Linux-Development-Sys Digest #962, Volume #6 Wed, 14 Jul 99 23:13:57 EDT
Contents:
Re: Abit bp6 dual-celeron and dual-display (linux/X11) (Ketil Froyn)
Re: Kernel Locks (Andi Kleen)
Re: Screen Machine II (Tristan Wibberley)
Re: Kernel Locks ([EMAIL PROTECTED])
Unix queues IPC? ("Davorin Rusevljan")
Help with RAID driver ("Michael L. DePolis")
Re: New OS development (Graffiti)
Re: new modem ...new problem (Scott Lanning)
Re: MICROSOFT LINUX DISTRIBUTION (Frank v Waveren)
glibc 2.0 accessing _cnt member in FILE struct (stdio.h) ("John E. de Valpine")
Re: How to compile C++ in Linux? (Johan Kullstam)
Re: Unix queues IPC? (Warren Young)
rh6.0 and dev permissions (Don Carroll)
DMA from/to user space? (Clark Williams)
HP Tachyon drivers? (William Ryder)
Re: New OS development (Christopher Browne)
OLE in Linux? (Habin)
Re: DMA from/to user space? (Arun Sharma)
Re: New OS development (Christopher Browne)
Re: The opportunities that I would like to see Linux address. [Not that anyone
reads these things] - rather long (Christopher Browne)
Re: Abit bp6 dual-celeron and dual-display (linux/X11) (Christopher Browne)
----------------------------------------------------------------------------
From: Ketil Froyn <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: Abit bp6 dual-celeron and dual-display (linux/X11)
Date: 14 Jul 1999 18:29:25 +0200
Bryan <Bryan@[EMAIL PROTECTED]> writes:
>
> In comp.os.linux.development.apps Bertrand Renuart <[EMAIL PROTECTED]>
>wrote:
> : Great !!!
> : This is exactly what I'm looking for...
> : We are using XFree86 and wonder if it supports multiples screens under
> : Linux/i386. If so, could someone point me to some documentation ?
> I'm told that not until xfree 4.0. for now, you have to pay the $39
> for metro-x or $300 for Xig (yuk!). I think metro-x is quite based on
> xfree (the config files surely are) and its not as fast as Xig claims
> their server to be, but for my xterms and emacs windows move opaquely
> fast enough for me so I'm happy. and the price is an order of
> magnatude less. and metro-x doesn't spam the newsgroups, either...
I'm running multiple screens at home right now with a Matrox Millenium
I card and a Voodoo2(!!!). I can drag windows from one screen to the
other :) To do this, you need GGI (http://www.ggi-project.org/) and
XGGI instead of XFree86. This is all free software. I don't know any
other way to run 2D graphics on a 3D-only card.
Ketil Froyn
--
The angle of the dangle is proportional to the heat of the meat.
http://www.ifi.uio.no/~ketilf/
mailto:[EMAIL PROTECTED]
------------------------------
From: Andi Kleen <[EMAIL PROTECTED]>
Subject: Re: Kernel Locks
Date: 14 Jul 1999 19:21:26 +0200
[EMAIL PROTECTED] writes:
> A question from a linux kernel novice...
> I have code running in three "contexts" in the kernel:
> - TCP/IP packet handler (ie, ip_input)
> - IOCTL commands from user space
> - timer pops
> I need a "mutex" type of lock to synchronize access
> to common memory structures (ie, init, lock, unlock).
> I tried the pthread_mutex routines with no luck.
> I could not get the spinlocks in asm/locks.h to compile.
> Can someone point me to the correct set of functions
> to use, and how to link the appropriate libraries?
> I am using RedHat 5.2, with the 2.0.36 kernel.
pthread_* is only available in user space. Kernel is completely different.
start_bh_atomic()/end_bh_atomic() as defined in linux/interrupt.h
block out bottom halves (that covers timers and net_bh [TCP]).This is
probably what you need. Fine grained locks against interrupts/bhs make
no sense in Linux 2.0 because of the giant kernel lock. You only need
to use them in the user context function, bottom halves are guaranteed
to be atomic to itself. In the kernel Linux uses cooperative
multitasking, in case your ioctl may sleep [e.g. in a user access] you
could synchronize using a semaphore, as defined in
asm/semaphore.h. Semaphores can be only used in user context, never in
interrupt/bh context.
Make sure you compile your module with -O2 to get the correct inlines, otherwise
you may get unresolved symbols.
-Andi
--
This is like TV. I don't like TV.
------------------------------
From: Tristan Wibberley <[EMAIL PROTECTED]>
Subject: Re: Screen Machine II
Date: Tue, 13 Jul 1999 17:07:51 +0100
Reply-To: [EMAIL PROTECTED]
Reinhard Arzberger wrote:
>
> I have a Screen Machine II that used to work without faults under
> Linux 5.3 with the kernel version 2.0.36. I now use Linux 6.1
> (kernel 2.2.5)
>From the version numbers I'll assume this is the Redhat distribution.
> and the SM II doesn't work anymore.
> I got the source code from IPC and compiled it. When I try to run the
> SMII i get the error message "Segmentation fault".
> Do you have any experience with the SM II modules compiled with
> kernel 2.2.5 or could you tell me, where in the internet to find any
> information about it ?
The binary of SM II that you have was probably compiled for libc5. The
source also probably relies on libc5 quirks. Your new installation
probably hasn't got libc5 compatibility installed. You need to either
find the libc5 compatibility packages from your cds, or from the redhat
website (www.redhat.com).
--
Tristan Wibberley
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Kernel Locks
Date: Wed, 14 Jul 1999 16:13:52 GMT
Josef,
I will use both single and multiple processor machines.
On both types of machines, I need a mechanism to
prevent code in one context from interrupting another.
For example, if my user space IOCTL code is running
and modifying a global shared linked-list, I want to
postpone a timer pop from interrupting and trying to
modify that same linked-list.
Could you elaborate on the 2.0 kernel?
Thanks, Andy Dingsor
------------------------------
From: "Davorin Rusevljan" <[EMAIL PROTECTED]>
Subject: Unix queues IPC?
Date: Wed, 14 Jul 1999 14:57:09 +0200
Hi,
I was wondering, does Linux support unix queues as a means of inter process
communication?
Thnaks,
Davorin Rusevljan
------------------------------
From: "Michael L. DePolis" <[EMAIL PROTECTED]>
Subject: Help with RAID driver
Date: Wed, 14 Jul 1999 11:12:06 -0500
I have a dell poweredge 4300, need to Install Red Hat linux on it (5.2
or 6.0). This server has a
PERC2 quad channel RAID controller, which even Dell doesn't have a
driver for. Does anyone know of one or a
substitute driver that may work?
Any help would be greatly appreciated. TIA
------------------------------
From: Graffiti <[EMAIL PROTECTED]>
Subject: Re: New OS development
Date: 14 Jul 1999 13:38:01 -0700
In article <[EMAIL PROTECTED]>,
Igor Zlatkovic <[EMAIL PROTECTED]> wrote:
[snip]
>The idea behind this programming is OO. But this is not and will never be a
>class or something equivalent. What you typed above is called ADT = Abstract
>Data Type, in C it is a struct. In order to become objects, C structs need
>exactly three more features:
>
>1. Inheritance
struct snuggle {
struct wibble *parent;
...
};
>2. Polymorphy
union vivisect {
int *foo();
int *bar();
int *baz();
};
struct hack_n_slash {
enum which_one_do_we_use vivisection_switch;
union vivisect fun;
...
};
>3. Late binding
As above.
(syntax might be off. don't use unions/enums/funcpointers much).
>Structs in C do not have these attributes. You seem to think that everything
>that has data members and some sort of methods is an object. You may be smart,
>but you are wrong.
He's right. Most "OO" languages just have "everything that has data members
and some sort of methods" represented differently in the language syntax.
The underlying concepts don't change from a language where you have to
explicitly construct those from language primitives, ala C.
>> OO can be
>> (and had been) done in any language, assembler included.
>
>Nonsense. Please, can you post some code in assembler where we can see the
>above three OO features language-supported? Yes, they *must* be
>language-supported. Without native support for most important OO features, we
>don't speak about OO _programming_, but about OO _design_.
Sure.
cat > foo.cc
#include <iostream>
int main()
{
cout << "Good-bye, cruel, cruel, uncaring world!" << endl;
return -1;
}
^D
g++ -S foo.cc
See? foo.s has it.
OO *is* "language-supported" in ASM. It's just different from C++.
"OO _programming_" and "OO _design_" go together, quite often. If you
use an "OO _design_", then whatever program you use will be quite easy
to do in "OO _programming_" style.
Just because, say, you have to do:
ld1 func1;
ld2 func2;
cmp r3, r4;
eqb [r3]
bra [r4]
(no, this isn't asm from any particular machine. the meaning should be
obvious, though. :-).
doesn't mean it's not OO.
-- DN
------------------------------
From: [EMAIL PROTECTED] (Scott Lanning)
Subject: Re: new modem ...new problem
Date: 14 Jul 1999 21:12:39 GMT
Ranger ([EMAIL PROTECTED]) wrote:
: windoze and on dev/tty1 on Mandrake linux and it keep telling
: me my modem is busy !
Probably there is a lock file. On Redhat, this file is called
/var/run/ppp0.pid . So,
# cd /var/run
# ls ppp0.pid
if the file exists, then
# rm ppp0.pid
Now your modem is no longer busy.
--
Scott Lanning: [EMAIL PROTECTED], http://physics.bu.edu/~slanning
"One should not confuse this craving for change and novelty with the
indifference of play which is in its greatest levity at the same time
the most sublime and indeed the only true seriousness." --Georg Hegel
------------------------------
From: [EMAIL PROTECTED] (Frank v Waveren)
Subject: Re: MICROSOFT LINUX DISTRIBUTION
Date: Wed, 14 Jul 1999 23:14:42 GMT
In article <7mgb1t$2mp$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Christer Sandvik) writes:
> ALAN COX IS A GOD!
Alan Cox is a beard. Nobody knows what's underneath....
--
Frank v Waveren
[EMAIL PROTECTED]
ICQ# 10074100
------------------------------
From: "John E. de Valpine" <[EMAIL PROTECTED]>
Subject: glibc 2.0 accessing _cnt member in FILE struct (stdio.h)
Date: Wed, 14 Jul 1999 19:16:16 -0400
Hi:
Glibc 2.0 has opaque stdio. How can I get access to the _cnt member of
of the FILE struct. Looking into the source for glibc 2.0.6 (?) shows
that stdio/glue.c appears to resolve this. But I can find no reference
to this in /usr/include/stdio.h. Is this something that needs to be
compiled into the library specially or accessed differently.
Or is there another way to query the input buffer status?
I am running Redhat 5.2 (2.0.36) with glibc 2.0.7.
Thank you for your time.
Regards,
-Jack de Valpine
------------------------------
Subject: Re: How to compile C++ in Linux?
From: Johan Kullstam <[EMAIL PROTECTED]>
Date: 14 Jul 1999 19:42:46 -0400
[EMAIL PROTECTED] (John McKown) writes:
> On Tue, 13 Jul 1999 22:30:54 -0400, Ryan Michaels wrote:
> >I'm fairly new to Linux development.. most of the time I just use Perl. Is
> >making a C/C++ program in Linux similar to that of Windows (DOS) where you
> >write the program in a file, then compile it, or does Linux interpret it
> >like Perl?
> >
> >I don't mean to ask such simple questions, but I can't seem to find any
> >documentation on the system about this (if so, could you please let me
> >know?).
> >
> >T.I.A.
> >
> >Ryan
> >
>
> C is generally compiled. I know of some interpreters that ran back in the
> old DOS days. I don't know of any for Linux. Every distribution of Linux
> that I've heard of comes with a C/C++ compiler. It is the GNU compiler
> called gcc. The standard used by gcc is that file ending in a small
> "c" are C and one ending in a upper case "C" are C++. I prefer the
> way Borland's compiler did it with "c" vs "cpp". But whatever.
feed C programs to *gcc*; feed C++ programs to *g++*. i call all my
source files <something>.c.
> You'd
> create your C source code using an editor. Whatever you use for perl
> should be OK. Once you've written your code, you'd compile it using gcc.
> For really simple programs, you'd enter something like:
>
> gcc -o program program.c
and for C++, use
g++ -o program program.c
(i am sure .C &c are also valid suffixes)
--
J o h a n K u l l s t a m
[[EMAIL PROTECTED]]
Don't Fear the Penguin!
------------------------------
From: Warren Young <[EMAIL PROTECTED]>
Subject: Re: Unix queues IPC?
Date: Wed, 14 Jul 1999 18:28:02 -0600
Davorin Rusevljan wrote:
>
> I was wondering, does Linux support unix queues as a means of inter process
> communication?
It supports them and they work well. We have a set of programs with
28,000 LOC between them that use message queues extensively. The
program was written on a SysVR4.2 system, but the IPC bits recompiled
without even a warning. (TLI versus sockets is another story, though.
B-> )
--
= Warren -- http://www.cyberport.com/~tangent/
=
= ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m
------------------------------
From: Don Carroll <[EMAIL PROTECTED]>
Subject: rh6.0 and dev permissions
Date: Thu, 15 Jul 1999 00:58:25 GMT
does anyone know what is changing /dev/ file permissions from what
I set them to , to another user that logs in
it is driving me crazy , I have settup groups and added the users to
those
groups , but when whatever is changing them changes /dev/dsp
it messes up what I have done
/dev/dsp* /dev/audio* and /dev/mixer* /dev/ttyS*
are the ones changing
------------------------------
Date: Wed, 14 Jul 1999 19:17:19 -0500
From: Clark Williams <[EMAIL PROTECTED]>
Subject: DMA from/to user space?
I'm sorry if this is a FAQ (I haven't seen the question asked), but is
there a recommended way to set up PCI busmastering DMA directly to and
from user space, without doing copyin() and copyout()? I have not found
a routine to lock down a set of user pages, which would seem to be a
requirement for this to work.
Since you haven't hit the delete button yet, I'll ask another question
;-)
I've got scatter-gather capability, so do I really need to use
__get_free_pages() to allocate contiguous pages for my DMA buffers?
Can't I just vmalloc() a buffer and copy_from_user() the data from a
write request, then translate the vmalloc'ed pages to bus addresses and
DMA from them, right?
Any answers/suggestions/hecklers would be welcome.
Clark
--
Clark Williams ([EMAIL PROTECTED])
Chief Architect
WireSpeed Communications
(256) 704-9256
------------------------------
From: William Ryder <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.networking
Subject: HP Tachyon drivers?
Date: Wed, 14 Jul 1999 09:43:32 -0400
Does anyone know of any drivers for HP Tachyon based boards that I could
down load source for to use as an example?
TIA,
Bill
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: New OS development
Reply-To: [EMAIL PROTECTED]
Date: Thu, 15 Jul 1999 02:03:36 GMT
On 14 Jul 1999 02:47:23 -0400, Alexander Viro <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>,
>Christopher B. Browne <[EMAIL PROTECTED]> wrote:
>>- If you are planning to build an OS kernel, then you should be detailing
>> the sorts of abstractions that you intend for the kernel to provide in
>> such areas as memory management, task management, security management,
>> and device management.
>>
>>Saying "it must be object oriented" is largely devoid of meaning, in and of
>>itself.
>
>If it applies to programming style - it isn't.
If the poster does not specify whether he's referring to system
design, language design, or programming style, then we don't know what
he's talking about being "OO."
>Linux and *BSD kernels *are*
>OO. VM layer, VFS, device drivers, line disciplines - you name it. Folks,
>whenever you see
>struct foo {
> ....
> struct foo_operations *foo_op;
>}
>struct foo_operations {
> int (*bar)(struct foo *, char *, mode_t);
> int (*foo_delete)(struct foo *);
> ...
>}
>look around - 9:1 that you are looking at the equivalent of class. OO can be
>(and had been) done in any language, assembler included.
<sarcasm> But Linux is written in C, and C is *obviously* not object
oriented, like C++ is... </sarcasm>
I agree with you, Alexander, but what you're talking about is a design
that *includes* aspects that are object oriented.
This is largely not what people are talking about when they run
through and say "Here! Let's write a Truely Object Oriented Operating
System."
--
Where do you want to go, toady?
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/oses.html>
------------------------------
From: Habin <[EMAIL PROTECTED]>
Subject: OLE in Linux?
Date: Thu, 15 Jul 1999 01:21:39 GMT
Hello,
I've studied OLE (Object Linking and Embedding) programming under
Windows 95 using VC++ 5.0. The OLE has many functions and Windows 95
support it as system level.
I'm interesting about OLE in Linux.
Is there any project or group which implement OLE functions
on Linux, especailly X-Window?
KDE or GNOME do the OLE stuff? I heard that KDE has OLE stuff which
almost the same as Microsoft's architecture: In-Place editing and
look.
Blue Skies
Habin
--
Software Engineer (VC++, Linux)
Skydiver (FAI certificate, total freefall time is
50m 45s)
Company URL is "http://www.mizi.co.kr"
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED] (Arun Sharma)
Subject: Re: DMA from/to user space?
Reply-To: [EMAIL PROTECTED]
Date: Thu, 15 Jul 1999 01:40:40 GMT
On Wed, 14 Jul 1999 19:17:19 -0500, Clark Williams <[EMAIL PROTECTED]> wrote:
> I'm sorry if this is a FAQ (I haven't seen the question asked), but is
> there a recommended way to set up PCI busmastering DMA directly to and
> from user space, without doing copyin() and copyout()? I have not found
> a routine to lock down a set of user pages, which would seem to be a
> requirement for this to work.
Locking user pages is deprecated by some kernel developers. The recommended
method is to mmap pages from the device driver.
>
> Since you haven't hit the delete button yet, I'll ask another question
> ;-)
>
> I've got scatter-gather capability, so do I really need to use
> __get_free_pages() to allocate contiguous pages for my DMA buffers?
> Can't I just vmalloc() a buffer and copy_from_user() the data from a
> write request, then translate the vmalloc'ed pages to bus addresses and
> DMA from them, right?
If your hardware has limitations on where it can DMA to, then you'll
probably want to use get_free_pages() to ensure allocation in low
memory.
-Arun
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: New OS development
Reply-To: [EMAIL PROTECTED]
Date: Thu, 15 Jul 1999 02:03:34 GMT
On Wed, 14 Jul 1999 08:05:21 +0000, Igor Zlatkovic <[EMAIL PROTECTED]> wrote:
>Alexander Viro wrote:
>> OO can be
>> (and had been) done in any language, assembler included.
>
>Nonsense. Please, can you post some code in assembler where we can see the
>above three OO features language-supported? Yes, they *must* be
>language-supported. Without native support for most important OO features, we
>don't speak about OO _programming_, but about OO _design_.
Of course, this is heading down the usual "slippery slope" away from
the original subject of the issue.
Without an OS *design,* there can be no OS.
The critical issue in creating a new OS, or, frankly, of any system,
is of designing the major facilities of that system.
Programming techniques, which is precisely what Object Oriented
Programming is, represent matters to be considered *after* determining
what a system is supposed to do.
--
Rules of the Evil Overlord #34. "If my supreme command center comes
under attack, I will immediately flee to safety in my prepared escape
pod and direct the defenses from there. I will not wait until the
troops break into my inner sanctum to attempt this."
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/oses.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: The opportunities that I would like to see Linux address. [Not that
anyone reads these things] - rather long
Reply-To: [EMAIL PROTECTED]
Date: Thu, 15 Jul 1999 02:03:33 GMT
On 14 Jul 1999 20:46:42 GMT, Darren Winsper
<[EMAIL PROTECTED]> wrote:
>On Wed, 14 Jul 1999 14:05:06 -0500, Michael L. DePolis
<[EMAIL PROTECTED]> wrote:
>> 1. a better printing system.
>> lets face it LPR (and the other print spoolers) are a bunch of
>> crap. Once there is a better printing subsystem, I think we would be
>> more likely to see "drivers"
>> from the printer vendors.
>
>Would be nice.
LPD is the crucial "piece of crap."
It is useful, for some purposes, but has been little improved on since
RFC 1179 came out.
(See: <http://sunsite.auc.dk/RFC/rfc/rfc1179.html>,
<http://www.hex.net/~cbbrowne/printing.html>)
LPD suffers from the problems that it:
- doesn't track documents from source to destination, and
- isn't itself aware of document format or of what transformations
take place. Add to that that
- it isn't transactional, providing guarantees of ordering, or that
documents are printed once and exactly once. (This matters a *lot*
if those documents happen to be preprinted cheque forms, or if the
"print job" is that of stamping signatures onto cheques...)
LPRng may have provided a moderately enhanced front end, but something
like the Common UNIX Printing System <http://www.cups.org/> is more
like Where Printing Needs To Go Tomorrow.
Another view would be for there to be something like CUPS, which also
provides a CORBA-based interface so that applications could ask for
capabilities, with something like LDAP to provide a "printer directory
service."
The point here being that there needs to be an ability for graphical
applications to push their output into a queue and have a reasonably
manageable way for that output to be sent to a printer.
>> 2.Easier software installation.
>> I'm not an MS fan but for crying out loud, who wants to configure,
>> then make, then make install.
>
>I tell you what, when I get around to finish porting and release
>Scrambler, you can provide me with some nice setup scripts. Happy?
There are two choices for handling binary packages:
a) Use Debian, where getting over the 'installation hump' can be a bit
of a challenge, or
b) Use Red Hat/SuSE, where there are not yet enough tools available to
ensure the validity of RPM configuration as to permit you to expect
that packages can *always* be installed safely.
>> This is handled to two ways - first,
>> don't be a lazyass programmer,
>
>Why not? It's not like I'm gonna get paid for things like Scrambler.
Actually, the *right* idea is to try to be as *lazy as possible,* at
the right places.
RPM wasn't lazy enough.
It does provide some limited automation of running "make," and
automates verifying certain things about making sure you aren't
installing files that conflict with other packages.
Unfortunately, it doesn't automate "making sure that things are
installed according to FHS" or other such things, which means that
porting an source RPM from one distribution/architecture to another is
a nontrivial task.
>> send binaries and make source available,
>> but don't just send source.
>
>I can't be arsed to recompile Scrambler for libc, glibc 2 and glibc
>2.1 along with statically linked binaries, not to mention binaries for
>every other OS I want Scrambler to run on whenever I make a new
>version. If someone wants to do that for me, they're welcome to.
Don't bother. Distribute sources, and let those that are building
distributions worry about making sure that the results comply with
their libraries/kernels.
Rumor has it that the Debian folk have a plan to have the next "major"
version of dpkg have the ability to preferentially look for sources,
with the benefit that you could have defaults set up to compile
programs in a manner tuned to your hardware. Thus, if you have a
PentiumPro, you might run something like:
# apt-get build-from-source glibc
and it would go off, grab sources to glibc from ftp.debian.org,
compile them in a manner tuned for your architecture, and install it.
This is, of course, an 'evil commandline' way of doing things; they
already have text-UI and GUI apps that put a "friendlier" frontend
onto this sort of thing.
[Or, better still, if this is run weekly from a cron job, the Dumb
User never has to see it, but merely sees that his system is kept up
to date.]
--
10.0 times 0.1 is hardly ever 1.0.
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/lsf.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.x
Subject: Re: Abit bp6 dual-celeron and dual-display (linux/X11)
Reply-To: [EMAIL PROTECTED]
Date: Thu, 15 Jul 1999 02:03:39 GMT
On Wed, 14 Jul 1999 16:10:52 GMT, Bryan
<Bryan@[EMAIL PROTECTED]> wrote:
>In comp.os.linux.development.apps Bertrand Renuart
<[EMAIL PROTECTED]> wrote:
>: Great !!!
>: This is exactly what I'm looking for...
>: We are using XFree86 and wonder if it supports multiples screens under
>: Linux/i386. If so, could someone point me to some documentation ?
>
>I'm told that not until xfree 4.0. for now, you have to pay the $39
>for metro-x or $300 for Xig (yuk!). I think metro-x is quite based on
>xfree (the config files surely are) and its not as fast as Xig claims
>their server to be, but for my xterms and emacs windows move opaquely
>fast enough for me so I'm happy. and the price is an order of
>magnatude less. and metro-x doesn't spam the newsgroups, either...
The situation is somewhat opposite to what you're suggesting; much of
XFree86 was based on code written by the people that founded
MetroLink.
Furthermore, the fabled "modularization scheme" to be in XFree86 4.0
represents a codebase contributed to XFree86 by MetroLink.
The folks at MetroLink are a little "less aggressive" than some of the
Xi Graphics staff have been wont to be; that *might* indicate that
they're nicer to work with, but I can't verify that...
--
"... Turns out that JPG was in fact using his brain... and I am inclined to
encourage him to continue the practice even if it isn't exactly what I
would have done myself." -- Alan Bawden (way out of context)
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/xwindows.html>
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
