I tracked this down on the KDE bug system and a temporary workaround is to put
RandomDevice=/dev/urandom
in the [General] section of /etc/kde3/kdm/kdmrc.
The reason for the failure to start is that by default /dev/random is used and
that will block if there isn't enough randomness available.
Package: udev
Version: 0.076-4
Severity: normal
I'm not sure what severity to set this bug at because on my system it
breaks bootlogd. Also I'm not sure the fix should be in udev.
My kernel does not have CONFIG_LEGACY_PTYS set so it relies on devpts to
provide ptys. Bootlogd works with both
Hi Roland,
The conversion script does sort the file in correct order but if the
user twiddles with it things may get run in incorrect order (e.g.
put networking at the end and see how many daemons fail to start).
This is an ugly situation, but I don't think that it justifies a
critical
OOPS, I hit ctrl-enter before I typed in the good news part.
Now the good news:
All you have to do is replace xxx $configfile with sort
$configfile | xxx. Actually, you can't since sort is in /usr/bin,
you'll have to do a sort with bash functions.
I don't have a sh function available
Client Update - 4 New Matches:
Hey there, just wanted you to know I have matched your profile with 4 girls
within a few miles of your City who are seriously interested in hooking up with
you, yes I've done my homework ;-) now its time for you to bring your A game.
I'm sure they will probably
Hi All,
I just blundered into this bug and had a thought about this error
message. It
looks weird because everything is running as root in the initrd boot
sequence.
...except with NFS root doesn't get root access to remote files unless the
server is set to allow it. If the NFS server
Package: libc6
Version: 2.3.5-6
Severity: normal
/etc/init.d/glibc.sh is called by init before anything else in the boot
sequence. When it is called there is no guarantee that /usr is mounted
so using dpkg may not work. /proc and /sys won't be mounted either and
/dev may not be fully populated.
Package: mailutils
Version: 1:0.6.90-3
Severity: normal
For one reason or another I tried to remove mailutils but dpkg
complained and exited with an error since I was also purging the
dependencies too. It was one of these I was purging so try thos to
reproduce:
libgsasl7
libmailutils0
Package: hplip-data
Version: 1.6.7-2
Severity: normal
This is causing me grief with trying to build dependancy trees. hplip
correctly depends on hplip-data and I don't think there is anything in
hplip-data apart from data.
-- System Information:
Debian Release: testing/unstable
APT prefers
On Tue, 2006-09-12 at 12:05, Henrique de Moraes Holschuh wrote:
severity 387093 wishlist
thanks
On Tue, 12 Sep 2006, Tim Phipps wrote:
This is causing me grief with trying to build dependancy trees. hplip
Please expand on that. The reverse recommends is not a strong dependency
loop
Package: unionfs-modules-2.6.17-2-686
Version: 1.3+debian-4+2.6.17-2
Severity: normal
Title say it all, I couldn't build without this package.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh
Package: usplash
Version: 0.3e
Severity: normal
After a usplash timeout (e.g. fsck) I'm left looking at vt8 which
doesn't have much on it and I need to alt-F1 to get to the real console
messages. Does usplash need to switch to vt8? If so can it switch back
on timeout.
-- System Information:
On Tue, 2006-09-19 at 02:33, Debian Bug Tracking System wrote:
From: Daniel Baumann [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Subject: Re: /usr/share/doc/syslinux/*.doc is empty.
Date: Tue, 19 Sep 2006 03:24:47 +0200
This was fixed already when I first uploaded a new syslinux package.
Package: icecast2
Version: 2.3.1-5
Severity: wishlist
The init script says Starting icecast2. and the icecast2 also says the
smae thing. Adding /dev/null to the start-stop-daemon commands for
start and restart fixes it, and doesn't stop error messages.
-- System Information:
Debian Release: 4.0
I'm changing this to grave since it looks like it's not something I'm
doing wrong and the driver crashes the server.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: xserver-xorg
Version: 1:7.1.0-5
Severity: normal
The package xserver-xorg has a circular dependency with
xserver-xorg-core. The package does nothing for me and will only do
something once anyway (generate an X config). Would you consider moving
the only file in this package to
The bug is in this section:
if [ ${NFSROOT#:*} = $NFSROOT ]; then
NFSROOT=${ROOTSERVER}:${NFSROOT}
fi
The :* is the wrong way around so if $NFSROOT=192.168.0.4:/ the :*
pattern won't match (NFSROOT doesn't start with :) and so NFSROOT gets
Package: initscripts
Version: 2.86.ds1-20
Severity: important
Setting this to important since it won't break simple systems.
I'm using the initramfs boot=nfs aproach to run a diskless system. I
have to remove eth0 from the auto line in /etc/network/interfaces
otherwise ifup -a hangs the system. I
On Friday 22 September 2006 14:17, Petter Reinholdtsen wrote:
[Tim Phipps]
Why do ifup -a hang on your system? This seem to be the most obvious
bug to fix to solve this issue.
It's hard to tell since it locks totally. I think ifup is deconfiguring the
interface before it starts to up
On Friday 22 September 2006 15:04, Petter Reinholdtsen wrote:
[Tim Phipps]
What's the problem with having the old systems of an init.d script
running after ifup?
Some times it take too long for the network to start working, so
mountnfs.sh would always fail. See #360123.
Hmm, tough
Package: ifupdown
Version: 0.6.7-0.1
Severity: important
Diskless boot is broken for me at the moment. initscripts has moved
nfsmount.sh to /etc/network/if-up.d for good reasons but that means my
nfs mount of /home doesn't happen (meaning I can't log on etc.). I have
eth0 not marked as auto in
Package: initscripts
Version: 2.86.ds1-20
Severity: normal
I have verbose=no in /etc/default/rcS and invoke-rc.d rc.local start
says this:
invoke-rc.d: initscript rc.local, action start failed.
This happens even if /etc/rc.local has an exit 0 at the end. I think
it's this line:
[ $VERBOSE !=
Package: initramfs-tools
Version: 0.80
Severity: normal
I'm still getting the timeouts on root mount refered to in bug 359926.
With this version of initramfs-tools the retry fails without really
retrying and it enters a loop. After 180s it times out and drops to a
panic shell. IP is configured
Package: aptitude
Version: 0.4.1-1.1
Severity: minor
Running in a chroot with home directories mounted read-only I get this
error message in a read dialog box:
E: /home/root/.aptitude/config - Unable to open %s for writing (13 Permission
denied)
I think the %s should be removed.
-- System
Package: bible-kjv-text
Version: 4.20
Severity: normal
bible-kjv and bible-kjv-text depend on each other. If there really is
such adependancy they should be made into one package. I think that
bible-kjv-text doesn't depend on bible-kjv though. The data is fine
without bible-kjv, it just sits
On Mon, 2006-08-14 at 14:15, Matthew Vernon wrote:
Hi,
bible-kjv and bible-kjv-text depend on each other. If there really is
such adependancy they should be made into one package. I think that
They're separate packages because the text is architecture-
independent; that's a saving on
Package: udev
Version: 0.093-1
Severity: normal
Feel free to change this to minor or wishlist. I'd like vol_id to not
change the case of FAT partitions labels. I don't believe this will affect
many people since FAT partitions are usually created with upper case labels
anyway. It was quite hard to
On Mon, 2006-08-14 at 21:01, Kay Sievers wrote:
Huh, what do you mean? Libvolume_id does not translate anything. What
piece of software is showing a label with lowercase chars?
Sorry Kay,
False alarm. I had changed the label at the start of the volume but the
real label is stored in
Package: microcode.ctl
Version: 1.12-2
Severity: normal
I know it seems like nonsense but I do want the installation of
microcode.ctl to download the microcode even though my CPU is a VIA C3!
It won't harm the VIA C3 since the program checks at run time if the CPU
is suitable. I run a Celeron 4
Package: xvfb
Version: 2:1.1.1-21etch5
Severity: normal
The man page suggest ths as an example:
Xvfb :1 -screen 0 1600x1200x32
and it crashes with this error:
Fatal server error:
Couldn't add screen 0
The problem is the depth, x24 works fine but xvncviewer doesn't like
that and moves down to
Package: less
Version: 394-4
Severity: normal
Can't believe this hasn't been found before. You can't kill the loop
with ctrl-c or ctrl-Z. I had to kill it from another terminal.
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386
Package: checkinstall
Version: 1.6.1-5
Severity: wishlist
The --provides option would be useful so I could remove some duplicate
stuff that other packages depend on.
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i686)
Shell:
Package: apt-build
Version: 0.12.4
Severity: important
And --patch-strip doesn't work, you have to use --patch_strip.
I've got a patch to utils-linux that enables me to run a diskless system
by having /etc/mtab be a symlink to somewhere else. Without the patch
some programs don't work since they
I get this too, I'm using mpd to stream to icecast and ogg123 to play the
stream. I know it's not mpd or icecast since I'm playing the same stream on
two computers and they start corruption at different times. When it does
start the input buffer is very low and it stays low. Well, it would
This is a known bug upstream:
http://trac.xiph.org/cgi-bin/trac.cgi/ticket/471
There's a patch that works for me which could be applied now:
--- vorbis-tools-1.0.1.orig/ogg123/buffer.c 2003-09-01 21:15:19 +0100
+++ vorbis-tools-1.0.1/ogg123/buffer.c 2005-04-21 17:28:23 +0100
@@ -165,7
Package: ncmpc
Version: 0.11.1-4
Severity: normal
ncmpc looks for the default config file in /usr/etc/ncmpc/config. There
isn't even a /usr/etc on this machine so I think it should be looking in
/etc/ncmpc.
-- System Information:
Debian Release: 3.1
APT prefers testing
APT policy: (101,
Package: lvm-common
Version: 1.5.20
Severity: important
For whatever reason /home could not be unmounted (actually a USB DVB-T
tuner had gone awol and the process attached could not be killed, so
that processes cwd could not be unmounted). vgchange hangs with the
message 'Can't deactivate vol
Package: mplayer
Severity: wishlist
I don't know if there's ever going to be a 1.0, so can we have rc2
packaged please?
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux
Package: aide
Version: 0.13.1-2
Severity: normal
I have /tmp symlinked to /dev/shm/tmp and the config script to create
rules under /dev does now quote filenames with spaces. aide then quits
with an error when it gets that config line.
-- System Information:
Debian Release: 4.0
APT prefers
Package: x11vnc
Version: 0.8.2-1
Severity: minor
Running with -quiet I get this output:
==
OK: x11vnc -forever -nofilexfer -usepw -quiet
-permitfiletransferpermit file transfer support
The VNC desktop is: hal:0
PORT=5900
==
I'm guessing that the -permitfiletransfer help line doesn't have
Package: mysql-server
Version: 5.0.32-7
Severity: minor
Now that /etc/mysql/my.cnf supports the include directory
/etc/mysql/conf.d I've moved all the customizations I had made into a
file. All except one: bind-address. I still have to comment this out in
/etc/mysql/my.cnf and I can't find a way
Package: mdadm
Version: 2.5.6-9
Severity: minor
mdadm.conf(5):
symlinks=no
Normally when creating devices in /dev/md/ mdadm will create
a matching symlink from /dev/ with a name starting md or md_.
Give symlinked=no to suppress this symlink creation.
This is now fixed with the latest evdev driver.
Thanks,
Tim.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: inputlirc
Version: 11-1
Severity: wishlist
Just started using this and it's very nice. So much easier than farting
around with the evdev driver for xorg and wrestling xmodmap.
Anyway, I wasted a fair amount of time wondering why the OK key of my
remote was the only one not working. I
On Friday 16 February 2007 5:30 pm, Guus Sliepen wrote:
It's a little bit different. Without arguments, inputlircd opens all
input event devices. This allows you to have multiple remotes, and also
process the multimedia keys on contemporary keyboards with lirc. Because
it reads the main
Package: xine-ui
Version: 0.99.4+dfsg+cvs2006-1
Severity: grave
Justification: renders package unusable
Without these packages xine from xine-ui just quits. Turning on
--verbose lists two failures in linking in libraries from these
packages. Installing these packages makes xine work again.
This is what i get with xine --verbose:
This is xine (X11 gui) - a free video player v0.99.5cvs.
(c) 2000-2006 The xine Team.
Built with xine library 1.1.2 (1.1.2)
Found xine library version: 1.1.2 (1.1.2).
Plateform informations:
--
system name : Linux
Package: aptitude
Version: 0.4.4-1
Severity: wishlist
It's only small and it makes it so much easier to work out what changes
are taking place. I struggled for ages until I found this lib.
-- System Information:
Debian Release: 4.0
APT prefers testing
APT policy: (500, 'testing')
Package: libmpd-dev
Version: 0.12.0-2
Severity: important
The second __cplusplus in this file is spelt _cpluplus.
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux
Package: debianutils
Version: 2.14.3
Severity: normal
Here's the description of what savelog does:
1) Version numbered files are cycled. Version cycle-2 is moved to
ver-
sion cycle-3, version cycle-1 is moved to version cycle-2 , and
so
on.
Package: file-rc
Version: 0.8.7
Severity: critical
Justification: breaks the whole system
The script /etc/init.d/rc does not use the sort field of
/etc/runlevel.conf to select the order of execution of the init.d
scripts. The conversion script does sort the file in correct order but
if the user
Package: kdm
Version: 4:3.3.2-1
Severity: grave
Justification: renders package unusable
There is crappage in /etc/init.d/kdm that causes that script to exit
without an error message. The lines in question 58 59:
# get the server text into the cache
wc -l /usr/bin/X11/XFree86
Package: libc6
Version: 2.3.5-6
Severity: normal
/etc/init.d/glibc.sh is called very early on in the boot process. /usr
may not be mounted so using /usr/bin/dpkg may fail for some people. So
would using /proc since the may not be mounted either. /bin/uname should
be there and /bin/sh can do
Package: aptitude
Version: 0.2.15.9-2
Severity: minor
When you press u and all your sources are from one server aptitude
estinates the size of each file to retrieve as zero before it gets the
size from the server. For the first files the file progress and the
total progress match each other
Looking at this some more I've found a bug in config.c. This code is bad:
/* Search the first */
for (i = 0; i = strlen (file_content) file_content[i] != ''; i++);
file_content += i + 1;
/* Search the last */
for (j = strlen (file_content); j = 1 file_content[j - 1] != ''; j--);
Package: apt-build
Version: 0.12.9
Severity: important
I installed apt-build 0.12.9 and after that edited
/etc/apt/apt-build.conf to suit. I tried to apt-build install aalib1-dev
and it fails with a mysterious error. When i restore the original
apt-build.conf it works. Here's my apt-build.conf:
Package: roundcube-mysql
Version: 0.1-4~bpo40+1
Severity: normal
Would it be possible to include a cron job to dump the database to
/var/backups? Something like savelog folowed by mysqldump.
-- System Information:
Debian Release: 4.0
APT prefers stable
APT policy: (500, 'stable')
Package: xserver-xorg
Version: 6.9.0.dfsg.1-6
Severity: normal
This version of xserver-xorg has removed the evdev man page, I'm sure it
was there in previous versions.
-- Package-specific info:
Contents of /var/lib/xfree86/X.roster:
xserver-xorg
/etc/X11/X target unchanged from checksum in
Package: xserver-xorg
Version: 6.9.0.dfsg.1-4
Severity: normal
I'm using the evdev input driver to get a Hauppauge remote control to
work. All the buttons work but two of the buttons end up with the same
keycode which is not much use. Using evtest the kernel device driver is
providing the correct
I have the same problem but not when the DVB drivers are modules. When the
drivers are modules udev manages to load the modules and create the device
files.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Package: initramfs-tools
Version: 0.60
Severity: minor
/usr/share/initramfs-tools/scripts/functions has a panic() function to spawn an
emergency or debug shell. It does so with this line
FS1='(initramfs) ' /bin/sh /dev/console /dev/console 21
FS1 should be PS1 if you want to set the
Package: lsb-base
Version: 3.1-10
Severity: normal
if TERM is set to xterm-mono the function log_use_fancy_output() returns
true and the log_* functions the try to use `tput setaf 1` and `tput op`.
These two fail with TERM=xterm-mono and scripts like /etc/init.d/apache2 exit
at that point since
Package: initramfs-tools
Version: 0.67
Severity: important
For some reason my diskless system seems to fail to boot every other time.
I get time-out messages from nfsroot and then the /init script dumps me
into a debug shell when it can't find /root/sbin/init.
I don't know the cause of the mount
Hi Peter,
On Friday 26 May 2006 14:59, Peter S Galbraith wrote:
Tim Phipps [EMAIL PROTECTED] wrote:
Package: xtide-data
Version: 20040203-1
Severity: normal
There's a circular dependancy here.
How so?
I think xtide-data shouldn't depend
Package: pmount
Version: 0.9.13-1
Severity: minor
pmount --help gives this in the first paragraph:
Mount device to a directory below /media/ if policy requirements
are met (see pmount(1) for details). If label is given, the mount point
will be /media//label, otherwise it will be
You can work around this with mount --bind /var/media /media or this
patch will make pumount use the real path of /media for the mount point
check.
Cheers,
Tim.
--- pmount-0.9.13.orig/src/pumount.c2006-08-29 14:20:05.0 +0100
+++ pmount-0.9.13/src/pumount.c 2006-08-29
Package: hal
Version: 0.5.7.1-1
Severity: normal
I though I'd screwed something up so I purged hal and reinstalled this
version and I still get some warnings. For installing I always get this:
Selecting previously deselected package hal.
(Reading database ... 202065 files and directories
Package: initramfs-tools
Version: 0.76
Severity: minor
The top part of init has this:
mkdir /sys
mkdir /proc
mkdir /tmp
mkdir -p /var/lock
These can all be put in the initramfs image can't they? Also (not a
bug), I can't figure out how /root gets created, it's not in the image
and there's no
Package: initramfs-tools
Version: 0.77
Severity: minor
The new mountroot() in scripts/nfs works nicely in that it does retry if
the first mount fails but it always says Retrying even though it
should say Trying the first time. Very minor bug though.
-- Package-specific info:
-- /proc/cmdline
Package: initramfs-tools
Version: 0.77
Severity: important
The new method of creating /dev/null is causing usplash to not work. At
the top of the init script /dev/null is created if it doesn't exist and
then a tmpfs is mounted on top of /dev. This hides the /dev/null and
/dev/console so any
Package: usplash
Version: 0.3e
Severity: normal
In the function log_end_msg, line 104 has 'printf \r' which is used to
keep the ok/fail message on the same line as the log_begin+msg text.
This confuses the output of bootlogd since it uses its own primitive line
buffer. Replacing the \r with \b
Package: ethtool
Version: 3-1
Severity: wishlist
It would be nice if the ethtool packages installed a few files to get
ethtool to run at boot time. e.g. /etc/init.d/ethtool that would read
/etc/default/ethtool so I could get it to switch on WOL on my card.
-- System Information:
Debian Release:
Package: udev
Version: 0.093-1
Severity: minor
I've traced a couple of error messages down to upslash_write being used
after initramfs.bottom has run. usplash_write tries to use stuff in /dev
but initramfs.bottom has moved /dev to $rootmnt/dev. The following patch
replaces the empty /dev with a
Package: e2fsprogs
Version: 1.39-1
Severity: minor
fsck always prints out its version and date when run. It's the only
program run at boot time that does this (in my setup) so it looks
a bit odd. Would you add the -T option to the initscripts?
-- System Information:
Debian Release:
On Wed, 2006-09-06 at 21:51, maximilian attems wrote:
From: Tim Phipps [EMAIL PROTECTED]
To: Debian Bug Tracking System [EMAIL PROTECTED]
I've traced a couple of error messages down to upslash_write being used
after initramfs.bottom has run. usplash_write tries to use stuff in /dev
Package: usplash
Version: 0.3e
Severity: normal
I have the kbd package installed (it handles bigger fonts than
console-utilities) and its initscript is called
/etc/init.d/console-screen.kbd.sh. The usplash init.d script is not
going to use this since it's looking for
/etc/init.d/console-screen.sh
Package: klibc-utils
Version: 1.4.19-1
Severity: wishlist
Could readlink be extended to include the -f option to canonicalize
paths?
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh linked to
On Tue, 2006-08-22 at 16:20, maximilian attems wrote:
it be nice if you would add an example where you need that?
Hi Maks,
I'm using readlink -f in a script in init-top from the
initramfs-tools to change the root device name from
/dev/disk/by-label/root to the real root device
On Tue, 2006-08-22 at 17:53, maximilian attems wrote:
first i'm a bit surprised why you would need to do that,
initramfs-tools boots just fine with
root=/dev/disk/by-label/root
Well, I don't _need_ to do it, I just want to so that the output of df
looks nicer. It's purely cosmetic.
well if
You can consider this bug work-arounded if you use the panic=0 boot
parameter. Getting a timeout would then reboot the machine thus giving a
retry (although a bit slower).
Cheers,
Tim.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL
Package: console-tools
Version: 1:0.2.3dbs-63
Severity: normal
The man page for openvt documents the -w option but it is not mentioned
in the SYNOPSIS section. It's an upstream bug since the same feature is
in RHEL3.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
Package: libchipcard2-libgwenhywfar38-plugins
Version: 2.1.4-1
Severity: normal
I have a script that runs aptitude to produce an aptfoster like list,
your package comes out as libchipcard2-libgwenhywfar38-pl. This might be
an aptitude bug or I may need to set COLUMNS=1000, either way a shorter
Package: libaqbanking-plugins-libgwenhywfar38
Version: 2.0.0-3
Severity: normal
I have a script that uses aptitude to produce an aptfoster like list,
Aptitude gave out a truncated version of your package name. Reportbug
also only got as far as libaqbanking-plugins-libgw in an 80 column
screen.
Package: syslinux
Version: 3.11-3
Severity: normal
/usr/share/doc/syslinux/*.doc just contains one line:
No catdoc available
I don't know if this is a problem on my system or on the build machine.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy:
Package: xtide-data
Version: 20040203-1
Severity: normal
There's a circular dependancy here. I think xtide-data shouldn't depend on
xtide. The invoke-rc.d stuff will work fine if the xtide package isn't
installed.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
Package: gpm
Version: 1.19.6-22
Severity: normal
I've got /dev/input/mice as my GPM mouse device (I think it was the default)
and I'm using udev. If a mouse isn't plugged in udev does not load the
mousedev driver at boot (which is probably correct) and so there is no
/dev/input/mice for gpm to
Package: kino
Version: 0.90-1
Severity: normal
kinoplus depends on kino so there's a dependency loop here.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux
Package: kgeography-data
Version: 4:3.5.2-1
Severity: normal
kgeography already depends on kgeography-data so it's redundant.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh linked to
Package: keurocalc-data
Version: 0.9.6-1
Severity: normal
keurocalc already depends on keurocalc-data so in all probability
keurocalc-data will never be installed on it's own.
-- System Information:
Debian Release: testing/unstable
APT prefers testing
APT policy: (500, 'testing')
Below is a very hacky script that implements the retry loop. The script
should go in /etc/initramfs-tools/scripts/nfs-bottom/retry. It adds a test
in /conf/param.conf and that script is sourced by the main /init. It needs
to do this since mountroot is only available to /init. The real fix should
Package: initramfs-tools
Version: 0.69b
Severity: normal
I created a script called /etc/initramfs-tools/scripts/nfs-bottom/retry-nfs.
There were plenty of hyphens in the path leading to script name so I
didn't think the name would be a problem. Imagine my suprise when I
rebooted and got some
Package: linux-image-2.6.16
Version: hippo.3
Severity: normal
drivers/media/video/cx88/*.c has this in places:
#define dprintk(level,fmt, arg...) if (debug = level) ...
I think that = should be a so that when debug is set to 0 no messages
are printed by the kernel. At the moment I have
The patch in 2.6.24.4 and the soon to be 2.6.25 fixes this problem but
only if you build with CONFIG_X86_GENERIC set.
The current Debian build for the 686 kernel does not set this and the
long NOPs will still be there.
This e-mail was sent from a group e-mail system of ARC International Plc.
Package: initramfs-tools
Version: 0.85h
Severity: normal
You can't use break=??? if you use panic=0. This is because the
maybe_break function calls the panic function to get a shell. It should
be simple to fix by just setting panic=1 inside maybe_break() before
calling panic().
--
Package: xserver-xorg-input-evdev
Version: 1:1.1.2-3
Severity: important
Just trying out the new evdev driver, here's what prints in the
terminal:
==
OK: X :2 vt6 -layout test
X Window System Version 7.1.1
Release Date: 12 May 2006
X Protocol Version 11,
Package: initscripts
Version: 2.86.ds1-61
Severity: normal
I looks like the pioodl() function is an attempt to run sort -r on
/proc/mounts. I think a better purpose would be to reverse the lines in
that file. Either you can use the program /usr/bin/tac or the following:
rev() {
local line
Here's a patch, I've tested it on my system and tried a few contrived mount
scenarios and it works better than pioodl(). Here's a way to break
pioodl():
mkdir -p /tmp/1/2/3
mount LABEL=disk3 /tmp/1/2/3
mount LABEL=disk2 /tmp/1/2
mount LABEL=disk1 /tmp/1
You get a /proc/mounts that looks like
Package: libreplaygain-dev
Version: 1.0~r412-2
Severity: important
I have been using gain_analysis.{c,h} copied from the source code for
mp3gain. That assumes the input is 16-bit twos-comp integers converted
to double so their range is -2^15 to +2^15. If I use that with
libreplaygain I just get a
Package: chrony
Version: 1.23-6+lenny1
Severity: normal
This fixes chronyc access to remote hosts:
From: Thomas Zajic zlatko zlatko fdns net
Date: Tue, 29 Jul 2008 23:35:42 +0100
Subject: [PATCH] Fix IP addressing in chronyc
Thomas wrote:
I found a bug in the chrony client (chronyc) that affects
Package: chrony
Version: 1.23-6+lenny1
Severity: normal
It uses FLAGS+= string which is a bashism.
-- System Information:
Debian Release: 5.0.4
APT prefers stable
APT policy: (500, 'stable')
Architecture: i386 (i686)
Kernel: Linux 2.6.32-1-686 (SMP w/2 CPU cores)
Locale: LANG=en_GB.UTF-8,
1 - 100 of 119 matches
Mail list logo