Linux-Misc Digest #396, Volume #20 Sat, 29 May 99 09:13:19 EDT
Contents:
Re: Commercially speaking (NEWS)
Re: core dump (NEWS)
Re: email bridge (NEWS)
Re: SuSE 6.1 as a PPP ser (NEWS)
Re: Rebuilding SRPMs (NEWS)
Re: Netscape 4.6 .rpm, .d (NEWS)
Re: How to get multiple r (NEWS)
Re: System Admin (NEWS)
Re: Diald dials out every (NEWS)
Re: first/second/third wo (NEWS)
CONFIG_NET_ALIAS (NEWS)
Re: Commercially speaking (NEWS)
Cisco 4000M for Sale (NEWS)
----------------------------------------------------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: Commercially speaking
Date: 28 May 1999 11:17 GMT
Erik Olson wrote:
>
> In comp.os.linux.advocacy gus <[EMAIL PROTECTED]> wrote:
> > Erik, I recommend that you read the license again....
> > less /usr/src/linux/COPYING.
> > Read point 0 and 1.
>
> Thanks for the /path. I read 0) and 1) and I don't see how they relate
> to GPL'd Linux Kernel calls being OK from closed source commercial apps.
> Point 1) does say you can charge money for it, by my issue was about
> the "closed source" nature.
>
> > There is nothing wrong with writing commercial applications for linux
> > just so long as you do not copy / use / base any code on the *source*
> > *code* of a GPL program. There is *nothing* wrong with using a GPL
> > program for what it is desiged for.
>
> OK, if this is the case then how is calling a GPL'd Linux kernel function
> any different than calling a GPL'd library function? Being called is what
> they both were designed for.
gcc is GPL, is anything compiled with gcc GPL? *no*!!!
So, where is the gap ... ;-)
StarOffice, Oracle, Sybase, SAP all run on Linux (and call kernel
functions ...) and none are GPL, have they broken a license? *no*!
Either they have made a mistake, or we mis-understand the license.
>
> RMS's main point in his "use GPL for libraries and not the LGPL" rant
> was to lessen library usefullness for closed source commercial products.
> This I believe was RMS's whole point of the article.
>
My take on the "spirit" of GPL is to do one and only one thing. To allow
people / organisations to contribute intellectual material (in the form
of software) to the general public. It is the request of the owner that
there is no charge to be made for any of his/her intellectual material.
The GPL entrenches that because it says that any use of the material
(The *source code*) must also be released under the same license as the
original code. Thus, the desire of the code "owner" perpetuates with the
license. The intellectual material he/she contributed remains freely
accessible to all who use it. The GPL protects the code.
Let me say it again. The GPL entrenches the rights of *everyone* to be
able to access [for free, or reasonable distribution cost] the *source
code* to any application which includes any source code derived from GPL
applications. There is no requirement to release your "product" under
GPL *unless* it, or some of it, is derived from the *source code* of
another GPL "product"...
Let me say it again. The GPL entrenches the rights of *everyone* to be
able to access the *source code* to any application which includes any
source code derived from GPL applications. There is no requirement to
release your "product" under GPL *unless* it, or some of it, is derived
from the *source code* of another GPL "product"... ;-) ;-)
basically, it is wrong to try to restrict access to any code which is
previously GPLed.
> > If I were to wrap a GPL program in a
> > non-gpl suite, there is nothing stopping me from charging for the
> > wrapper, just so long as I do not charge for the GPL program.
>
> Charging is OK, but that wrapper had better be open source and GPL'd!
>
No, the wrapper does not restrict access to intellectual property (the
object it wraps) which is under the GPL. If I write the wrapper (without
including any GPL source code), it is *my* property. I can do with it
what I want, including charge a million for it, and withold the source.
It may not make financial sense, but I can still do it. Just so long as
if I also distribute the application it wraps, I make the source
available for that application with that pplication, and do not charge
for *that* applicaion (other than reasonable distribution costs).
> So are you saying you can charge if you use GPL code and your app is
> open source. But you can't charge for the app if it isn't open, but
> you can charge for the wrapper? If that is what you are saying then
> I missed that in my interpretation of the GPL. Time for me to re-read
> it for the Nth time today.
>
I think you got a wire crossed here.
I am saying that you can charge for an application which does *not*
contain any source code derived from GPL software regardless of what
calls it makes to *other* applications even if that other application
*is* GPL.
Let me say it again ;-), the GNU Public License does not restrict the
manner in which an application is used, it restricts the manner in which
the source code can be used and distributed.
My understanding of the LGPL is far more sketchy. But, my understanding
of it is that it is a compromise between certain parties which want to
make an application available *free-of-charge* for distribution (like a
driver), without releasing the source code (because it includes
intellectual property which is valuable, and a trade-secret), but to
make the library / driver useful they have to incorporate code which
would otherwise be GPL. Their choice (without the LGPL) is to loose
secrecy of their intellectual property by releasing under GPL, or to not
release at all in which case there is one less driver/library available.
I could be way off base on that one, and thinking of another license.
> erik olson
> [EMAIL PROTECTED]
Anyway, been fun
gus
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: core dump
Date: 28 May 1999 11:17 GMT
On Fri, 28 May 1999 [EMAIL PROTECTED] wrote:
> Getting familure with Linux (RH 5.2), tried to
> look at the memory allocation in memory. I found
> "/dev/mem" and tried to read the file with 'joe',
> It was unreadable. Exited 'joe' and got a report
> of 'dump' or core dump'. I found "core" file in
> my root directory. Illedgable with any editor I
> use. "core" file is greater than 5 megs.
>
> Question:
> 1. What did I do wrong?
> 2. Is this file a 'core dump' of /dev/mem?
> 3. Is it necessary to do any repairs to
> /dev/mem?
> 4. How should I do any repairs, outside of
> reloading RH.
> 5. Just what is 'core dump'? I remember the
> statement from back in the 60s. But then
> we had 'magnetic cores'.
Core dumps are the process images dumped to disk by the kernel when the
process receives a fatal exception (eg Segmentation fault etc).
I your case the core file is from the editor joe.
Just delete it.
David
--
David Vrabel
Engineering Undergraduate at University of Cambridge, UK.
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: email bridge
Date: 28 May 1999 07:17 GMT
On Wed, 26 May 1999 15:24:12 GMT, Jack Cheng <[EMAIL PROTECTED]> wrote:
>Hello,
>
>I have a domain name hosted in my ISP but use the dial-up account.
>
>I use the Rh5.2 Linux box and 10 Win95 box with different email
>address, I want to use the Linux server to collect all the emails send
>to this domain every 1 hour.
>
>Is it possible to use single dial-up account to send/receive emails for
>more then one email address?
You have several options, offhand I can think of these:
1. Get multiple email-accounts with your ISP, let your ISPs mail-server
collect these for you, by accepting mail to your domain, and fetch them
periodically with fetchmail.
2. Ask your ISP if they will provide an UUCP feed to you. This is
a very old technology, but still of good use in cases like this. Look at
the UUCP-HOWTO for a description.
3. Subscribe your users to free/cheap email-services, have your ISP create
aliases to forward mails to the correct accounts, and use fetchmail to get
the mail from the free email-services.
I guess the most flexible solution for you, would be to use UUCP - this way,
your ISP will not have to do anything at all when you add new users. I would
guess very many ISPs are reluctant to do this, though, perhaps they have not
even heard about UUCP. Because of this, don't take no for an answer at once,
but ask them the reason, and explain them what UUCP is (which you'll find
out after having read through the UUCP HOWTO).
- Vegard
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: SuSE 6.1 as a PPP ser
Date: 27 May 1999 12:17 GMT
Technical Computing Services wrote:
>
> The subject says it all!
>
> My 2nd puter is NOT dialing into my SuSE box with a modem; I'm able
> to telnet directly to the Linux box via ethernet cards (both are
> connected
> via hub & ethernet cards, and only one can have a PPP connection at a
> time...
> unless, of course
> I can figure out how to make SuSE a PPP server)
>
> This is, again, a direct connection from a Win95 machine to SuSE 6.1
> Linux. No modem shall be used for such!
>
> Any way to do this?
>
> Much thanks!
I am unsure about what you wish to accomplish here. You state that
you cannot dial into your Linux box from Win95 via a modem, then you
state that no modem shall be used for your direct connection from Win95
to Linux!?
If you want to dial set up a PPP dialup to your Linux box, you need to
check mgetty how-to's. If you wish to access the internet from your
Win95 box through your Linux box, you need to check the IP-Masquerading
(or IP Chains--depending on the kernel), and IP-forwarding. Both of
these need to be compiled in to the kernel and some set up needs to be
done. Check the Net-3 how-to, it is very inforamtive :)
David
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: Rebuilding SRPMs
Date: 28 May 1999 14:32 GMT
"Thomas Svenson" <[EMAIL PROTECTED]> writes:
> On 26-Maj-99 21:59:47, Johan Kullstam wrote about Re: Rebuilding SRPMs
> >"Thomas Svenson" <[EMAIL PROTECTED]> writes:
>
> >> Do you have any suggestions about the best default option for K6, Pentium
> >> and Pentium II/III?
>
> >the classic pentium requires some bizzare scheduling stunts.
> >compiling for a pentium target will hurt performance on the
> >p-pro/ii/iii and vice versa. if you want the *same* binary to run on
> >*both* classic pentium and ppro, just compile for i486.
>
> Well, I don't have a Ppro.
i was using ppro as a catch-all for pentium-pro, pentium-ii, celeron
and pentium-iii. they all share the same processor core. apart from
variations in cache size, clock speed and those useless mmx
instructions, they are similar.
> Right now the machines used for Linux is one K6-200 (the first
> model) and a P120. I will probably add other boxes later, and those
> will either be PII/III or K6-x.
> Right now I just use i586 -O2 as optflags for the compiling.
nod.
my only point was that the choice of optimizing for pentium-classic or
optimizing for pentium-pro et al. is an either or choice. you do not
get both at the same time unfortunately.
as long as you don't use -march=pentiumpro to enable new op-codes, the
programs will still run and performance variation is fairly small.
what i am trying to say is this
1) if it's mission critical - compile specifically for the target host
processor. recompile for each machine architechure.
2) it it's *not* mission critical, then it just doesn't matter does
it? just use a happy middle such as i486 target.
there is no point in trying to optimize in two directions at once.
--
johan kullstam
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: Netscape 4.6 .rpm, .d
Date: 28 May 1999 10:47 GMT
Michel wrote:
>
> Matt O'Toole wrote:
> >
> > Has anyone seen Netscape 4.6 rpms or debs yet? Where? I'm too lazy to
> > install it otherwise... plus, it's nicer to have it packaged and ready for
> > an easy upgrade or uninstall if necessary..
> >
>
> The US version is only available in libc5 for both 4.51 and 4.6
> and they are tar.gz files
I just saw this in linux.redhat.announce yesterday:
New netscape packages are available. While these are not
specifically security updates, among the changes listed
are 'Fixes to improve security'; therefore it is recommended
that users update to the new packages.
SPARC packages will become available when SPARC binaries are
available from Netscape.
Red Hat Linux 6.0:
==================
i386:
rpm -Uvh
ftp://updates.redhat.com/6.0/i386/netscape-common-4.6-1.i386.rpm
rpm -Uvh
ftp://updates.redhat.com/6.0/i386/netscape-communicator-4.6-1.i386.rpm
rpm -Uvh
ftp://updates.redhat.com/6.0/i386/netscape-navigator-4.6-1.i386.rpm
Source rpm:
rpm -Uvh
ftp://updates.redhat.com/6.0/SRPMS/netscape-4.6-1.src.rpm
Red Hat Linux 5.2:
==================
i386:
rpm -Uvh
ftp://updates.redhat.com/5.2/i386/netscape-common-4.6-0.i386.rpm
rpm -Uvh
ftp://updates.redhat.com/5.2/i386/netscape-communicator-4.6-0.i386.rpm
rpm -Uvh
ftp://updates.redhat.com/5.2/i386/netscape-navigator-4.6-0.i386.rpm
Source rpm:
rpm -Uvh
ftp://updates.redhat.com/5.2/SRPMS/netscape-4.6-0.src.rpm
--
-John ([EMAIL PROTECTED])
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: How to get multiple r
Date: 27 May 1999 15:17 GMT
In article <7ijkku$n6p$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] wrote:
> I have posted several messages in the past but I rarely ever get
> replies. Since I am a newbie my questions are usually straight forwrd
> and to the point. Why don't I get replies? Any advise would be
> apreciated.
>
Posts tend to get ignored for a few reasons. Either you're posting in
the wrong group, or you give a vague subject line ("i need help" is a
bad subject line), but most likely, your question has already been
answered, be it in a FAQ, HOWTO, or countless times in the newsgroups.
I always do a power search on dejanews before I post a question, and 9
times out ot 10, I either find an answer to my question, or glean enough
information to figure it out for myself.
I looked at your posts, and they were X questions, so they belong in
comp.os.linux.x. While I'm up, I'll give a couple quick answers to your
posts:
video card problem: re-run whatever X configuration program you like
(Xconfigurator, XF86Setup, xf86config).
apps menu problem: this is a window manager-specific problem.
different window managers have different ways of adding items to menus,
so be sure to specify your window manager when you post in c.o.l.x or
better yet, do a power search in dejanews - you may find you answer
without needing to post.
--
fred anger
http://members.home.net/twist/
--
--== Sent via Deja.com http://www.deja.com/ ==--
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: System Admin
Date: 28 May 1999 10:32 GMT
_Essential Unix System Administration_, Frisch and Vol 8
of the X Windows System series, both O'Reilly of course.
-ckm
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: Diald dials out every
Date: 27 May 1999 03:17 GMT
marco tephlant wrote:
>
> Im pleased to say i've got IP masquerading and diald working this
> weekend, one problem though is that diald spontaneously dials out. I've
> looked through the man pages and checked theres no cron job causing it
> but can't figure out whats making it happen, I have two Win98 PC's
> connected to the network, neither of them were running any network app
> or anything. As a test I left the server and one PC switched on for a
> couple of hours and didnt touch anything, but logs still showed it
> connecting every fifteen minutes.
>
> Any tips as to what this could be?
>
Did you set it to be active with control-panel? or KDE?
If that is the case you are technically connected 24hrs a day. If something
disconnects you the redialer will happily reconnect you. It looks like your
ISP disconnects you after 15 minutes of inactivity. Linux is nice enough to
reconnect you.
--
Tired of Windows' rebootive multitasking?
then try Linux's preemptive multitasking
http://www.netonecom.net/~bbcat/
We have software, food, music, news, search,
history, electronics and genealogy pages.
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: first/second/third wo
Date: 28 May 1999 23:47 GMT
pspc wrote:
> No, it's got nothing to do with the idea of the "New World," which is
> how the Americas have been described since Columbus. "Third World" is
> a more recent idea. This is a scheme that some academic, whose name I
> can't remember at the moment, came up with to describe how he divided
> the post-WWII world into various economic blocs. As I recall it was:
>
> First World: Industrialized nations (US, Canada, Western Europe).
> Second World: The Communist bloc (Soviet Union and East European
> satellites).
> Third World: Developing nations (for example, Brazil, India).
> Fourth World: Undeveloped nations (for example, Bangladesh).
My interpretation:
First World: Us
Second World: Them
Third World: Where you won't drink the water
--
Cheers,
Bev
===================================================
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: CONFIG_NET_ALIAS
Date: 27 May 1999 10:17 GMT
Hi all,
I wish to upgrade my 2.0.36 kernel to the 2.2.9 kernel.
The problem is that my machine got an IP alias (it have ifcfg-eth0
and ifcfg-eth0:0) and the 2.2.9 kernel doesn't have the
CONFIG_NET_ALIAS option in the kernel configuration. It only
have CONFIG_IP_ALIAS but it doesn't seem to be enougth.
Can someone tell me how to keep my alias even with this kernel ?
I tried to launch ifup-aliases eth0:0 but it didn't work.
Maybe a script need to be modified but wich one and how ?
Thanks
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Re: Commercially speaking
Date: 27 May 1999 10:02 GMT
Erik Olson <[EMAIL PROTECTED]> writes:
>>> Now this has confused me, making system calls to the Linux kernel by non
>>> open source commercial programs should be illegal, or at least my
>>> interpretation of the GPL makes me think so. But apparently it is not.
>>> I don't understand. Can you explain why?
>> Why would it be?
>Because the Linux kernel is GPL and not LGPL, and the GPL states "programs
>that make library calls (aka kernel calls) of a GPL'd program, whether
>static or dynamically linked must have the source code freely available" or
>something like that. This I believe is the main distinction between the
>GPL and the LGPL.
I hate to pull a proof by intimidation here, but the best answer to
your question is "because Linus said so." To quote from the COPYING
file in the Linux source code distribution:
NOTE! This copyright [GNU] does *not* cover user programs that use kernel
services by normal system calls - this is merely considered normal use
of the kernel, and does *not* fall under the heading of "derived work".
So he explicitly releases programs that make system calls from the
requirements of the GPL. He has also allowed kernel modules to be
distributed sans source as well, as long as the modules do not require
modifications to the kernel itself. And since Linus and kernel
contributors who presumably agree with his interpretation own the
copyright to the Linux kernel, it is not up to you or me to say that
the Linux license doesn't allow this software or that to be run under
Linux.
You might feel this is dumb, but I think this is internally consistent.
>Here is my logic:
>
>1) Any program can be twisted and bent into the form of a library.
Fine.
>2) So if indeed the "dynamic linkage of libraries" (as opposed to static
> linkage) falls apart in count, then the GPL is basically downgraded to
> the equivilent of the LGPL.
The reason you made this statement in the first place, I think, is
because the Linux kernel doesn't infect the programs it runs with
GPL. As I mentioned above, Linus has explicitly granted users immunity
from that nonsense, so this part of the argument is now specious.
No matter what you do to a GPL'ed source, it stays GPL. You cannot ever
make it LGPL without convincing the copyright holders to do it for you.
> Then I can legally turn any GPL program
> into a LGPL library and use it as commercial closed source.
No.
>3) A wrapper, ignoring the inefficencies and pointer manipulation
> limitations for a moment, is functionally the same as a library call.
OTOH, a wrapper does not give you nearly as much flexibility as a
library call or tweaking the original source code. It merely lets
you use the program in the same way you would have if it were closed
source to begin with.
>4) So with the wrapper case, I don't even need to wait for the GPL to get
> beaten up in court inorder to use a GPL program commercial closed source
> as HP's gdb based GUI debugger demonstrates.
Yes. If this is what interests you, go ahead. Sun has been using
XEmacs as part of its developer suite (definitely not open source) for
some time. Netscape Communicator for Unix is distributed with GNU
movemail, but it doesn't provide anything under GPL except for
movemail itself. Metrowerks is working on an (presumably
closed-source) IDE for Linux with gcc as its back end. There is a free
IDE called Code Crusader and Code Medic which uses gcc and gdb as
back ends. Their sources are freely available, but they cannot be GPL'ed
because they depend on a closed-source widgets library (xforms).
These are just off the top of my head; I'm sure there are many other
examples of non-GPL'ed wrappers around GPL'ed software.
>Game over man.
Not so fast.
I don't see why you are so worked up over wrappers. Closed-source
vendors using wrappers around open-source software aren't taking any
more advantage of open-source software than open-source software that
are wrappers around closed-source software take advantage of the
closed-source software. Any GPL'ed Java application that is run on
Sun's JVM fall under this category. Running GNU bash on Solaris and
typing 'ls' in it would qualify, too. So far I haven't heard any
argument from Sun that either of these actions would put the user
applications under whatever licenses Sun would like.
It is worth pointing out that people have been writing non-GPL'ed
wrappers around GPL'ed software for ages. The last time I checked,
GPL'ed software is hardly in a "game over" state.
>> A wrapper can not behave like a library.
>
>Yes it can.
Here is a challenge: wrap up GNU readline so that it can be used from
arbitrary proprietary programs. If you can get it to work, some people
might be interested in it, as (AFAIK) there really isn't any good
non-GNU alternative to readline. I don't doubt it is possible, but I
don't think it will be particularly simple, efficient, or feature-
complete either.
>My point was shell scripting languages can call programs like "rmdir",
>this is what I mean as a wrapper (calling a program instead of a function).
>They can do this legally even if the program is GPL'd. I know rmdir is
>BSD and not GPL, work with me here, just pretend that rmdir is GPL. So in
>the case of scripting languages the source is freely available since that
>is the way shell scripts are. Now throwing in the existence of a perl or
>csh script compiler, you now can have closed source perl scripts.
>
>So please tell me, how is the wrapper for "rmdir" (calling it as a program)
>functionally any different that doing a rmdir(const char *pathname); from
>your program?
rmdir(1) itself is little more than a simple wrapper for rmdir(2) with
a small option parser. So no, a wrapper for the GPL'ed rmdir binary
is not very different from making the rmdir() call from your program.
OTOH, you didn't save yourself very much work, and you did lose
performance with the extra fork-and-wait burden, which is probably
insignificant in this case.
For more complex programs, the difference may become more
obvious. Let's say you are writing some sort of program that needs an
embedded command line and a script parser, and you decide to use bash
for this purpose. If you are willing to GPL your program, you just rip
apart bash's source code, embed it into your oown program, and maybe
add a few internal commands for your purposes. If you want to thumb
your nose at RMS by using wrappers, well, good luck. You might get
something usable out of it, but it certainly won't win you any awards
in the performance department.
--
Shimpei Yamashita <http://www.submm.caltech.edu/%7Eshimpei/>
------------------------------
From: NEWS <[EMAIL PROTECTED]>
Subject: Cisco 4000M for Sale
Date: 28 May 1999 04:17 GMT
serious only email [EMAIL PROTECTED]
4000M full memory & flash, mint.
Just replaced with 7500 series router.
$2000 U.S. firm.
IOS 11.x
You pay UPS. Serious only.
------------------------------
** 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.misc) 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-Misc Digest
******************************
