Thanks Ben; I expected that was the case based on policy.
Contrary to my expectations, this is unresolved upstream for my hardware
combination in STA mode. I've included some details on my additional
tests upstream.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
wit
On Thu, 2010-10-21 at 04:09 +0100, Ben Hutchings wrote:
> I don't see any fixes there except for disabling power-saving by
> default, and we have that already.
After disabling power management (which I suspect may have already been
turned off by default), no change with this issue.
Oct 20 18:18:
Package: linux-2.6
Version: 2.6.32-5
Severity: important
After upgrading to Squeeze, my ath9k is dropping its connection
intermittently.
My hardware and issues matches a bug that is apparently resolved in
current linux vanilla (per upstream comments), however persists in the
current squeeze 2.6.3
I would like to clarify a few points made in my earlier comments; In
particular that it was blkid that had the incorrect UUID output and that
additional steps I performed to retain my raid data in addition to
correcting the UUID issue that may be relevant to others with this issue
on RAID1.
In my
I have seen this problem when migrating systems from a single ext[234]
disk to RAID1. The problem seems to be that when an ext filesystem is
overwritten with an md device, the UUID of the original ext file system
is preserved, or continues to be detected.
In my cases the md device represents the e
The same version of lvm2 is used in Ubuntu Karmic, and exhibits similar
behaviour. My system is running 2.6.31-20-generic #57-Ubuntu SMP AMD64
(current karmic-proposed).
Ubuntu bug is filed here, however I suspect this is the same issue.
https://bugs.launchpad.net/ubuntu/+source/lvm2/+bug/5302
see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=545224 on current xen and
grub2 boot hang
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=545224 for this issue
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Package: grub-pc
Version: 1.97~beta3-1
Followup-For: Bug #545224
We have been using a recipe Lenny+RAID+Crypto+LVM+Xen+iSCSI with quite a bit of
sucess on legacy servers with large disks.
1.0TB, 1.5TB were not an issue. Now with 2.0TB and larger, we are facing GPT
parititioning forcing us away
This message,
(XEN) Early fatal page fault at e008:828c801da4a2
seems to be a result of Grub2 not supporting Xen kernels.
More info here:
http://lists.gnu.org/archive/html/grub-devel/2009-06/msg00219.html
And a related thread here:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=505517
c
Although the kernel loads as per our comments, there is a problem in
that it is not yet supported by Grub2. Only Xen 3.4 or newer are
supported. There is a grub2 patch for i386, as noted below.
http://lists.gnu.org/archive/html/grub-devel/2009-06/msg00219.html
Some evidence of others with this too
Package: grub-pc
Version: 1.97~beta3-1
Followup-For: Bug #505517
Using the patches (second set) above was a bit non-trivial. In the end our
system now adds Xen entries correctly.
Below is the output,
-- BEGIN OUTPUT --
endor:/usr/src# cat grub-pc_00_allow_xen.patch | patch -p0
patching file /et
Package: grub-installer
Followup-For: Bug #548156
Our team encountered this issue as well with 2TB disks (Intel SATA on Dell
Poweredge).
It caught us by surprise since 1.5TB disk recipes were not an issue (not GPT
also)
Latest 5.0.3 installer; We ran both normal, and again in expert in the p
enter the name of the package in which you have found a problem, or type
'other' to report a more general problem.
> locales
*** Welcome to reportbug. Use ? for help at prompts. ***
Detected character set: us-ascii
Please change your locale if this is incorrect.
Using 'Ian Mac
Package: linux-image-2.6.25-2-amd64
Version: 2.6.25-6
Severity: normal
Although my architecture is different(AMD/MCP51), I am experiencing exactly the
problem described in #482153.
My broadcom device is
03:00.0 Network controller: Broadcom Corporation BCM4312 802.11a/b/g (rev 01)
After hibern
Package: util-linux
Followup-For: Bug #426171
After installing nVidia drivers with patches supplied in #476504, there is no
hwclock hang issues and the
nvidia-kernel-source package provides graphical interface. For the first time
ever since 2006 my dv9000z is
running without any issues related t
Package: nvidia-kernel-source
Version: 169.12-1
Followup-For: Bug #476504
After reviewing the full bug report, and patching the source and
Makefile.modpost everything is working great.
A big win for HP laptop users with MCP51/nVidia who are experiencing the
hwclock issues reported in #426171 t
Package: nvidia-kernel-source
Version: 169.12-1
Followup-For: Bug #476504
Same error after testing with the most recent linux-kbuild-2.6.25 in sid.
-- System Information:
Debian Release: lenny/sid
APT prefers unstable
APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86
Package: kvm-source
Followup-For: Bug #467260
After a little hint from Jan, thought m-a clean; m-a a-i kvm-source did not
work,
after removing /usr/src/modules/kvm the same commands worked just fine.
thanks,
iMac
-- Package-specific info:
selected information from lshal(1):
/proc/cpuinfo:
Package: kvm-source
Version: 65+dfsg-2
Followup-For: Bug #467260
Experiencing the same build errors as Mike Hommey, even after m-a clean; m-a
a-i kvm-source
iMac
system.firmware.release_date = '11/22/2007' (string)
system.firmware.vendor = 'Hewlett-Packard' (string)
system.firmware.ve
Package: util-linux
Followup-For: Bug #426171
Testing out the new 2.6.25-rc2 kernel images from trunk at
kernel-archive.buildserver.net and the hwclock --hctosys hang
is no longer present. Expecting the official 2.6.25 Sid packages
to resolve this long standing issue for HP AMD z series laptops.
Package: util-linux
Version: 2.13-13
Followup-For: Bug #426171
This hwclock problem still exists even w/ the latest 2.6.24-rc5 images and new
F.3D BIOS for my device. With software RAID-1 it seems to trigger a rebuild
almost every time it happens.
For myself, commenting out all the 'hwclock --h
Package: vnc4server
Version: 4.1.1+X4.3.0-21
Followup-For: Bug #381274
Further on this bug that still exists, I am hoping to provide some clarity for
other users stumbling on a gdm that cannot load xvnc4server.
All -dpi -geometry and -depth options fail if they use the '=' rather than
space sep
Package: madwifi-source
Version: 1:0.9.2+r1784.20061027-1
Severity: important
The following is from the buildlog after executing #m-a a-i madwifi
dh_testroot
dh_clean
/usr/bin/make -C /usr/src/modules/madwifi clean \
KERNELPATH=/lib/modules/2.6.18-1-powerpc/build
KERNELRELEASE=2.6
is unable to do the completion for flash video files (flv). I
> attach a patch (diff -ruNp) to modify /etc/bash_profile in order to add
> this feature.
Fixed for next release.
Ian
--
Ian Macdonald | When this load is DONE I think I'll wash it
[EMAIL PROTECTED] | AGAIN ...
http://www.caliban.org/ |
|
|
On Sun 23 Apr 2006 at 16:37:55 +0200, you wrote:
> The mplayer completion misses the vro extension (used by DVD-recorders
> for DVD-Video files).
Added for next release.
Ian
--
Ian Macdonald | You know it's going to be a bad day when
[EMAIL PROTECTED] |
, Radu Spineanu wrote:
> Ian MacDonald wrote:
> > I use resolvconf to configure my resolv.conf on my Dom0. When I create
> > a new image using xen-tools, it creates a resolve.conf symlinked to a
> > non-existent file, something it would appear to have obtained from the
Hi Steve,
Sorry bout the delay; Shoud'a run the --debug myself first :)
I notice the first error is cat: /etc/resolv.conf: input file is output
file. This is bug 355910, actually unresolved - headed there to update
that. I manually patched my resolv.conf to be normal and ran a second
pass with
Package: xen-tools
Version: 1.2-1
Severity: minor
Recently I had some problems with how the /tls.disabled was created; I
noted an upstream patch that had resolved the problems with creating a
/lib/tls directory that could not be removed. I had made some debug to
the script to try and figure out th
Package: gallery2
Version: 2.0.4-1
Severity: minor
Upon upgrade, the installer reports the following file integrity as
invalid, yet the file was just freshly dropped by the 2.0.4-1 packages.
Probably just a minor internal checksum update or whatever gallery2
uses to validate itself.
cheers,
Ian.
Package: xen-tools
Version: 1.2-1
Severity: normal
For some reason, I cannot build a successful Sarge box with xen-tools
without the --cache=no parameter. Whenever I use the --cache=yes I get
the following output, followed by the default screen on how to
mount/start the DomU image. The passwd pr
Package: gnome-blog
Version: 0.9-3
Severity: grave
Something changed. I miss gnome-blog, its just so convienent.
~$ gnome-blog-poster
Traceback (most recent call last):
File "/usr/bin/gnome-blog-poster", line 3, in ?
pygtk.require('2.0')
File "/usr/lib/python2.3/site-packages/py
Package: xen-tools
Version: 1.1-1
Severity: normal
I had trouble finding a browseable upstream changelog(again), so this may be in
CVS.
I used
xen-duplicate-image --from=hostA --ip=192.168.20.2 --hostname=hostB
to replicate an newly created sarge image. Upon boot, the new image came
up without
Package: xen-tools
Version: 1.1-1
Severity: normal
Tags: none
Resolvconf is great, as it allows you to configure your nameservers from
/etc/network/interfaces using the dns-nameservers directive.
I use resolvconf to configure my resolv.conf on my Dom0. When I create
a new image using xen-tools,
s right. Anything that contains absolute
position references is going to have trouble when used in combination
with sudo, nice, etc.
Ian
--
Ian Macdonald | Perhaps the biggest disappointments were
[EMAIL PROTECTED] | the ones you expected
On Sat 25 Feb 2006 at 21:21:01 -0300, you wrote:
> El dom, 26-02-2006 a las 01:16 +0100, Ian Macdonald escribió:
> > On Mon 05 Sep 2005 at 14:21:56 +0200, you wrote:
> >
> > > Trying to complete on "sudo invoke-rc.d" causes the following error
> > > me
t; +
> +case $prev in
> +invoke-rc.d)
> +files=$( find /etc/init.d/ -perm +111 | sed 's|^/etc/init.d/||' )
> +COMPREPLY=( $( compgen -W "${files}" -- $cur ) )
> +return 0
> +;;
> +
> +*)
> +COMPREPLY=( $(
bash
> Version: 3.0-15
> Severity: wishlist
>
> Please include Mercurial complations with bash. See full bash script
> at:
>
>http://www.selenic.com/pipermail/mercurial/2005-August/003378.html
Added to contrib for next release.
Ian
--
Ian Macdonald | A pa
wing will not complete to "pkg-config",
>
> DEB_BUILD_OPTIONS=nostrip apt-get -b source pkg-confi
>
> but the following will
>
> apt-get -b source pkg-confi
This is outside of the control of the bash completion project, I'm
afraid. The variable assignment at the
here is no package-completion for "aptitude show", like
> "apt-cache show" has.
Thanks. Someone else has already reported this, so it's fixed for the
next release.
Ian
--
Ian Macdonald | To lead people, you must follow behind.
[EMAIL PROTECTED]
[: ==: unary operator expected
>
> Plainly completing on invoke-rc.d doesn't cause this problem, same for
> sudo with other commands (xine, ls, synaptic, etc.).
I can't reproduce this here. sudo's completion function doesn't even use
" will not complete at all.
This can't be fixed, due to a limitation in bash itself. For cd
completion to treat pathnames correctly and not append a space to them,
'complete -o filenames' must be used. Unfortunately, this has the
effect of treating $foo as a filename, too,
On Sat 25 Feb 2006 at 15:03:53 +, you wrote:
> I believe I was talking about the former, but I just tested it and it
> seems to be fixed in the version I have...
Well, that's all that matters :-)
Ian
--
Ian Macdonald | You can tell the ideals of a nation by
:
set_no_funcs() { ( set -o posix; set ) }
That way, there's no need to store the current state of the 'posix'
option and restore it afterwards.
Ian
--
Ian Macdonald | In spite of everything, I still believe
[EMAIL PROTECTED] | that pe
On Fri 24 Feb 2006 at 19:41:33 -0500, you wrote:
> Eh? Brandon wasn't talking about vulnerabilities; are you sure this
> is the right bug?
Yes. I was using the word in a broad sense and meant that the code is
almost certainly vulnerable to this bug in other places.
Ian
--
Ia
2,13 +6762,13 @@
> _physicalvolumes()
> {
> COMPREPLY=( $( pvscan | \
> - awk '/PV/ {if ($2 ~ /^'$cur'/) print $2}' ) )
> + awk '/PV/ {if ($2 ~ /^'${cur//\//\\\/}'/) print $2}' ) )
> }
>
> _logical
}
> [ "${have:-}" ] && complete -F _dhclient dhclient
>
>
> Doc Nielsen <[EMAIL PROTECTED]>
>
> -- System Information:
> Debian Release: 3.1
> APT prefers unstable
> APT policy: (500, 'unstable
bash: _interfaces: command not found
>
> I'm able to reproduce it both on an unstable box and on a testing one.
Which version of bash completion is this? I cannot reproduce this here.
In fact, I see no references to a function called _interfaces at all.
_dhclient() calls _configured_interfac
> Could the DEBUG variable at the top of /etc/bash_completion please be
> renamed to something a little less generic, given that any program is
> likely to want to use DEBUG?
The next release will use $BASH_COMPLETION_DEBUG.
Ian
--
Ian Macdonald
is a space in the
> executable path. Causes problems especially on Mac OS X with its
> quirky pathnames.
To be clear, are you talking about command completion (i.e. completion
of the first token on the command line) or completion of a subsequent
token?
The former case is handled by bash i
iving you the opportunity to select the 2nd file.
It's that way by design, since this is the desired behaviour after you
have performed the initial check-in, which, by definition, is only
needed once.
Thanks,
Ian
--
Ian Macdonald | Kliban's First Law of Dining: Never eat
On Fri 24 Feb 2006 at 05:59:27 -0800, you wrote:
> On Fri, Feb 24, 2006 at 01:07:09PM +0100, Ian Macdonald wrote:
> > On Mon 05 Sep 2005 at 14:27:33 +0200, you wrote:
> > > If I try to complete a file name before the file is created, once the
> > > file is created,
s foobar' is useful on its
> own to find the contents of the foobar alias.
>
> Solution
>
>
> Include neither an equals nor a trailing space in the completion.
This is partially fixed for the next release.
A single completion will now complete to 'foo=' (i.e.
be clever enough to dynamically eval command substitutions. That's way
too hard to write in shell code :-)
I can't even reproduce your partial success here. I have no idea how
you're getting that to find your debian subdir.
Ian
--
Ian Macdonald | Out of the crooke
b-24
> Severity: minor
>
> [EMAIL PROTECTED]:/tmp/gcc-3.4-3.4.3$ du --max-depth\= 1 -h debian/
> du: invalid maximum depth `'
> Try `du --help' for more information.
>
> The extra space is rejected by the du argument parser.
Unfortunately, there's no way to instru
libncurses5 5.4-4Shared libraries for terminal
> hand
> ii passwd 1:4.0.3-30.8 Change and administer password
> and
>
> -- no debconf information
Fixed for next release.
Ian
--
Ian Macdonald | The American nation in t
for GNU tar. To retain compatibility with the largest
number of systems, I will have to leave the completion as it is for now.
On the other hand, we could detect the version of tar at source time and
install a modified version of the function if tar >= 1.15. The question
is: is it worth it? Peo
foo
> tty1:
>foo
>$ cat f
>
> It would be better if it completed the file name in this case.
I cannot reproduce this here.
Ian
--
Ian Macdonald | As President I have to go vacuum my coin
[EMAIL PROTECTED] | collection!
http://www.caliban.o
est case, this is just leftovers and
> something that hasn't yet been updated. Worst case, bash2 completion
> is actually inconsistent with bash3 completion. That doesn't seem to
> be the case, since I also see:
>
> # features supported by bash 3.0 and higher
Fixed
le, less will correctly view gzipped files,
RPM packages and other file types, thanks to LESSOPEN. On the other
hand, zless won't handle most of these, because it is itself just an
application of the LESSOPEN facility.
I suggest that this be tackled locally by each distribution, as it's
|asf|ASF|vob|VOB|bin|BIN|dat|DAT|vcd|VCD|ps|PS|pes|PES|fli|FLI|viv|VIV|rm?(j)|RM?(J)|ra?(m)|RA?(M)|yuv|YUV|mov|MOV|qt|QT|mp[34]|MP[34]|og[gm]|OG[GM]|wav|WAV|dump|DUMP|mkv|MKV|m4a|M4A|aac|AAC|m2v|M2V|dv|DV|flac)'
> ;;
> esac
Fixed for next release.
Ian
--
Ian
f
> ii libc6 2.3.2.ds1-20 GNU C Library: Shared libraries
> an
> ii libncurses5 5.4-4Shared libraries for terminal
> hand
> ii passwd 1:4.0.3-30.9 change and administer password
> and
>
> --
e -f -X
> '!*.@(zip|ZIP|jar|JAR|exe|EXE|pk3|war|wsz|ear|zargo|xpi|sxw|ott)' unzip
> zipinfo
> complete -f -X '*.Z' compress znew
> complete -f -X '!*.@(Z|gz|tgz|Gz|dz)' gunzip zcmp zdiff zcat zegrep zfgrep
> zgrep zless zmore
> complete -f -X &#
t;
> Versions of packages bash depends on:
> ii base-files 3.1.4Debian base system miscellaneous
> f
> ii libc6 2.3.2.ds1-22 GNU C Library: Shared libraries
> an
> ii libncurses5 5.4-6Shared libraries fo
the best I can come up with at the moment.
Sorry for the absurdly long delay in replying. I've been completely out
of touch for six months as I relocated back to Europe with a new baby,
bought a house, etc.
Cheers,
Ian
--
Ian Macdonald | * JHM wonder
Package: gringotts
Version: 1.2.8+1.2.9pre1-10
Severity: grave
Problem: Gringotts does not allow me access to my .grg file via ususal
String on unstable.
Solution: Copy .grg over to Sarge, install Gringotts and export the
DISPLAY (In my case to the same X server where the local Gringotts
failed).
o interference from the completion).
There's not much that can be done about this, I'm afraid. The set
built-in displays the current environment, and that includes any
functions that have been defined. bash completion makes heavy use of
shell functions.
It would be nice if there were a
work in the
> way it was original intended to be used again. Having to type the name
> yourself or doing some silly workarounds like writing ls first, doing
> the tab completion and then changing the ls to w3m shouldn't really be
> needed.
Fixed for next release.
Ian
--
Ian Macdona
account.
> ls > foo
> or
> cat > foo
> autocomplete to all filenames (directories and non-directories), as it
> should.
This is a limitation of the completion mechanism. There's no easy way to
detect redirection at arbitrary positions on the command line and act
accordingly
;
>
> patch works for me using an commercial ssh2 environment.
>
> Anyway, for the beef (sorry about the unnecessary "cleanup" in some of
> the function - I guess I can go through my CVS and try to remove the
> cruft I changed if you don't want those parts):
Than
> Since sudo and all the other functions that become meta commandlines
> in themselve all accept switches,we should pass over them:
Integrated for next release.
Ian
--
Ian Macdonald | Increased knowledge will help you now.
System Administrator| Have mate's phon
On Sun 16 Jan 2005 at 23:34:37 +, you wrote:
> Hello,
>
> On Sun, 2005-01-16 at 10:57 -0800, Ian Macdonald wrote:
> > I can't reproduce this here.
> >
> > Everything in your /etc/bash.bashrc is fine, so your
> > /etc/bash_completion mus
On Sun 16 Jan 2005 at 20:04:36 +0100, you wrote:
> On Sun, Jan 16, 2005 at 10:47:22AM -0800, Ian Macdonald wrote:
> > This problem is not in the upstream release. 'cvs annotate' says:
> >
> > 1.435(ianmacd 10-Oct-02): !(python|-?))
> >
>
eturn 0
> ;;
> - !(*python|-?))
> + !(python|-?))
> [[ ${COMP_WORDS[COMP_CWORD-2]} != -@(Q|W) ]] && _filedir
> ;;
> esac
This problem is not in the upstream release. 'cvs annotate' says:
1.435(ianmacd 10-Oc
ing in your /etc/bash.bashrc is fine, so your
/etc/bash_completion must be at fault. Please e-mail me a copy.
There's almost certainly something wrong with the code that checks for
which version of bash is installed, but it all works fine for me here.
Ian
--
Ian Macdonald | By pers
ttp-passwd=bar options. Unfortunately, when I do
>
> [EMAIL PROTECTED]:~$ wget --http-u
>
> I get
>
> [EMAIL PROTECTED]:~$ wget --http-user\=
>
> instead of the desired
>
> [EMAIL PROTECTED]:~$ wget --http-user=
>
> i.e., I get escaped equals sign and a
s really are zip archives. So it would be great
> to enable filename completion for sxw with unzip utility. Patch
> attached.
Fixed for next release.
Ian
--
Ian Macdonald | Religion is what keeps the poor from
System Administrator| murdering the rich. -- Napoleon
on a prayer.wav'
This is a tough one to fix. If I add '-o filenames' to the complete
command for dd, bash will start to prefix options with a
backslash, e.g. conv\=. This is cosmetic, however. Things will still
work.
I'll make the change for the next release and we'l
> Probably not very important, as running lvm utilities as non root is
> meaningless.
> Still, using bash_completion with commands like lvresize when using non
> root will fail with an error.
Fixed for next release.
Ian
--
Ian Macdonald | Dr. Jekyll had something to Hyde.
79 matches
Mail list logo
