Re: Problem with packages

[Peter Nelson]

Read the logs, the archive has moved:
http://amd64.debian.net/README.mirrors.html
deb http://bach.hpc2n.umu.se/debian-amd64/debian/ sid main contrib
-Peter
Ronny Wikh wrote:
Hi!
After a period of inactivity I needed to update my pure64 installation.
Looking through the (massive) backlog of messages it turns out that
the organisation had changed, so I modified my apt sources.list to
point to
deb http://bach.hpc2n.umu.se/debian-pure64 sid main contrib non-free
deb-src http://bach.hpc2n.umu.se/debian-pure64 sid main contrib non-free
as per instructions. Now for the upgrade:
-
 

apt-get -fu dist-upgrade
   

Reading Package Lists... Done
Building Dependency Tree... Done
Calculating Upgrade... Done
The following packages will be REMOVED:
 apache2 apache2-common apache2-mpm-prefork cpp cpp-3.3 g++ g++-3.3 gcc
 gcc-3.3 kernel-image-2.6-amd64-k8 kernel-image-2.6.8-10-amd64-k8
 libapache2-mod-auth-pam libapache2-mod-php4 libapache2-svn
 libstdc++5-3.3-dev resolvconf x-window-system-core xbase-clients xdm
 xfonts-100dpi xfonts-75dpi xfonts-base xfonts-scalable xutils
The following NEW packages will be installed:
 lsb-release
The following packages have been kept back:
 apache2-utils bind9-host bzip2 chkrootkit cpp-3.4 dash devfsd dmsetup
 dnsutils e2fslibs e2fsprogs fontconfig gcc-3.4 gettext gettext-base hdparm
 ifupdown imagemagick lbxproxy lesstif2 libapr0 libbind9-0 libblkid1
 libbz2-1.0 libc-client2002edebian libc6 libc6-dev libcomerr2
 libdevmapper1.01 libdm0 libdns20 libdps1 libexpat1 libfontconfig1
 libfreetype6 libgcc1 libgd2-noxpm libgimpprint1 libgpmg1 libice6 libisc9
 libisccc0 libisccfg1 libkpathsea3 libkrb53 libldap2 liblwres1 libmagick6
 libneon24 libnewt0.51 libsm6 libssl-dev libssl0.9.7 libstdc++5 libsvn0
 libtasn1-2 libtext-charwidth-perl libusb-0.1-4 libuuid1 libx11-6 libxaw7
 libxext6 libxft1 libxi6 libxmu6 libxmuu1 libxp6 libxpm4 libxrandr2
 libxrender1 libxt6 libxtrap6 libxtst6 libxv1 locales login logrotate
lvm2 mc
 mlock modutils net-tools nis nmap openssl passwd portmap postfix ppp
 proxymngr python2.3 python2.3-dev python2.3-subversion python2.4 rcs rsync
 sensible-mda strace subversion subversion-tools sudo tar tcpdump tetex-bin
 twm unzip usbutils uw-imapd wget whiptail xdu xfs xfwp xlibmesa-gl
 xlibmesa-glu xnest xserver-common xserver-xfree86 xterm xvfb
The following packages will be upgraded:
 apache2-doc ca-certificates gcc-3.3-base ia32-libs initrd-tools
 libtext-wrapi18n-perl mime-support php4-common python-pyopenssl
 python-twisted tetex-base tetex-doc xfree86-common xlibmesa-dri xlibs
 xlibs-data
16 upgraded, 1 newly installed, 24 to remove and 120 not upgraded.
Need to get 62.2MB/62.2MB of archives.
After unpacking 116MB disk space will be freed.
Do you want to continue? [Y/n] n
-
Now, this is all rather unwelcome news to me as I'm dependant on
having apache2 running.
In the listing above it seems that libc6 is part of the upgrade,
so I figured that perhaps I'd better go for that and apache2 to
see what happens. This is what I get:
---
 

apt-get install apache2 libc6
   

Reading Package Lists... Done
Building Dependency Tree... Done
The following extra packages will be installed:
 apache2-common apache2-mpm-prefork apache2-utils libapache2-mod-php4
libapr0
 libc6-dev locales php4-common
Suggested packages:
 php4-pear
The following packages will be REMOVED:
 aide base-files bash deborphan hotplug mailx
The following packages will be upgraded:
 apache2 apache2-common apache2-mpm-prefork apache2-utils
libapache2-mod-php4
 libapr0 libc6 libc6-dev locales php4-common
WARNING: The following essential packages will be removed
This should NOT be done unless you know exactly what you are doing!
 base-files bash
10 upgraded, 0 newly installed, 6 to remove and 144 not upgraded.
Need to get 15.9MB of archives.
After unpacking 4461kB disk space will be freed.
You are about to do something potentially harmful
To continue type in the phrase 'Yes, do as I say!'
 

NO!
   

--
The question is, what's causing this and how do I fix it? Will I
need to install everything from scratch again, or what's going on?
I'd very much appreciate any kind of advice.
Regards,
Ronny Wikh
 


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Any idea what's wrong?

[Peter Nelson]

Goswin von Brederlow wrote:
Vernon Dozier <[EMAIL PROTECTED]> writes:
 

--- Goswin von Brederlow
<[EMAIL PROTECTED]> wrote:
   

Vernon Dozier <[EMAIL PROTECTED]> writes:
 

I just did a apt-get update and apt-get
dist-upgrade... this is the first time I'd done
   

this
 

since the archive move, and I'm getting some error
messages relating to emacs.  Specifically...
   

Did you use gcc-3.4 before?
 

No
   

Does reinstalling emacs or xemacs (whatever you have installed) and
emacsen-common help?
 

I had the exact same problem.  Never used gcc-3.4 before.  A forced 
reinstall of just emacsen-common was enough to fix it.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: fglrx64 for ati 9700 pro

[Peter Nelson]

Use my debs instead of trying to alien them:
http://rufus.hackish.org/~rufus/files/fglrx/
-Peter
Christophe Dupré wrote:
Hi guys,
I've been trying to install fglrx64 in order to make
use of the 3D acceleration of my ATI 9700 pro.
I downloaded fglrx64_4_3_0-8.10.19-1.x86_64.rpm,
aliened it, and then got
fglrx64-4-3-0_8.10.19-2_amd64.deb.
I install it, recompiled my kernel (2.6.11), patched
fglrx, and compile the module.
My problem is when I run XFree86.
I don't understand why some libs and modules are
placed in /usr/X11R6/lib64 while all others are placed
in /usr/X11/lib/
dpkg -L fglrx64-4-30
/usr/X11R6/lib64
/usr/X11R6/lib64/libGL.so.1.2
/usr/X11R6/lib64/libfglrx_gamma.a
/usr/X11R6/lib64/libfglrx_gamma.so.1.0
/usr/X11R6/lib64/modules
/usr/X11R6/lib64/modules/dri
/usr/X11R6/lib64/modules/dri/fglrx_dri.so
/usr/X11R6/lib64/modules/drivers
/usr/X11R6/lib64/modules/drivers/fglrx_drv.o
/usr/X11R6/lib64/modules/linux
/usr/X11R6/lib64/modules/linux/libfglrxdrm.a
When I start X, it complains that the driver fglrx is
not found.
Can you please help.
Thanks a lot,
Christophe

Send instant messages to your online friends http://uk.messenger.yahoo.com 

 


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Time drift in amd64

[Peter Nelson]

Bharath Ramesh wrote:
I just installed a fresh installation of amd64 on another athlon box. I
find that the clock seems to be drifting by 8-10 seconds every day. Any
idea how I can fix this drift. This drift seems to be large.
Just install ntp-simple and it'll keep you within a few milliseconds off 
official atomic time.  If it asks you were to point use pool.ntp.org.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: flash player for firefox

[Peter Nelson]

Erik Norman wrote:
Hello,
I have tried several times (using the "install missing plugins" feauture
of mozilla-firefox) to install flashplayer, no way.
It does not work.
Any hint?
 

There is no flash player for amd64.  From 
http://sluglug.ucsc.edu/macromedia/faq.html:

   *10. Where is Flash for PowerPC or x86-64 Linux?*
   It does not exist probably because Macromedia management feel there
   is no demand. Contact Macromedia
   <http://www.macromedia.com/macromedia/contact/> and express your
   opinion if you truly care about this. Also buy more x86-64 hardware
   (AMD64, not the crap Intel stuff) to make the platform truly
   lucrative for software makers in a business sense.
There is libflash-swfplayer, an open-source one, but it's very, very buggy.
-Peter Nelson
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: afs client

[Peter Nelson]

Charles Leggett wrote:
Has anyone managed to get the AFS client working with a 2.6 kernel on
an opteron? I'm running a 2.6.10 kernel, and have tried recent CVS
snapshots, and while I'm no longer getting kernel oopses, when trying
to start the openafs client I get messages like:
Starting AFS services: afsd: WARNING: cache probably too small!
afsd: malloc() failed for cache file inode table with -25 entries.
BTW, the cache is set to 95000.
I'm using a compile of 1.3.79 on 2.6.8 using the Debian/ directory from 
the 1.3.74 experimental debs.  The debs build up nicely except that 
Debian/sysname checks for x86_64 instead of amd64 (I just reported that 
bug to the package maintainer).

Also if you experience any more oopses please report them to 
openafs-devel@openafs.org as the amd64 port isn't all that well tested.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: dchroot script (was Re: running openoffice from chroot)

[Peter Nelson]

Alexandru Cabuz wrote:
The suggestion with the 
openoffice \"%u\" worked. But if you say there is a cleaner solution...

 

Yes, but a better thing is to fix dchroot's / su's stupidity with a
script (make this do_chroot):
 

I don't understand. What should I do with this script?
 

From the FAQ:
   You can make things slightly simpler even than this. Write a small
   shell script like the following, and place it in your PATH: eg.
   /usr/local/bin/do_dchroot
[Use the new script from the mailing list]
 

   Make it executable:
chmod 755 /usr/local/bin/do_dchroot
 

   Now create a symlink to this script from each command you want to
   run inside the chroot to this shell script:
cd /usr/local/bin
ln -s do_dchroot openoffice
ln -s do_dchroot oowriter
ln -s do_dchroot oocalc
 

   Now you can execute openoffice by simply typing openoffice or oowriter.
-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

dchroot script (was Re: running openoffice from chroot)

[Peter Nelson]

Adam Skutt wrote:
Yes, but a better thing is to fix dchroot's / su's stupidity with a 
script (make this do_chroot):
#!/bin/bash

# Work around dchroot brain-deadness
ARGS=""
for i in "$@" ; do
ARGS="$ARGS '$i'"
done
exec dchroot -c ia32 -d -q "`basename $0`" "$ARGS"

Thank you for the corrected script.  I'd been planning on hacking up 
some sed-fu to do it, but his is a clean solution.  Probably should be 
added to the FAQ.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: unstable fglrx 8.10.19?

[Peter Nelson]

Andrei Mikhailovsky wrote:
Hello everyone!
After upgrading the fglrx drivers from 8.8.25 to the latest 8.10.19,
i've noticed that my system became very very unstable. I am experiencing
crashes about twice a day, sometime system frizzes without being able to
reboot using Alt-SysRr. 

Try using the drivers I built using flavio's packages here:
http://rufus.hackish.org/~rufus/files/fglrx/
If that doesn't fix it, I'd suggest reporting of rage3d.net's forums, as 
some ATI developers hang out there.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: ldd failing on fglrx-installer

[Peter Nelson]

Mark Kennedy wrote:
Hi there,
I just started installing the debian-pure64-3.4 port
on my machine, and ran into a couple of problems
building & installing an ATI driver according to
.
 

Make sure you're using the amd64 version from 
.  
It looks like you're trying to compile the 32bit drivers.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Shared library error compiling ATI drivers

[Peter Nelson]

Pedro I. Sanchez wrote:
I'm following the steps in
http://xoomer.virgilio.it/flavio.stanchina/debian/fglrx-installer.html
Those are the 32-bit drivers.  He just posted today about the 64-bit 
drivers at:
http://xoomer.virgilio.it/flavio.stanchina/debian-experimental/fglrx-experimental.html

You can find my compiles of these drivers here:
http://rufus.hackish.org/~rufus/files/fglrx/
The only thing is that currently you have to create the link  
"/usr/X11R6/lib64 -> lib".  The next version should fix this one way or 
another.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

gcc-3.4 creates a bad libgcc_s_32.so link.

[Peter Nelson]

I was having trouble compiling with -m32 because of errors with 
libgcc_s_32.so.  The link setup was this:

/usr/lib/gcc/x86_64-linux/3.4.4/libgcc_s_32.so -> 
../../../../lib32/libgcc_s.so.1
However the package lib32gcc1installes into:
/emul/ia32-linux/usr/lib/libgcc_s.so.1
Fixing the link to this fixes the compilation.
-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

GPG error mirroring the archive

[Peter Nelson]

Hi, my mirror of the amd64 archive recently broke becaues of the 
following errors:

gpg: Signature made Thu Jan 20 17:40:22 2005 EST using DSA key ID 2FE487B0
gpg: Can't check signature: public key not found
gpg: Signature made Thu Jan 20 17:40:22 2005 EST using DSA key ID 2FE487B0
gpg: Can't check signature: public key not found
Failed to download some Package, Sources, Contents or release files!
I've gotten around this by using the option --ignore-release-gpg but I'm 
wondering how to properly get this key.  I installed debian-keychain, 
but am still seeing the error.

And sometime soon now (yeah, I've said this before...) I'll open up my 
mirror for Internet 2 users, who should be able to max out their I2 
connections getting from me.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: AMD64 patch for flavio's fglrx debs

[Peter Nelson]

Peter Nelson wrote:
Cyril Chaboisseau wrote:
BTW, I have kernel-patch-debian-2.6.10 but that shouldn't make any
difference for the compilation of the module
There are known problems compiling against 2.6.10.  Look at this 
thread on rage3d for some patches:
http://rage3d.net/board/showthread.php?t=33798874
Sorry, I'm an idiot.  From Flavio's site:
   The kernel module doesn't compile out-of-the-box with kernel 2.6.10,
   however an official patch is available in this Rage3D forum
   <http://www.rage3d.org/board/showthread.php?t=33798874>. This patch
   is already included in my packages, you don't need to apply it yourself.
You must be seeing something else.  I'm not sure what error you're 
seeing, but try posting at rage3d as there are more knowledable people 
there.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: AMD64 patch for flavio's fglrx debs

[Peter Nelson]

Cyril Chaboisseau wrote:
Le 19 janvier vers 22:33, Peter Nelson écrivait:
 

Here's a patch I've made against flavio's latest fglrx packages to 
generate amd64 debs.  I *think* it does the right thing placing the 
32bit drivers in /emul/ia32-linux/usr/X11R6/lib/ and the 64bit ones in 
/usr/X11R6/lib/, with the correct diversions.  It works for me, but I 
don't have any real games to test the 32bit compatability with.
   

the compilation of the packages went almost smoothly (I just had to
modify a couple usr/X11R6/lib to usr/X11R6/lib64) 

Don't change any of the lib/lib64's.  Other than symlinking lib64 to 
lib, my patch includes all the changes that are needed.

BTW, I have kernel-patch-debian-2.6.10 but that shouldn't make any
difference for the compilation of the module
There are known problems compiling against 2.6.10.  Look at this thread 
on rage3d for some patches:
http://rage3d.net/board/showthread.php?t=33798874

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: AMD64 patch for flavio's fglrx debs

[Peter Nelson]

Quoting Johannes Pleikies <[EMAIL PROTECTED]>:

> Thank you very much, this worked well for me.
You're welcome.

> I used Peter's patch and
> http://xoomer.virgilio.it/flavio.stanchina/debian/fglrx-installer.html .
yeah, I forgot to link to (and thank!) Flavio's page.

> I used the kernel-agpgart.
> After all my XFree started without any errors, but my DRI was not yet
> working.
> I found the solution at an Ubuntu-Forum:
>  export LIBGL_DRIVERS_PATH=/usr/X11R6/lib/modules/dri/
> as normal user.
I have a feeling that the drivers are hard-coded to use the lib64 X11 directory,
though I can't find where they are.  Try making a symlink from /usr/X11R6/lib64
to lib and see if it works without that export.

If that works should the symlink be added to this package, left up to the user
to fix, or added to an other package?

-Peter


This message was sent using IMP, the Internet Messaging Program.


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

AMD64 patch for flavio's fglrx debs

[Peter Nelson]

Here's a patch I've made against flavio's latest fglrx packages to 
generate amd64 debs.  I *think* it does the right thing placing the 
32bit drivers in /emul/ia32-linux/usr/X11R6/lib/ and the 64bit ones in 
/usr/X11R6/lib/, with the correct diversions.  It works for me, but I 
don't have any real games to test the 32bit compatability with.

As a side note, module-assistant kicks ass.
-Peter
diff -ur old/fglrx-installer-8.8.25/debian/control 
fglrx-installer-8.8.25/debian/control
--- old/fglrx-installer-8.8.25/debian/control   2004-12-26 17:25:50.0 
-0500
+++ fglrx-installer-8.8.25/debian/control   2005-01-19 18:01:41.0 
-0500
@@ -6,7 +6,7 @@
 Standards-Version: 3.6.1
 
 Package: fglrx-driver
-Architecture: i386
+Architecture: amd64 i386
 Depends: ${shlibs:Depends}, xserver-xfree86 (>= ${XVERSION}), xserver-xfree86 
(<< ${XVERSIONMAX})
 Recommends: fglrx-kernel
 Description: Video driver for the ATI graphics accelerators
@@ -28,7 +28,7 @@
 
 Package: fglrx-driver-dev
 Section: non-free/devel
-Architecture: i386
+Architecture: amd64 i386
 Depends: fglrx-driver
 Description: Video driver for the ATI graphics accelerators (devel files)
  Video driver for the ATI Radeon and FireGL graphics accelerators.
@@ -37,7 +37,7 @@
  and the FGLRXGAMMA extension interface library.
 
 Package: fglrx-kernel-src
-Architecture: i386
+Architecture: amd64 i386
 Depends: debconf, debhelper (>= 4), kernel-package, make
 Recommends: fglrx-driver
 Description: Kernel module source for the ATI graphics accelerators
@@ -46,7 +46,7 @@
  This package provides the kernel module build environment.
 
 Package: fglrx-control-qt3
-Architecture: i386
+Architecture: amd64 i386
 Depends: ${shlibs:Depends}, fglrx-driver
 Conflicts: fglrx-control-qt2
 Description: Control panel for the ATI graphics accelerators
@@ -55,7 +55,7 @@
  This package provides the control panel comnpiled for Qt3.
 
 Package: fglrx-sources
-Architecture: i386
+Architecture: amd64 i386
 Recommends: fglrx-driver, libqt3-mt-dev, xlibs-dev
 Description: Source code for the ATI graphics accelerators control panel
  Video driver for the ATI Radeon and FireGL graphics accelerators.
diff -ur old/fglrx-installer-8.8.25/debian/postrm 
fglrx-installer-8.8.25/debian/postrm
--- old/fglrx-installer-8.8.25/debian/postrm2004-03-22 03:35:44.0 
-0500
+++ fglrx-installer-8.8.25/debian/postrm2005-01-19 20:51:23.0 
-0500
@@ -15,6 +15,16 @@
 
 case "$1" in
 remove)
+   if [ `arch` = 'x86_64' ]; then
+   dpkg-divert --remove --package fglrx-driver \
+   --divert /usr/share/fglrx/diversions/emul/libGL.so.1.2 \
+   --rename /emul/ia32-linux/usr/X11R6/lib/libGL.so.1.2
+
+   if [  -d /usr/share/fglrx/diversions/emul ]; then
+   rmdir /usr/share/fglrx/diversions/emul || true
+   fi
+   fi
+
dpkg-divert --remove --package fglrx-driver \
--divert /usr/share/fglrx/diversions/libGL.so.1.2 \
--rename /usr/X11R6/lib/libGL.so.1.2
diff -ur old/fglrx-installer-8.8.25/debian/preinst 
fglrx-installer-8.8.25/debian/preinst
--- old/fglrx-installer-8.8.25/debian/preinst   2004-03-22 03:35:30.0 
-0500
+++ fglrx-installer-8.8.25/debian/preinst   2005-01-19 20:51:38.0 
-0500
@@ -20,6 +20,16 @@
dpkg-divert --add --package fglrx-driver \
--divert /usr/share/fglrx/diversions/libGL.so.1.2 \
--rename /usr/X11R6/lib/libGL.so.1.2
+
+   if [ `arch` = 'x86_64' ]; then
+   if [ ! -d /usr/share/fglrx/diversions/emul ]; then
+   mkdir -p /usr/share/fglrx/diversions/emul || true
+   fi
+
+   dpkg-divert --add --package fglrx-driver \
+   --divert /usr/share/fglrx/diversions/emul/libGL.so.1.2 \
+   --rename /emul/ia32-linux/usr/X11R6/lib/libGL.so.1.2
+   fi
 ;;
 
 upgrade)
diff -ur old/fglrx-installer-8.8.25/debian/rules 
fglrx-installer-8.8.25/debian/rules
--- old/fglrx-installer-8.8.25/debian/rules 2005-01-18 15:50:14.0 
-0500
+++ fglrx-installer-8.8.25/debian/rules 2005-01-19 20:42:38.0 -0500
@@ -46,7 +46,12 @@
 
 # These file names don't always match XVERSION+PVERSION:
 # look on the driver info page and update accordingly.
-RPMNAME := fglrx_$(subst .,_,$(XVERSION))-$(PVERSION)-1.i386.rpm
+ifeq ($(shell arch),x86_64)
+   RPMNAME := fglrx64_$(subst .,_,$(XVERSION))-$(PVERSION)-1.x86_64.rpm
+else
+   RPMNAME := fglrx_$(subst .,_,$(XVERSION))-$(PVERSION)-1.i386.rpm
+endif
+
 PATCHDIR := debian/patches-$(PVERSION)
 
 # Package names
@@ -107,11 +112,23 @@
dh_installdirs
 
# Create the directories to install into
+   # the amd64 package includes 32bit compatability libraries
+ifeq ($(shell arch),x86_64)
dh_installdirs -p$(PKG_driver) \
+   emul/ia32-linux/usr/X11R6/lib \
+   emul/ia32-linu

Re: New ATI fglrx drivers work with AMD64

[Peter Nelson]

Nick Hemsley wrote:
OK scrap that, I was using gcc3.3 instead of 3.4. changing the symlink 
worked for me, although I think you can do it in the make.sh.

The module now starts, starting XFree86 however freezes the system 
(not sure if I can ssh), and /var/log/mesages shows:
Jan 19 14:00:02 localhost kernel: Call 
Trace:{:fglrx:firegl_init_aperture+160}
Jan 19 14:00:02 localhost kernel:
{:fglrx:firegl_agp_acquire+0} 
{:fglrx:__ke_agp_available+50} 
well, certainly something agp related.  Try setting
Option "UseInternalAGPGART" "no"
In /etc/X11/XF86Config-4.  I think that's the same crash I got and 
that's how I fixed it.  If not, try various permutations of related AGP 
things.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New ATI fglrx drivers work with AMD64

[Peter Nelson]

I had similar issues on i386 when using both the fglrx drivers and the 
radeonfb console frame buffer.  Try disabling your console frame buffer 
(or maybe loading vesafb instead of radeonfb?) and it should be fine.  
Obviously a sub-optimal solution, but I think accelerated X11 is far 
more important than accelerated console.

-Peter
Andrei Mikhailovsky wrote:
I've managed to get the 64bit drivers running fine, games and
screensavers are now very fast ))
However, i am experiencing screen corruption when I switch to the
console.
Does anyone have similar issues?
Thanks
On Mon, 2005-01-17 at 23:08 -0500, Peter Nelson wrote:
 

Hi, I'd just like to give the list a heads up that the new ATI fglrx 
drivers actually work with amd64.  They're only available (as usual) in 
rpm's but I hacked together Flavio's debian packages and with a bit of 
manual copying around it works.  Hopefully someone with a bit more 
knowledge of how things like /usr/X11R6/lib vs lib64 are supposed to 
work can make a clean package (I just copied stuff around untill it worked).

Get from http://www.ati.com/support/drivers/linux/radeon-linux64.html
-Peter
   


--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: apt wants to remove 526 packages!!?

[Peter Nelson]

Alexandru Cabuz wrote:
Hello,
I usually do a 
apt-get update
apt-get upgrade

every couple of days. When there are bugs I just put packages on hold
and wait it out, and after a month I try them individually, and if
they are still buggy, then I leave them on hold.
I decided tonight I was sick of this, so I would do apt-get dist-upgrade.
 

Do note that anything that's on hold won't get upgraded by 
dist-upgrade.  I have ssh on hold to a version I built myself and it 
doesn't get upgraded with dist-upgrade.  You probably have a package on 
hold that's been updated and causes other things to break.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

New ATI fglrx drivers work with AMD64

[Peter Nelson]

Hi, I'd just like to give the list a heads up that the new ATI fglrx 
drivers actually work with amd64.  They're only available (as usual) in 
rpm's but I hacked together Flavio's debian packages and with a bit of 
manual copying around it works.  Hopefully someone with a bit more 
knowledge of how things like /usr/X11R6/lib vs lib64 are supposed to 
work can make a clean package (I just copied stuff around untill it worked).

Get from http://www.ati.com/support/drivers/linux/radeon-linux64.html
-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: amd64(gcc4): aptitude dumps core

[Peter Nelson]

Harald Dunkel wrote:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Harald Dunkel wrote:
| Hi folks,
|
| Since the most recent upgrade of apt (0.5.27.2.0.0.1.gcc4)
| aptitude dies with a core dump immediately.
|
PS: Going back to the old version built with gcc 3.4 (AFAIK)
fixes the problem.
Folks, I need my PC. amd64(gcc-3.4) is on Unstable, but not
Experimental. Don't use an experimental compiler, please.
The gcc-3.4 (well, I guess it's now 4.0 but still has the old name) 
distrobution should be considered experimental.  If you just want 
something that works use the pure64 version.

-Peter
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: Kernel-2.6.10 and 32bit emulation

[Peter Nelson]

Aritz Beraza Garayalde [Rei] wrote:
If you google a bit, you'll be able to find some amd64 packages for
mplayer, also from marillat. Unfortunately last time i checked, there
was no debian repository, so you can't install them by adding a few
lines to your sources.list and using apt. You'll have to get all
needed packages and install them with dpkg.

   Christian Marillat has maintained an unofficial repository of Debian
   packages relating to multimedia applications for several years. He's
   now set up an amd64 version of the repository.
-Peter

Re: upgrading base-files problem

[Peter Nelson]

Not sure if this is directly related to the other errors but libc6 and 
base-files are unhappy with each other.  Here's the error:

Preparing to replace libc6 2.3.2.ds1-19 (using 
.../libc6_2.3.2.ds1-19.0.0.1.pure64_amd64.deb) ...
Unpacking replacement libc6 ...
dpkg: error processing 
/afs/hackish.org/data/debian-amd64/pure64/pool/unstable/main/amd64/g/glibc/libc6_2.3.2.ds1-19.0.0.1.pure64_amd64.deb
 (--unpack):
trying to overwrite `/usr/lib64', which is also in package base-files
Errors were encountered while processing:
/afs/hackish.org/data/debian-amd64/pure64/pool/unstable/main/amd64/g/glibc/libc6_2.3.2.ds1-19.0.0.1.pure64_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
When I did --force-overwrite it works, giving these warnings:
Preparing to replace libc6 2.3.2.ds1-19 (using 
.../libc6_2.3.2.ds1-19.0.0.1.pure64_amd64.deb) ...
Unpacking replacement libc6 ...
dpkg - warning, overriding problem because --force enabled:
trying to overwrite `/usr/lib64', which is also in package base-files
dpkg - warning, overriding problem because --force enabled:
trying to overwrite `/lib64', which is also in package base-files
Setting up libc6 (2.3.2.ds1-19.0.0.1.pure64) ...
Thanks,
-Peter

Re: AMD64 VS EM64T

[Peter Nelson]

Jin Zhao wrote:
I am currently faced with choosing one of them as our forthcoming 64 
bit platform. So far I read a couple of reviews, most of which seems 
favor AMD64 a little bit. I also did some initial testings on an 
opteron box with Debian pure64 unstale. So far it looks good.
Here's a review comparing amd64 to emt64 under SuSE 9.1 Pro with Linux 
2.6.4.  Slightly dated and limited review, but should give you a good feel.
http://anandtech.com/linux/showdoc.aspx?i=2163

-Peter

Kernel 2.6.9?

[Peter Nelson]

I'm just wondering where kernel-image-2.6.9 is.  It's been in i386 
unstable for almost a month now, and fixes quite a few issues I have.

Thanks,
-Peter

Re: Kernel 2.6.9?

[Peter Nelson]

Peter Nelson wrote:
I'm just wondering where kernel-image-2.6.9 is.  It's been in i386 
unstable for almost a month now, and fixes quite a few issues I have.
I guess I should have been more specific.  kernel-image-2.6.9-amd64 
exists in /pure64, but not in /gcc-3.4 which is what I'm using.  Is it 
safe to install the pure64 (gcc-3.3) kernel on gcc-3.4.  And when will 
the kernel make its way to 3.4?

Thanks again,
-Peter

Bug in libc6?

[Peter Nelson]

I have come across the strangest problem.  About a week or two ago 
heimdal kerberos simply stopped working.  I don't think it's a heimdal 
problem because I tried multiple versions and none of them worked, but 
it did use to work.  Here's a stack trace from it crashing:

Program received signal SIGFPE, Arithmetic exception.
0x003279b094ed in _dl_allocate_static_tls () from /lib/ld-linux-x86-64.so.2
(gdb) bt
#0  0x003279b094ed in _dl_allocate_static_tls ()
  from /lib/ld-linux-x86-64.so.2
#1  0x003279b0a4ae in _dl_relocate_object_internal ()
  from /lib/ld-linux-x86-64.so.2
#2  0x002a95e6c7fb in dl_open_worker () from /usr/lib/debug/libc.so.6
#3  0x003279b0baa0 in _dl_catch_error_internal ()
  from /lib/ld-linux-x86-64.so.2
#4  0x002a95e6cfef in _dl_open () from /usr/lib/debug/libc.so.6
#5  0x002a95e6e2f1 in do_dlopen () from /usr/lib/debug/libc.so.6
#6  0x003279b0baa0 in _dl_catch_error_internal ()
  from /lib/ld-linux-x86-64.so.2
#7  0x002a95e6e38b in __libc_dlopen_mode () from /usr/lib/debug/libc.so.6
#8  0x002a95e4d528 in __nss_lookup_function ()
  from /usr/lib/debug/libc.so.6
#9  0x002a95e4d684 in __nss_lookup () from /usr/lib/debug/libc.so.6
#10 0x002a95e52c18 in getservbyname_r@@GLIBC_2.2.5 ()
  from /usr/lib/debug/libc.so.6
#11 0x002a95e529f4 in getservbyname () from /usr/lib/debug/libc.so.6
#12 0x002a95690b04 in krb5_getportbyname (context=0x2a959ec478, 
   service=0x2a956a22d2 "kerberos", proto=0x2a956a27d7 "udp", default_port=88)
   at get_port.c:47
#13 0x002a95695d9a in krb5_krbhst_init (context=0x2a959ec478, 
   realm=0x50a4b0 "HACKISH.ORG", type=0, handle=0x7fbfff8948) at krbhst.c:633
#14 0x002a9569d603 in krb5_sendto_kdc2 (context=0x505010, 
   send_data=0x7fbfff8b30, realm=0x50a208, receive=0x7fbfff8b20, master=0)
   at send_to_kdc.c:386
#15 0x002a9568fdda in krb5_get_in_cred (context=0x505010, options=266, 
   addrs=0x0, etypes=0x0, ptypes=0x0, preauth=0x0, 
   key_proc=0x2a95690730 , keyseed=0x7fb3c0, 
   decrypt_proc=0x40, decryptarg=0x40, creds=0x7fb1a0, 
   ret_as_reply=0x7fb010) at get_in_tkt.c:694
#16 0x002a95691421 in krb5_get_init_creds_password (context=0x505010, 
   creds=0x7fb4c0, client=0x509fe0, password=0x7fb3c0 "", 
   prompter=0x402000 , data=0x0, start_time=0, 
   in_tkt_service=0x0, options=0x7fb560) at init_creds_pw.c:438

The line that it crashes on in get_port.c (and the first line run in the 
function) is:

if ((sp = getservbyname (service, proto)) == NULL)
I tried to reproduce this using the following stupid program, but this 
workes fine:

#include 
int main() {
 struct servent *sp;
 const char *service = "kerberos";
 const char *proto = "udp";
 sp = getservbyname(service, proto);
 return 0;
}
I tried building libc6 setting DEB_BUILD_OPTIONS=nostrip, but I still 
didn't get debug info from libc.so.6.  Any suggestions on how to track 
this down?

-Peter

Re: blackdown Java, amd64 and java-package

[Peter Nelson]

Robert King wrote:
What am I doing wrong here?
solzhenitsyn:/usr/local/package# make-jpkg j2re-1.4.2-rc1-linux-amd64.bin
Creating temporary directory: /tmp/make-jpkg.mFAQu4
Loading plugins: blackdown-j2re.sh blackdown-j2sdk.sh common.sh 
j2re.sh j2sdk.sh j2se.sh sun-j2re.sh sun-j2sdk.sh

No matching plugin was found.
Removing temporary directory: done 

Add the following to /usr/share/java-package/blackdown-j2re.sh in the 
middle where it lists other files:

  "j2re-1.4.2-rc1-linux-amd64.bin")
  j2se_version=1.4.2+rc1
  j2se_expected_min_size=53 # 54748 kB
  found=true
  ;;
The make-jpkg program does file-name comparisons to figure out what 
package the file is, but doesn't recognize the amd64 version by default.

-Peter

Re: some processes having large memory footprints

[Peter Nelson]

Richard Wohlstadter wrote:
I'm new with the AMD64 debian port and had some questions I'm hoping 
someone out there can answer for me.  One thing I noticed after 
installing the AMD64 port on our opterons is that some processes seem 
to have very large memory footprints on the system.  For example, sshd 
instances are taking 26m virtual, bash taking 8m, ntpd taking 11m.  
Does this have to do with how these binaries are allocating their data 
structures and since I'm pure 64 bit they suck up more memory?  Just 
seems like alot of waste(although resident memory usage is more in 
line with what 32bit machines use).  Thanks for any input anyone can 
share regarding this.
It's most likely the program makes a lot of use of pointers.  Remember 
that all pointers are now 64-bit as opposed to 32-bit, so if a program 
is very pointer heavy it can take close to 2x as much memory.  A perl 
program of mine that builds a gigantic (~5million entry) hash table 
takes >1.5x as much memory on AMD64 as i386.  It's just one of the facts 
of life about amd64.  The thing is, though perl was taking up so much 
more memory it actually ran a bit faster on my amd64 than a xeon.

-Peter

Re: Working ia32-linux chroot image

[Peter Nelson]

Bob Proulx wrote:
Peter Nelson wrote:
 

After a bit of tweaking I've gotten what I think is the smallest
chroot that will still allow you to use apt-get and dpkg.
   

You did not say but what was the final amount of disk space that you
were able to reduce your minimum system down to?  My barebones
debootstrap install is around 110MB.  But of course after I install
various libraries and packages such as mozilla-firefox and
openoffice.org the size is much larger.  (I am running the 32-bit web
browsers to allow me to use 32-bit binary plugins.  And similar
reasons for openoffice.org.)
 

The compressed size is 20mb and it extracts out to ~70mb but it balloons 
up to ~100mb once you do the first `apt-get update` and it builds the 
package lists.  My current chroot is ~200mb after installing a bunch of 
libraries for mplayer.

Since it took me a while to build (debian depends don't work when
you're removing half the base system)
   

Does this statement mean that depends is broken in the resulting
image?  Or only that it was difficult to get to the result because of
the dependencies?
 

Depends work in the resulting image.  My problem was that packages only 
depend of binarys required to run the program, not side programs that 
are called.  At first I just installed bash, apt, and dpkg with their 
direct dependencies.  Dpkg ran but couldn't install anything because it 
calls tar and gzip and some packages use grep or perl in their install 
scripts, but don't depend on them.  Overall I think this only shows up 
when doing as minimal of an install as I did so I don't think it's worth 
the effort to fix.

I've decided to share it.  You can download a tar.bz with
instructions on how to use it from my site here:
<http://rufus.hackish.org/wiki/I386Chroot>
   

Thanks for sharing your work.
 

No problem, thanks for the feedback,
-Peter

Working ia32-linux chroot image

[Peter Nelson]

I've been playing around with the /emul/ia32-linux area a bit lately and 
I finally got tired of pulling down library packages debs and extracting 
them into that area manually.  I first used debootstrap to install a 
full system into a chroot there, but realized that's a heck of a lot 
wasted space on unneeded packages.  After a bit of tweaking I've gotten 
what I think is the smallest chroot that will still allow you to use 
apt-get and dpkg.  Since it took me a while to build (debian depends 
don't work when you're removing half the base system) I've decided to 
share it.  You can download a tar.bz with instructions on how to use it 
from my site here:

<http://rufus.hackish.org/wiki/I386Chroot>
Any feedback is welcome,
-Peter Nelson

Re: firefox crash

[Peter Nelson]

was was wrote:
Hi, I have a problem with a mozilla-firefox. When I
write in search engines like google, altavista or
alltheweb the mozilla crash. Is the package of firefox
wrong? Is it a problem of my computer?
 

It's a problem with form autocomplete.  Turn off Privacy > Saved Form 
Information and it'll go away, but obviously is a sub-optimal solution.

-Peter

Re: Java and others on amd64

[Peter Nelson]

Tobias Krais wrote:
as for I am new in this list I wanna ask some more questions. Up to 
now I found not very much information how you solve problems with 
missing packages on debian amd64. I search for following packages (or 
an other way how you got it to work):

- Java
Use the backdown version 
ftp://metalab.unc.edu/pub/linux/devel/lang/java/blackdown.org/JDK-1.4.2/amd64/rc1/

- Openoffice
Install ia32-libs-openoffice.org to get the needed 32bit libraries then 
manually install the i386 openoffice debs (using --force-architecture)

- Eclipse
The official Linux (AMD 64/GTK 2) build from eclipse.org works fine for me.
- Realplayer
No clue, nor do I particularly care =)
-Peter

Re: Do you know this mirror?

[Peter Nelson]

Bob Proulx wrote:
Brett Viren wrote:
 

2) Preservation of hard links and the resulting preservation of disk
space.
   

I can't think of any files in a depot that would be hard linked.  All
of the files look to be unique files to me.  What am I missing?
 

I know for a fact that Debian has used hard-links in the mirror tree in 
the past because one of my friends ran into the problem when trying to 
mirror into AFS (which doesn't support hard links).  However I can't 
think of anything that would still be hard linked so it's probably 
irrelevant.

-Peter

Re: modules

[Peter Nelson]

was was wrote:
Hi, I have a problem with the modutils, when I use
modconf or update-modules the shell show:
Architecture-specific modutils configuration not
found, using defaults.
What´s wrong?
 

It's because /etc/modutils/arch only has i386, no x86_64 in it.  As far 
as I can tell the error is harmless since 2.6 (which I assume you're 
using) uses /etc/modprobe.d anyway.  If you really care though you can 
copy (or symlink) i386 over to x86_64.

-Peter

Re: Java working

[Peter Nelson]

David Liontooth wrote:
Sun distributes a version of java for amd64 at 
http://java.sun.com/j2se/1.5.0/download.jsp and it works great.

Cheers,
David
cd /usr/local
//jre-1_5_0-rc-linux-amd64.bin
vi /etc/profile
add /usr/local/jre1.5.0/bin
Only problem is it doesn't include a java plugin for mozilla.  
Blackdown's new 1.4.2 build includes the plugins.

ftp://metalab.unc.edu/pub/linux/devel/lang/java/blackdown.org/JDK-1.4.2/amd64/rc1/
-Peter

Re: Do you know this mirror?

[Peter Nelson]

Brett Viren wrote:
Kim Dong-ju <[EMAIL PROTECTED]> writes:
 

is it http://debian-amd64.alioth.debian.org/ ?
but this doesn't support rsync. 
   

Yes, it does.
 

Maybe it does, but I couldn't get it to work.  It always asked (and I 
just checked, still does) for a login name / password and a few generic 
guesses of mine (guest, anonymous, etc) didn't work.  I'd like to use 
rsync instead of http if possible.

-Peter

Mirroring Problems

[Peter Nelson]

Hello, I'm creating a mirror of pure64/gcc-3.4 and every time I run to 
update I get the following errors:

Download of 
pool/unstable/main/amd64/a/aspell-ukr/aspell-ukr_0.51-0-1_amd64.deb 
failed: 404 Not Found Download of 
pool/unstable/main/amd64/d/diasce2/diasce2_1.3.1-2_amd64.deb failed: 404 
Not Found Download of 
pool/unstable/main/amd64/liba/libapache-mod-tsunami/libapache-mod-tsunami_2.0-2_amd64.deb 
failed: 404 Not Found Download of 
pool/unstable/main/amd64/t/totem/totem-gstreamer_0.99.15.1-2_amd64.deb 
failed: 404 Not Found Download of 
pool/unstable/main/amd64/t/totem/totem-xine_0.99.15.1-2_amd64.deb 
failed: 404 Not Found Download of 
pool/unstable/main/amd64/t/totem/totem_0.99.15.1-2_all.deb failed: 404 
Not FoundFailed to download files (6 errors)!

I don't use any of said packages, but I'm curious as to why they're 
listed in the package file if they're not actually there.  Also, my 
mirror will (Real Soon Now) be open for anyone on Internet2.

-Peter

Re: Is AMD64 Debian port suitable for production servers

[Peter Nelson]

Pavol Luptak wrote:
Hello,
I need to choose a stable software/hardware server solution based on Debian 
distribution. I'm planning to buy a 64-bit system (MB GigaByte K8VT800,
AMD Athlon 64, socket 754, chipset VIA K8T800+8237 with SATA disks).
According to the compatibility list 
http://alioth.debian.org/docman/view.php/30192/27/mainboards.html 
the above-mentioned motherboard should work in Linux without problems.
I would like to ask if Debian AMD64 port is stable enough for production
environments (with the latest kernel). Could you describe your experiences and
stability problems with your AMD64 production servers?
Do you think Intel P4 (and IA32 MB) is still more suitable for stable servers?
All help would be highly appreciated.
 

The AMD64 is more than stable enough (with a good motherboard) to run as 
a server, but the Debian AMD64 port is not close.  The port is somewhere 
between Debian unstable and Debian experimental, so unless you'd 
consider those suitable for production use neither is this port.  That 
said, just normal (i386) Debian will work perfectly fine on said system.

-Peter

Re: /etc/apt/apt.conf APT::Get::Compile flags

[Peter Nelson]

Jean-Michel POURE wrote:
Dear all,
I would like to try compiling ftp://ftp.nerim.net/debian-marillat packages to 
AMD-64. Are there special flags needed in /etc/apt/apt.conf to compile using:
apt-get source package_name --build?

You shouldn't need any flags anywhere as long as you have a deb-src line 
for nerim and the dpkg-dev package (which has dpkg-buildpackage).  I 
actually just built this program for amd64 and what I ended up doing is 
a loop of  `apt-get build-dep mplayer`, it complains it can't get 
foo-bar, apt-get source --build foo-bar, repeat. 

Speaking of which, sometime soon I'm both going to setup an 
amd64/gcc-3.4 mirror and a repository of things like mplayer custom built.

-Peter