41 552 4400 X2205    Fax: +44 41 552 2487 | |__._)|  |__| \| || | | \_/ 
     <duncan@spd.eee.strathclyde.ac.uk>       | -==- lingvo internacia -==- 

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

From: krl@myhost.subdomain.domain (Klaus Liedl)
Subject: floating point benchmarks (486 vs. Pentium)
Date: 21 Mar 1994 22:08:02 GMT
Reply-To: Klaus.Liedl@uibk.ac.at

Has anybody compared floating-point performance of 486 vs. Pentium using f2c 
on a Linux box ?
Especially Benchmarks as Linpack, Whetstone and SPEC would be very 
interesting.


Thanks

Klaus
--
(Klaus.Liedl@uibk.ac.at)
======================================
LinuX the choice of the GNU-generation

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

Date: Tue, 22 Mar 1994 06:40:00 MET
From: scoti@p46.keru.chg.imp.com (Enrico Scotoni)
Subject: Xconfig: truecolor/pseudocolor/staticcolor what does it mean ?

Hi all,

I searched through almost every X... man page and the other awailable doc's for
XFree86 but I did not find the answer to my question:

In Xconfig you can set options like `truecolor' `pseudocolor' `staticcolor' and
also some `greyscale' etc. options. Well although the grey-options seem to be
clear (I loose color display) I would like to have an explanation of the color
options, what they exactly mean and when to set which. Especially I was
THINKING I had to set truecolor because I have a ET4000 chip and a 24bit DAC
(whatever this means) and the manual of the card says: "The card supports true
colors with 32K colors (TARGA compatible) 64K colors (XGA compatible) and 16M
colors in 640x480". But it turned out to be the wrong setting (displaying some
scanned gif picture did not give nice pictures, when truecolor was set).

Can somebody please enlighten me about all that technical stuff please ?

Sorry, that I post this on this news-group, but I have no access to the right
(X-related) newsgroup and I don't know whom to ask directely (at least my
question isn't god-related, devil-related or pronounceation-related and not
cross-posted to all the groups) :-))))

Thanks for your kind help in advance

Regards

Enrico.

---

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

From: mlord@bnr.ca (Mark Lord)
Subject: Re: STRAW POLL RESULT: Linux groups automonitoring
Date: 22 Mar 1994 16:43:57 GMT


In article <2mk21e$3dd@lyra.csx.cam.ac.uk> iwj@cam-orl.co.uk writes:
...
Yet another net.facist is born.
...
>I think that this gives me a fairly clear mandate to go ahead; I'm not
>going to apply the overmajority rules that apply to group creation.

Really?  I came to exactly the opposite conclusion!
>
>As I said during the discussion period, I've been convinced that
>Subject line tags are more technically feasible than Keywords at the
>moment (shame on you, newsreader authors).

Sure, let's change the whole way usenet works for the sake of this shmuck.

>I'll therefore arrange for the monitoring program to expect Subject
>lines of the form
>  Subject: [keyword] rest of subject

A real pity this scheme ain't good enough for his own posts.

>Note that discussion of this probably belongs in news.groups;

Logical.  Since maybe nobody will notice it until "too late" there. :)

An "unmoderated" newsgroup is exactly that:  unmoderated.

Get a life!
-- 
mlord@bnr.ca    Mark Lord       BNR Ottawa,Canada       613-763-7482

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

From: odoncaoa@panix.com (Douglas Donahue)
Crossposted-To: comp.os.linux.admin,alt.uu.comp.os.linux.questions
Subject: Incorrect IRQ detect of mouse on boot
Date: 22 Mar 1994 11:50:37 -0500


Hello again,

The linux boot kernel that was placed on the linux boot disk  as the result of
the install on my system now looks for the mouse at IRQ 11. The mouse is set
to use IRQ 5 and the (GUS) sound board is set for IRQ 11 & 2. Is there a way
to modify the booting kernel (ala 'rdev') so that it finds the IRQ for the
mouse at the proper setting? The error message that I get on  boot looks
something like the following:

        Probing 82C710 mouse port device.
        mcd init failed: no mcd device at 0x300 IRQ 11

Upon reflection, I think the when I boot the Base system disks, they don't have
this problem. It doesn't seem to find the mouse at all? Now that I look at the
'SetConfig' script that gets executed at installation time, I see that there
are some mouse manipulations that take place. However, the manipulations take
place after the boot disk gets cut.

Please respond to me by mail as I am not able to monitor the newsgroup.

Cheers,

Doug 


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

From: michaelv@iastate.edu (Michael L. VanLoon)
Crossposted-To: comp.os.386bsd.misc
Subject: Re: Impressions: FreeBSD vs Linux
Date: 22 Mar 94 16:38:00 GMT

In <Cn1KJ1.9pr@boulder.parcplace.com> imp@boulder.parcplace.com (Warner Losh) writes:

>In article <CMzw69.92K@tower.nullnet.fi> Ismo.Peltonen@tower.NullNet.FI writes:

>>> * FreeBSD has a more polished look and feel.  Linux definately looks and
>>> feels like a beta product.  FreeBSD seems to have consistancy where
>>> Linux does not.

>>What do people mean with this (`looks and feels like a beta/not finished')?
>>What in Linux makes that unfinished look'n'feel?

>From my point of view it is the building of a system.  On FreeBSD, all
>I type is "make world," then go out for the night.  When I come back,
>all my user level utilities have been build and installed (in addition
>to libraries, include files, etc).  For Linux I must have missed
>something because I've never seen a source distribution I could do
>this with (feel free to prove me wrong).  This is due, I think, to the
>fact that there is exactly one core distribution and an central group
>running the show that is responsible (as a group) for the entire
>system.
[...]
>Finally, I can get the latest sources to FreeBSD every night and
>rebuild w/minimal effort, since there is one place for the sources for
>the entire system.  I just sup new sources, and type make and I'm off.
>I usually get and install new sources about once a week, however,
>because a build does take quite a while.
[...]
>I've also seen various nits wrt files and file placement on Linux that
>may have gone away.
[...]

>P.S.  the usual disclaimer about FreeBSD v NetBSD: They are likely the
>same, but I haven't used NetBSD and I indent it no slight by my
>comments.

Everything you've stated does apply exactly the same to NetBSD as it
does FreeBSD -- with one addition.  You mentioned the nits about weird
locations of files on Linux, and that being better on FreeBSD.  NetBSD
not only has the cleanness of FreeBSD for the main system, but also
has made significant effort to clean up the machine-specific parts of
the sources and move them into architecture-specific sub-trees.

All common parts of the the kernle, libraries, etc. are in the normal
locations, and the machine-specific parts have their own sub-trees.
The build scripts automagically use the right machine-specific
sub-trees to build your system with so you could have one source
repository mounted on several different architecture machines and they
could all build from the one location without conflicts.


-- 
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 Michael L. VanLoon                 Iowa State University Computation Center
    michaelv@iastate.edu                    Project Vincent Systems Staff
  Free your mind and your machine -- NetBSD free Un*x for PC/Mac/Amiga/etc.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

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


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: Linux-Misc-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.misc) via:

    Internet: Linux-Misc@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Misc Digest
******************************
