Dec 132017
 
Linux news from the Internet.
File INTLX026.ZIP from The Programmer’s Corner in
Category UNIX Files
Linux news from the Internet.
File Name File Size Zip Size Zip Type
INTLX026.TXT 141958 49514 deflated

Download File INTLX026.ZIP Here

Contents of the INTLX026.TXT file


From: [email protected] (Mike Me)

Subject: Gcc problem with libm.a Solved!!!!
Message-ID: <[email protected]>
Originator: [email protected]
Sender: [email protected]
Organization: George Washington University
Date: Wed, 3 Jun 1992 16:19:01 GMT
Lines: 17



Well, the libm.a problem has been solved, thanx.


>I think this should be a new FAQ. It screwed me yesterday.
>rename libhard.a or libsoft.a to libm.a depending on whether you want
>software or hardware 387 emulation.
>
>--
>Jay Pfaffman, Computer Coordinator [email protected]
>Dwight-Englewood School
>315 E. Palisade Ave. 201-569-9500
>Englewood, NJ 07631-0489


Thanx Jay.


[next article]
From: [email protected] (Brian Edmonds)

Subject: How to use AF_UNIX sockets?
Message-ID: <[email protected]>
Keywords: linux, sockets, af_unix
Sender: [email protected] (Usenet News)
Organization: Computer Science, University of B.C., Vancouver, B.C., Canada
Date: Wed, 3 Jun 92 17:31:36 GMT
Lines: 8

I'm trying to port some internet software over to my pc for single
user testing (MOO), but am unsure of how AF_UNIX sockets work. I've
used AF_INET sockets many times, and understand how to create INET
connections. Could someone post a brief summary of how to set up a
AF_UNIX server, and how to connect to it with a client?

----------------------------------------------------------------------
Brian Edmonds (MSc CompSci) [email protected]


[next article]
From: [email protected] (Leslie Griffiths (Griff))

Subject: Re: SERIOUS SECURITY PROBLEM (I think)
Message-ID:
Date: 3 Jun 92 09:08:17 GMT
References: <[email protected]>
Sender: [email protected] (USENET News Owner)
Organization: Coventry Polytechnic, Coventry, UK
Lines: 28
Nntp-Posting-Host: cc_sysh

In article <[email protected]> [email protected] writes:
>I was just logged in as a regular user, and there was a file in my home
>directory that belongs to root, and I did an rm to it
> rm -f test.c
>and this is what I saw:
> rm: remove 'test.c', overriding mode 0644?
>to which I answered yes, and when i did an ls, the file was gone!!
>This didn't happen under .95, .95a, .12, .95c++, but it is happening
>under .96 (don't have the .96a yet)
>I think this is a security risk, but I am not an expert. ๐Ÿ˜‰
>I just wanted everyone to be aware that this can happen, and maybe
>it shouldn't?
>Jim
>[email protected]

If you have write (and execute) access on a directory, then you can delete
files within it, _whoever_ those files are owned by. It is a common error
to assume that the perms on the file or its ownership determine who can
remove it, but this is not the case - what is important are the perms on the
directory containing the file.

Cheers


--
Griff \\ [email protected]
^^^^^ :-= ^^^^^^^^^^^^^^^^
//


[next article]
From: [email protected] (Karl R. Buck)

Subject: Re: Installation problem with gcc2.12
Date: 3 Jun 1992 11:46:00 -0500
Organization: Kansas State University
Lines: 29
Message-ID: <[email protected]>
References: <[email protected]>
NNTP-Posting-Host: matt.ksu.ksu.edu

[email protected] (Heinz Christoph Schuerch) writes:

>I tried to install gcc 2.12; this failed with different .12 versions. Now I
[...]
> compiling and linking with the "-static"-switch
> -> ld: file or directory for libgcc.a
> if I have look, libgcc.a is in /usr/lib/static and is softlinked to
> /usr/lib/gcc-lib/i386-linux/2.1/Libgcc.a
[...]
> Afther all that ld complains about not finding "crt0.o".

I ended up just making a link from /usr/lib/gcc-lib/i386-linux/2.12c
to /usr/lib/gcc-lib/i386-linux/2.12 after changing the LIBDIR
environment variable in the ./inst2.x to point to the 2.12c directory.

I've almost got the kernel compiled with this version. I had to
comment out the declaration of execve in unistd.h and change all the
Makefiles so that they find the include files, but now it dies
right near the end of the make with the message:

gcc -static -Wall -O6 -fomit-frame-pointer -o tools/build.o tools/build.c

/usr/lib/static/libgcc.a(_builtin_new.o): Undefined symbol
___new_handler referenced from text segment

--
Karl Buck
Grain Marketing ---===PEROT FOR PRESIDENT===---
(913)776-2745


[next article]
From: [email protected] (David L Orman)

Subject: Anyone working on PPP?
Message-ID:
Sender: [email protected] (USENET News System)
Organization: Iowa State University, Ames IA
Date: Wed, 3 Jun 1992 15:23:21 GMT
Lines: 3

Well, I cant usr SLIP and I was wondering if anyone was working on PPP,
or (hopefully) already had it implemented. Any takers?



[next article]
From: [email protected] (Chris)

Subject: Paradise working at last !!
Message-ID:
Date: 3 Jun 92 16:11:15 GMT
Sender: [email protected] (USENET News Owner)
Organization: Coventry Polytechnic, Coventry, UK
Lines: 32
Nntp-Posting-Host: cc_sysh


HA HA !!
I got my Paradise 512k SVGA (WD90C30) working at last !!

Since many people mailed me asking to forward any results I got, here
is my Xconfig.....

----------------------------------------------------------------------
RGBPath "/usr/lib/X11/rgb"
FontPath "/usr/lib/X11/fonts/misc/"

Keyboard
AutoRepeat 500 5
ServerNumLock

Microsoft "/dev/ttys1"
BaudRate 1200
Emulate3Buttons

vga256
Chipset "pvga1"
Modes "800x600"
Clocks 36

ModeDB
# clock horzontal timing vertical timing
"800x600" 36 800 816 952 1056 600 608 610 633
----------------------------------------------------------------------

Hope this works for you ๐Ÿ™‚


-Chris


[next article]
From: [email protected] (I Reid)

Subject: Vgalib (non X graphics) questions?
Message-ID: <[email protected]>
Date: 3 Jun 92 15:03:10 GMT
Distribution: comp
Organization: Edinburgh University
Lines: 43

I was using X to provide the graphical component of some image analysis tools I
am writing at the moment but got fed up with the slow speed (crawling on a
386-33 with 8Mb of Ram) so fished out a copy of vgalib which I picked up ages
ago from banjo (I think). This used to work quite well with my old Oak svga
board, worked ok with a Paradise vga board that was temporarily in my machine
but didn't seem to work with my recently purchased ET4000 board with Sierra
HiColor DAC. I wasn't too bothered at the time since I was going to use X
(what a joke) but I have started looking at it again and have found that it
does actually work (with a few provisos). All that said, some questions....

1. Why does vgalib work properly with text modes up to 100 cols wide but not
with the 132 col modes? I would guess something to do with the character set
but haven't explored this yet.

2. Has anyone written anything beyond the reasonably standard vga stuff. 800
by 600 by 256 would be nice (and I could probably extend it to 800x600x32K).

3. More worryingly, what are the known bugs in this code 'cos I have a problem
that seems to stem from it. I have a 512x341 image that I have been displaying
in 640x480x16 colour mode by doing

vga_setmode (G640x480x16)
for (y = 0 to 341) {
for (x = 0 to 512) {
vgasetcolour (adjusted value of byte at (y * 512) + x)
vga_drawpixel (x, 480 - y);
}
}
vga_setmode (TEXT);

First time I run this code it is fine but if I call it again with the same
(unchanged) block of memory it will fail. The screen blanks and will not return
but keyboard input is still possible.

If the block of memory involved is smallish (i.e. less than about 50K) then the
image will redisplay but anything bigger causes almost guaranteed failure.
Nothing changes between the two calls so I'm at somewhat of a loss to explain
this behaviour. If I quit my program and restart then I can display my image
ok. Occasionally the character set is corrupted though.

Any ideas anyone?

Iain


[next article]
From: [email protected] (Hans B. Wiezorke)

Subject: what are the correct device numbers and names for ptys?
Date: 3 Jun 1992 15:18:52 GMT
Organization: University of Karlsruhe (FRG)
Lines: 25
Distribution: world
Message-ID:
NNTP-Posting-Host: irav23.ira.uka.de
X-News-Reader: VMS NEWS 1.22

hi all,

i just installed X and it works just ticketyboo... except that i can't
start an xterm :-(.

it seems to be missing ptys. at bootup linux says it has pty's, i just
don't have the correct device files. i tried 4/128 and 4/192 but
that didn't work.

what are the names and numbers for the pty devices?

thanks,

hans


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
I __ I Hans Bernhard Wiezorke (at Uni. of Karlsruhe) I
I /opus\ _____(@ \ I E-Mail : [email protected] I
I \sndo/ /. ! I or [email protected] (for large mails) I
I ------) ! I- - - - - - - - - - - - - - - - - - - - - - - - - -I
I X~~~~! I W A N T E D I
I- - - - - - - - - - - - -I Schroedinger's Cat I
I std. disclaimers apply I dead or alive I
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -


[next article]
From: [email protected] (Peter MacDonald)

Subject: X README.prog correction
Message-ID: <[email protected]>
Sender: [email protected]
Nntp-Posting-Host: sanjuan.uvic.ca
Organization: University of Victoria, Victoria, BC, CANADA
Date: Wed, 3 Jun 92 16:46:49 GMT
Lines: 34



In the X README.prog, Orest says in the requirements section:

o Gcc2.11c, available from a whole bunch of places. Later releases may or
may not be acceptable.


Unfortunately, later (or earlier for that matter) are not acceptable.
Shared libs are actually images, which are linked to eliminate all
unresolved references. In the case of libc, no problem, cause there
are no dependencies. But the X libs depend upon libc. So references
to strcpy and the like are resolved when the shared lib is built, to
the current shared libc. In this case that is 2.11c. If you try now
to compile against say 2.12b, then your image will load the libc_v2_12b,
but the X shared lib will try calling the entry points for 2.11c.

Shudder. Ker-chunk. ๐Ÿ™

You can try rebuilding the X shared libs, but that may be dangerous.
Alternatively, you could use

gcc -V 2.11c

But you have to have the libc link libs around. Maybe gcc should be
modified to detect which libc version X was using and try to
put in the -V 2.11c for your. Only Hlu can answer that one.

Peter
[email protected]






[next article]
From: [email protected] (Heinz Christoph Schuerch)

Subject: Installation problem with gcc2.12
Message-ID: <[email protected]>
Date: 3 Jun 92 09:36:49 GMT
Organization: Ascom Hasler AG, Berne, Switzerland
Lines: 32

I tried to install gcc 2.12; this failed with different .12 versions. Now I
saw, that people are refering to this version, so can anybody advice me how
to install.
The installations for the 2.11 versions went on smoothly; I had to untar the
two files and run "inst2.x".
I applied the same procedure for gcc 2.12 and 2.12c. The result was the
following:
- gcc 2.12
compiling and linking with shared libs
-> worked ok.
compiling and linking with the "-static"-switch
-> ld: file or directory for libgcc.a
if I have look, libgcc.a is in /usr/lib/static and is softlinked to
/usr/lib/gcc-lib/i386-linux/2.1/Libgcc.a

- gcc 2.12c with binutils from banjo.concert.net
compiling and linking with or without shared libs doesn't work.
First gcc complains about not finding cpp. Afther adding "/usr/lib" to
my path (this was not necessary with gcc 2.11c) this problem is solved.
But then gcc complains about not finding cc1; this can be solved by adding
"/usr/lib/gcc-lib/i386-linux/2.12c" to the search-path (how strange !).
Afther all that ld complains about not finding "crt0.o".

So I gave it up and reinstalled gcc 2.11c. But can those people who are using
gcc 2.12 give me some hints about the mistakes I made.

heinz

-------------------------------------------------------------------------------
ascom tech ag |
switzerland | [email protected]
-------------------------------------------------------------------------------


[next article]
From: [email protected] (Mark Evans)

Subject: Re: Ethernet
Message-ID: <[email protected]>
Date: 3 Jun 92 16:17:57 GMT
References: <[email protected]>
Sender: [email protected] (Usenet administrator)
Organization: Aston University
Lines: 11
Nntp-Posting-Host: uhura

[email protected] (the MOST SIGNIFICANT bit) writes:
: (and I'm still trying to figure out how to get linux to see the
: ethernet card)
Could try looking at the sources for Clarkson Telnet (CUTCP),
and seeing how this accesses the ethernet.

--
-------------------------------------------------------------------------
Mark Evans |[email protected]
+(44) 21 565 1979 (Home) |[email protected]
+(44) 21 359 6531 x4039 (Office) |


[next article]
From: [email protected] (Satish Chittamuru)

Subject: Re: More questions about rawrite
Message-ID: <[email protected]>
Date: 02 Jun 92 19:24:48 GMT
Organization: Microsoft Corporation
References: <[email protected]> <[email protected]
.ksu.edu>
Lines: 37

In article <[email protected]> [email protected] (Karl R. Buck) w
rites:
> [email protected] (Mike McBain) writes:
>
> >Further to my earlier question about problems mounting a rawritten disk:
>
> >How is the hopeful system builder supposed to transfer files to a
> >hard disk for uncompression and untar-ing? Don't tell me to use
> >mtools, because that's the program I'm trying to transfer.
>
> Do: "tar cvf diskfile.tar file.tar.Z"

NO! NO! NO!
Well, yes, you could do it that way. But you don't have to.

Just rawrite the .tar.Z file to disk.
Then on the Linux machine use the z option while untarring.

tar xvzf /dev/at0 (or whatever device you are using).

It will untar just fine. The extra tar step and the double untar
required to get it out is not really necessary. I do it all the
time and I know for a fact that it works.

> Then rawrite the file to the disk. Now since the file is in tar format
> jut take it to the destination machine and type:
>
> "tar xvf /dev/at0"
>
> This is of course assuming you have a 5.25" HD drive.
> --
> Karl Buck
> Grain Marketing ---===PEROT FOR PRESIDENT===---
> (913)776-2745


--
Satish K. Chittamuru [email protected]
Microsoft Corporation Redmond, WA
=====
Veni! Vidi! Visa! (I came! I saw! I purchased!). -- Julius, Sieze Her!


[next article]
From: [email protected] (Hamish Macdonald)

Subject: Re: X README.prog correction
In-Reply-To: [email protected]'s message of Wed, 3 Jun 92 16:46:49 GMT
Message-ID: <[email protected]>
Lines: 19
Sender: [email protected] (Usenet News Admin)
Organization: Bell Northern Research
References: <[email protected]>
Date: Wed, 03 Jun 92 17:28:24 GMT

>>>>> On Wed, 3 Jun 92 16:46:49 GMT,
>>>>> In message <[email protected]>,
>>>>> [email protected] (Peter MacDonald) wrote:

Peter> Unfortunately, later (or earlier for that matter) are not
Peter> acceptable. Shared libs are actually images, which are linked
Peter> to eliminate all unresolved references. In the case of libc,
Peter> no problem, cause there are no dependencies. But the X libs
Peter> depend upon libc. So references to strcpy and the like are
Peter> resolved when the shared lib is built, to the current shared
Peter> libc. In this case that is 2.11c. If you try now to compile
Peter> against say 2.12b, then your image will load the libc_v2_12b,
Peter> but the X shared lib will try calling the entry points for
Peter> 2.11c.

Peter> Shudder. Ker-chunk. ๐Ÿ™

Perhaps what is needed is a new shared library mechanism with some
sort of pseudo-dynamic linking mechanism.


[next article]
From: [email protected] (Jerry Weiler)

Subject: Re: How to use AF_UNIX sockets?
In-Reply-To: [email protected]'s message of Wed, 3 Jun 92 17: 31:36 GMT
Message-ID:
Originator: [email protected]
Sender: [email protected]io-state.edu (NETnews )
Organization: NAIVE - New And Improved Virtual Environment
References: <[email protected]>
Date: Wed, 3 Jun 1992 19:58:25 GMT
Lines: 18

In article <[email protected]> [email protected] (Brian Edmonds) wr
ites:

I'm trying to port some internet software over to my pc for single
user testing (MOO), but am unsure of how AF_UNIX sockets work. I've
used AF_INET sockets many times, and understand how to create INET
connections. Could someone post a brief summary of how to set up a
AF_UNIX server, and how to connect to it with a client?

----------------------------------------------------------------------
Brian Edmonds (MSc CompSci) [email protected]

There's sample code on tsx-11.mit.edu in pub/linux/sources/libs

it's in the file socket.tar.Z

it sets up a simple server and client. Check it out...

- Jerry.


[next article]
From: [email protected] (Kristian Ejvind)

Subject: Re: linux-0.96a .. keyboard freezes +
In-Reply-To: [email protected]'s message of 2 Jun 92 14:43:10 PDT
Message-ID:
Sender: [email protected] (Usenet)
Nntp-Posting-Host: gaia.electrum.kth.se
Organization: The Royal Institute of Technology, Kista, Sweden
References: <[email protected]>
Date: Wed, 3 Jun 1992 19:19:13 GMT
Lines: 9

In article <[email protected]> [email protected]
m (David.L.) writes:

Linux-0.96a, made from 0.96 source & the 0.96->0.96a patch, compiled with
gcc2.11c .. boots O.K. runs for ~10-12 seconds .. then the keyboard freezes.

I have experienced the same problem when using agetty instead of getty
on any device. Perhaps this is the problem for you too.

/Kristian


[next article]
From: [email protected] (Clarence Wilkerson)

Subject: VFS for OS/2 HPFS
Keywords: OS/2, read only, HPFS
Message-ID: <[email protected]>
Date: 3 Jun 92 19:10:31 GMT
Sender: [email protected]
Lines: 14

I use a hard disk on which I have a big piece devoted to OS/2 and TeX.
In particular I have about 10 megs of TeX fonts stored on a HPFS. Is there
some way to access these as a whole under Linux?

I guess this would involve a virtual file system for mounting HPFS partitions.
I'm mainly interested in read-only access.


--
Clarence Wilkerson \ Bitnet: wilker%[email protected]
Prof. of Math. (topology)\ Internet: [email protected]
Dept. of Mathematics \ messages: (317) 494-1903, FAX 494-0548
Purdue University, \ office: (317) 494-1955 (voice/modem)
W. Lafayette, Indiana 47907 \


[next article]
From: [email protected] (Thierry Priol)

Subject: Re: Dexxa mice
Message-ID: <[email protected]>
Date: 3 Jun 92 10:57:48 GMT
References: <[email protected]>
Sender: [email protected]
Organization: IRISA, Rennes (Fr)
Lines: 20

I have a 3 button dexxa mouse on my PC.

It works fine with:

* Microsoft mouse driver under DOS
* Microsoft mouse driver under Windows
* X386 1.1 under UNIX System V

It defitively does not work with:

* X386 1.2 under Linux

I do not know why it works with X386 1.1 under Unix System V and
does not work with X386 1.2 under Linux

--
Thierry PRIOL Phone: 99 84 72 10
IRISA / INRIA U.R. Rennes Fax: 99 38 38 32
Campus Universitaire de Beaulieu Telex: UNIRISA 950 473F
35042 RENNES CEDEX - FRANCE E-mail: [email protected]


[next article]
From: [email protected] (Bernard Steiner)

Subject: Re: What files to grab ?
Message-ID: <[email protected]>
Date: 3 Jun 92 11:35:02 GMT
References: <[email protected]> <[email protected]
ava.Helsinki.FI>
Organization: EUnet Backbone, Dortmund, Germany
Lines: 30
NNTP-Posting-Host: walhalla.germany.eu.net

In article <[email protected]>, [email protected]
FI (Linus Benedict Torvalds) writes:
> > Also, if Linux more reliable and stable than
> > BSD386 ?
>
> Haven't tested 386BSD, but I assume it will be pretty good in a month or
> two, and it's probably quite useable already. If you want NFS and full
> networking, it's the way to go. Comments, anyone?

Yupp -- a comment: 386bsd release 0.1 has been coming out next week ever
since it was released on April 17th (or so).
The system & sources supplied do NOT eneable you to re-compile a working
kernel. There is no X11 support, you need a co-processor (or one of the
hacked root disks that don't use floats).

A program like main() { for(;;) } will execute forever and you don't have a
chance to interrupt it by ^C. The kernel crashes about every two minutes
while some program waits for disk i/o.
On the other hand, your file names are of arbitrary length...

I'll see what Linux is like. Hope there are enough READMEs ๐Ÿ™‚

Ciao,
Bernard
--
Bernard Steiner, FB Informatik/IRB, Uni Dortmund, vox +49 231 755 2444
Postfach 500500, D-W-4600 Dortmund 50, Germany fax +49 231 755 2386
[email protected] ...!uunet!unido!bs

*III And they gave it Instructions, but knew it not. } From The Book of Nome,
*IV It is, they said, a Box with a Funny Voice. } Mezzanine v.III-IV


[next article]
From: [email protected] (Darren Senn)

Subject: Re: SERIOUS SECURITY PROBLEM (I think)
Message-ID: <[email protected]>
Date: 3 Jun 92 19:48:57 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected]
Organization: Curiosity Confederacy
Lines: 11

In article <[email protected]>, [email protected] (Theodor
e Ts'o) writes:
> Adding this functionality to Linux would not be difficult. A few lines
> of code to fs/namei.c.

To my understanding, this already exists in Linux: I've been using it since
0.95c+, and it seems to work.

--
Darren Senn Phone: (408) 479-1521
[email protected] Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo. Santa Cruz, Ca 95062


[next article]
From: [email protected] (David Giller)

Subject: Re: VFS for OS/2 HPFS
Message-ID: <[email protected]>
Keywords: OS/2, read only, HPFS
Organization: Occidental College, Los Angeles, CA 90041
References: <[email protected]>
Date: Wed, 3 Jun 1992 19:43:34 GMT
Lines: 19

[email protected] (Clarence Wilkerson) wrote:
>I use a hard disk on which I have a big piece devoted to OS/2 and TeX.
>In particular I have about 10 megs of TeX fonts stored on a HPFS. Is there
>some way to access these as a whole under Linux?
>
>I guess this would involve a virtual file system for mounting HPFS partitions.
>I'm mainly interested in read-only access.

Well, HPFS isn't really a publicly-available spec.

More likely, you should hope to see a UFS IFS for OS/2, and a real UFS
filesystem for Linux, then both can share a partition.

-Dave
--
David Giller, Box 134 | Q: How many Oregonians does it take to screw in a light
Occidental College | bulb? A: Three. One to replace the bulb, and two to
1600 Campus Road | fend off all the Californians trying to share the
Los Angeles, CA 90041 | experience. [email protected]


[next article]
From: [email protected] (Leslie Griffiths (Griff))

Subject: Re: SERIOUS SECURITY PROBLEM (I think)
Message-ID: <[email protected]>
Date: 3 Jun 92 09:31:49 GMT
References: <[email protected]> <[email protected]
.oxy.edu> <[email protected]>
Sender: [email protected] (USENET News Owner)
Organization: Coventry Polytechnic, Coventry, UK
Lines: 75
Nntp-Posting-Host: cc_sysh


In article <[email protected]> [email protected]
du (Oren Kastner) writes:
>In article <[email protected]> [email protected] (D
avid Giller) writes:
>
>As far as I know, when it comes to directories, the permissions are as
>follows:
>r - you may `read' the directory, meaning, you can do an `ls' on it and see
> the files in it.
>

But cannot 'ls -l' unless you have _both_ read _and_ execute permissions, since
the x perm allows you to stat the files, which is what ls -l will do.


>w - you can write to the directory, meaning you can create or delete files
> in it. (which is what happened in this case)

To create and remove files in a directory, you need _both_ w and x perms.

>x - you can `pass thru' the directory, meaning you can `cd' to it or anything
> below it (providing the permissions on the object that follow allow to be
> `cd'ed to.
>

See above. Execute permission allows you to cd or chdir (2) to that directory,
and also to stat (2) files within it. If you attempt any operation that
requires a stat but do not have execute permission, you will get some strange
errors.

For example, consider a file called 'f' in a dir called 'test'. You own '.',
test and f. Permissions on test are

drw-r-xr-x

and you attempt the operation

$ rm test/f
rm: test/f nonexistent

You also attempt the operation

$ ls -l test
test/f not found
total 0

$ ls test
f

Now give yourself x permission. Your permissions become:

rwx

Now, if we 'ls -l'

$ ls -l test
-rw-r--r-- 1 ccx018 0 Jun 3 10:11 f

and

$ rm test/f
$

Sorry if this sounds like a lecture, but I recently discovered the above myself
and was shocked to find that my understanding of permissions left a lot out.

On a similar thread, am I right in assuming that Linux doesn't use the
setgid bit on directories (a la SunOS)? I know that it uses the sticky bit
(which is more than our version of Dynix does!), it would be nice to see that
Linux is making full use of all the file permission bits.

Cheers

--
Griff \\ [email protected]
^^^^^ :-= ^^^^^^^^^^^^^^^^
//


[next article]
From: [email protected] (Toshiyasu Morita)

Subject: Re: X README.prog correction
Message-ID:
Date: Thu, 04 Jun 92 02:34:49 GMT
Organization: Netcom - Online Communication Services (408 241-9760 guest)
References: <[email protected]> <[email protected]
r.ca>
Lines: 13

Sorry, but I'm running under termprog that's very weird.

Somebody mentioned something about pseudo-dynamic linking, and I thought I'd
mention I'm currently working on a port of dld-3.2.3 - it's the dynamic linker
which is used for smalltalk-1.1.1.

I'm currently having a rough time porting it - if anyone has ported it before
please let me know! I'm getting general protection errors...


Toshi Morita
[email protected]



[next article]
From: [email protected] (Theodore Ts'o)

Subject: Re: X README.prog correction
Message-ID: <[email protected]>
Sender: [email protected] (Mr Background)
Reply-To: [email protected]
Organization: The Internet
Date: Thu, 4 Jun 1992 03:47:10 GMT
Lines: 28

From: [email protected] (Hamish Macdonald)
Date: Wed, 03 Jun 92 17:28:24 GMT

Peter> Unfortunately, later (or earlier for that matter) are not
Peter> acceptable. Shared libs are actually images, which are linked
Peter> to eliminate all unresolved references. In the case of libc,
Peter> no problem, cause there are no dependencies. But the X libs
Peter> depend upon libc. So references to strcpy and the like are
Peter> resolved when the shared lib is built, to the current shared
Peter> libc. In this case that is 2.11c. If you try now to compile
Peter> against say 2.12b, then your image will load the libc_v2_12b,
Peter> but the X shared lib will try calling the entry points for
Peter> 2.11c.

Peter> Shudder. Ker-chunk. ๐Ÿ™

Perhaps what is needed is a new shared library mechanism with some
sort of pseudo-dynamic linking mechanism.

Or what's needed is libraries with jump tables, which provide a fixed
entry point for the routines. Then when libraries change, you can
preserved the fixed entry point, and you won't lose. You still need to
worry about global variables, such as errno, but it becomes easier to
updated a shared library without breaking the rest of the world.

- Ted




[next article]
From: [email protected] (Darren Senn)

Subject: Re: .xdefaults
Message-ID: <[email protected]>
Date: 3 Jun 92 20:02:05 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected]
Organization: Curiosity Confederacy
Lines: 24

In article <[email protected]>, [email protected] wri
tes:
> Can anyone tell me what the general protocol is for adding entries to
> one's .defaults file is? When I get a new program do I just add its
> defaults in to the end of my file? I've also noticed that a lot of
> defaults conflict, when something of the form *background is specified.
> This is bad from isn't it? Am I suppose to edit this and then put it
> into my defaults file? Thank for any help in advance.

When you get a new program, usually it comes with an App-defaults file.
The info in this file does _not_ go into your ~/.Xdefaults file. Instead,
you should put it in /usr/lib/X11/app-defaults/Programname. An example
is the xtetris applications defaults which should be
/usr/lib/X11/app-defaults/Xtetris
(notice that the first letter is capitalized, and the rest isn't). Actually,
the name should be the same as the program itself uses to request resources
from the server, but you can't find that out without a complete man page
or the source code... usually it's the program name capitalized. The
file needs to be mode 644. If you want to customize the program on a
per-user basis, _then_ you use your ~.Xdefaults.

--
Darren Senn Phone: (408) 479-1521
[email protected] Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo. Santa Cruz, Ca 95062


[next article]

Subject: Modplay
Message-ID:
From: [email protected] (Erik Green)
Date: 3 Jun 92 15:02:58
Distribution: comp.os.linux
Organization: Long polymers in double-helix formation
Nntp-Posting-Host: att2.cs.mankato.msus.edu
Lines: 48


In an earlier article,[email protected] (Mark Cox) writes:

>[email protected] (Erik Green) writes:

>>B: It works on the parallel port, so we don't need to write a new kernel
>> driver for it.

>Unfortunately all my sound players and things (I wrote Modplay, and I also
>use linux) need to hook the timer interrupt and they reprogram it to run
>at many kHz. This isn't something thats going to work or be of use in
>Linux.

I kinda had the idea that writing a program like modplay to work with this
wouldn't work with Linux's ideas about timing, but I thought that since the
circuit diagram was widely available, it would be a good place to start.
As far as programming the timer to run faster, I presume you do this to
pump data out to the parallel port at a constant (fast) rate. There just
has to be some other (probably _not_ nice) way to do this under Linux.
Even if it won't work under Linux, I'd recommend that anyone who wants good
sound off of MS-DOS cheap build the circuit.

>I was thinking that it would be *really* easy to get a /dev/audio device
>to output through sound cards that have DMA - like the Soundblaster (yuk)
>and Covox Sound Master etc... I have code that will do DMA but I wouldn't
>have a clue how to get a device driver working for linux... anyone interested
>in having a go with me?

To my knowledge, you just write the routines to control the board into the
kernel, with appropriate #defines for the IRQs and port addresses, and write
system calls into the libraries to access the kernel routines. I'm not sure
how a /dev/sound fits into this, other than as a convenient alias for the
board,(or rather, the data registers _on_ the board).

I'd like to do this, but with two problems:
1) I may be going home for the summer, and be out of touch with the net -
ask me in about 2 weeks.
2) I only have an Adlib(Don't laugh, I got it secondhand cheap).

-Erik

>Mark

--
Erik "Longshot" Green
[email protected](vax1 | att2.cs | krypton | theory.cs).mankato.msus.edu
"Ash nazg durbatuluk, ash nazg gimbatul, ash nazg
thrakatuluk agh burzum-ishi krimpatul!"


[next article]
From: [email protected] (Jeffrey L Newbern)

Subject: 2 quick questions...
Message-ID: <[email protected]>
Keywords: pcomm, accounts
Sender: [email protected] (News system)
Nntp-Posting-Host: w20-575-93.mit.edu
Organization: Massachusetts Institute of Technology
Date: Wed, 3 Jun 1992 20:06:51 GMT
Lines: 15

i just have 2 quick questions.

1) when i run pcomm, i connect, but i find that things don't get written to the
screen until i hit return. for instance, the prompt does not appear, but i can
type in a command. when i hit return, the prompt and command appear, and the
command is executed. please help!

2) when i login as anyone but root (user for example) i get errors like "getwd:
cannot access parent directories" two or three times when i execute a command.
could someone explain what is going on?

thanx,
jeff newbern
[email protected]



[next article]
From: [email protected] (Darren Senn)

Subject: Re: a problem with /dev/kmem
Message-ID: <[email protected]>
Date: 3 Jun 92 20:16:01 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected]
Organization: Curiosity Confederacy
Lines: 33

In article <[email protected]>, [email protected] (t
he MOST SIGNIFICANT bit) writes:
> I've looked in the FAQ, and didn't find anything, so...
>
> Last night I did something quite brilliant, I accidently nuked
> /dev/kmem and so far haven't found any documentation that lists
> the major/minor numbers. So, could anyone tell me what to do to
> re-make it?

From the FAQ:

VI. INSTALLATION HINTS
======================
[...]
QUESTION: What are the device minor/major numbers?

ANSWER:
Memory devices: Major = 1 (characted devices) minor
0 /dev/ram
1 /dev/mem
2 /dev/kmem - not implemented (easy, but I haven't done it)
3 /dev/null
4 /dev/port (implemented, but untested - don't play with it)

example: "mknod /dev/null c 1 3"
[...]

Now, that little parenthetical comment ("not implemented [...]") is
wrong: it's quite implemented, and I've used it for quite some time.

--
Darren Senn Phone: (408) 479-1521
[email protected] Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo. Santa Cruz, Ca 95062


[next article]
From: [email protected] (Darren Senn)

Subject: Re: ESDI disk ? anybody
Message-ID: <[email protected]>
Date: 3 Jun 92 20:23:15 GMT
References: <[email protected]> <[email protected]
.edu>
Sender: [email protected]
Reply-To: [email protected]
Organization: Curiosity Confederacy
Lines: 11

In article <[email protected]>, [email protected] (Hubert Bahr) wr
ites:
> Does anybody use or have tried to use an ESDI disk under Linux.

I use a 120Mb Micropolis HD under Linux. I had to do a low-level format
to remap all my bad sectors, but it worked fine afterwards. Make sure
that the controller you get supports bad sector remapping (most do).

--
Darren Senn Phone: (408) 479-1521
[email protected] Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo. Santa Cruz, Ca 95062


[next article]
From: [email protected] (David.L.)

Subject: Re: linux-0.96a .. keyboard freezes +
Message-ID: <[email protected]>
Date: 3 Jun 92 20:37:13 GMT
References: <[email protected]> ectrum.kth.se>
Organization: Security Pacific Automation Co., LA, CA
Lines: 13

In article , [email protected] (Krist
ian Ejvind) writes:
> In article <[email protected]> [email protected]
com (David.L.) writes:
>
> Linux-0.96a, made from 0.96 source & the 0.96->0.96a patch, compiled with
> gcc2.11c .. boots O.K. runs for ~10-12 seconds .. then the keyboard freezes
.
>
> I have experienced the same problem when using agetty instead of getty
> on any device. Perhaps this is the problem for you too.
>
No .. using the standard getty .. and the rest of the /bin /usr/bin suite
--
----> [email protected]
David L.


[next article]
From: [email protected] (Brendan B. Boerner)

Subject: Re: 2 quick questions...
Message-ID: <[email protected]>
Keywords: pcomm, accounts
Sender: [email protected] (The Netnews Manager)
Nntp-Posting-Host: ca
Organization: Novell, Inc. --Austin
References: <[email protected]>
Date: Wed, 3 Jun 1992 22:35:17 GMT
Lines: 13

In article <[email protected]> [email protected] (Jeffr
ey L Newbern) writes:
>1) when i run pcomm, i connect, but i find that things don't get written to the
>screen until i hit return. for instance, the prompt does not appear, but i can
>type in a command. when i hit return, the prompt and command appear, and the
>command is executed. please help!

I've seen then same thing happen with Kermit-5A. Any ideas?

Brendan
--
Brendan B. Boerner Phone: 512/346-8380
Internet: [email protected] MHS: [email protected]
Please use ^^^^^^^^^^^^^^^^^ if replying by mail.


[next article]
From: Andrew McCallum

Subject: Questions
X-To: ALL
Organization: The Zoo of Ids
Date: Wed, 3 Jun 1992 14:37:00 GMT
Message-ID: <[email protected]>
Sender: [email protected] (Andrew McCallum)
Lines: 17

I'm sorry if I seem a bit behind on some information, but I REALLY am.

I just recently got my hands on a copy of Linux. It was v0.12, and it
seems like there have been some more versions since then.

What is currently the latest version?
What comes with it (v0.12 had hardly anything with it.)?
Where can I get it? Is is on any Toronto boards? What FTP site is it on?

Thanks a whole bunch!


This article has NOT been selected by Brad Templeton, and none of my articles
probly never will!
Who: Andrew McCallum What: I dunno. When: I Dunno.
Where: [email protected] Why: I dunno. When:



[next article]
From: [email protected] (Gregg Harris)

Subject: Re: Gcc , a quick question on libraries?
Message-ID: <[email protected]>
Date: 3 Jun 92 20:41:10 GMT
References: <[email protected]>
Organization: Research and Evaluation Dept. Portland Public Schools, Portland Or
egon.
Lines: 26

[email protected] (Mike Me) writes:

# Hey folks I have a program that tries to link to a libm.a library.

# My gcc (2.12b) didn't come with one. (I think), so whats so special
# about libm.a??
it is the math library.

# Is it called something else in gcc?

# Thanx,
# MikeE



--
**************************************************************
* To poor to be sued.
* [email protected] or [email protected]
* Gregg Harris
**************************************************************
--
**************************************************************
* To poor to be sued.
* [email protected] or [email protected]
* Gregg Harris


[next article]
From: [email protected] (Ajay Shah)

Subject: gcc installation hassles, and suggestions for ftp-able gcc
Date: 3 Jun 1992 23:13:47 -0700
Organization: University of Southern California, Los Angeles, CA
Lines: 24
Sender: [email protected] (Ajay Shah)
Message-ID:
NNTP-Posting-Host: almaak.usc.edu

I've had quite some hassles installing gcc, using files from tsx. I'm
not clear on the exact list of files which need to be installed.

Right now I seem to have a gcc -v on hello.c working, except that he
gives up when he sees no assembler. I tried using as86 but that is
clearly not it. After as he's going to need a linker and I don't see
one.

I also saw him complain about something in stdio.h when I used flags
-Wall -ansi -pedantic.

The difficulties in making a gcc work are clearly because of the
clutter of files in the gcc directory. IMHO it can be much better
done using a distinct directory for each release, containing each and
every file needed to install gcc from scratch. (Most of the files in
such directories would just be sym links into files in other
directories, so there is no wastage of disk space).

Eagerly waiting for a slick release of a easy-to-install v1.0....

-ans.
--

Ajay Shah, (213)749-8133, [email protected]


[next article]
From: [email protected] (Rik Faith)

Subject: Re: SCSI and Future Domain controler
Message-ID: <[email protected]>
Date: 3 Jun 92 12:52:20 GMT
References: <[email protected]>
Sender: [email protected]
Organization: UNC-Chapel Hill Computer Science
Lines: 17

In article <[email protected]>, [email protected] (Claude Lecommandeur) writ
es:
> I'have just read in the SCSI FAQ that patches for the Future
> Domain controler exists. Where can i find them ?

I'm finishing up beta testing, and will post the patches to tsx-11 soon.

For now, you can get them from:

omicron.cs.unc.edu:/pub/faith/fd-1.6.tar.Z --- the patches
fd-1.6.boot.Z --- a Future Domain-only kernel


If anyone tries these patches or the kernel, please send me e-mail and
tell me how they worked.

--
Rik Faith: [email protected]


[next article]
From: [email protected] (David Tucker)

Subject: Xterm scrollbar
Message-ID: <[email protected]>
Date: 4 Jun 92 00:11:58 GMT
Sender: [email protected]
Organization: Digital Equipment Corporation
Lines: 6


Is there a parameter to turn on in .Xdefaults or somewhere to
get a scroll bar on an xterm? How do I get TERM=xterms when I
startx instead of xterm?

David Tucker @ATO / [email protected] / KC4ZGO


[next article]
From: [email protected] (Richard Sharman)

Subject: Re: additional keys (Alt cursor etc...)
Message-ID:
Date: 1 Jun 92 03:33:21 GMT
References:
Sender: [email protected]
Organization: Mitel. Kanata (Ontario). Canada.
Lines: 46
In-reply-to: [email protected]'s message of 28 May 92 22:53:15 GMT

Greg said:

How about adding a little more keys to linux?

I'd like to be able to use 'Alt Cursor up' or 'Ctrl Home' etc. in emacs.
Linux doesn't recognize those combinations, though it would be
trivial to add them. The question is: what key codes should be used?
Is there any standard? If not, is anyone else interested so we can
set up a standard of our own? Otherwise I'm just going to configure
my system as I need it.

One suggestion: let the ALT or CTRL key modify the "printable"
characters in the sequence given by the normal key. Eg, since the up
arrow generates ESC [ A
have ALT-up generate ESC [ 'A'|0x80
and CTL-up generate ESC [ 'A'&~0x40
(or should the third byte above the 'A'&~0x60 ?).
Since nearly all of the "printable" characters seem to be upppercase
already, perhaps SHIFT with the character would downcase it. So
shift-control-up would generate ESC [ a.

A "pro" of this is that it seems nice and consistent. However, a
"con" is that these "clash" with some existing ANSI (? or VT100
anyway) codes. E.g. Home is ESC [ H, but ESC [ h
is used heavily as an output control function.


Another suggestion would be to reserve a "lead in" character. I'll
assume '!' here. This is followed by a value which depends on the
shift state of the ALT, CONTROL and SHIFT keys. These characters are
then followed by the "normal" sequence the key produces.

For the up arrow example:

shift up arrow ESC [ ! 1 A
control up arrow ESC [ ! 2 A
alt up arrow ESC [ ! 4 A

This scheme would allow combinations, such as:
shift-control up arrow ESC [ ! 3 A
shift-control-alt up arrow ESC [ ! 7 A

(You could even distinguish between left and right come to think of it
if you really wanted to.)

Richard


[next article]
From: [email protected] (Paul Lahaie)

Subject: Help needed with floppy driver
Message-ID:
Date: 29 May 92 23:23:44 GMT
Sender: [email protected]
Organization: International Shared Information Service (Ottawa)
Lines: 21


I have just installed the new 0.96a bootimage, and I'm unable to use the
floppy disk. Whenever I access /dev/PS0 (the same /dev/PS0 as on 0.12) I
get a Reset-floppy called (multiple times) and an error message about block
7 (it's ALWAYS block 7, no matter if I cp Image /dev/PS0, or mkfs /dev/PS0
1200). This has gotten to the point that making a new bootimage is nearly
impossible, as I must upload the compressed binary to another system, reboot
into DOS, and download it to go thru rawrite.exe. Oddly enough, the
0.95 bootimage (which used to work) doesn't work anymore. Neither does the
0.12 bootimage. Since the floppies haven't had any problems under DOS, it
intrigues me that it no longer works!

On another note, I wanted to recompile the 0.96a sources to remove SCSI
support, since I only have the GCC 1.4 compiler, I modified the Makefilke
according to the instructions. After doing make a few times (it would die
every once in awhile), I got myself a boot image (I stripped the scsi.a line
and removed all references to scsi_dev_init). After going thru the
upload/download process, I rebooted from the new bootimag,e and got a kernel
panic. Any ideas? Also, is there any documentation for Taylor UUCP?

- Paul


[next article]
From: [email protected] (David J. Hughes)

Subject: Re: 386BSD
Message-ID: <[email protected]>
Organization: ITS, Bond University
References: <[email protected]> <[email protected]>
Date: Thu, 4 Jun 1992 00:31:29 GMT
Lines: 17

[email protected] (Jim Burns) writes:

> Not a
>whole lot on 386BSD right now, since there hasn't been a new release in two
>months.


Wait for the announcement. It's just around the corner.


bambi - who maintains kirk.bu.oz.au (Linux and 386BSD archive)

___ David J. Hughes [email protected]
/ \ / / /
/ __/ __ __ ____/ / / __ Senior Network Programmer
/ \ / \ / \ / / / / / \ / Comms Development & Operation
\____/ \__// / \__/ \___/ / / / AUSTRALIA (+61 75 951450)


[next article]
From: [email protected] (Brendan B. Boerner)

Subject: another permissions problem
Message-ID: <[email protected]>
Date: 4 Jun 92 00:36:33 GMT
Sender: [email protected] (The Netnews Manager)
Reply-To: [email protected] (Brendan B. Boerner)
Organization: Novell, Inc. --Austin
Lines: 24
Nntp-Posting-Host: ca

Hello,

I've been having some problems with permission also.

Suppose I have a directory "/src", owned by root, grouped (is that a
word?) by bin and permissions are rwxr-x--- and all files in src are
rw-r--r--. As bin, I can't do an ls on src unless I change the
permissions to rwxr-xr-x. Odd, so I checked to see who I was (bin)
and ran groups which reported root. ??? Even odder is that if I log
in as user, it reports that I am bin and am in group root.

Furthermore, if I try to run tcsh as either bin or user where tcsh is
rwxr-xr-x bin bin, it reports permission denied.

The passwd and group files are from Linux root disk 0.95a.

Can anyone clue me in?

Thanks in advance,
Brendan
--
Brendan B. Boerner Phone: 512/346-8380
Internet: [email protected] MHS: [email protected]
Please use ^^^^^^^^^^^^^^^^^ if replying by mail.


[next article]
From: [email protected] (eric.h.baller)

Subject: Compiler bug or illegal asm statement?
Message-ID: <[email protected]>
Date: 4 Jun 92 01:18:45 GMT
Organization: AT&T
Lines: 16

I've gotten pretty far in my compilation of .96a, with gcc 2.11c,
but I still get the following error when compiling in kernel/chr_drv:

> gcc -I../../include -Wall -fomit-frame-pointer -c -o console.o console.c
console.c: In function 'scrup':
console.c:288: fixed or forbidden register was spilled.
This may be due to a compiler bug or impossible asm statements.
make[1]: *** [console.o] Error 1

That line is just a close bracket, and looks a lot like other functions.
If it is a compiler bug, which files did I forget to ftp? I thought I
had everything for gcc 2.11c

Thanks for any help!
Eric
[email protected]


[next article]
From: [email protected] (A. V. Le Blanc)

Subject: Re: I can't boot Linux: why?
Message-ID: <[email protected]>
Date: 3 Jun 92 06:35:03 GMT
Followup-To: [email protected]
Organization: Computing Centre, University of Manchester
Lines: 15

In article <[email protected]> [email protected] (Du
perval Laurent) writes:
>That's when i decided to get the mcc- version (what's mcc stand for anyway?).

MCC stands for Manchester Computing Centre.

>That one booted fine... But (there always is one) I couldn't do a thing with
>it: no ls, rm, or anything of that kind. According to what I read, I should
>have gotten a login.

The MCC boot disk does NOT work with a root disk, since its root disk
is included on the boot disk. However, it does require the MCC utilities
disk, which contains ls, rm, and so forth.

-- Owen
[email protected]


[next article]
From: [email protected] (Doug Evans)
Newsgroups: comp.os.linux,comp.sources.wanted

Subject: qic-80 driver source wanted
Message-ID: <[email protected]>
Date: 4 Jun 92 00:26:08 GMT
Reply-To: [email protected]
Followup-To: poster
Organization: Edmonton, Alberta
Lines: 7
X-Newsreader: Tin 1.1 PL3

I'm looking for the sources of a qic-80 tape device driver from which to base
a Linux driver for my CMS Jumbo 250 on.

Thanks in advance.
--
Doug Evans | "You're just supposed to sit here?"
[email protected] | - Worf in a mud bath.


[next article]
From: [email protected] (Doug Evans)
Newsgroups: comp.os.linux,comp.unix.xenix.sco

Subject: Usefulness/Legality of Xenix filesystem for Linux
Message-ID: <[email protected]>
Date: 4 Jun 92 00:37:39 GMT
Followup-To: comp.os.linux
Organization: Edmonton, Alberta
Lines: 20
X-Newsreader: Tin 1.1 PL3

I have SCO Xenix running on my 486 and would like to be able to use
either Xenix or Linux to access news, sources, etc. IE: I'd like to
have one news spool partition and be able to access it from Xenix or Linux.
Secondly, pending a driver for the CMS Jumbo 250, if most of my Linux

stuff was on Xenix partitions, I could back up most of Linux from Xenix.

Does anyone have any idea on the general usefulness of this?

How about the legality? As an example, I presume I can't distribute Xenix
headers, but I don't see that (in itself) as a problem. Anyone who wants this
would already have these headers. But there must be more issues, like ... ?

I would not, repeat not, suggest we go willy-nilly adding various filesystems
to the standard Linux distribution, but that doesn't mean Joe Bloe can't have
them as plug-n-play add-ons.

Followups to comp.os.linux.
--
Doug Evans | "You're just supposed to sit here?"
[email protected] | - Worf in a mud bath.


[next article]
From: [email protected]
Date: 02 Jun 92 20:15 CDT

Subject: Re: Orphaned Response
Message-ID: <[email protected]>
References: <[email protected]>
Lines: 89


/* Written 11:01 am May 20, 1992 by [email protected] in tisdec.tis.tandy.co
m:comp.os.linux */
In article I wrote:

> Second, I think sigprocmask() should return 0 on success. Currently
> it seems to return old mask. Are there different practises on this?

Ok, seems I was right. Thanks, Chris!

Here's the source for sigprocmask() from libc 2.11c:

int sigprocmask(int how, sigset_t * set, sigset_t * oset)
{
sigset_t old;
int rval = 0;

old = ___sgetmask();

if (set) switch(how) {
case SIG_BLOCK:
rval = ___ssetmask((*set) | old);
break;
case SIG_UNBLOCK:
rval = ___ssetmask((~(*set)) & old);
break;
case SIG_SETMASK:
rval = ___ssetmask(*set);
break;
}

if (oset)
*oset = old;
return rval;
}

As you can see, sigprocmask() returns whatever is returned by the
system call ssetmask(). ssetmask(), however, always returns the
old mask, so either the library function or the system call should
be changed.

> Third, rusage support is still incomplete, right? Since the library
> interface is already there, I decided to compile tcsh with rusage
> support, so that it could be linked against newer libraries in the
> future. However, after executing a commad, tcsh immediately hangs,
> leaving the executed process as a zombie. I assume it hangs on
> wait3(), but I haven't really tried to pinpoint this yet. Tcsh works
> fine without the rusage stuff.

Ok, now I have. It indeed hangs on wait3(). I replaced the wait with
a waitpid()/getrusage() combination, which does the trick. I'm not
entirely happy about this hack, since it's not atomary, but the code
isn't too critical.

Here's the source for wait3():

pid_t wait3(int *wait_stat, int options, int *reserved)
{
if (reserved) {
errno = EINVAL;
return -1;
}
return waitpid(0, wait_stat, options);
}
^^
Shouldn't this be -1 ?!? I thought wait3() operates on _all_
the children, not just those in the same process group.
Well, I guess we need a new system call for this, anyway...

> Fifth, is locale support already finished? I have no idea, it's kind of
> hard to test. The library interface is there, but the sources I have
> (outdated) only have stubs.

Ok, locale support is _not_ there. Working on it?

> And last, but not least... H.J. PLEASE, make it a practise to include
> library sources to each major release, at least. They are really the
> only way to make sure of what is supported and how, what is buggy and
> what is not. Especially since there are include files for stuff that
> is not even implemented.

Thank you, H. J. ๐Ÿ™‚

Mika
--
Mika Liljeberg Email: [email protected]
Helsinki University [email protected]
Dept. of Computer Science
/* End of text from tisdec.tis.tandy.com:comp.os.linux */



[next article]
From: [email protected] (Joe Waters)

Subject: ack. xdvi can't find fonts - help pleeze
Message-ID:
Date: 4 Jun 92 03:14:30 GMT
Organization: The University of New Mexico, Albuquerque, NM
Lines: 36

Well, I've recompiled xdvi three times, twice from sources on tsx-11 and once
from the base distribution (I got it off of gatekeeper.dec.com)... The compile
goes fine, the program appears to run, but when I try to run it on a simple
test dvi file, it can't find the fonts it needs. I untarred the
texpk300.tar.Z file into /usr/local/lib/tex/fonts/pk.300, and untarred
textfm.tar.Z into /usr/local/lib/tex/fonts/tfm. In my Makefile for xdvi, I
set the following:

DEFAULT_FONT_PATH=.:/usr/local/lib/tex/fonts/pk.300:/usr/local/lib/tex/fonts/tfm

and recompiled. But I get this error message when I try to xdvi test.dvi:

Can't find font cmex10.
Can't find font cmmi10.
Can't find font cmr7.
Can't find font cmr10.
xdvi: Not all pixel files were found

and then it quits. Obviously it doesn't like where I put the fonts. Could
somebody either a) tell me what to set my compile flags to so this works
properly or b) tell me where to put the fonts, or c) send my a statically-
linked binary that WORKS and tell me where they have their fonts and which
ones they have? Thanks a million.

P.S. - I compiled gnuplot from the standard distribution so that it works
with the color xterm terminal types x11 and X11 - if there's enough
interest, I'll upload statically linked binaries to tsx-11 or somewhere.

Thanks in advance for any help/suggestions,

Joe
--
/---------------------------+-------------------------------------------------\
| Joe Waters ("Falc") | "Most any problem can be solved with duct tape |
|[email protected]| and Froot Loops . . . or a good heavy club." |
\---------------------------+-------------------------------------------------/


[next article]
From: [email protected] (Hubert Bahr)

Subject: GCC 2.12 & shared Libraries
Message-ID: <[email protected]>
Date: 4 Jun 92 02:25:37 GMT
References: <[email protected]>
Sender: [email protected] (News system)
Organization: engineering, University of Central Florida, Orlando
Lines: 18

GCC 2.12? were a set of beta releases being tested. They have since
been moved. GCC 2.12c has been declared a dog and is no loger being
tested. If you want more info on the GCC beta activities you need
to join the GCC channel. Most of the work going on in Linux is truly
beta testing. As a result there are inconviences such as multiple
shared libraries. Let us not make work trying to find fixes for things
that are only temporary in nature such as "multiple libraries" but
concentrate on true long term advances. The GCC effort is making great
strides torward becoming an excellent compiler for ANSI C and C++. The
compactness of shared libs has and will be a great boon that doesn't need
to be corrupted to avoid an inconvience during beta testing. After all
it is this extensive beta testing that allows LINUX to make the rapid
strides torward an extremely useful operating system with X windows.

Lets Note bugs. Keep trying new things. Have patience with beta
testing. Thanks all for bringing Linux.

Hubert Bahr [email protected]


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: backups, how to do?
Message-ID: <[email protected]>
Sender: [email protected] (USENET News system)
Nntp-Posting-Host: next.wam.umd.edu
Organization: University of Maryland, College Park
Date: Thu, 4 Jun 1992 02:22:43 GMT
Lines: 17

I've finally started doing real work with Linux, so I have to start
worrying about backups. What are people doing? What I need is a
reliable (!!!) way to do both full and incremental backups. Is there
a script already written, floating around somewhere? How are
incremental backups handled with Unix?

On a side note, I'm running 0.95a+patches, and I managed to kill one
of the VC's somehow. I was logged in a root, logged out, and the VC
was dead. I don't know what happened, or why this time was different
than 100 other times I've logged out. According to ps, /bin/getty is
running on it. I even tried killing that getty, but the new one
didn't work either. ``echo testing > /dev/tty1'' produced no results
either. Neither did setterm -reset > /dev/tty1. It's dead. Any
ideas?

-Joel
([email protected])


[next article]
From: [email protected] (Hongjiu Lu)

Subject: X11 and gcc
Message-ID: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
Date: Thu, 4 Jun 92 00:49:15 GMT
Lines: 43

From: [email protected] (Peter MacDonald)

In the X README.prog, Orest says in the requirements section:

o Gcc2.11c, available from a whole bunch of places. Later releases may or
may not be acceptable.


Unfortunately, later (or earlier for that matter) are not acceptable.
Shared libs are actually images, which are linked to eliminate all
unresolved references. In the case of libc, no problem, cause there
are no dependencies. But the X libs depend upon libc. So references
to strcpy and the like are resolved when the shared lib is built, to
the current shared libc. In this case that is 2.11c. If you try now
to compile against say 2.12b, then your image will load the libc_v2_12b,
but the X shared lib will try calling the entry points for 2.11c.

Shudder. Ker-chunk. ๐Ÿ™

You can try rebuilding the X shared libs, but that may be dangerous.
Alternatively, you could use

gcc -V 2.11c

But you have to have the libc link libs around. Maybe gcc should be
modified to detect which libc version X was using and try to
put in the -V 2.11c for your. Only Hlu can answer that one.

Peter
[email protected]


------------------------
The next release of gcc will not have this problem. If it couldn't find the
compatible X11 shared libs, it will say no such a file or something like that.

You have to use

gcc -V 2.xxx

to tell gcc which version you want to use to link X11.

H.J.


[next article]
From: [email protected] (Scott R. Johnson)

Subject: Re: Anyone working on PPP?
Message-ID: <[email protected]>
Date: 4 Jun 92 01:14:43 GMT
References:
Organization: Amdahl Corporation, Advanced Systems, Fremont CA
Lines: 13

In article , [email protected] (Davi
d L Orman) writes:
> Well, I cant usr SLIP and I was wondering if anyone was working on PPP,
> or (hopefully) already had it implemented. Any takers?

Is SLIP implemented? I could sure use it... I'm looking for an alternative to
an Xterminal
for work, and linux would be a h*ll of a lot nicer overall...


--
--srj Scott R. Johnson [email protected]
o_O Amdahl Advanced Systems Development
=( ) = 46525 Landing Parkway, M/S 580
U Ack! Fremont, CA 94538 510-623-2333


[next article]
From: [email protected] (Andrew Stevens)

Subject: Misbehaving Pcomm
Message-ID: <[email protected]>
Date: 4 Jun 92 07:35:26 GMT
Sender: [email protected]
Reply-To: [email protected] (Andrew Stevens)
Organization: Oxford University Computing Laboratory, UK
Lines: 16

Dear all,

has anyone else had problems with pcomm's echo-ing of incoming data?
My copy (gcc2.1, 0.96) works fine *except* that incoming stuff is only
echoed after a line-feed. This looks like a bug in the libraries
compiled into the binaries resulting in stuff not being flushed from
buffers. However: the same problem occurs both with a locally compiled
version *and* the distributed binary.

Any suggestions? A swift change of libs?

Andrew
Andrew Stevens
Programmming Research Group JANET: [email protected]
11 Keble Road, Oxford, England UUCP: ...!uunet!mcvax!ukc!ox-prg!as
OX1 3QD +44 0865 272563


[next article]
From: [email protected] (Sakari Aaltonen)

Subject: fdisk vs. fdisk
Summary: novice question
Keywords: partitions
Message-ID: <[email protected]>
Date: 4 Jun 92 07:51:32 GMT
Sender: [email protected] (Usenet pseudouser id)
Reply-To: [email protected] (Sakari Aaltonen)
Organization: Helsinki University of Technology, Finland
Lines: 24
Nntp-Posting-Host: sandman.cs.hut.fi

I have a problem with hard disk partitions. I can boot from the bootimage
floppy; then, when I start fdisk from the rootimage one, it reports that
none of the existing partitions should be used.

I have one primary DOS partition and one extended (with 4 logical drives).
That's the way the DOS fdisk seems to want it. I don't know how I could
make 2 primary partitions, for example.

I would appreciate advice on how to run the DOS fdisk so that the resulting
partitions - or one of them - would be recognized as usable by the Linux
fdisk.

Or should I do it with another program? The DOS2Linux guide mentions PMK32;
however, the copy I found at nic.funet.fi is broken (it fails the CRC checks).

Thanks in advance,
s.a


--
-----------------------------------------------------------------------------
Sakari Aaltonen Helsinki University of Technology
Email: [email protected]
--- You can't keep a Finn down without him getting all red in the face ----


[next article]
From: [email protected]

Subject: Re: Floating point problem (Where did it go? ๐Ÿ™‚
Message-ID: <[email protected]>
Date: 3 Jun 92 09:41:28 +1000
References: <[email protected]>
Organization: Computer Centre, Monash University, Australia
Lines: 25

In article <[email protected]>, [email protected]
(Dan Miner) writes:
>
> double scalar;
> [...]
> scalar = 480 / 640;
>
> You would expect a value between 0 and 1. You
> print it.. What do I get? 0.00000
> What is going on here?
--
The result is correct. The rhs of your assignment is all integer, therefore
the division will be integer with the correct truncated result of 0, which
gets converted to 0.00000 upon assignment. To force floating point arithmetic
try:
scalar = 480. / 640;
or
scalar = 480. / 640.;
or etc, etc..

---------------
Bill Metzenthen
Mathematics Department
Monash University
Australia


[next article]
From: [email protected] (Daniel Brahneborg)

Subject: Man page for termios?
Message-ID:
Date: 4 Jun 92 09:34:51 GMT
Sender: [email protected] (News Administrator)
Organization: Dep. of Info.Proc, Umea Univ., Sweden
Lines: 17

Hi all!
Are there any man pages for the flags in the termio[s]-structure
anywhere? Up to now I've been looking at the man pages for SunOs
and HP-UX, but that doesn't feel quite right. Especially not
since the flags aren't the same. I've taken a peek at the kernel
(0.96a), but since the flags are used at a lot of different locations,
it didn't help very much.

/Basic

/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\ When hacking : Basic /\ The best way to accelerate /
/ In the real world: Daniel Brahneborg \/ a M*c is with \
\ On the net : [email protected] /\ 9.82 m/s^2 /
/ Phone: +46 90 198198 Fax: +46 90 197085 \/ \
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
Hi! I am a .signature virus. Copy me into your .signature to join in!


[next article]
From: [email protected] (Fabien Coutant)

Subject: Re: Starting X386
Message-ID: <[email protected]>
Date: 4 Jun 92 09:31:55 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected] (Fabien Coutant)
Lines: 20
Nntp-Posting-Host: uapu

In article <[email protected]>, [email protected] writ
es:
|>
|> We *finally* got X386 working on our paradise board with a taxan multisync
|> monitor. After spending all day (12+ hours) staring at video timings, we
|> happened to boot up into the 80-column text mode, and type 'startx'. It worke
d.
|> I can't remember reading of anyone having mentioned this before - using the
|> 132-column modes can screw your graphics mode registers.
|>

I had the same problem with the vgalib. It seems like it has
something to do with horizontal or vertical frequency of the
monitor not being correctly set... If I run in 80-culomns mode,
then I can use 320x200x256, 320x240x256, ... 640x480x16, but If
I run in 132-column mode (no matter how much lines), then I can
use none of these, but I can use 800x600x16, which I can't use in
80-column mode. Any ideas ?

--
Fabien COUTANT
[email protected]


[next article]
From: [email protected] (Aldo Serraglio)

Subject: Man pages
Message-ID: <[email protected]>
Sender: [email protected] (USENET Superuser)
Organization: Magneti Marelli, Electronics Division (ITALY)
Date: Thu, 4 Jun 1992 09:37:20 GMT
Lines: 4

Are there any man pages for "fsck" and "mkfs" (0.95c+ or 0.96a) ?

Aldo Serraglio
[email protected]


[next article]
From: [email protected] (Mark Evans)

Subject: Re: Help needed with floppy driver
Message-ID: <[email protected]>
Date: 4 Jun 92 10:04:38 GMT
References:
Sender: [email protected] (Usenet administrator)
Organization: Aston University
Lines: 15
Nntp-Posting-Host: uhura

[email protected] (Paul Lahaie) writes:
:
: I have just installed the new 0.96a bootimage, and I'm unable to use the
: floppy disk. Whenever I access /dev/PS0 (the same /dev/PS0 as on 0.12) I
: get a Reset-floppy called (multiple times) and an error message about block
: 7 (it's ALWAYS block 7, no matter if I cp Image /dev/PS0, or mkfs /dev/PS0
: 1200).
The /dev/PS0 is for a 1.44M 3.5" floppy in the A: drive.
Your example mkfs suggests you have 5.75" disks.
Try using cat Image >/dev/at0
--
-------------------------------------------------------------------------
Mark Evans |[email protected]
+(44) 21 565 1979 (Home) |[email protected]
+(44) 21 359 6531 x4039 (Office) |


[next article]
From: [email protected] (& Stevens)
Newsgroups: comp.os.linux,comp.sys.transputer

Subject: Re: Using transputers under Linux, current state
Message-ID:
Date: 4 Jun 92 09:43:09 GMT
References: <[email protected]>
Sender: [email protected]
Organization: Dept. Engineering Science, Oxford University, UK
Lines: 14
In-reply-to: [email protected]'s message of 31 May 92 18
:51:38 GMT

Both michael in Germany and me (michael in UK) seem to be have a great
deal of success using transputers under linux. I think michael recent
posting summaries the current state of things.

As you would expect the performance compared to DOS doing file serving
is much better. If anyone else is interested or even better is also
doing something I'm sure we would both be pleased to here from you.

How about it,

--
Michael Stevens, Robotics Research Group,Dept of Engineering Science,Oxford,UK
JANET: [email protected]
INTERNET: michael%[email protected]


[next article]
From: [email protected] (Gabor Herr)

Subject: Subject: Running X11 with 0.96a as a normal user
Sender: [email protected] (The Usenet-News System)
Message-ID: <[email protected]>
Date: Thu, 04 Jun 1992 10:41:33 GMT
Nntp-Posting-Host: sun2.iti.informatik.th-darmstadt.de
Organization: TH Darmstadt
Lines: 73

Path: iti.informatik.th-darmstadt.de!herr
Newsgroups: comp.os.linux
Distribution: world
Followup-To:
From: [email protected] (Gabor Herr)
Organization: TU Darmstadt
Keywords: X11, kernel, kill, permissions

Hi Linuxers,

yesterday I installed the new version of the X binaries (from nic.funet.fi)
with kernel 0.96a + paches. Everything worked fine ๐Ÿ™‚ as long as I was logged
in as root. As a normal user I could start up a session, but it was impossible
to end it :-(, the keyboard was completely dead and the machine hung.

After a couple of times of rebooting, I redirected the error messages of
startx in a file and got the following:

> X386 Version 1.2 / X Windows System [linux v0.2]
> (protocol Version 11, revision 0, vendor release 5000)
>
> VGA256: et4000 (mem: 1024k clocks: 25 28 32 36 40 45 50 65)
>
> xinit: Operation not permitted (errno 1): Can't kill X server
>

I think the problem is caused by a bug in the kill() system call. In the
corresponding kernel routine send_sig() (linux/kernel/exit.c), only the
effective userids of the killing and victim processes are compared.
So a normal user can't kill a setuid-program, that he has started. And that
is the situation we have under X: the server is running setuid-root (effective
userid 0) and xinit with normal permissions. After the last client died,
xinit tries to kill the server, but it fails and the server keeps running.

I don't know how kill() is defined in POSIX, but I think every user should be
able to kill all processes he has started, regardless of their effective user
id's.

The following patch fixes this bug (if this is one), by allowing a process to
send signals to processes with the same uid.

--------------------------- cut here -----------------------------------------

*** linux/kernel/exit.c.orig Mon May 18 18:10:06 1992
--- linux/kernel/exit.c Wed Jun 3 23:38:30 1992
***************
*** 22,28 ****
if (!p || (sig < 0) || (sig > 32))
return -EINVAL;
if (!priv && ((sig != SIGCONT) || (current->session != p->session)) &&
! (current->euid != p->euid) && !suser())
return -EPERM;
if (!sig)
return 0;
--- 22,28 ----
if (!p || (sig < 0) || (sig > 32))
return -EINVAL;
if (!priv && ((sig != SIGCONT) || (current->session != p->session)) &&
! (current->euid != p->euid) && (current->uid != p->uid) && !suser())
return -EPERM;
if (!sig)
return 0;

--------------------------- cut here -----------------------------------------

Gabor


------------------------------------------------------------------------------
Gabor Herr
Computer Science Department
Technical University Darmstadt
Germany


[next article]
From: [email protected] (Gabor Herr)

Subject: Subject: Xmodmap for German keyboards
Sender: [email protected] (The Usenet-News System)
Message-ID: <[email protected]>
Date: Thu, 04 Jun 1992 10:44:46 GMT
Nntp-Posting-Host: sun2.iti.informatik.th-darmstadt.de
Organization: TH Darmstadt
Lines: 57

Path: iti.informatik.th-darmstadt.de!herr
Newsgroups: comp.os.linux
Distribution: world
Followup-To:
From: [email protected] (Gabor Herr)
Organization: TU Darmstadt
Keywords: xmodmap, key, keyboard, X, X11

Hi Linuxers,

here are some xmodmap key bindings, that map the raw keycodes for a
German MF-2 keyboard. To install just save them in a file and copy it to
'/usr/lib/X11/xinit/.Xmodmap'. Or if you want to test it first, call
'xmodmap file' directly.

-------------------------------- cut here ------------------------------------
! Keycode definitions for a German MF-2 keyboard ** gh92
keycode 11 = 2 quotedbl
keycode 12 = 3 paragraph
keycode 13 = 4 dollar
keycode 14 = 5 percent
keycode 15 = 6 ampersand
keycode 16 = 7 slash braceleft
keycode 17 = 8 parenleft bracketleft
keycode 18 = 9 parenright bracketright
keycode 19 = 0 equal braceright
keycode 20 = ssharp question backslash
keycode 21 = apostrophe grave
keycode 24 = q Q at
keycode 29 = Z
keycode 34 = udiaeresis Udiaeresis
keycode 35 = plus asterisk asciitilde
keycode 47 = odiaeresis Odiaeresis
keycode 48 = adiaeresis Adiaeresis
keycode 49 = asciicircum
keycode 51 = numbersign asciicircum
keycode 52 = Y
keycode 59 = comma semicolon
keycode 60 = period colon
keycode 61 = minus underscore
keycode 94 = less greater bar
keycode 113 = Mode_switch
add mod5 = Mode_switch

-------------------------------- cut here ------------------------------------

Gabor


------------------------------------------------------------------------------
Gabor Herr
Technical University Darmstadt
Germany






[next article]
From: [email protected]

Subject: I can't build static images w/ gcc2.12c
Message-ID: <[email protected]>
Date: 4 Jun 92 11:04:11 GMT
Organization: University of St.Gallen, Switzerland
Lines: 13

I can't build static images with gcc 2.12c. The linker aborts with the
following error:

/usr/lib/static/libgcc.a(_builtin_new.o):
Undefined symbol ___new_handler referenced from text segment

Is it a bug?

(I run Linux 0.96a on a 386 cpu with 387 co-processor)

Thanks for your help!
-Fabian Mueller .__________________________________.
_________________________________________| e-mail: [email protected] |__


[next article]
From: [email protected] (William E. Harvey)

Subject: Re: Help needed with floppy driver
Message-ID: <[email protected]>
Sender: [email protected]
Organization: Los Alamos National Laboratory
References:
Date: Thu, 4 Jun 1992 13:19:27 GMT
Lines: 11


I was having the same problem (or I guess I really still am) with
Reset-floppy being called when accessing /dev/PS0. By chance I ejected
the floppy and shoved it back in without doing anything else. The
system started reading the floppy then without any problems. I haven't
had any reason to write anything on the floppy from Linux, so I don't know
if that will work. However so far this kludge works every time.

Billy Harvey
[email protected]



[next article]
From: [email protected] (Stephen A Uhler)

Subject: Re: MGR and X
Keywords: MGR
Message-ID: <[email protected]>
Date: 4 Jun 92 13:26:35 GMT
References:
Sender: [email protected]
Reply-To: [email protected] (Stephen A Uhler)
Distribution: comp
Organization: Bellcore, Morristown NJ
Lines: 55
Nntp-Posting-Host: sleepy

In article [email protected] (Kevin C. Dorff)
writes:
>I know what X is and what it gives you, but if someone could quickly
>compare and contrast MGR and X for me, along with their speed, disk
>requirements, capabilites, memory requirements, etc, I would REALLY
>appreciate it. ...

Although I'm probably not the best person to talk objectively about X (I wrote
MGR), I can say a few things about MGR.

MGR is a window system I originally wrote for SUN 100's and SUN 120's
in 1984-1985. It manages asynchronous updates of overlapping windows.
"Backing-store" is built in. MGR has the basic capabilities of the X
server, window manager, and xterm all rolled into one. Each window
acts like an intelligent "terminal", using escape-sequences to effect
all of the windowing capabilities. The C library interface consists primarily o
f
printf's that emit the proper escape sequences. MGR clients have been
written in C, shell, lisp, prolog, emacs, awk, perl -- any language
that can send ASCII data over a serial connection.

MGR tries to adhere to the traditional UNIX philosophy of doing only
one job, but doing it well. MGR has no networking code built in: it
uses whatever networking capabilities already exist (rlogin, kermit,
etc.). MGR doesn't do database management, or authentication; it just
manages windows.

The source to MGR is 16k lines of C. The complete distribution, including the
fonts, icons, manuals and demos is about 1.2meg (compressed). The MGR process
that is currently running the 17 windows on my SPARC-2 display is up to 600k
bytes of total size. On comparible hardware, running ordinary tasks, MGR is
usually anywhere from 2 to 10 times faster than X (don't believe me -- try it fo
r
yourself.)

This is an excerpt from the MGR manual (available from bellcore.com):

... In addition to terminal emulation, MGR provides each client window
with: graphics primitives such as line and circle drawing; facilities
for manipulating bitmaps, fonts, icons, and pop-up menus; commands to
reshape and position windows; and a message passing facility enabling
client programs to rendezvous and exchange messages. Client programs
may ask to be informed when a change in the window system occurs, such
as a reshaped window, a pushed mouse button, or a message sent from
another client program. These changes are called events. MGR notifies
a client program of an event by sending it an ascii character string in
a format specified by the client program. Existing applications can be
integrated into the windowing environment without modification by
having MGR imitate keystrokes in response to user defined menus or
other events. The user interface provides a simple point-and-select
model of interaction using the mouse with pop-up menus and quick access
to system functions through meta-keys on the keyboard. MGR also
provides a cut and paste function that permits a user to sweep out and
copy text from any window and paste it into any other.

I hope this helps.
Stephen Uhler
[email protected]


[next article]
From: [email protected] (Stephen A Uhler)

Subject: Re: MGR and X
Keywords: MGR
Message-ID: <[email protected]>
Date: 4 Jun 92 13:14:15 GMT
References:
Sender: [email protected]
Reply-To: [email protected] (Stephen A Uhler)
Distribution: comp
Organization: Bellcore, Morristown NJ
Lines: 56
Nntp-Posting-Host: sleepy

In article [email protected] (Kevin C. Dorff)
writes:
>I know what X is and what it gives you, but if someone could quickly
>compare and contrast MGR and X for me, along with their speed, disk
>requirements, capabilites, memory requirements, etc, I would REALLY
>appreciate it.
> ...

Although I'm probably not the best person to talk objectively about X (I wrote
MGR), I can say a few things about MGR.

MGR is a window system I originally wrote for SUN 100's and SUN 120's
in 1984-1985. It manages asynchronous updates of overlapping windows.
"Backing-store" is built in. MGR has the basic capabilities of the X
server, window manager, and xterm all rolled into one. Each window
acts like an intelligent "terminal", using escape-sequences to effect
all of the windowing capabilities. The C library interface consists primarily o
f
printf's that emit the proper escape sequences. MGR clients have been
written in C, shell, lisp, prolog, emacs, awk, perl -- any language
that can send ASCII data over a serial connection.

MGR tries to adhere to the traditional UNIX philosophy of doing only
one job, but doing it well. MGR has no networking code built in: it
uses whatever networking capabilities already exist (rlogin, kermit,
etc.). MGR doesn't do database management, or authentication; it just
manages windows.

The source to MGR is 16k lines of C. The complete distribution, including the
fonts, icons, manuals and demos is about 1.2meg (compressed). The MGR process
that is currently running the 17 windows on my SPARC-2 display is up to 600k
bytes of total size. On comparible hardware, running ordinary tasks, MGR is
usually anywhere from 2 to 10 times faster than X (don't believe me -- try it fo
r
yourself.)

This is an excerpt from the MGR manual (available from bellcore.com):

... In addition to terminal emulation, MGR provides each client window
with: graphics primitives such as line and circle drawing; facilities
for manipulating bitmaps, fonts, icons, and pop-up menus; commands to
reshape and position windows; and a message passing facility enabling
client programs to rendezvous and exchange messages. Client programs
may ask to be informed when a change in the window system occurs, such
as a reshaped window, a pushed mouse button, or a message sent from
another client program. These changes are called events. MGR notifies
a client program of an event by sending it an ascii character string in
a format specified by the client program. Existing applications can be
integrated into the windowing environment without modification by
having MGR imitate keystrokes in response to user defined menus or
other events. The user interface provides a simple point-and-select
model of interaction using the mouse with pop-up menus and quick access
to system functions through meta-keys on the keyboard. MGR also
provides a cut and paste function that permits a user to sweep out and
copy text from any window and paste it into any other.

I hope this helps.
Stephen Uhler
[email protected]


[next article]
From: [email protected] (Michael Staats)

Subject: Is there a passwd program?
Summary: looking for passwd
Keywords: passwd
Message-ID:
Date: 4 Jun 92 14:29:43 GMT
Organization: Rechenzentrum Uni-Duisburg
Lines: 24


Hello Linuxers,

I searched through some ftp servers for linux programs but I couldn't
find passwd. Does it exist? If yes, how can I get it?

I'm quite new to Linux and just for testing I installed it on a 10 MB
partition on a 1 MB (yes, one!) 386SX Laptop with 2 MB swapspace ---
it works, but of course not for real work. I'll get a 4 MB RAM Card
in two months or so (when I saved enough money) and if like linux
it will be the only OS on my PC (well, in fact it is the only OS right now,
the other thing running is MS DOS :-).

So you see - I need binaries. I wouldn't even try to do a gcc with
this configuration.

Thanks for your help,

Michael
--
Michael Staats
Theoretische Tieftemperaturphysik
Universitaet Gesamthochschule Duisburg
[email protected]


[next article]
From: [email protected] (Matt)

Subject: Re: ``du'' reports double the real disk usage. Hmmm.
Message-ID: <[email protected]>
Date: 4 Jun 92 15:49:48 GMT
References: <[email protected]>
Sender: [email protected] (USENET News Owner)
Organization: abUSENET, Coventry Polytechnic, Coventry, UK
Lines: 31
Nntp-Posting-Host: cc_sysh

In article <[email protected]> [email protected] (Joel M. Hoffman
) writes:
>I've noticed that ``du'' (and, for that matter, ls -l) report double
>the actual disk usage. Is this a bug or are they reporting 512K units?
>
>For example:
>
> #ls -l
> total 50
> -rw-r--r-- 1 joel joel 25478 Jun 2 13:38 note.wd
> #du .
> 51 ./
> #
>
>Shouldn't the totals be 25 and 26?
>
>-Joel
>([email protected])


Don't forget your using the gnu-du which report sizes in 512byte blocks.

Try a 'du -k' to report disk usage in 1k bocks...


Cheers

--
---
Matthew Roderick /Email: [email protected] | \ //\ |`. |
"""""""""""""""" /Interests: Fine Art & Good Times | \/\//--\|_| |
/Quote: If I felt any worst I'd be dead, If I felt any better I wouldn't be me


[next article]
From: [email protected] (david.e.wexelblat)

Subject: Call for Beta-Testers for X386 1.2E (X11R5, enhanced)
Keywords: X386, X11R5, beta-test
Message-ID: <[email protected]>
Date: 4 Jun 92 16:26:09 GMT
Organization: AT&T
Lines: 91

[ This was originally posted to comp.windows.x and comp.unix.sysv386. The
response has been overwhelming. I have not gotten any responses from the
Linux community, though. Since I believe that X11R5 (X386 1.2) has been
brought up on Linux (since I'm not a devotee, I may be wrong :->), this
might interest some of you. We will accept at most 3-4 volunteer for
this beta-test, so read on if you are interested.

--dwex
]

We are currently looking for beta-testers for an enhanced X386 1.2 (X11R5)
(to be called X386 1.2E). If you are interested in helping with this
beta-test, and meet the following conditions, please send email to
one of the addresses in the 'Contacts' section below.

1) Running Intel 3/486 Unix of some flavor
2) Using an SVGA card currently supported by X386
3) Have the disk space and experience to build X11R5 from scratch (*)
4) Have access to anonymous FTP and Internet email
5) Have the Unix/X/C experience to be a good beta-tester

(*) Only a limited binary distribution (SVR4) is available at this time;
if you are short on disk space, but still want to get involved, send
email, and we'll see if we can accomodate you.

What's Included:
1) SpeedUp from Glenn Lai, for compatible hardware
2) fX386 from Jim Tsillas, for when SpeedUp is not being used
3) Support for newer VGA cards with clocks that the current X386
can't deal with.
4) LOCALCONN for both SVR3 and SVR4

plus a whole host of bugfixes and enhancements

What We're Looking For:
Currently, we have tested with the following Unices:

- ESIX SVR4.0.3
- Microport SVR4 3.1 and 4.1
- Interactive SVR3 2.2.0

We currently have beta-testers with SVR4 from Dell and UHC.
We are looking for additional SVR4 (AT&T, MST, Consensys), and
SVR3 (any we can find) users who can help us ensure compatibilty.
Users of other flavors (BSD/386, 386BSD, Linux, Mach) may
volunteer, however, you will likely be asked to fix your own
bugs, unless the server itself is the problem. Please also note
that we may not accept all volunteers; we are looking for broad
coverage, not necessarily large numbers.

We expect beta-testers to build X386 1.2E distribution (which is
currently a large patch kit against the X11R5 PL12 base), to
provide bug reports (and hopefully fixes :->), and to stress the
system in their configuration.

There will be an FTP archive and an email list for communication
among the beta-testing community. When you contact us initially,
be sure to provide a sound email address.

Who's Involved:
David Dawes
Glenn Lai
Jim Tsillas
David Wexelblat

What does SGCS have to say about all this (disclaimer):
Thomas Roell and Mark Snitily have been consulted all throughout
this effort, and have provided both technical and moral support,
including allowing our use of the name 'X386 1.2E'. It is the
intention of all parties involved that this work will be merged
with the SGCS contribution to MIT's X11R6, when that becomes
a reality.

When will X386 1.2E be generally available:
This has not yet been decided; it will depend on how the beta-test
goes. This enhanced server has been running on the systems mentioned
above for a couple of months now. More information will be posted
when it is available.

Contacts:
To volunteer, send email to David Wexelblat
or David Dawes . During the window of
time from 6/9/92-6/22/92, I'll be in the process of moving (home, not
job), and will be unavailble for several days at a time. During
that interval, inquiries should be sent to David Dawes directly.

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
David Wexelblat | [email protected] | I asked her her name.
AT&T Bell Laboratories | ...!att!mtgzfs3!dwex | She said her name was
200 Laurel Ave - 4B-421 | | 'Maybe'
Middletown, NJ 07748 | (908) 957-5871 | --Damn Yankees


[next article]
Newsgroups: comp.sys.ibm.pc.hardware,comp.unix.bsd,comp.os.linux,comp.os.os2,asc
om.bsdi,comp.periphs.scsi
From: [email protected] (Julian Elischer)

Subject: Re: Bustek-742 EISA controller (compatible with Adaptec 1742 ?)
Message-ID: <[email protected]>
Keywords: OS/2 386BSD UNIX-V.4
Organization: TRW Financial Systems
References: <[email protected]>
Date: Thu, 4 Jun 1992 17:19:18 GMT
Lines: 49

In article <[email protected]> [email protected]
(Norbert Bladt) writes:
>Hi there,
>
>I saw an interesting offer for an EISA system which includes a 330 MB SCSI disk
>and the Bustek EISA-SCSI controller (742). Now, that there is a Adaptec-1742
>I would really like to get this one. However, this would add to the cost
>of the system. Now, my question is whether these controllers are compatible.
>I know that the Bustek and the Adaptec-1740 will work with OS/2 2.0. What I
>don't know is will these work with Dell or Interactive UNIX and 386BSD-UNIX
>and especially the Adaptec-1742 with OS/2 2.0. In other words, is it worth
>the trouble and cost of changing the Bustek-742 to an Adaptec-1742 ?

I don't know the compatibility betweenthe 742 and the 1742
but the 742 comes up emulating a 1542.
I have written drivers for both and run the 742 on the 1542 driver.

My scsi subsystem that some people got copies of in the first week of
the jolitz release included drivers for both the 742 and the 1542.

The 742 driver knows about the extended 32 bit address structures and can be
used with machines with more than 16MB of ram.

We run these drivers on net2, OSF/1 and mach2.6 without source changes, however
the setups are somewhat skewed towards mach and a little 'un-BSD-ish' in
style. (also the disk driver needs setting up with utilities from mach and
CMU have not clarified whether these are PD or not.)

I Have ordered the 1742 technical reference manual so in a couple of
weeks I should be able to either
1/ know it's compatible with the 742a
2/ if not, write a new one.
however, if you have less than 16MB of ram the 742a will run successfully
out of the box as a 1542.

performance for the 742 running as a 1542 is about 2.7MB/sec off a
segate elite ST41600N using 16k raw blocks. the 1542 achieves about 1.5MB/sec
for the same test, but of course that is an ISA card. The 1742 is probably
similar (or better than) to the 742a.



+----------------------------------+ ______ _ __
| __--_|\ Julian Elischer | \ U \/ / On assignment
| / \ [email protected] +------>x USA \ in a very strange
| ( OZ ) 2118 Milvia st. Berkeley CA. \___ ___ | country !
+- X_.---._/ USA+(510) 704-3137(wk) \_/ \\
v




[next article]
From: [email protected] (Werner Almesberger)

Subject: Re: fdformat
Message-ID: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Swiss Federal Institute of Technology (ETH), Zurich, CH
References: <[email protected]>

Date: Thu, 4 Jun 1992 15:36:18 GMT
Lines: 23

In article [email protected] (Steven M. Palm)
writes:
> On my system, fdformat works fine on already-formatted disks, but gives
> reset-floppy errors and dies on a never-formatted disk. Both 5.25" and 3.5"
> disks.

There is a bug in at least one of the 0.96 kernels: a superfluous disk read
request is generated when the floppy device is opened. This can be fixed by
removing the line

check_disk_change(inode->i_rdev);

in fs/minix/blkdev.c. (The disk change test is already in the floppy driver,
which handles the special case of formatting correctly.)

You might get some errors at the end of fdformat's verification pass too.
They're caused by the read-ahead bug and can safely be ignored.

- Werner
--
_________________________________________________________________________
/ Werner Almesberger, ETH Zuerich, CH [email protected] /
/ IFW A44 Tel. +41 1 254 7213 [email protected] /
/_BITNET:[email protected]__HEPNET/CHADNET:_[20579::]57414::ALMESBERGER_/


[next article]
From: [email protected] (Greg Wettstein)

Subject: Trouble with BASH on serial lines.
Message-ID: <[email protected]>
Date: 4 Jun 92 14:52:45 GMT
Organization: North Dakota Higher Education Computer Network
Lines: 38

We have been noting a consistent problem with BASH running on serial lines.
The problem occurs irrespective of kernel level (0.95a, 0.95c+, 0.96a etc)
and only occurs on serial lines, the console is unaffected.

Specifically when a user logs in on a serial port and has BASH specified
as their login shell, BASH will refuse to echo a newline after a command
is typed. This causes the BASH prompt for the next command line to overtype
the first line of whatever output was generated by the command. For example:

~> echo $TERM

Will cause VT100 to be printed and then immediately overtyped by the next
command line prompt.

The same login works flawlessly from any one of the virtual consoles. I
have made sure that the default terminal type is set to VT100 in the
inittab file. I have also compared stty dumps from the serial login to
logins on the virtual console and they appear to be identical. No amount
of munging with TERMIO settings via stty can get BASH to generate the
necessary linefeeds.

I suspect the problem may be linked to a TERMCAP definition or some such
thing but we are at a loss to explain the problem. Both the version of
BASH found on the 0.95 rootimage as well as 1.12 picked up from tsx-11
display the problem. Any suggestions would be greatly appreciated. If
possible please use the e-mail address in my sig. Thanks in advance.

As always,
Dr. G.W. Wettstein
Oncology Research Division Computing Facility
Fargo Clinic / MeritCare

UUCP: uunet!plains!wind!greg
INTERNET: greg%[email protected]
Phone: 701-234-2833

`The truest mark of a man's wisdom is his ability to listen to other
men expound their wisdom.'


[next article]
From: [email protected] (Wen-Chun Ni)

Subject: Dynamic Linking etc.
Message-ID: <[email protected]>
Sender: [email protected]
Organization: Brown University Department of Computer Science
Date: Thu, 4 Jun 1992 04:15:05 GMT
Lines: 33

After dealing with X11R5 programming, I got something in mind. Combining
the fast evolution of GCC in our Linux system with X-11, we just can't
decide what compiler to use, 2.11c or 2.12? The major problem may lie
in the shared library mechanism used. So the problem extends itself
to the loading/linking method used under Linux and it's not solely
solvable by gcc. We need something to link the running program without
caring about the current version of compiler or library, otherwise,
the problem will aggravate itself as Linux's application base becomes
larger. This seems to be of little problem with commercial systems since
they fix all the known problems during one release, while the gaps
between releases are much much longer than those of Linux.

The kernel needs to get something like "external loader" to do the dynamic
linking processes if we want to get rid of the above problem. Then the
modified gcc/ld should be able to generate some "stubs" to do the jobs
implicitly. The linking process should be as normal as we have done for
two decades under Unix, while the loading processes should be as flexible
as possible. I am not asking for this; I am just drawing the possible
directions to making Linux an "Open System." This topic may be worthy
of serious discussions, even though the decision to make a major change
is still in Linus's hands.

I hope that our Linux is not only a hacked system but an
experimental system with lots of new stuffs. I also think that FSF
should be supporting Linux since it is the first OS written
entirely in gcc plus some assembly code. In such a short interval,
Linux has evolved from a poor man's Unix to its current extraordinary
shape. The remaining things are not only SYSV-like but some well-planned
interesting features.



Wen-Chun Ni


[next article]
From: [email protected] (Dan Miner)

Subject: server/client ideas: Outline wanted.
Message-ID: <[email protected]>
Date: 4 Jun 92 17:05:02 GMT
Sender: [email protected]
Organization: University of Missouri - Rolla
Lines: 44


I have been watching Xwindowms and MGR being
ported. There is a server. There is a client.
Correct if I am wrong:

Server is a process (first process to be run?) that
basically forks (or spawn?) of itself to start other
process (not itselt here).

Client is the resulting process that performs a
function that is subject to the ablity of the server. The
input and output are reported to the server (how??).

I know all this is a little fuzzy but that is the
best I can come up with at this hour. ๐Ÿ™‚

I downloaded the sources of MGR and have been looking at
it. I see that is too is a server. But I can't see how

1) It can start other processes than itself.
2) How does a server "capture" the output and
input of a client?
3) Manages all the processes. They are all
running at the same time. *confusion*

I will be the first to admit that I don't know
much about Unix OS(es). I come from the DOS and Commodore
worlds.

I have been playing with fork() and it is interesting.
I have NO programmer's guides/references. Nor is there any
man pages for the kernel routines. Luckly, I have access to
a commerical 4.3 BSD that has many man pages.... So? What am I
getting at? Could someone explain what is going on under the
hood of these servers? I am really interested how a server
can obtain the input and output of another process.. ???

Thanks.
Dan (drowning) Miner
--
-------------------------------------------------------------------------
Dan Miner email:
[email protected] (prefered)
Computer Science [email protected]


[next article]
From: [email protected] (David.L.)

Subject: SIGBUS = ?
Message-ID: <[email protected]>
Date: 4 Jun 92 10:37:48 PDT
Organization: Security Pacific Automation Co., LA, CA
Lines: 9

Hi

Attempting to make TIN, in signal.c a constant SIGBUS is needed. What sould the
value be ?

I'm making it using 'sysv' as make param. correct ?
--
----> [email protected]
David L.


[next article]
From: [email protected] (John Beaven)

Subject: Genius GM-6 Mouse & X11
Message-ID: <[email protected]>
Date: 4 Jun 92 17:54:24 GMT
Sender: [email protected] (The news facility)
Reply-To: [email protected] (John Beaven)
Organization: U of Cambridge Computer Lab, UK
Lines: 26

Hi everyone

I'm having trouble setting up X11 with my Genius GM-6 Mouse.

The mouse was about the cheapest you could get about 2 years ago.
It is serial, has 3 buttons, and supposedly emulates either
MouseSystems or Microsoft mice. The way this is done is that
if you hold a mouse button down when powering the machine on, it
sets itself to Microsoft, otherwise it sets itself to MouseSystems.

Anyway, I start X the usual way, the usual things appear on the screen
including the X cursor, but I can't get it to respond to what
I do with the mouse. I've played around with the Baud Rate in Xconfig,
and I only get any response at all for 1200 and 2400, but
whenever I move the mouse, the cursor moves erratically,
and menus drop down, as if I was pressing buttons. Other baud rates
give no response at all.

So my question. Has anybody else got this mouse? If so, what
are the settings in Xconfig that you use? Do I need to put some
value in the Sample Rate (that line was commented out in the
standard distribution)? If so, what?

Thanks in advance,

John


[next article]
From: [email protected] (Darren Senn)

Subject: Re: ack. xdvi can't find fonts - help pleeze
Message-ID: <[email protected]>
Date: 4 Jun 92 19:07:18 GMT
References:
Sender: [email protected]
Reply-To: [email protected]
Organization: Curiosity Confederacy
Lines: 37

In article , [email protected]
du (Joe Waters) writes:
> texpk300.tar.Z file into /usr/local/lib/tex/fonts/pk.300, and untarred
> textfm.tar.Z into /usr/local/lib/tex/fonts/tfm. In my Makefile for xdvi, I
> set the following:
>
> DEFAULT_FONT_PATH=.:/usr/local/lib/tex/fonts/pk.300:/usr/local/lib/tex/fonts/t
fm

Check your permissions on the font files and the directory where they live

> and recompiled. But I get this error message when I try to xdvi test.dvi:
>
> Can't find font cmex10.
> Can't find font cmmi10.
> Can't find font cmr7.
> Can't find font cmr10.
> xdvi: Not all pixel files were found
>
> and then it quits. Obviously it doesn't like where I put the fonts. Could
> somebody either a) tell me what to set my compile flags to so this works
> properly or b) tell me where to put the fonts, or c) send my a statically-
> linked binary that WORKS and tell me where they have their fonts and which
> ones they have? Thanks a million.

If you want a precompiled binary, I uploaded the binaries for my own xdvi
a while back. It's on banjo and tsx-11. The file contains a dynamic
library, and a .a that you can link yourself. It's also got a pretty
verbose readme file for installation, recompiling from the original,
etc.

File:
tsx-11.mit.edu:/pub/linux/binaries/usr.bin.X11/xdvi-bin.tar.Z
banjo.concert.net:/pub/Linux/X11/Incoming/xdvi.bin.T.Z

--
Darren Senn Phone: (408) 479-1521
[email protected] Snail: 1785 Halterman #1
Wasurenaide -- doko e itte mo soko ni anata wa iru yo. Santa Cruz, Ca 95062


[next article]
From: [email protected] (Steven Samuel Jr. Seiden)

Subject: Games for X
Message-ID: <[email protected]>
Date: 4 Jun 92 16:51:16 GMT
Organization: Univ. of Calif., Irvine, Info. & Computer Sci. Dept.
Lines: 36
Nntp-Posting-Host: anita.ics.uci.edu

I've uploaded xgames.tar.Z to incoming at tsx-11.
This replaces xtetris.tar.Z. It includes xtetris
plus a few other games, all in portable .a format.
Enjoy!

Here's the README:
------------------------------------------------
This is a collection of games for X11.
All are in lib.a format, and thus can be used
with shared libs or not. Included are:

xtetris
xboard and gnuchessx - chess
othello
puzzle - sliding tile puzzle
xpuzzle - different puzzle
xchomp - pac man clone

To install, check the Makefile to see that
the paths are correct for your system.
You need sort and uniq to automatically
install the whatis lines. Certain paths
are compiled into the binaries and can't
be changed. Simply type:

make all

Sources can be obtained from
export.lcs.mit.edu and gatekeeper.dec.com
(in pub/comp.sources.x). Comments? Questions?
New additions or requests? Contact:

Steve Seiden
[email protected]
--
Steve Seiden Dept of Info & Computer Science UC Irvine


[next article]
From: Andrew Haylett (GEC-Marconi Research Centre)

Subject: selection-1.1 available, cut'n'paste for virtual consoles
Message-ID: <[email protected]>
Sender: [email protected] (Mr Background)
Reply-To: [email protected]
Organization: The Internet
Date: Thu, 4 Jun 1992 19:24:08 GMT
Lines: 25

I have released a new version of selection, the utility that allows you to
copy and paste text between virtual consoles using the mouse. There may still
be some out there like me who can't run X and who would like some consolation.

Changes from version 1.0:

- support for most common serial mouse types, hopefully.
- selection by word or line as well as by character.
- changes in the interface to make it behave more like xterm.
- patches are against 0.96a (with first patch from Linus installed).

This version comes with a test program so you can see if your mouse is
compatible without having to patch and rebuild the kernel. Perhaps it will
help some who are having problems with mice under X or MGR. I took some code
fragments from X386 to handle different mouse protocols.

Available at: banjo.concert.net in pub/Linux/Incoming/selection-1.1.tar.Z

Also sent to ftp-mail at tsx-11.mit.edu, will probably appear in
pub/linux/patches/selection-1.1.tar.Z at some point.

----
Andrew Haylett | Inet: [email protected] | Fax: +44 245 75244
GEC-Marconi Research | Tel: +44 245 73331 x.3283 | Telex: 995016 GECRES G



[next article]
From: [email protected] (Thierry Priol)

Subject: Logitech mouse with X11
Message-ID: <[email protected]>
Date: 4 Jun 92 18:45:13 GMT
Sender: [email protected]
Organization: IRISA, Rennes (Fr)
Lines: 16

I bought a Dexxa mouse with my PC and It did not work with X11. I
decided to buy a new one : a Pilot mouse from logitech which is
supported by X11. Unfortunatly, this mouse do not work with X11.

Is there someone who can help me ? (Its is maybe a FAQ!)

Thanks in advance.

Thierry Priol


--
Thierry PRIOL Phone: 99 84 72 10
IRISA / INRIA U.R. Rennes Fax: 99 38 38 32
Campus Universitaire de Beaulieu Telex: UNIRISA 950 473F
35042 RENNES CEDEX - FRANCE E-mail: [email protected]


[next article]
From: [email protected] (Mike Me)

Subject: A question on GCC header files......
Message-ID: <[email protected]>
Originator: [email protected]
Sender: [email protected]
Organization: George Washington University
Date: Thu, 4 Jun 1992 17:41:14 GMT
Lines: 25

Hi linuxers, whats up?

I'm trying to port some stuff to linux, and I'm using the gcc 2.12b.

I get an error because the program I'm trying to compile needs
2 header files. hdr.h and var.h

I looked in our HP-9000's libraries and I found vars.h

Does anyone know if it is the same thing as var.h, and where I may find
hdr.h?

I guess it should have been distributed with the GCC compiler, but
I dont know exactly what the deal is. Maybee GCC doesn't use these
headers, or names them differently.

I'm obviously a beginner when it comes to the gcc compiler and all its
tricks, I appreciate the help.

Thanx,
MikeE






[next article]
From: [email protected] (H.J. Lu)

Subject: Re: Orphaned Response
Message-ID: <[email protected]>
Sender: [email protected] (USENET News System)
Organization: Washington State University
References: <[email protected]> <[email protected]>
Date: Thu, 4 Jun 92 15:43:44 GMT
Lines: 65

In article <[email protected]> [email protected] writes:
>
>As you can see, sigprocmask() returns whatever is returned by the
>system call ssetmask(). ssetmask(), however, always returns the
>old mask, so either the library function or the system call should
>be changed.

Has been fixed in the next release along with other signal bugs.

>
>> Third, rusage support is still incomplete, right? Since the library
>> interface is already there, I decided to compile tcsh with rusage
>> support, so that it could be linked against newer libraries in the
>> future. However, after executing a commad, tcsh immediately hangs,
>> leaving the executed process as a zombie. I assume it hangs on
>> wait3(), but I haven't really tried to pinpoint this yet. Tcsh works
>> fine without the rusage stuff.
>
>Ok, now I have. It indeed hangs on wait3(). I replaced the wait with
>a waitpid()/getrusage() combination, which does the trick. I'm not
>entirely happy about this hack, since it's not atomary, but the code
>isn't too critical.
>
>Here's the source for wait3():
>
>pid_t wait3(int *wait_stat, int options, int *reserved)
>{
> if (reserved) {
> errno = EINVAL;
> return -1;
> }
> return waitpid(0, wait_stat, options);
>}
> ^^
>Shouldn't this be -1 ?!? I thought wait3() operates on _all_
>the children, not just those in the same process group.
>Well, I guess we need a new system call for this, anyway...
>

According to HP-UX man page, it should be 0. What doe the other
man pages say about this?

BTW, waitpid (-1, ...) -> wait ().


>> Fifth, is locale support already finished? I have no idea, it's kind of
>> hard to test. The library interface is there, but the sources I have
>> (outdated) only have stubs.
>
>Ok, locale support is _not_ there. Working on it?

Wait for glibc.a to support it.

>
>> And last, but not least... H.J. PLEASE, make it a practise to include
>> library sources to each major release, at least. They are really the
>> only way to make sure of what is supported and how, what is buggy and
>> what is not. Especially since there are include files for stuff that
>> is not even implemented.
>

I have done it since ..... You will get the picture if you are on the
GCC channel or wait for the official release.

H.J.


[next article]
From: [email protected] (Chris Flatters,208,7209,homephone)

Subject: Re: server/client ideas: Outline wanted.
Message-ID: <[email protected]>
Date: 4 Jun 92 19:42:17 GMT
References: <[email protected]>
Sender: [email protected]
Reply-To: [email protected]
Organization: NRAO
Lines: 35

In article [email protected], [email protected] (Dan Miner) write
s:
>
> I have been watching Xwindowms and MGR being
>ported. There is a server. There is a client.
>Correct if I am wrong:
>
> Server is a process (first process to be run?) that
>basically forks (or spawn?) of itself to start other
>process (not itselt here).
>
> Client is the resulting process that performs a
>function that is subject to the ablity of the server. The
>input and output are reported to the server (how??).

A server is a process that mediates access to some resource (eg. a screen or a
database) on behalf of other processes (the servers clients). A client is not
usually a child of the server process.

Clients and servers communicate using some form of interprocess communication (I
PC).
Examples include shared memory, message queues, named pipes and network
connections. The server must normally be running before a client attempts
to use it and must use a prearranged IPC address so that clients know
where to send their messages.

If you want to find out about this in detail, I recommend that you buy a copy
of the following book.

UNIX Network Programming
W. Richard Stevens
Prentice Hall, 1990
ISBN 0-13-949876-1


Chris Flatters
[email protected]


[next article]
From: [email protected] (peter.j.fillingham)

Subject: Bad Blocks
Organization: AT&T Bell Laboratories
Date: Thu, 4 Jun 1992 18:20:49 GMT
Message-ID: <[email protected]>
Keywords: bad,sectors,blocks
Lines: 16

As others have reported, my HD occasionally barfs, thus:
hard disk I/O error
dev 0301 block 1438
hard disk I/O error
dev 0301 block 1442
....
while emitting vague grinding noises. This is only stopped by frequent
^C's; Even F? carries this over to another login.
Is there some way to mark badblocks without reformating the HD ?
Is there (ala sysV) a badblock prg for linux ?
Occasionally one of my linux systems creates a subdir
under root with a name like /.badsectors with files which are apparently
linked to runnable prgs. Deleting the former 9i.e. those under /.bad...)
can be a disaster. Your help is solicited ....
grumble...[even CP/M had a find findbad for floppies!].
Keep the Aspidistra Flying. pete <>


[next article]
From: [email protected] (Joe Waters)

Subject: xdvi can't find fonts: SUMMARY
Message-ID:
Date: 4 Jun 92 22:04:23 GMT
Organization: The University of New Mexico, Albuquerque, NM
Lines: 15

Okay, thanks to the many people who mailed me to tell me how to set this thing
up. It turns out it's in the docs (actually, it's not in the installation doc
or the readme, but at the very bottom of the man page), but for a system setup
like the TeX binaries give you (where they put the pk fonts in
/usr/local/lib/tex/fonts/pk.300/... ) you need to set up an environment
variable XDVIFONTS=/usr/local/lib/tex/fonts/pk.%d/%f.%p
which will allow xdvi to find the fonts correctly. Thanks to everybody
who sent me answers to get this thing going.

Joe
--
/---------------------------+-------------------------------------------------\
| Joe Waters ("Falc") | "Most any problem can be solved with duct tape |
|[email protected]| and Froot Loops . . . or a good heavy club." |
\---------------------------+-------------------------------------------------/


[next article]
From: [email protected] (H. Peter Anvin N9ITP)

Subject: Re: passwords in linux and remote session
Message-ID: <[email protected]>
Date: 4 Jun 92 21:05:35 GMT
References: <[email protected]> <[email protected]>
Sender: [email protected] (Usenet on news.acns)
Reply-To: [email protected] (H. Peter Anvin)
Organization: You must be kidding!
Lines: 37

In article <[email protected]> of comp.os.linux,
[email protected] (Jim Burns) writes:
> in article <[email protected]>, [email protected] (Nathan Laredo)
says:
>
> > binaries) Also, consider be illiterate but what does
> > a password of * mean? Is it not possible to log on
> > with that user name or something?
>
> crypt() can't produce a password of '*', so it either means that login is
> disabled for that account, or it's also a common convention when shadow
> passwords are installed. There is a passwd binary in the mcc-interim 0.95c+
> stuff at ftp.mcc.ac.uk:dsk/pub/linux/mcc-interim/0.95c+. Make sure you
> install /lib/lib92.04.06 from the boot-U? image into your /lib world
> executable for any of the mcc stuff to work, if you just want to pick and
> choose utilities from the utils-U? image. See my previous post about
> installing the mcc stuff on a running system.

On all systems I have worked on, if shadow passwords are enabled, the
password in /etc/passwd is '##'. '*' is used for "no logins
allowed". I don't know about Linux since I never got the shadow password
program package to work properly, but it seems to me it should at least be
remotely similar.

The two first characters in the password is the 'salt', a 2-character
string encoding a 12-bit number using the charset [a-zA-Z0-9./]. '##' is
thus illegal (and so is '*') and used to signal 'look in the shadow
database'.

/hpa



--
INTERNET: [email protected] TALK: [email protected]
BITNET: [email protected] HAM RADIO: N9ITP, SM4TKN
IRC: Scalar NeXTMAIL: [email protected]
"Kissing a smoker is like licking an ashtray." -- R. Heinlein


[next article]
From: [email protected] (Joe Waters)

Subject: Gnuplot 3.1, color x11 emulation
Message-ID:
Date: 4 Jun 92 22:09:25 GMT
Organization: The University of New Mexico, Albuquerque, NM
Lines: 18

Okay, due to the number of responses I got about uploading a version of gnuplot
that works with x11 color emulation, I decided that instead of uploading
statically-linked binaries, I would upload either cdiffs from the gnuplot 3.1
source code OR a linkable .a file. The only thing is, I have no idea what
is required to make a .a file, so if somebody wants to mail me instructions
on how to create one, I'll upload that. Basically the only differences between
this version of gnuplot and the one that's already commonly available for Linux
is that this one was compiled with the extra options to pipe the gnuplot output
through a separate, external x11 client (called, strangely enough, gnuplot_x11)
to display color, resizable high-resolution graphs... It's a nice setup -- if
somebody will tell me how to make a .a file, I'll do it and upload to tsx-11...

Joe
--
/---------------------------+-------------------------------------------------\
| Joe Waters ("Falc") | "Most any problem can be solved with duct tape |
|[email protected]| and Froot Loops . . . or a good heavy club." |
\---------------------------+-------------------------------------------------/


[next article]
From: [email protected] (Klaus Weidner)

Subject: Re: getting linux to see internal modem on COM2
Message-ID: <[email protected]>
Date: 4 Jun 92 11:32:04 GMT
References: <[email protected]>
Organization: GSF
Lines: 34

From article <[email protected]>, by [email protected] (Vinc
e Skahan):
> I've been unable to get linux (mcc-interim) to be able to
> see (or use) the internal Hayes-y 2400 baud modem on COM2.

I have the same problem with my modem - somebody else asked the
same question on this group a few weeks ago but nobody seems to
have found a solution.

This is my configuration:
AST 386, internal 2400 modem on COM2 (works fine with Dos and Minix),
tested under Linux 0.12-0.96. The builtin COM2 is marked as disabled
in the setup menu to avoid a conflict. (For Linux this does not
change anything).

The modem is ignored totally, I do not even get garbage or anything,
so I do not think that this is an installation problem. (Yes, I
have set up /dev/ttys2 as 4/65). It acts rather like /dev/null
(except that it does not output any EOFs). The other serial line
works okay.

This is the output of `stty | speed 2400 baud; line = 0;
| min = 1; time = 0;
| -cread
| -brkint -icrnl -imaxbel
| -opost -onlcr
| -isig -icanon -iexten -echo -echoe -echok -echoctl -echoke

I hope that somebody finds a solution (other than buying a new
modem ๐Ÿ™ ), because my hard disk is rather crowded with three
operating systems, and I would like to move completely to
Linux. (Thanks to everybody involved!)

Klaus Weidner ([email protected])


[next article]
Newsgroups: comp.sys.ibm.pc.hardware,comp.unix.bsd,comp.os.linux,comp.os.os2,asc
om.bsdi,comp.periphs.scsi
From: [email protected] (Petri Helenius)

Subject: Re: Bustek-742 EISA controller (compatible with Adaptec 1742 ?)
In-Reply-To: [email protected]'s message of 2 Jun 92 19:14:53 GMT
Message-ID:
Sender: [email protected] (Petri Helenius)
Reply-To: [email protected]
Organization: Fimeko-Data Oy, Helsinki, Finland
References: <[email protected]>

Date: Thu, 4 Jun 1992 22:31:26 GMT
Lines: 18

In article [email protected] (Paul A Vi
xie) writes:
the adaptec 174x has a 154x emulation mode which seems quite compatible
with the various 154x drivers. if the bustek 74x is as similar to the
adaptec 174x as the bustek 54x is to the adaptec 154x, it will have this
emulation mode and you can just use it. perhaps a call to bustek is in
order?
--
The BUSTEK controller works fine with ISC 3.0. There is no support (?)
for it in previous versions.

Pete
--
--
Petri Helenius, Fimeko-Data Oy
Phone +358-0-458 2421, Telefax +358-0-458 2425
Internet [email protected], Snail: Hollantilaisentie 36, 00330 HKI, FINLAND
X.400 /C=fi/ADMD=fumail/PRMD=inet/O=fidata/S=Helenius/G=Petri/
Looking for Unix(r) FAX-system ? Mail queries to [email protected]


[next article]
From: "David W. Metcalfe"

Subject: Adaptec 1542B and 0.96a
Reply-To: "David W. Metcalfe"
Organization: Gated to News by demon.co.uk
Date: Thu, 04 Jun 1992 20:59:47 GMT
Message-ID: <[email protected]>
Lines: 68
Sender: [email protected] (C-News Owner)

I have recently installed an Adaptec 1542B controller and SCSI disk
into my computer which already has an ESDI controller and disk
installed as the primary drive. I had been using Linux without any
problems on a 32MB partition on the ESDI drive right up to the time
that I installed the SCSI drive.


After booting the 0.96a kernel with the SCSI drive installed I get
the following after the sock_init: message :-

status=30 intrflags=0
status=10 intrflags=0
Host 0 is detected as a(n) Adaptec 1542.
Adaptec 15421 host adaptors detected.
Detected scsi disk at host 0, ID 0, lun 0
Detected scsi disk at host 0, ID 1, lun 0
Detected scsi disk at host 0, ID 2, lun 0
Already have detected maximum number of SCSI disks Unable to
add drive at SCSI host Adaptec 1542, ID 2, lun 0
Detected 2 disks 0 tapes total.
70 00 06 00 00 00 00 28 00 00 00 00
Warning: SCSI drive at host 0, id 1, lun 0 failed READ CAPACITY
status=1, message=00, host=00, driver=00
Sense not available
Block size assumed to be 512 bytes, disk size 1GB.
harddisk I/O error
dev 300, block 0
Unable to read partition table of device 0300
Partition table ok.
harddisk I/O error
dev 0304, block 1
bread failed
Kernel panic: Unable to mount root

At this point the system stops. The following should be noted:

a) I only have one SCSI disk, not 3 as the above messages seem to
indicate. This is presumably why the second one fails READ
CAPACITY.

b) The SCSI disk contains both a primary DOS partition and an
extended DOS partition plus a 64MB partition which I was
intending to devote to Linux. Is the extended partition the
problem, even though I am not using it for Linux?

c) The Linux root partition on the ESDI drive is still accessible if
I boot with the 0.95c+ kernel with no SCSI support code. Thus it
is the presence of the SCSI support code that is preventing
access to the root partition on dev 0304.

Does anyone have any idea what is going on? From other people's
experiences this should work. Fortunately I have managed to get around
the problem temporarily by recompiling the 0.96a kernel without SCSI
support.

David


--------------------------
David Metcalfe
[email protected]
--------------------------


[next article]
From: [email protected] (Dan Miner)

Subject: Re: 2 quick questions...
Keywords: pcomm, accounts
Message-ID: <[email protected]>
Date: 4 Jun 92 22:08:01 GMT
References: <[email protected]>
Sender: [email protected]
Organization: University of Missouri - Rolla
Lines: 35

In article <[email protected]> [email protected] (Jeffr
ey L Newbern) writes:
>i just have 2 quick questions.
[deleted stuff]
>
>2) when i login as anyone but root (user for example) i get errors like "getwd:
>cannot access parent directories" two or three times when i execute a command.
>could someone explain what is going on?
>
>thanx,
>jeff newbern
>[email protected]

Sounds like you have permission problems. I have things like
that too. Now the shell and inital start programs need to be all
executables.. Second, you need to make should that the home directory
is theirs and that they too can write to it.. ๐Ÿ™‚ (I did this once.)

ei. Person: foodude

(You are root)
mkdir /usr/home/foodude
chown pid.gid foodude
(I have been using the account name as the pid and a group of student)
chown foodude.student foodude

Be sure you passwd file is set up correctly. I think you need to
look at the group file also..

Hope this helps.
Dan
--
-------------------------------------------------------------------------
Dan Miner email:
[email protected] (prefered)
Computer Science [email protected]


[next article]
From: [email protected]

Subject: Help installing on an Adaptec 1542BK
Message-ID: <[email protected]>
Date: 4 Jun 92 22:56:32 GMT
Sender: [email protected] (Notes Person)
Organization: NYU Medical Center, New York, NY 10016, USA
Lines: 13
Nntp-Posting-Host: mcclb0.med.nyu.edu

Hello *,

I would like some assistance installing Linux on an
Adaptec 1542BK SCSI host adapter with 2 Quantum SCSI HD. I tried making the
devices using the 'mknod' command but it does not work. I know that I need a
device driver but I don't know which driver I need nor from where can I get it
or How would I get it onto the root disk.
I hope that somebody will be able to help me out here. Thanks in
advance.
Edwin
[email protected]
^
that's a zero.


[next article]
From: [email protected] (Ken Corey - Operator)

Subject: fsck flags to fix ailing filesystems?
Message-ID: <[email protected]>
Date: 4 Jun 92 22:24:01 GMT
Sender: [email protected] (UNT USENet Adminstrator)
Organization: University of North Texas
Lines: 54


Interesting, fsck reports that larvsm are the flags it understands,
but my man page on my (Sun4.1.1a) Unix machine a t work, and on
my (mac)Minix don't show all those arguements. The problem is that
when I first got Linux, I also downloaded the unzip program.
In the course of unzipping something, it trashed by /usr/root
directory. Messing up the pointers somehow, so that the /usr/root
directory could not be romoved.

rmdir & rm rf returns:

warning - bad directory on dev 0811
rmdir: /usr/root: Directory not empty

ls from within /usr/root gives:
ls:.

ls root from outside gives:
?.kermrc ?.profile ?ro
?.bash_history ?.mcwd ?man.tar

Finally, fsck -r /dev/sd11 returns:
/root: bad directory: '.' isn't first
/root: bad directory: '..' isn't second
Inode 1, i_nlinks=24, counted=21.Set i_nlinks to count (y/n)?y
----------------------------
FILE SYSTEM HAS BEEN CHANGED
----------------------------

But, the next time fsck is run, it happens again. I can't figure out
what the deal is here. vile will open the directory file, but can't
write it back out (/usr/root is chmodded to 777).

Finally, today, I was using kermit in the first VC, and compiling in a two
other VC's when all of a sudden, the date command didn't work. I went to go tak
e a look at it, and the file type was ']' in a ls -al::
]r--S--S--
Now, I'm new to Linux, but is this valid? When I fsck'd the partition that
time, I had a whole slew of errors. (I had not logged out for 2 hours before
this and everything was working just fine up to this point.)

Am I doing something extremely stupid here? Can someone tell me what's going
on with a bad directory that won't go away, and files turning to mush when I
*should* be okay?

My setup is Linux .96a, 386-25, ST01 adapter with an ST296N: 32 MB Dos
40 MB Linux
8 MB swap
ST125N: 21 MB Linux

swap was enabled.

Thanks for any suggestions.

-Ken


[next article]
From: [email protected] (Drew Eckhardt)

Subject: Re: Help needed with floppy driver
Message-ID: <[email protected]>
Date: 4 Jun 92 22:29:41 GMT
References:
Sender: [email protected] (The Daily Planet)
Organization: University of Colorado at Boulder
Lines: 22
Nntp-Posting-Host: kinglear.cs.colorado.edu

In article [email protected] (Paul Lahaie)
writes:
>
>
> On another note, I wanted to recompile the 0.96a sources to remove SCSI
>support, since I only have the GCC 1.4 compiler, I modified the Makefilke
>according to the instructions. After doing make a few times (it would die
>every once in awhile), I got myself a boot image (I stripped the scsi.a line
>and removed all references to scsi_dev_init). After going thru the
>upload/download process, I rebooted from the new bootimag,e and got a kernel
>panic. Any ideas? Also, is there any documentation for Taylor UUCP?
>
> - Paul


1. Removing SCSI :
Go into include/linux/config.h, and #undef CONFIG_DISTRIBUTION. Make sure
the SCSI options are all #undef'd, and scsi support will be omitted.

2: Using gcc 1.40
I was unable to get gcc < 2.0 to compile a kernel that would not die
at some point. The problems go away with gcc 2.x, you could also
try removing the -O flag in CFLAGS everywhere it appears.


[next article]
From: [email protected] (Joel M. Hoffman)

Subject: Looking for sources (diffs, actually) for TeX, MF, and GNU Emacs
Message-ID: <[email protected]>
Sender: [email protected] (USENET News system)
Nntp-Posting-Host: next.wam.umd.edu
Organization: University of Maryland, College Park
Date: Thu, 4 Jun 1992 23:01:29 GMT
Lines: 8

Where can I find the sources used to build GNU Emacs, TeX, and MF.
What I really need is c-diffs, but I'll take a huge tar.Z file it
that's all there is.

Thanks in advance.

-Joel
([email protected])


[next article]
From: [email protected]

Subject: a simple question, and a favor needed. ๐Ÿ™‚
Message-ID: <[email protected]>
Date: 4 Jun 92 22:45:34 GMT
Sender: [email protected] (Mr Background)
Reply-To: [email protected]
Organization: The Internet
Lines: 22

first, the question.(showing my ignorance here.) ๐Ÿ˜‰
uhh, errr.... Is crontab supported? if so, how should it look, is there
a special program I need to get? (what format is the crontab file supposed
to have?)
also, what are the various *tab files that can be put in /etc, or for that
matter, any files that should be in /etc. Thanks for any answers. ๐Ÿ˜‰

Now for the favor: (if you are afraid of your mailer, skip the rest..)
I need people to send me some test mail to [email protected]
I have incoming mail working (mostly), but the outgoing mail is screwed up.
turns out I defined the wrong name in the conf/EDITME file for smail.
I can't re-compile it until I get bash, expr, uniq, tar, and compress.
Watch the rm * command folks!! (as in meaning to rm *~, and miss tilde key,
but don't notice it.) ๐Ÿ™

The good thing about rm'ing bash, I now use tcsh, and I remember why I
chose it over bash to begin with(on the sun I was on at UGA last year).

Thanks, and I am looking to hear from everyone soon!!
Jim Gifford
[email protected]
[email protected]


[next article]
From: [email protected] (Steven M. Palm)

Subject: Re: What files to grab ?
Message-ID:
Date: 4 Jun 92 03:15:45 GMT
References: <[email protected]> <[email protected]
ava.Helsinki.FI> <[email protected]> <2027.9[email protected]
or.cf.ac.uk>
Organization: Currently Lacking Organization
Lines: 39
X-NewsSoftware: Amiga GRn V1.29 4/25/92 (Beta) by Mike Schwartz

In article <[email protected]> [email protected] (Paul Richards)
writes:
>There is a version of 386bsd that does not need a co-processor (ask in
>comp.unix.bsd for details). The coprocesor is not essaential to the
>kernel but some of the utilities to set up the HD needed it. Somebody
>has fixed this for systems without a coprocessor.
>
>It's already very usable. I've got a system running NFS connected to a
>SUN 3/60 and ALMOST everything works. Telnet and ftp work fine but I've
>had a few NFS problems. If you have to have networking (which I did)
>then it's worth a look. The kernel is not as stable as Linux but it's
>getting there very quickly. It does however come with most unix
>utilities as part of the distribution.

Is this the same BSD version that is a beta-test for a soon-to-be
commrecial product, or is there a TRULY free BSD, that will remain
free for time to come. ๐Ÿ™‚

>Linux users may be interested to know that Linux has been used as an
>example on many occasions for bug fixes etc. I hope this swapping of
>info increases as both systems have an important role to play in the
>future of free unix systems.

I've had more fun under Linux than I've had in a long time. Thanks
loads to Linus and H.J. (right?) for Linux and the GCC work respectively.
You guys have put a smile on my face. And then, of course, thanks to all
the countless hands at keyboards that have made it so useable and fun to
be a part of.

I was under the impression that the 386 BSD was going to be free only as
long as it was beta, but hope Linux continues it's free status. It is one
great piece of work.

> [email protected] Internet: spedpr%[email protected]
> UUCP: [email protected] or ...!uunet!mcsun!ukc!cf!thor!spedpr

--
-- [email protected] -----------------------------------
-- Steven M. Palm - This whole world is umop ap!sdn -
-- Milwaukee, WI -----------------------------------




 December 13, 2017  Add comments

Leave a Reply