Linux-Misc Digest #212, Volume #20 Sat, 15 May 99 03:13:09 EDT
Contents:
Re: GNU reeks of Communism (Peter Seebach)
Re: well, what about Debian? (Marco Anglesio)
Re: Netscape 4.51 suddenly exits ????????????? ("Gero H. Marten")
PCI modems and Linux: still there... (Andrew Comech)
Re: good, free ORB w/ C++ & Java lang support (Robert Lynch)
Re: What happened to fdformat (Rob Komar)
Re: Control Panel problem (Jeremy Nickolet)
Re: Pro-Unix vs anti-WinTel ([EMAIL PROTECTED])
Chat server (S P Arif Sahari Wibowo)
Re: Pro-Unix vs anti-WinTel (Alexander Viro)
Re: page faults (Rob Komar)
Re: Most stable kernel? (brian moore)
Re: kernel compile error (jason)
Re: Enough already! Re: Pro-Unix vs anti-WinTel (Alexander Viro)
FIXED(?): Staroffice/glibc problem (Peter Englmaier)
----------------------------------------------------------------------------
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.linux.advocacy,gnu.misc.discuss
Subject: Re: GNU reeks of Communism
From: [EMAIL PROTECTED] (Peter Seebach)
Date: Sat, 15 May 1999 05:11:59 GMT
In article <8a4%2.4163$[EMAIL PROTECTED]>,
Christopher Browne <[EMAIL PROTECTED]> wrote:
>This compares nicely to the situation we see where "decent colour TVs"
>are being produced in great quantity; the value, and thus the price,
>of a TV set, has fallen considerably.
True.
However, one of the things this means is that our current standards of
poverty are *much* different from the ones we used to have. We take quite
a lot for granted...
>30 years ago, having a big black+white TV was a sign of being quite
>wealthy.
>Today? The kids may have a 27" set in the playroom, and the fact of
>having a TV indicates *nothing* about one's economic well-offness.
Pretty much.
>The unfortunate part is that this implies that there *isn't* a fixed
>definition of what poverty truly means. Which makes life difficult
>for those trying to figure out if they're helping *and* for those
>with less gracious "agendas" both.
There's two kinds of "poverty". There's "less than everyone else", and
there's "really very little". About the same number of people have "less than
everyone else" most of the time in free market economies. It'll fluctuate,
but it'll wobble back.
"really very little" is a lot rarer than it used to be, but one thing
that we do is, as we get used to more luxuries, we adapt the rules. So,
you're now in "grinding poverty" if you can only get the level of nutrition
and medical care that the wealthiest people could just barely afford a couple
generations back.
Which isn't to say that it doesn't suck; it's just to say that part of that
is a change in expectations. Think about literacy. Not that long ago, 20%
literacy would have been an amazing accomplishment. Our standards are moving,
with the result that even the poorest of the poor are, century after century,
noticably better off.
-s
--
Copyright 1999, All rights reserved. Peter Seebach / [EMAIL PROTECTED]
C/Unix wizard, Pro-commerce radical, Spam fighter. Boycott Spamazon!
Will work for interesting hardware. http://www.plethora.net/~seebs/
Visit my new ISP <URL:http://www.plethora.net/> --- More Net, Less Spam!
------------------------------
From: [EMAIL PROTECTED] (Marco Anglesio)
Subject: Re: well, what about Debian?
Reply-To: [EMAIL PROTECTED]
Date: Sat, 15 May 1999 04:25:50 GMT
On 15 May 1999 01:00:49 GMT, Andrew Chung <[EMAIL PROTECTED]> wrote:
>Debian is "traditionally" seen as a distro geared more towards programmers and
>such.. I think someone told me that there's more deb packages than any other
>distros but I'm not sure.. I've also heard people say that it's difficult to
>install, although I've installed it without any problems..
That's a yes and no situation. There are more Debian-supported debs than
Red Hat-supported rpms, because nearly every deb is supported by a member
of the Debian team. That makes for excellent qc, since nearly every
package available is controlled for such purposes, but it makes for slower
introduction of new software.
marco
--
,--------------------------------------------------------------------------.
> Marco Anglesio | One of me stayed on the ground, <
> [EMAIL PROTECTED] | without provisions or hope or <
> http://www.the-wire.com/~mpa | sight or legs, and refused to leave. <
`--------------------------------------------------------------------------'
------------------------------
From: "Gero H. Marten" <[EMAIL PROTECTED]>
Subject: Re: Netscape 4.51 suddenly exits ?????????????
Date: Fri, 14 May 1999 14:33:08 +0200
Julio De Gregorio wrote:
> What can I do????????
Nothing, but to hope that they fix all the bugs in the next version.
--
Gero H. Marten
<http://www.provi.de/gmarten/index.html>
--
------------------------------
From: [EMAIL PROTECTED] (Andrew Comech)
Crossposted-To: comp.os.linux.hardware
Subject: PCI modems and Linux: still there...
Reply-To: [EMAIL PROTECTED]
Date: 2 May 1999 21:13:08 -0500
Hi,
Below is the email which I received from Richard Nelson from Actiontec
Electronics regarding their particular PCI modem (the email contains
installation details). Richard says that that modem indeed worked
under Linux, and that it may be available as early as in June.
The copy of that email lives at
http://www.math.sunysb.edu/~comech/tools/PCImodems.html
where I am trying to keep the latest development in this area.
I am doing this not because there is anything exciting about using PCI
modems, but in attempt to give the precise meaning to the rumor that
"there are PCI modems work under Linux".
AFAIK, PCI modems which are available today _do not_ work under Linux.
Best,
Andrew
PS. In my opinion, this modem does not have much of the future: It is to
cost around $100, while the same speed ISA modems are already available
for as low as $40; and, what is even more important, Cables are coming...
=======================================================================
>From [EMAIL PROTECTED] Mon Apr 26 16:24:43 1999
1) I have talked with RedHat on numerous occasions, and have sent
them the FAQ documents Friday. We will also be sending them a test
unit soon as well.
2) The reason it [Linux] isn't mentioned with OS/2
[at http://www5.pc.ibm.com/us/products.nsf/$wwwPartNumLookup/_33L4618 ]
is because I had just finished testing it under OS/2 and hadn't tested
it under Linux yet.
3) Attached is the FAQ document I wrote up, it has been sent to
RedHat and a few other's as well. Feel free to post it, Thomas Dodd
also helped greatly, along with Michael ( forgot the last name,
sorry ) and a few others. ( Tom, if you're reading this, call me at
work, I forgot your number!:)
4) Believe me, I COMPLETELY understand the need for not giving
newbies in particular the wrong advice, or the wrong idea. The last
thing we want is for everyone to go out and buy a bunch of PCI WIN
MODEMS ( such as the desklink Pro PCI ) and expect them to work under
Linux.
Regards
Richard Nelson
***INSTALLATION OF V90 PCI LUCENT VENUS BASED MODEM***
I. After physically installing the modem, log into your "Root" account.
II. To determine resource settings.
Type the following command: cat /proc/pci
An example of the outcome will look like this;
5. Non-prefetchable 32 bit memory at 0x00000000.
Bus 0, device 17, function 0:
Communication controller: Unknown vendor Unknown device (rev 0).
Vendor id=11c1. Device id=480.
Medium devsel. Fast back-to-back capable. IRQ 11. Master Capable.
No bursts.
Min Gnt=252.Max Lat=14
Non-prefetchable 32 bit memory at 0xe4000000.
I/O at 0x6400.
I/O at 0x6800.
I/O at 0x6c00.
III. Take note of the IRQ, and FIRST I/O address, then type in this command;
setserial /dev/modem uart 16550A port 0x6400 irq 11
NOTE: THE IRQ AND PORT MAY VARY, IT IS DEPENDANT UPON THE RESULTS FROM STEP II
IV. To set Linux to configure the modem upon boot-up.
There are several options. One method is going to the /etc/rc.d/
directory, and using an editing program such as "jed" edit the
rc.local file, and insert
setserial /dev/modem uart 16550A port 0x6400 irq 11
as the last line.
_________________________________________________________
*TEST CONDITIONS:
OS: RedHat Linux 5.2 Boxed CD with Linux kernel 2.0.36
BIOS: AWARD
MOTHERBOARD: Shuttle Hot 557 w/i437 VX chipset and latest BIOS update.
CPU: Intel Pentium 166MMX cpu
MEMORY: 32 MB EDO RAM
VIDEO: PCI S3 Virge DX with 2mb memory
STORAGE: 820 Quantum EIDE Hard Drive, ATAPI CD-ROM, 3.5" floppy
SOUND: Creative Labs Sound Blaster AWE 32 PNP
* As each user with Linux is more than likely to have a unique setup,
the details listed above should comply with other versions of Linux as
well, ie Slackware, Redhat, etc. But are NOT 100% guaranteed to have
an identical setup method as far as exact commends, etc. are concerned.
Richard Nelson
Actiontec Electronics, INC
--
Looking for a Linux-compatible V.90 modem? See
http://www.math.sunysb.edu/~comech/tools/CheapBox.html#modem
------------------------------
Date: Fri, 14 May 1999 21:41:44 -0700
From: Robert Lynch <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux,linux.redhat.misc
Subject: Re: good, free ORB w/ C++ & Java lang support
Salman Ahmed wrote:
>
> I am looking to learn CORBA, and want to know if someone
> can recommend to me a good, free implementation of
> CORBA for Linux (RedHat 5.2 w/ kernel 2.0.36).
>
> I intend to write CORBA apps in both C++ and Java, and
> won't be using any of the advanced features of the CORBA
> spec.
I'm reading an article in the latest Linux Journal, part 1 of a 3 part
series.
They are using omniORB available from:
http://www.uk.research.att.com/omniORB/omniORB.html
FWIW. Bob L.
> Also, what is the recommended JDK for Linux ? I know that
> Blackdown's JDK is pretty highly recommended. Are there
> any other good Java implementations for Linux ?
>
> Thanks for any info.
>
> --
> Salman S. Ahmed
> ssahmed AT interlog DOT com
>
> Remove the "nospam." portion from my email address
> to reply to this message.
--
Robert Lynch-Berkeley CA [EMAIL PROTECTED]
http://www.best.com/~rmlynch/
------------------------------
From: [EMAIL PROTECTED] (Rob Komar)
Crossposted-To: linux.debian.user,linux.redhat.misc,comp.os.linux.hardware
Subject: Re: What happened to fdformat
Date: 15 May 1999 05:52:40 GMT
[EMAIL PROTECTED] wrote:
: I know this sounds stupid, but I used to think that you can low-level
: format a floppy in Linux using "fdformat". Well, on my Debian system
: this is what I get:
: ----------------------------------
: histria ~ # fdformat
: bash: fdformat: command not found
: histria ~ # man fdformat
: No manual entry for fdformat
: -----------------------------------
: (as root). How can I format a floppy ?
fdformat is part of the util-linux-2.9X package. You could download
the source code and build it yourself if you're feeling adventurous.
Cheers,
Rob Komar
------------------------------
From: Jeremy Nickolet <[EMAIL PROTECTED]>
Subject: Re: Control Panel problem
Date: Sat, 15 May 1999 05:05:54 GMT
[EMAIL PROTECTED] wrote:
>
> I am running Linux RedHat 5.0 on a 486/66 with 32 mg
> ram, a generic video card in VGA mode with 512k
> memory, 640x480x16.
>
> When I run startx and come up to my menu system,
> control panel is not there (name is, but no icon). I
> can go to xterm and input this command 'control-panel
> & ' and up comes an error message ' unable to find
> usable depth '.
>
> Is this a video problem?
>
> What makes me suspect is that I get a similar message
> after starting Netscape, and the colors go haywire on
> the screen, Netscape works fine, but the colors flash
> when going back to the desktop.
>
> Do I need more memory on the card or is there a
> config file that needs to be edited?
>
> Thanks for any help in advance
>
> Terry
>
> --== Sent via Deja.com http://www.deja.com/ ==--
> ---Share what you know. Learn what you don't.---
Is your "16" referring to 16 colours or 16 bit colours? I believe
control panel needs at least 16 bit colours.
--
Jeremy
http://members.home.com/nickoljt/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Pro-Unix vs anti-WinTel
Crossposted-To: comp.os.linux.development.apps,comp.unix.bsd.freebsd.misc
Date: Sat, 15 May 1999 05:52:18 GMT
G. Sumner Hayes <[EMAIL PROTECTED]> wrote:
: [EMAIL PROTECTED] wrote:
>snip<
:> I'll take source installs any day of the week, thanks.
:
: Then ports is definitely a better overall system as far as I can tell. I
: do think that you're missing a lot of the point of rpm and dpkg, though.
: They have different aims from ports (and even different aims from each
: other) and are quite useful in their domains.
I never said Debian wasn't a good system (although RPM is trash:-).
Debian is one of the best package systems going.
Any FUD spewing fool that pops up and says that ports is total crap
however, needs to be put in there place. :-)
:>: For me, ports would be a great thing.
:> www.FreeBSD.org You could be up and running minutes. :-)
: vi is better than emacs. ;-)
And Joe's Own Editor is better then both of them. :-)
--
-Zenin ([EMAIL PROTECTED])
My code is filled with comments! It's just that my comments are
written in Perl.
------------------------------
From: S P Arif Sahari Wibowo <[EMAIL PROTECTED]>
Crossposted-To: linux.redhat.misc
Subject: Chat server
Date: Fri, 14 May 1999 23:58:44 -0500
Hi!
I like to install a chat server in my Linux box, something with multi room
capability and emote feature.
Do you have any suggestion?
Thanks.
S P Arif Sahari Wibowo
_____ _____ _____ _____
/____ /____/ /____/ /____ [EMAIL PROTECTED]
_____/ / / / _____/ http://spas.8m.com/
------------------------------
From: [EMAIL PROTECTED] (Alexander Viro)
Crossposted-To: comp.os.linux.development.apps,comp.unix.bsd.freebsd.misc
Subject: Re: Pro-Unix vs anti-WinTel
Date: 14 May 1999 23:15:53 -0400
In article <ia4%2.4165$[EMAIL PROTECTED]>,
Christopher Browne <[EMAIL PROTECTED]> wrote:
>
>I've not build any Debian packages from source, so I can't speak for
>it; I *have* built some RPMs, and RPM has the "lack" that you can't
>just "fiddle 'til you get it right," but rather, in order to use a
>.src.rpm file, need to do a "big bang compile pretty much all of it in
>scripted manner."
With Debian packages:
$ dpkg-source -x nethack_3.2.2-15.dsc
dpkg-source: extracting nethack in nethack-3.2.2
% untared pristine source from nethack_3.2.2.orig.tar.gz and pathched it
% with nethack_3.2.2-15.diff.gz
% cd nethack-3.2.2
$ debian/rules build
% building the thing. debian/rules starts from #!/usr/bin/make -f and it
% invokes main Makefile from nethack (plus wades into doc/ and does make
% there).
$ fakeroot debian/rules binary
% bunch of sanity checks, generation of package metadata and building of
% binary package.
$ ls ../nethack*
../nethack_3.2.2-15_i386.deb ../nethack_3.2.2.orig.tar.gz
$ ls debian/tmp
DEBIAN etc usr var
% IOW, final tree is left there. debian/tmp/DEBIAN/* contains package
% metadata.
% OK, I don't like something
$ vi include/dungeon.h src/dbridge.c
% hack on...
$ rm build-stamp
$ debian/rules build
% accurate rebuild follows. make(1) is make(1) - rebuilding only the stuff
% that depends on the modified files.
$ fakeroot debian/rules binary
% building the package again...
% OK, satisfied this time.
$ fakeroot debian/rules clean
$ cd ..
$ dpkg-source -b nethack-3.2.2
dpkg-source: building nethack using existing nethack_3.2.2.orig.tar.gz
dpkg-source: building nethack in nethack_3.2.2-15.diff.gz
dpkg-source: building nethack in nethack_3.2.2-15.dsc
$
So there... You've got a binary package (could just say fakeroot debian/rules
binary - it would invoke build) and debian-source constructed modified patch
and signature file. Minimal build being
$ dpkg-source -x <somewhere>/foo-bar.dsc; cd foo
$ fakeroot debian/rules binary
$ cd ..; rm -rf foo foo.orig.tar.gz
>If there's some obscure thing broken, it would be nice to just go in
>and do repeated "makes," patching things as needed until it works.
>RPM isn't too happy with that approach...
Works with dpkg and with ports - both are built on make, so it's not
a big wonder...
--
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid. Get yourself a better computer" - Dilbert.
------------------------------
From: [EMAIL PROTECTED] (Rob Komar)
Subject: Re: page faults
Date: 15 May 1999 06:17:17 GMT
Bob Tennent ([EMAIL PROTECTED]) wrote:
: On Fri, 14 May 1999 21:32:39 GMT, NF Stevens wrote:
: >> >
: >>OK, I've tested my RAM using memtest and it detected no errors.
: >>So what *else* might be causing apparently random page faults?
: >
: >I had a couple of those page fault things immediately after
: >upgrading to kernel 2.2.5. I cleaned all the dust out of the
: >fan on my processor. It hasn't occurred again yet. (Touch wood).
: >
: Good suggestion. In fact my CPU fan wasn't even moving. Since
: replacing it, I haven't had a page fault. Thanks.
Well, that's good news, but to be pedantic, you've probably had
many thousands of page faults since then. Not all of a process'
pages may be in memory when the program is running, and when it
tries to access a page that isn't there, a page fault is generated.
If all goes well, the required page is fetched into memory and
the program continues to run. So, page faults are a normal
occurrence when a process runs. Now, those failed kernel paging
requests you saw are a different matter!
Cheers,
Rob Komar
------------------------------
From: [EMAIL PROTECTED] (brian moore)
Subject: Re: Most stable kernel?
Date: 15 May 1999 06:20:23 GMT
On Sat, 15 May 1999 04:20:13 GMT,
Alan Fried <[EMAIL PROTECTED]> wrote:
> jik- <[EMAIL PROTECTED]> wrote:
> Yes I had the same problem as well and in addition I am unable to get sound support
> going as well as the printer. This is also why I hesitate to upgrade from Red Hat
> 5.1 to Red Hat 6.0.
>
> Any other opinions?
I dunno, I upgraded my Slack machine at work from what I call "Slack
3.0+++++" (since it started at 3.0 a long time ago :)) to the beta Slack
4 without a hitch. Runs better than ever.
My latest system came with RH5.2+stuff (ie, not a stock RH :)) and
already had a 2.2 kernel on it and it runs great as well.
Only time either has been rebooted is because I needed to. (I added a
sound card to my work machine, so now I can use the CD burner on my home
machine to burns reams of mp3's onto a CD to listen to at work when it
gets too noisy.)
--
Brian Moore | "The Zen nature of a spammer resembles
Sysadmin, C/Perl Hacker | a cockroach, except that the cockroach
Usenet Vandal | is higher up on the evolutionary chain."
Netscum, Bane of Elves. Peter Olson, Delphi Postmaster
------------------------------
From: jason <[EMAIL PROTECTED]>
Subject: Re: kernel compile error
Date: Sat, 15 May 1999 02:17:47 -0400
James Chang wrote:
>
> Hi there
> During kernel compiling, I met the error message as following
> "cs4234.o(.text+0x81): undefined reference to 'ad1848_detect'"
> Could anybody tell me the meaning of error message and how to solve the problem?
Hmm, just a guess... try backing up your /usr/src/kernel/.config, do a
'make mrproper', then copy back your .config, and re-compile.
Good luck,
-jason
(to reply via email, make the appropriate substitution in my email address)
------------------------------
From: [EMAIL PROTECTED] (Alexander Viro)
Crossposted-To: comp.os.linux.development.apps,comp.unix.bsd.freebsd.misc
Subject: Re: Enough already! Re: Pro-Unix vs anti-WinTel
Date: 15 May 1999 02:22:31 -0400
In article <[EMAIL PROTECTED]>,
Donn Miller <[EMAIL PROTECTED]> wrote:
>Alexander Viro wrote:
>
>> Folks, could we fscking stop this idiocy? All packaging systems in
>> question are documented. STFW and RTFM. It is there. None of those systems
>> is ideal. IMO RPM is the weakest of them, but you may have different needs
>> and YMMV. Could we bloody stop this holy war and RTFM instead of imitating
Wow. Fear: when you generate blocktext and do not notice it...
>Agreed. But how about Slackware's tgz package format? I don't
>know if it handles dependencies or not; I always thought that
>Slackware just installed the package "forcibly" ala dpkg --force.
Well, duh ;-) .tgz is hardly a package format, even less so a Slackware's
anything. It's .tar.gz, spelled that way for keeping things on braindamaged
medium (DOS diskettes). What Slack used to have (disclaimer: I didn't look
into more or less recent releases) was a tarball which contained /install.sh
I.e. to install the thing you could say cd /; tar zxf wherethefsckitsits.tgz;
/install.sh. Now, there is a tricky part: there is more that one might want from
the packaging system.
* binary dependencies. I.e. to install foobar 2.37 you should have
libmumblefrob 12.x where 4<=x<97, or you'll screw the system. It is actually
even trickier: there are different kinds of contraints. The simplest being:
to have this version of this package you should <boolean expression involving
presence of packages and ranges of versions>, or the package will not work.
Nastier: don't even think of installing this version of package unless
<expression>. Not only it will not work, it will break the system (e.g. with
the upgrades of critical libraries order of unpacking *really* matters).
Another case: installation of foo breaks bar, but foo is a valid replacement
(again, bar may very well be an expression).
* source dependencies. Notice that they may differ from the binary
dependencies between the same packages. E.g. you may need a modula-3 compiler
to build foo, but not to install and run it.
* Notion of the base system (different for source and binary stuff):
minimal set that is in all dependencies by default. IOW, bootstrap subset.
* ability to actually install the thing for binary packages and
ability to build for source ones. Not so trivial as it sounds: the point
being to have a uniform set of commands for that.
* ability to keep enough state (i.e. to remember what is where, etc.)
and to maintain it across the upgrades/installs/uninstalls.
* ability to find (and verify) packages.
* decent bandwidth/computational cost of upgrades.
* preserving configs across the upgrade. Includes letting admin a
chance to choose between keeping the old config/installing precanned one/
merging them by hands.
* ability to survive gradual updates (as opposed to backup/nuke/
reinstall/restore config)
* sane handling of alternatives. E.g. on the box I'm sitting on
right now (telnetted from it to Sun box in Uni) I have both nvi and vim.
/usr/bin/vi being a symlink to /etc/alternatives/vi which points to
/usr/bin/nvi (two-level scheme, since /usr may be shared by many boxen).
If I'll remove nvi vim will take its place. Oh, and man vi will give vim
manpage instead of nvi one. Not too critical wrt editors (who wants to
switch them, anyway? Besides vim is wimpy to the point of creeping
EMACSism ;-), but when it comes to default C compiler or ftpd or yacc, etc.
things become more serious. It's a matter of preserving the defaults
across the upgrade and handling them in decent way across install/uninstall.
* various kinds of search in the package database.
* ability to clone the setup (jumpstart and equivalents).
* ability to put as much as possible into the cron tasks.
* decent package selection stuff (IOW *not* "here is the list of
<bignum> packages, scroll and enjoy" vs. "plonk everything on the disk
and let the devil sort his stuff").
* pre-canned variants of set and ability to adjust them (mostly useful
for initial install).
* wimpy GUI to crash-test the swap partition size and mess the
things up ;-)
That's more or less random choice - there are other things to look
for. And that - not counting set of packages and policies wrt records in
/etc/inetd.conf, UID and GID allocation, handling of that .info abomination
(what to do with the dir.info upon install/uninstall/upgrade/downgrade),
yodda, yodda.
Moreover, packaging system DSW is pretty similar to benchmark DSW -
everything depends on the job mix you are going to deal with (and on your
tastes, indeed - I couldn't care less for point'n'drool, but there are people
who can't live without it. Sweet memories of SH^HMIT... <shudder>)
>Slack's package format seems to be a lot like FreeBSD's package
>format, and even "pkgtool" and "setup" resembles FreeBSD's
>"sysinstall". I get the impression that Slackware's setup and
>package management tools were modeled after FreeBSD'.
Nope. Look and feel of setup/sysinstall/dinstall is more or less
the same, but it's simply look and feel of dialog(1). Slack's package format
predates both Linux and 4.4BSD - gzip is older than both and tar... well, is
older than v7.
>Another scenario is that FreeBSD inherited sysinstall and package
>management from the original 4.4BSD admin. tools, and Slack
>borrowed either directly from original BSD or FreeBSD or a
>combination of both.
Nope, again. Slack has package manager, all right, but it's *really*
different from *BSD one. Look into <ouch> /var/packages and you'll see.
>How about a web-based system for doing upgrades, package
>management, and installs? I think Solaris does it this way,
>imitating Microsoft's patented method for Windows upgrades.
>(BTW, does Windows NT have the same web-based upgrade?)
Could you spell "pretty much useless and calls for security problems
*really* big way"?
>I can't believe people are flaming each other over which distro
>and/or package management is best.
Welcome to USENET. Now, could we discuss something less
flame-attracting? Choice of editors, for example... Or STREAMS vs
sockets... Or guns contr^#..!{y{w{{...O{{w
NO CARRIER
------------------------------
From: Peter Englmaier <[EMAIL PROTECTED]>
Subject: FIXED(?): Staroffice/glibc problem
Date: Sat, 15 May 1999 02:11:45 -0400
Yet another solution for the Staroffice/glibc problems.
RH6.0 has problems with old glibc-2.0.7 binaries. Staroffice and
many other programs do not work with the new glibc. Here is
a workaround based on the 'misc. binary format' feature of
the linux kernel. Even printing seems to work (at least for me).
Please post if this does or does not work for you. Staroffice
still gives some error message at startup, but works fine.
Note: This may also fix other problems reported with the glibc.
0) do *not* modify any staroffice scripts or binaries... (the easy
step first)
1) install the rpm's for
compat-glibc-5.2-2.0.7.1
compat-libs-5.2-1
2) install the appended wrapper file in
/usr/i386-glibc20-linux/lib/ld-wrapper
and do
chmod a=rx /usr/i386-glibc20-linux/lib/ld-wrapper
This file will be used to run old applications.
3) install the appended rc.binfmt file below in /etc/rc.d/
make it execute:
chmod a=rx rc.binfmt
run rc.binfmt to activate it. If it produces error
messages, you probably have no misc. binary format
compiled in the kernel. You want to call the
file from /etc/rc.d/rc.serial as well (add a line
saying '/etc/rc.d/rc.binfmt'.
This works as follows: when the kernel is asked to 'execute' a
binary with extension '.bin', it runs the wrapper file instead.
The wrapper loads the binary with the 'right' loader. To run
other old binaries, e.g. 'goodie', simply do:
mv goodie goodie.bin
ln -s goodie.bin goodie
BTW, if somebody finds out the 'magic' bytes of old binaries,
the script could be triggered using the magic fingerprint.
Peter.
>>>>>>>>>>>>>>>>>>>ld-wrapper:
#!/bin/sh
# wrapper for old glibc binaries
p=`basename $1`
if [ "$p" = "soffice.bin" -o "$p" = "psetup.bin" -o "$p" = "setup.bin"
]; then
exec /usr/i386-glibc20-linux/lib/ld-linux.so.2 --library-path
/usr/i386-glibc20-linux/lib:$LD_LIBRARY_PATH "$@"
fi
exec /lib/ld-linux.so.2 "$@"
>>>>>>>>>>>>>>>>>>rc.binfmt
#!/bin/sh
if /sbin/modprobe binfmt_misc.o ; then
# clear register first
echo -1 >/proc/sys/fs/binfmt_misc/status
REGISTER=/proc/sys/fs/binfmt_misc/register
# old glibc
echo ':oldglibc:E::bin::/usr/i386-glibc20-linux/lib/ld-wrapper:'
>${REGISTER}
fi
------------------------------
** 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
******************************
