Bug#501692: [Pkg-libvirt-maintainers] Bug#501692: libvirt: Trouble with qemu 0.8.2 (etch)
On Thu, Oct 09, 2008 at 10:13:49PM +0200, Marc Fargas wrote: Also etch is neither my priority. I was bitten by this because I have etch + some stuff from lenny. As libvirt is not on etch maybe the sane thing to do would be add a versioned Recommends (if possible). I don't know if Recommends lines can be versioned, conflicting with qemu 0.9 is another option ;) I'll add a versioned recommends = 0.9.1 then, I'll also move the Recommends from libvirt0 to libvirt-bin while doing that. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500260: [Pkg-libvirt-maintainers] Bug#500260: [virt-manager] The VM creation wizard refers to the inexistant path /var/lib/libvirt/images
tags 500260 forwarded-upstream On Wed, Oct 08, 2008 at 01:43:34AM +0200, Laurent Léonard wrote: Mmm I didn't saw that dialog. But if I remove the default storage pool and I add a new one, the VM creation wizard always uses the default path /var/lib/libvirt/images, even if I restart /etc/init.d/libvirt-bin. There's a difference between qemu:///session and qemu:///system. The former can't write to the pool. I've send a patch upstream to fix this. I'll also include this in the next upload. Cheers and thanks for reporting back, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501999: ITP: ruby-libvirt -- libvirt ruby bindings
Package: wnpp Severity: wishlist Owner: [EMAIL PROTECTED] * Package name: ruby-libvirt * URL : http://libvirt.org/ruby/ * License : LGPL Programming Lang: C, Ruby Description : libvirt ruby bindings Needed for oVirt server. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501824: [Pkg-libvirt-maintainers] Bug#501824:
On Sun, Oct 12, 2008 at 07:03:11PM +0200, Francesco Frassinelli wrote: I don't know, it's the default installation. How could I know if the problem is that? Is the user in the libvirt group? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501824: [Pkg-libvirt-maintainers] Bug#501824:
On Sun, Oct 12, 2008 at 08:05:29PM +0200, Francesco Frassinelli wrote: Shouldn't be better ask during virt-manager installation for what users should be active those permissions? Fixed with virt-manager in experimental. You need to start the network with virsh manually (can be done as a normal user if he's in the libvirt group). -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502029: RFP: ruby-krb5 -- krb5 ruby bindings
Package: wnpp Severity: wishlist * Package name: ruby-krb5 * URL : http://rubyforge.org/projects/krb5-auth/ * License : LGPL Programming Lang: C, Ruby Description : basic kerberos ruby bindings Needed for Ovirt packaging. Would the ruby team be interested in packaging/maintaining this? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502117: [Pkg-libvirt-maintainers] Bug#502117: libvirt-bin: Doesn't create /etc/xen/vm_name after installing a new domU
severity 502117 wishlist On Mon, Oct 13, 2008 at 02:39:18PM -0300, André Luís Lopes wrote: [..snip..] There is, however, a file under /var/lib/xend/domains/dom_id called config.sxp, which seems to have the guest's config but which is not using the same syntax accepted by Xen's xm command. I assume virsh --all shows the machine fine as shutoff and you can start it with virsh start dom? I'm attaching the /var/lib/xend/domains/dom_id/config.sxp file for reference. Let me know if you need any other file. That's perfectly fine. Libvirt (or Xen actually) puts them there (at least it did when I last used Xen). Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502253: git-buildpackage: please allow gbp.conf in ./debian subdir
On Wed, Oct 15, 2008 at 12:47:49AM +0200, sean finney wrote: the ability to have per-branch gbp.conf is nice, but unfortunately it requires having a top-level file .gbp.conf. this unfortunately creates an additional file in the diff that is not in the debian subdir, which causes alarms by lintian and otherwise lowers the aesthetic value of having everything else in ./debian I kind of do like this. We could also call it debian/gbp.conf (no hidden files involved) which would make things more transparent. Thanks for the suggestion. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#416204: closed by Guido Guenther [EMAIL PROTECTED] (Bug#416204: fixed in python-dateutil 1.4.1-2)
On Wed, Oct 15, 2008 at 07:57:01PM +0200, Kurt Roeckx wrote: * [a5de481] don't ship timezone information (Closes: #416204) [...] The description of python-dateutil says: * internal up-to-date world timezone information based on Olson's database. Maybe you would like to remove that from the description as well. Mayby you should also add a Depends on tzdata since it's not actually essential (but util-linux Depends on it, and that is.) Yes of course. Updated and pushed into git. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502253: [PATCH] add debian/gbp.conf to list of searched conffiles
Hi Sean, On Wed, Oct 15, 2008 at 09:20:16PM +0200, Sean Finney wrote: as requested in #502253. this allows to keep all debian-specific changes (including gbp.conf) in the debian subdir, which provides both aesthetic and technical advantages. the file was added to the end of the list of locations searched, so it should hopefully not intefere with anyone who is using any of the previous locations. thanks for the patch but I think we have to push debian/gbp.conf in front of .git/gbp.conf - otherwise the user has no way of overriding configurations pushed onto him by the remote repo. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502439: please add libvirt plugins
Package: munin-plugins-extra Severity: wishlist Hi, please add libvirt plugins. Network and Interface I/O plugins are available here: http://honk.sigxcpu.org/projects/libvirt/monitor/ A plugin to monitor the load of virtual machines is available here: http://packages.ubuntu.com/intrepid/admin/virt-goodies This would then need a: Recommends: python-libvirt, python-libxml2 Or would you prefer a separate munin-plugins-libvirt? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502253: [PATCH] add debian/gbp.conf to list of searched conffiles
On Thu, Oct 16, 2008 at 07:48:18PM +0200, Sean Finney wrote: hi guido, On Thu, Oct 16, 2008 at 09:12:42AM +0200, Guido Günther wrote: thanks for the patch but I think we have to push debian/gbp.conf in front of .git/gbp.conf - otherwise the user has no way of overriding configurations pushed onto him by the remote repo. that's a good point. pretty trivial to fix anyway... do you need another patch or can we leave it at that? :) No patch needed. I just wanted to point this out in case you expected a different behaviour. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502565: git-buildpackage: [git-import-orig] cannot upgrade after uscan (watch file)
retitle 502565 git-import-orig: should handle orig.tar.gz symlinks gracefully severity 502565 minor On Sat, Oct 18, 2008 at 01:07:27AM +0300, Jari Aalto wrote: 1. Run uscan to update to newest according to debian/watch 2. Run git-import-orig to try to upgrade the newest tar.gz This of course works if you pass foo_2.6.orig.tar.gz and not the foo-2.6.tar.gz. = Cannot symlink '../foo-2.6.tar.gz' to '../foo_2.6.orig.tar.gz': File exists No it shouldn't. We could check if the orig.tar.gz symlink points to the tarball you passed to git-import-orig and stop otherwise. But we shouldn't simply ignore the fact that there's a symlink already. Pristine tar wants to use the orig.tar.gz and if that's different from the tarball you passed things end up badly. welcome. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502737: [openoffice.org-writer] Missing mime-mapping entry for .oxt
On Sun, Oct 19, 2008 at 05:58:32PM +0200, Rene Engelhard wrote: And why should it be mapped to -writer? -core would make more sense (there's also non-writer extensions). But those extensions might need e.g. draw (pdfimport), impress (present*) or base (report-builder) or writer (wiki-publisher) to work. Not having packages installed is another issues. This bug talks about people downloading extensions and not knowing what to do with it. This works: /usr/bin/openoffice bla.oxt so mapping this to /usr/bin/openoffice (openoffice.org-common) seems to do the trick for all kinds of extensions. So what appluication do you assign it to? To all. etc. See above. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502737: [openoffice.org-writer] Missing mime-mapping entry for .oxt
On Mon, Oct 20, 2008 at 01:25:12AM +0200, Rene Engelhard wrote: And it's now a Type=MimeType one (thus no Exec=), so I wonder where it belongs. (And it probably won't fix this bug at all, as you want oxt /usr/share/mimelnk? associated with somethig executable so that users don't need to go to the extension manager themselves) This depends. Is the MimeType specified in the ooo-extension.desktop mapped to one of the applications .desktop files? If so we just need to install that one to /usr/share/mimelnk. If not something like: [Desktop Entry] Version=1.0 Encoding=UTF-8 Terminal=false Type=Application NoDisplay=True Exec=openoffice %U MimeType=Mime-Type Name=OpenOffice.org in /usr/share/applications/ will do the trick. Mime-Type would be the mime type defined in the other desktop file you found. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#502737: [openoffice.org-writer] Missing mime-mapping entry for .oxt
On Mon, Oct 20, 2008 at 09:30:12AM +0200, Rene Engelhard wrote: Hi, Guido Günther wrote: On Mon, Oct 20, 2008 at 01:25:12AM +0200, Rene Engelhard wrote: And it's now a Type=MimeType one (thus no Exec=), so I wonder where it belongs. (And it probably won't fix this bug at all, as you want oxt /usr/share/mimelnk? Oh, that still exists? It's at lease used by KDE3 so it won't hurt. Having the appropriate info in /usr/share/mime for Gnome/KDE4 is more important though - but maybe it's already there - which Mimetype is given for the extension? associated with somethig executable so that users don't need to go to the extension manager themselves) This depends. Is the MimeType specified in the ooo-extension.desktop mapped to one of the applications .desktop files? If so we just need to install that one to /usr/share/mimelnk. If not something like: [Desktop Entry] Version=1.0 Encoding=UTF-8 Terminal=false Type=Application NoDisplay=True Exec=openoffice %U unopkg gui %U if we get an own .desktop for this. Yeah, even better. Regards, Rene P.S.: I am still not sure I want people make it that easy to run into problems this way when they try to install binary extensions (see e.g. see #502812, and no, it's IMHO not an option to build against an internal, ancient version of STLport again) but that would be no regression to 2.4.1... I think we should. The extension I used worked perfectly (and I might end up packaging it) but we shouldn't try to block users from installing extensions by not adding a mapping. In the worst case there should be a warning in the unopkg gui. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#503007: [Pkg-libvirt-maintainers] Bug#503007: virtinst tries to create a temporal file in inexistent directory /var/lib/xen and dies
Hi Nahuel, On Tue, Oct 21, 2008 at 04:32:15PM -0200, Nahuel Greco wrote: Package: virtinst Version: 0.300.3-5 Severity: normal virtinst dies when he tries to create a temporal file in the inexistent directory /var/lib/xen, when I manually create that directory everything goes ok: Could you check if 0.400.0-3 works as expected? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#503007: [Pkg-libvirt-maintainers] Bug#503007: virtinst tries to create a temporal file in inexistent directory /var/lib/xen and dies
fixed 503007 0.400.0-4 thanks On Tue, Oct 21, 2008 at 05:21:33PM -0200, Nahuel Greco wrote: 0.400.0-3 has another kind of errors: morron:/tmp# virt-install Traceback (most recent call last): File /usr/bin/virt-install, line 34, in module import virtinst File /var/lib/python-support/python2.5/virtinst/__init__.py, line 28, in module from Guest import Guest, VirtualNetworkInterface, XenGuest, \ File /var/lib/python-support/python2.5/virtinst/Guest.py, line 326 if os.getuid() == 0: ^ This is fixed now, thanks. TabError: inconsistent use of tabs and spaces in indentation morron:/tmp# After fixing that and a couple of other TabError's, I found that running virt-install without parameters as before gave me no interactive creation of the virtual machine: morron:/tmp# virt-install ERRORA name is required for the virtual machine. morron:/tmp# Yes. 0.400.0 dropped most of the interactive code. You can use --name, --ram, etc. Looking at the code your problem is inded fixed in 0.400 since it uses .virtinst/boot if /var/lib/xen doesn't exist. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#503007: [Pkg-libvirt-maintainers] Bug#503007: virtinst tries to create a temporal file in inexistent directory /var/lib/xen and dies
tags 503007 +patch thanks On Tue, Oct 21, 2008 at 04:32:15PM -0200, Nahuel Greco wrote: OSError: [Errno 2] No such file or directory: '/var/lib/xen/virtinstmnt.mpUlvF' morron:~/virtual_machines# morron:~/virtual_machines# ls -d /var/lib/xen* Here's a patch for 0.300.0 that should fix the problem. Can you try it? Cheers, -- Guido diff --git a/virtinst/Guest.py b/virtinst/Guest.py index c94c186..fc87b70 100644 --- a/virtinst/Guest.py +++ b/virtinst/Guest.py @@ -477,7 +477,8 @@ class Installer(object): def get_scratchdir(self): if self.type == xen: -return /var/lib/xen +if os.path.exists(/var/lib/xen): +return /var/lib/xen return /var/tmp scratchdir = property(get_scratchdir)
Bug#495975: please package current SVN
On Sat, Aug 23, 2008 at 09:17:36PM +0300, Riku Voipio wrote: On Thu, Aug 21, 2008 at 07:48:43PM +0200, Guido Günther wrote: Package: libtinymail Severity: wishlist Since tinymail is nearing a 1.0 release it would be nice to have the soon to be stable API in experimental. Especially since tmut and most other applications rely on something more recent than 0.0.9 anyway. Since nothing rdepends on libtinymail (yet), that should be fine. do you want to join pkg-maemo to push the uploads? Jonny Lamb who's done most of our tinymail work is bit busy at the moment.. Same here unfortunately. I'll just keep filing bugs and attach patches. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#496551: customization is hard
Package: zhone-session Version: 0.2 Severity: wishlist Currently everything is hardcoded in /usr/bin/zhone-session. There's now way in getting added programs via .Xsession or similar. One either has to modify /etc/init.d/zhone-session or /usb/bin/zhone-session. I'd be nice if the later would at least source e.g. /etc/zhone-startup or something similar. -- Guido -- System Information: Debian Release: lenny/sid APT prefers testing APT policy: (500, 'testing'), (50, 'unstable') Architecture: powerpc (ppc) Kernel: Linux 2.6.26-2-g53be6b6 Locale: LANG=de_DE.UTF-8, LC_CTYPE=de_DE.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#495290: [Pkg-libvirt-maintainers] Bug#495290: virt-manager: Can't send Ctrl+Alt+Del to guest
Hi Per, On Sat, Aug 16, 2008 at 12:14:34AM +0200, Per Andersson wrote: Menu entry Send Key - Ctrl+Alt+Del won't work -- unless line 197 in /usr/share/virt-manager/virtManager/console.py is changed from keys = [Control_L, Alt_L, Del] to keys = [Control_L, Alt_L, Delete] Sorry for the late reply. I'll prepare an upload as soon as possible. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#496927: Please package new upstream snapshot
On Thu, Aug 28, 2008 at 05:30:16PM +0200, Jelmer Vernooij wrote: Hi Günther, Ahem. Any chance you can package a new upstream snapshot of python-kerberos? My authGSSWrap changes have been added since the currently packaged snapshot. The current snapshot has your patch already. See: 029edc3f87a8a28f4864069dfb1502c6c1c81771 -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#497079: allow per distribution keyrings
Package: mini-dinstall Version: 0.6.25 Severity: wishlist Tags: patch Hi Christoph, attached patch allows (as you know) per distribution keyrings. Please apply. -- Guido From 1b3c7277683e9588d88aa8e215f3b91624e04318 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Fri, 29 Aug 2008 19:41:00 +0200 Subject: [PATCH] allow keyrings and extra_keyrings per repository this allows for fine grained permissions, like key 0x... is only allowed to upload into the experimental archive. --- mini-dinstall | 14 +++--- 1 files changed, 11 insertions(+), 3 deletions(-) diff --git a/mini-dinstall b/mini-dinstall index 08f1ffc..063967e 100755 --- a/mini-dinstall +++ b/mini-dinstall @@ -365,6 +365,8 @@ class DistOptionHandler: self._optionmap['experimental_release'] = ['bool', 0] self._optionmap['release_description'] = ['str', None] self._optionmap['release_signscript'] = ['str', None] +self._optionmap['keyrings'] = ['list', None] +self._optionmap['extra_keyrings'] = ['list', None] def get_option_map(self, dist): ret = self._distributions[dist] @@ -382,9 +384,10 @@ class DistOptionHandler: return self._configp.getint(dist, key) elif type == 'str': return self._configp.get(dist, key) + elif type == 'list': +return re.split(', ?', self._configp.get(dist, key)) elif type == 'bool': return self._configp.getboolean(dist, key) - assert(None) @@ -711,8 +714,13 @@ class ArchiveDir: self.__dict__['_' + key] = configdict[key] do_mkdir(dir) self._batch_mode = batch_mode -self._keyrings = keyrings -if not extra_keyrings is None : +if configdict['keyrings']: +self._keyrings = configdict['keyrings'] +else: +self._keyrings = keyrings + if configdict['extra_keyrings']: +self._extra_keyrings = configdict['extra_keyrings'] +elif extra_keyrings: self._extra_keyrings = extra_keyrings else: self._extra_keyrings = [] -- 1.5.6.3
Bug#467504: git-buildpackage: git-import-* is confused by .gitignore in the tarball
Hi Ansgar, On Sat, Aug 30, 2008 at 11:46:07AM +0200, Ansgar Burchardt wrote: Hi, I stumbled over this problem as well and was wondering why the `configure' script would not be imported. If you do not want to add the --force option to `git add' calls by default, it would be nice if there was an option to make git-import-orig do this. Could you confirm the attached patch does what you want. -- Guido From 91aee1c45910671d8738bbff6a1f1b6e7863f009 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Sat, 30 Aug 2008 13:54:27 +0200 Subject: [PATCH] Don't honor .gitignore during replace_source_tree this only leads to missing files during git-import-orig --- gbp/git_utils.py |2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/gbp/git_utils.py b/gbp/git_utils.py index 13845d5..f04beca 100644 --- a/gbp/git_utils.py +++ b/gbp/git_utils.py @@ -138,7 +138,7 @@ def replace_source_tree(repo, src_dir, filters, verbose=False): old = set(repo.index_files()) new = set(copy_from(src_dir, filters)) -GitAdd()(['.']) +GitAdd()(['-f', '.']) files = [ obj for obj in old - new if not os.path.isdir(obj)] if files: GitRm(verbose=verbose)(files) -- 1.5.6.3
Bug#496830: any packages yet
Hi Free, did you already start packaging python-augeas? I have packages ready at: http://git.debian.org/?p=users/agx/python-augeas.git;a=summary which I could just upload. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#496830: any packages yet
On Sun, Aug 31, 2008 at 02:48:09PM +0200, Free Ekanayaka wrote: Sorry, I really didn't notice it it :/ I've already made a package, No problem, I just stitched the package together this morning since I wanted to try the python bindings. it's in the new queue now: http://ftp-master.debian.org/new.html If you wish we could maintain it together. If you don't mind I'll simply add patches to the BTS for now (if necessary at all) - the package seemps quiet simple. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#497335: git-buildpackage: Don't use deprecated command names
tag 497335 +pending thanks On Mon, Sep 01, 2008 at 01:35:38AM +0200, Michael Biebl wrote: Package: git-buildpackage Version: 0.4.37 Severity: important Hi, as you can see, I'm using git 1.6 (from experimental). The usage of git-$command has been deprecated in 1.6, and the binaries are no longer available in $PATH. This makes git-buildpackage fail, as it e.g. uses git-archive and not git archive. I fixed all of these in gbp/ but some in the git-* commands slipped through. I just pushed a fix into current git. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500254: git-buildpackage: doesn't read ~/.gbp.conf
On Fri, Sep 26, 2008 at 06:28:19PM +0200, Luca Capello wrote: [EMAIL PROTECTED]:/mnt/lab/Openmoko_GTA02/git/fso-frameworkd$ grep builder /home/luca/.gbp.conf builder = /home/luca/bin/gismo-pdebuild.sh can you post the whole file? Maybe this is being overriden by a .gbp.conf in the repo itself? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499963: [Calendarserver-maintainers] Bug#499963: calendarserver: caldavd fails to authenticate and autocreate principal when running with NssDirectoryService
On Fri, Sep 26, 2008 at 10:07:29AM -0700, Ben Poliakoff wrote: ..pretty sure you meant /var/spool/caldavd. Permissions seem fine: Sure. Thanks. [EMAIL PROTECTED] ~]$ sudo su -s /bin/bash caldavd [EMAIL PROTECTED]:/home/benp$ touch /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ ls -l /var/spool/caldavd/test -rw-r--r-- 1 caldavd caldavd 0 2008-09-26 10:01 /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ rm /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ ls -l /var/spool/caldavd/test ls: cannot access /var/spool/caldavd/test: No such file or directory [EMAIL PROTECTED]:/home/benp$ This is getting weird. Did you check if the user benp is in the valid uid range [firstValidUid-lastValidUid]? If he is, it might make sense to try out the XML backend instead of NSS for testing. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500438: postgrey patterns
Package: aide-common Version: 0.13.1-10 Severity: wishlist Hi, please find aide patterns for postgrey attached. It simply marks the database files as VarFiles. -- Guido From 85613feffb5ff0b3a822478866baba23d3f5ee81 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Sun, 28 Sep 2008 10:45:38 +0200 Subject: [PATCH] postgrey aide patterns --- debian/aide.conf.d/31_aide_postgrey |4 1 files changed, 4 insertions(+), 0 deletions(-) create mode 100644 debian/aide.conf.d/31_aide_postgrey diff --git a/debian/aide.conf.d/31_aide_postgrey b/debian/aide.conf.d/31_aide_postgrey new file mode 100644 index 000..2859261 --- /dev/null +++ b/debian/aide.conf.d/31_aide_postgrey @@ -0,0 +1,4 @@ +/var/lib/postgrey$ VarDir +/var/lib/postgrey/postgrey(|lock)\.db$ VarFile +/var/lib/postgrey/log\.[0-9]{10}$ VarFile +/var/lib/postgrey/__db\.[0-9]{3}$ VarFile -- 1.6.0.1
Bug#500458: git-buildpackage: can't import history of native package into empty git repo
retitle 500458 git-import-dsc: allow to import into empty repos severity 500458 wishlist thanks git-import-orig imports into *empty* repos while git-import-dsc doesn't and wants to create a new one. Should be fixed. -- Guido On Mon, Sep 29, 2008 at 12:59:19AM +0400, Nikita V. Youshchenko wrote: Nikita V. Youshchenko schrieb: Package: git-buildpackage Version: 0.4.39 Severity: normal I've tried to import previous releases of debian native package (no orig.tag.gz) into an empty git repo. No need to create a git repo first, just pass the dsc to git-import-dsc and it will do all the rest (including repository creation). Does this produce the expected result? Yes, this works. But anyway if repo already exists then operation fails, which is not good :) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#416204: [Bug 271680] Re: MIR for python-dateutil
On Mon, Sep 29, 2008 at 02:22:09PM +0200, Martin Pitt wrote: Guido, that means for the Debian package we don't need the patch, just the packaging changes (plus a dependency to tzdata). O.k., I'll drop that from the next upload. Thank you both for sorting this out. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500499: git-buildpackage: git-import-orig has problems with whitespaces
reassign 500499 pristine-tar thanks On Sun, Sep 28, 2008 at 11:00:38PM +0200, Michael Gebetsroither wrote: Package: git-buildpackage Version: 0.4.39 Severity: normal git-import-orig can't handle whitespaces in upstream tarballs. % git-import-orig --no-dch --pristine-tar ../TrueCrypt\ 6.0a\ Source.tar.gz -u 6.0a Upstream version is 6.0a Warning: Can't symlink orig.tar.gz due to missing debian/changelog Initial import of '../TrueCrypt 6.0a Source.tar.gz' ... tar: ../TrueCrypt: Cannot open: No such file or directory ^^^ tar: Error is not recoverable: exiting now pristine-tar: tar tf exited nonzero /usr/bin/pristine-tar returned 2 Couldn't run '/usr/bin/pristine-tar' Import of ../TrueCrypt 6.0a Source.tar.gz failed That's pristine-tar failing: $/usr/bin/pristine-tar commit /var/scratch/debian/git-buildpackage/testsuite/500499/data//TrueCrypt 6.0a Source.tar.gz refs/heads/master tar: /var/scratch/debian/git-buildpackage/testsuite/500499/data//TrueCrypt: Cannot open: No such file or directory tar: Error is not recoverable: exiting now pristine-tar: tar tf exited nonzero -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499963: [Calendarserver-maintainers] Bug#499963: calendarserver: caldavd fails to authenticate and autocreate principal when running with NssDirectoryService
On Tue, Sep 30, 2008 at 02:37:41PM -0700, Ben Poliakoff wrote: * Guido Günther [EMAIL PROTECTED] [20080928 01:23]: On Fri, Sep 26, 2008 at 10:07:29AM -0700, Ben Poliakoff wrote: ..pretty sure you meant /var/spool/caldavd. Permissions seem fine: Sure. Thanks. [EMAIL PROTECTED] ~]$ sudo su -s /bin/bash caldavd [EMAIL PROTECTED]:/home/benp$ touch /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ ls -l /var/spool/caldavd/test -rw-r--r-- 1 caldavd caldavd 0 2008-09-26 10:01 /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ rm /var/spool/caldavd/test [EMAIL PROTECTED]:/home/benp$ ls -l /var/spool/caldavd/test ls: cannot access /var/spool/caldavd/test: No such file or directory [EMAIL PROTECTED]:/home/benp$ This is getting weird. Did you check if the user benp is in the valid uid range [firstValidUid-lastValidUid]? If he is, it might make sense to try out the XML backend instead of NSS for testing. -- Guido Using the XML backend seems to work fine (tested both with the example 'test' user and with a newly defined user): Did you check if the user benp is in the valid uid range [firstValidUid-lastValidUid]? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500865: git-dch submodule support
Package: git-buildpackage Severity: wishlist Manoj keeps debian/ in a git submodule (do others also use this workflow?). In this case git-dch can't detect all the changes, here's what could be done (as suggested by Manoj): first find the top of the tree : , | base_dir=$(git-rev-parse --show-cdup 2/dev/null) || return 1 | if [[ -n $base_dir ]]; then | base_dir=$(readlink -f $base_dir) | else | base_dir=$PWD | fi ` Then it can explicitly look for $base_dir/.gitmodules, and for path = debian in there, to see if it is a module (alternately, simply look for $base_dir/debian/.git, which is less clever, but still works). Then it can just cd $base_dir/debian/, and gather more things to add. The problem probably is keeping track of a second set of commits; we need not just the build branch, we now need a ./debian submodule branch, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499963: [Calendarserver-maintainers] Bug#499963: calendarserver: caldavd fails to authenticate and autocreate principal when running with NssDirectoryService
On Wed, Oct 01, 2008 at 04:24:58PM -0700, Ben Poliakoff wrote: Alright I see what's going on. The NssDirectoryService is required by the DirectoryService class to support three methods: recordTypes() listRecords() recordWithShortName() Thanks for debugging this! Now that we knew that the xml service works I was about to let you add debug code that prints out the users found in listRecords but you found out yourself already. My server is configured to use files and LDAP for NSS calls. We have several thousand users in our LDAP directory and implement the default limit of 500 search results. As a result 'getent passwd' returns a subset of all valid accounts (not including the 'benp' account). Yes, the nss service is basically meant for smaller installations as a quick means of not having double account maintenance, it's far to slow for that many users (calendarsever itself will have problems with this itlself - at least in 1.2). This can be used if only a few users need a calendar: add an hasCalendar attribute to every PosixAccount and filter in libnss-ldap by using the nss_base_passwd base?scope?hasCalendar=True (in case you use a separate machine for the calendar server). I'm doing something similar to cut down on the number of groups being looked at. I think I might take a stab at writing a generic LDAPDirectoryService using your NssDirectoryService as an example. There's already code in the calendarserver.org bugtracker for that. It might be worth testing it out and reporting back. So in the end this isn't really a bug with NssDirectoryService; but it's probably worth noting in the documentation that NssDirectoryService will only work properly within an environment where *all* valid users can be retrieved via the equivalent of 'getent passwd'. I'll add that, thanks. Sorry for the trouble, and thanks for your time! Thanks for debugging this. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500893: please build against heimdal to allow for pkinit
Package: libpam-krb5 Version: 3.11-2 Severity: wishlist Hi, MIT doesn't support pkinit atm, so I'd be great if you could build against heimdal or are there other features we'd miss out on then? Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#498920: fix for lenny
Hi Eric, can we get this bug fixed for lenny? It's a trivial typo and prevents a whole class of token to work. I can NMU if this helps. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500263: [Pkg-libvirt-maintainers] Bug#500263: [virt-manager] Impossible to attach a physical optical drive to an already created VM
On Fri, Sep 26, 2008 at 07:41:53PM +0200, Laurent Léonard wrote: --- Please enter the report below this line. --- When a VM is already created it is impossible to attach a physical optical drive to it, the drop list in the CD-ROM hardware details section is empty, even if I eject and reinsert the optical media. The drop list in the VM creation wizard is working without any problem. Works here (although the dialog keeps on getting new entries of the same name). Do you have hal installed? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500260: [Pkg-libvirt-maintainers] Bug#500260: [virt-manager] The VM creation wizard refers to the inexistant path /var/lib/libvirt/images
On Fri, Sep 26, 2008 at 07:26:04PM +0200, Laurent Léonard wrote: In the VM creation wizard the path /var/lib/libvirt/images is used as default path for the disk image but this path doesn't exist. Maybe this happened before you upgraded libvirt to 0.4.5? I'm having an autocreated storagepool there. See Edit-Connection Details-Storage Pools. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499720: [Pkg-libvirt-maintainers] Bug#499720: same with 0.6.0?
Hi Laurent, On Fri, Sep 26, 2008 at 06:49:56PM +0200, Laurent Léonard wrote: The behaviour seems to be similar in 0.5.4 and 0.6.0. It works here with libvirt 0.4.6 and virt-manager 0.6.0 - I do have hal installed though. Could you check with the above versions? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500263: [Pkg-libvirt-maintainers] Bug#500263: [virt-manager] Impossible to attach a physical optical drive to an already created VM
On Thu, Oct 02, 2008 at 03:27:08PM +0200, Guido Günther wrote: On Fri, Sep 26, 2008 at 07:41:53PM +0200, Laurent Léonard wrote: --- Please enter the report below this line. --- When a VM is already created it is impossible to attach a physical optical drive to it, the drop list in the CD-ROM hardware details section is empty, even if I eject and reinsert the optical media. The drop list in the VM creation wizard is working without any problem. Works here (although the dialog keeps on getting new entries of the same name). Do you have hal installed? Scratch that. The bug actually shows when you insert the CD before opening the dialog. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500263: [Pkg-libvirt-maintainers] Bug#500263: [virt-manager] Impossible to attach a physical optical drive to an already created VM
On Thu, Oct 02, 2008 at 04:35:54PM +0200, Laurent Léonard wrote: Le jeudi 02 octobre 2008 à 15:48, Guido Günther a écrit : On Thu, Oct 02, 2008 at 03:27:08PM +0200, Guido Günther wrote: On Fri, Sep 26, 2008 at 07:41:53PM +0200, Laurent Léonard wrote: --- Please enter the report below this line. --- When a VM is already created it is impossible to attach a physical optical drive to it, the drop list in the CD-ROM hardware details section is empty, even if I eject and reinsert the optical media. The drop list in the VM creation wizard is working without any problem. Works here (although the dialog keeps on getting new entries of the same name). Do you have hal installed? Scratch that. The bug actually shows when you insert the CD before opening the dialog. -- Guido By default, the ISO source is selected in the dialog, if I select physical media source there is nothing in the drop list and if I select again ISO source the drop list change its size. If I select again physical media source I can see 4 entries in the drop list with No media present (/dev/scd0). At this time if I insert a physical media in the drive each entries gets the volume name of the physical media. If I select again the ISO source the drop list change its size. If I select again the physical media source I get 8 entries in the drop list : 4 with only the volume name and 4 with volume name + (/dev/scd0). If I repeat the operation I get 4 entries more in the drop list. I've uploaded a fix for that and similar issues. It should work in -3. I'd be great to hear if it works for you too then. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#465498: still present in 0.6.0
Hi, 0.6.0 is in experimental, could you check if the bugs is still present? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: More info
On Fri, Oct 03, 2008 at 10:47:55AM +0200, Rik Theys wrote: Hi, It seems the -n option is only applied to LOGICAL VOLUME products, according to a strings on /sbin/multipath. When I add a device section for 'LOGICAL VOLUME' and specify a getuid_callout, it uses that one. I would expect multipath to use the defaults unless explicitely configured in /etc/multipath.conf There's a set of default for different classes of devices. I'll remove the -n but I don't think cciss is well supported by the multipath version in lenny. Current upstream git has it's own patch checker for cciss (which we can't backport easily since the whole structure of the path checkers changed). This bug can be downgraded because it will work with a config change, but should the -n option not be removed from the built-in defaults for LOGICAL VOLUME? Does it multipathing work cciss if you remove -n? Do the patchs get detected correctly. Could you attache the output of multipath -l? -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: incorrect regular expression?
On Fri, Oct 03, 2008 at 10:56:16AM +0200, Rik Theys wrote: devnode ^cciss!c[0-9]d[0-9]*(p[0-9]*)? instead of devnode ^cciss!c[0-9]d[0-9]*[p[0-9]*] The partition on the cciss disks should be optional? I'd say you don't multipath partitions at all but only the disks but then I don't know how multipath on cciss is supposed to work at all? What machine is this? What kind of storage? I'm inclined to blacklist cciss altogether for lenny since but please report back first how well it works for you. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: multipath-tools: Multipath-tools uses invalid scsi_id option, breaks multipath
severity 500991 normal retitle 500991 cciss not well supported thanks Hi Rik, On Fri, Oct 03, 2008 at 10:32:44AM +0200, Rik Theys wrote: [..snip..] cciss!c0d0: getuid = /lib/udev/scsi_id -n -g -u -d /dev/%n (controller setting) /lib/udev/scsi_id: invalid option -- n error calling out /lib/udev/scsi_id -n -g -u -d /dev/cciss!c0d0 The -n option is not supported by /lib/udev/scsi_id. scsi_id shifted quiet some options around -s got dropped (which is fixed) but probably -n too. Adding the getuid_callout option to the defaults section of /etc/multipath.conf without the -n option does not seem to help. It seems the section is ignored? No it isn't. Something like: defaults { getuid_callout /lib/udev/scsi_id -g -u -d /dev/%n } should work. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#497686: version in lenny supports -o
severity 497686 normal tag 497686 pending thanks Hi, the version in lenny supports -o so there's little point in adding a versioned dependency but maybe we should do so for the sake of etch updates. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: incorrect regular expression?
On Fri, Oct 03, 2008 at 11:50:01AM +0200, Rik Theys wrote: The machine is an HP DL380G5 with an HP smart array P400i raid controller. I was just testing the multipath support because I will need it later on. The RAID disk is not available using multiple paths, but multipath should still list it as one path (IMHO). With the default blacklist info and the patch checker configured for LOGICAL VOLUME devices, the callout still fails, because it checks for /dev/cciss!c0d0 instead of /dev/cciss/c0d0: You'll later need multipath support with cciss or with something else? As I wrote, for proper cciss support current multipath git is needed. This is AFAIK the only hardware that is supported in upstream git but not in any released version. All the rest (FC, iSCSI) should work well. I've just uploaded a new version of multipath-tools that blacklists cciss. I'd be great if you could test if your cciss disks get correctly blacklisted. In case you desperately need cciss, please file a wishlist bug to package up current git. I will try to do it soon then - help appreciated of course. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501017: ITP: oVirt -- Web based application for managing virtual machines
Package: wnpp Severity: wishlist Owner: Guido Günther [EMAIL PROTECTED] * Package name: oVirt Version : 0.93 * URL : http://ovirt.org * License : GPL Programming Lang: Ruby Description : Web based application for managing virtual machines oVirt consists of two major parts: the appliance for managing the virtual machines and the nodes where those VMs run on. Both are currently being shipped as Fedora 9 based os images. Building a Debian based node as well as packaging the web application will be quiet some work, so help is certainly welcome. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: incorrect regular expression?
On Fri, Oct 03, 2008 at 01:12:16PM +0200, Rik Theys wrote: I won't need cciss support in the near future as I will be connecting FC disks to the server. I will test the new version of multipath-tools O.k., this should work without problems then. (with cciss blacklisted), if you give me the location of the .deb. Or will it enter unstable? It will enter unstable. I've uploaded for ppc so i386 and amd64 might not be available before tomorrow. If I still have a spare server with an cciss card when the git version is packaged, I will help test/debug it. Great. Thanks for your feedback, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499243: [git-svn]: git-svn info fails
On Fri, Oct 03, 2008 at 02:36:23PM +, Gerrit Pape wrote: forcemerge 490400 499243 quit On Wed, Sep 17, 2008 at 11:41:00AM +0200, Guido G?nther wrote: Hi, git-svn fails like: $ git-svn info Use of uninitialized value within @_ in localtime at /usr/bin/git-svn line 4277. No such file or directory at /usr/bin/git-svn line 897 on a svn clone created with: git-svn clone https://svn.calendarserver.org/repository/calendarserver/PyKerberos/trunk/ Hi Guido, please see #490400, this has been fixed in 1.6.0-1. Confirmed, works here with 1.6.0-1. Thanks, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501155: [Pkg-libvirt-maintainers] Bug#501155: Please add script for automatic restart of kvm-based instances
Hi Andreas, On Sat, Oct 04, 2008 at 11:13:53PM +0200, Andreas Barth wrote: this scripts works for me, I added it with update-rc.d libvirt-suspendonreboot defaults 21 19 Nothing too great, but it makes sure kvm-domains survive a reboot of the host machine (and it doesn't check for possible symlink attacks in /var/lib/libvirt/autosuspend, so make sure that directory is root-only). thanks a lot - this is an area where libvirt currently isn't doing too well! I wonder if we should hook this directly into /etc/init.d/libvirt-bin in the long run. We could even offer to either shutdown or suspend the vms via (/etc/default/libvirt-bin) shutdown_action=shutdown|suspend then. I'm currently discussing with upstream what would be the best thing to do - e.g. in case of a restart we should simply keep the VMs running. Below is some minor nitpicking: [..snip..] for domain in ${suspenddir}/*dump; do if [ -f $domain ]; then domain=$(basename $domain .dump) echo resuming $domain ... virsh restore ${suspenddir}/${domain}.dump rm ${suspenddir}/${domain}.dump Do we want to keep the dump file around in case the restore failed? Should we add some lsb logging in case the restore failed? fi done ;; stop) for domain in /etc/libvirt/qemu/*xml; do domain=$(basename $domain .xml) state=$(virsh domstate $domain) if [ $state == running ]; then We probably want to pass the URI here like virsh -c qemu:///system so we don't clash with people using xen and we also don't handle transient domains (virsh create). -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501214: contact.*.commands mustn't trigger when going from not there - UNKNOWN
Package: munin Version: 1.2.6-6 Severity: normal Hi, the check for values above the warning or critical threshold levels seems to have changed from 1.2.5. When e.g. the »df« plugin detects a new mount point (say a plugged in USB stick) it triggers the »contact.*.command«'s with: example.com :: foo.example.com:: Disk usage (in percent) UNKNOWNs: /media/usb0 is unknown just to report 5 minutes later that the free disk space for that mount points is OK. The same happens when completely enabling new plugins - the state for all values is UNKNOWN, then OK. If munin doesn't find any values if shouldn't trigger with UNKNOWN but instead wait until it has a defined status - otherwise it generates lots of unnecessary UNKNOWN-OK flip flop mails, SMS, nagios events or whatever the contact.*.commands are wired to. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#499405: NMU uploaded to delayed 3-days
Hi Brian, due to lack of response and since we'd want to release lenny with working pkinit I've prepared an NMU and uploaded to delayed 3-days (so plenty of time to cancel it). Attached is the NMUdiff. Hope that's o.k., -- Guido diff --git a/debian/changelog b/debian/changelog index 3c59b7f..423c9bd 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,10 @@ +heimdal (1.2.dfsg.1-2.1) unstable; urgency=low + + * Non-maintainer upload. + * fix segfaults when using pkinit with wrong PIN. Closes: #499405 + + -- Guido Günther [EMAIL PROTECTED] Sun, 05 Oct 2008 15:12:05 +0200 + heimdal (1.2.dfsg.1-2) unstable; urgency=low * Fix library version symbols. Again. Closes: #492427. diff --git a/debian/patches/040_pkinit b/debian/patches/040_pkinit new file mode 100644 index 000..f82af2c --- /dev/null +++ b/debian/patches/040_pkinit @@ -0,0 +1,47 @@ +From 483babe37d5f85175260d21d8b603ae439f0fcec Mon Sep 17 00:00:00 2001 +From: =?utf-8?q?Guido=20G=C3=BCnther?= [EMAIL PROTECTED] +Date: Sun, 5 Oct 2008 15:06:00 +0200 +Subject: [PATCH] fix segfaults when using pkinit with wrong PIN + +upstream SVN revisions 23858, 23859 +Closes: #499405 +--- + lib/hx509/ks_p11.c |8 +++- + 1 files changed, 3 insertions(+), 5 deletions(-) + +diff --git a/lib/hx509/ks_p11.c b/lib/hx509/ks_p11.c +index bf46e66..659432d 100644 +--- a/lib/hx509/ks_p11.c b/lib/hx509/ks_p11.c +@@ -419,8 +419,6 @@ p11_get_session(hx509_context context, + char pin[20]; + char *str; + +- slot-flags |= P11_LOGIN_DONE; +- + if (slot-pin == NULL) { + + memset(prompt, 0, sizeof(prompt)); +@@ -454,16 +452,16 @@ p11_get_session(hx509_context context, + Failed to login on slot id %d + with error: 0x%08x, + (int)slot-id, ret); +- p11_put_session(p, slot, slot-session); + return HX509_PKCS11_LOGIN; +- } ++ } else ++ slot-flags |= P11_LOGIN_DONE; ++ + if (slot-pin == NULL) { + slot-pin = strdup(pin); + if (slot-pin == NULL) { + if (context) + hx509_set_error_string(context, 0, ENOMEM, + out of memory); +- p11_put_session(p, slot, slot-session); + return ENOMEM; + } + } +-- +1.6.0.1 + diff --git a/debian/patches/series b/debian/patches/series index 3f826e3..ee94d0d 100644 --- a/debian/patches/series +++ b/debian/patches/series @@ -10,3 +10,4 @@ 026_posix_max 027_rsh_use_ktelnet 030_autotools +040_pkinit
Bug#501243: Processed: Re: Bug#501243: dpkg-source: should ignore .gbp.conf files
On Mon, Oct 06, 2008 at 07:09:24PM +0200, Guido Günther wrote: forcemerge 445344 501243 thanks Hi Luca, You can put configuration into .git/gbp.conf -- Guido The main reason for allowing a $REPO/.gbp.conf is to store gbp information in the repository (so you can store per branch configuration and share the configuration with other team members). As far as I know this isn't possible with `git config` since only ~/.gitconfig and .git/config are being looked at. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#500991: incorrect regular expression?
Hi, On Mon, Oct 06, 2008 at 10:15:19AM +0200, Rik Theys wrote: [..snip..] I've tested the version from unstable (it also upgraded kpartx) and it now blacklists the cciss devices without a config file. So it works :-). Thank you very much for testing. I just mailed the release team so they hopefully allow the package into lenny. Thanks! -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#501692: [Pkg-libvirt-maintainers] Bug#501692: libvirt: Trouble with qemu 0.8.2 (etch)
severity 501692 normal thanks On Thu, Oct 09, 2008 at 06:15:24PM +0200, Marc Fargas wrote: Running libvirtd with qemu 0.8.2 will lead to a nice error message: Cannot determine QEMU argv syntax /usr/bin/qemu After talking in #virt it seems that libvirt runs: qemu -help | head -3 To look for the qemu version installed, but qemu 0.8.2 doesn't understand this argument hence the meaningfull error. qemu -h That works. Apparently it has something to do with Ubuntu and KVM, as per IRC, the command to check qemu was qemu until Ubuntu did something bad and now it is qemu -help which breaks with qemu 0.8.2 -h seems to be supported by 0.9.X and 0.8.2 so we could use this instead. Someone would have to do some digging if there's a difference between -h and -help in different versions of qemu. Supporting etch isn't very high on my agenda, so help would be welcome. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489850: multipath-tools-boot: fails to boot from a slow devices
On Fri, Jul 25, 2008 at 04:03:01AM +0200, Janusz Dziemidowicz wrote: [..snip..] Unfortunately it does not work. /dev/mapper/wwid entry is not created. When I launch this manually from within initrd shell with -v2 I get: : failed to get sysfs information Hmmm.weird - seems there's s.th. in multipath's sysfs code broken - I'll check that once I have access to a MP system myself again. What about: ACTION==add, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 %N This should hopefully do the trick. As a last resort: ACTION==add, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 /dev/%k should work too. Could you try that? and this even does not work on the working system, I see the same error. Quick strace shows: stat(/sys/block/sda, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0 lstat(/sys/block/^Z, 0x7fff65f44df0) = -1 ENOENT (No such file or directory) fstat(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 0), ...}) = 0 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2af744b82000 write(1, \32: failed to get sysfs informati..., 35) = 35 It's a bit too late for me to dig into this today. No hurries. Thanks a lot for testing! -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492254: [Pkg-libvirt-maintainers] Bug#492254: specifying virtio disks doesn't work
On Fri, Jul 25, 2008 at 01:20:58PM +0200, Guido Trotter wrote: Yes, this morning it works, yesterday it wasn't. The only thing that changed is a reboot, and yesterday I had just upgraded (to try the new version), is it possible that libvirtd didn't get restarted on upgrade and was still the old version? If it's this, there's an upgrade bug, otherwise I'm not sure how to explain the issue!! See libvirt-bin postinst: we can't restart the daemon since this would kill off running kvm/qemus. Now that it works for you I'm closing this report. If you consider the not restarting of libvirtd a bad thing (it seems to cause problems when you don't know about it), please file a bug. Cheers and thanks for retesting! -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492340: don't comment smartd.conf
On Fri, Jul 25, 2008 at 08:44:19PM +0800, [EMAIL PROTECTED] wrote: I have an idea. Instead of adding or adjusting the commented out examples in /etc/smartd.conf, which causes the user to have to deal with: Configuration file `/etc/smartd.conf' == Modified (by you or by a script) since installation. == Package distributor has shipped an updated version. What would you like to do about it ? Your options are: Why not just keep all the comments on the man page, that way the user would not need to repair the file upon upgrade. I agree here. A smartd.conf like: # The file gives a list of devices to monitor using smartd, with one # device per line. Text after a hash (#) is ignored, and you may use # spaces and tabs for white space. You may use '\' to continue lines. # You can usually identify which hard disks are on your system by # looking in /proc/ide and in /proc/scsi. # The word DEVICESCAN will cause any remaining lines in this # configuration file to be ignored: it tells smartd to scan for all # ATA and SCSI devices. DEVICESCAN -m root -M exec /usr/share/smartmontools/smartd-runner # For many more detailes examples please see: # /usr/share/doc/smartmontools/examples/smartd.conf would ease upgrades a lot. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#128059: closed by Charles Fry [EMAIL PROTECTED] (upstream won't fix)
On Fri, Jul 25, 2008 at 12:51:07PM +, Debian Bug Tracking System wrote: As already indicated upstream is not willing to fix this. As such, I don't see much point in keeping the bug open. Feel free to reopen if you strongly disagree. Shouldn't we at least leave this as wontfix then? Maybe if enough folks complain upstream will have another look? -- GUido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#492340: don't comment smartd.conf
On Fri, Jul 25, 2008 at 02:51:42PM -0500, Bruce Allen wrote: Sorry Guido, I just saw this. I suggest that you just make a custom smartd.conf for Debian. THanks for the comments! It's nicer to have upstream's blessing. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#434567: ...or at least warn that there are unsynced folders?
Hi, It would be nice if offlineimap could take new local folders and push them to the IMAP server, ideally controlled by an option. I wonder if it isn't actually a bug that offlineimap ignores locally created folders silently? Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489850: multipath-tools-boot: fails to boot from a slow devices
On Mon, Jul 28, 2008 at 09:39:25PM +0200, Janusz Dziemidowicz wrote: 2008/7/25 Guido Günther [EMAIL PROTECTED]: On Fri, Jul 25, 2008 at 04:03:01AM +0200, Janusz Dziemidowicz wrote: [..snip..] Unfortunately it does not work. /dev/mapper/wwid entry is not created. When I launch this manually from within initrd shell with -v2 I get: : failed to get sysfs information Hmmm.weird - seems there's s.th. in multipath's sysfs code broken - I'll check that once I have access to a MP system myself again. What about: Maybe those would help: http://git.kernel.org/gitweb.cgi?p=linux/storage/multipath-tools/.git;a=commit;h=88f88d11ecee7d753e9b43d67001d58c56789eb1 http://git.kernel.org/gitweb.cgi?p=linux/storage/multipath-tools/.git;a=commit;h=3aaecad682efef27ce6196c6d29db4cb93c9f2fb Didn't check it myself, will try next time I have some MP server for tests. Yes, we should include these (intended to do so since quiet some time) but not before the lenny release since this needs lots of testing. So help with testing these would (of course) be welcome again. This should hopefully do the trick. As a last resort: ACTION==add, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 /dev/%k should work too. Could you try that? This works (although I see that the cleanest solution would be to use major:minor). Thanks a lot for testing! I'll use this for lenny then. It's not as nice as major:minor but the advantage is rather academic - having stable multipath in lenny is more important. I just noticed that We also need to care about change events: ACTION==add|change, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 /dev/%k If I understand correctly, you want to get rid of /usr/share/initramfs-tools/scripts/local-top/multipath completely. I've tried to do that (with the /dev/%k trick) and it isn't so simple. There is a problem with loading dm-multipath module, it is not loaded automatically when needed, so /sbin/multipath fails if it was not loaded before udev event handlers from RUN were called. One solution was to add dm-multipath to /etc/initramfs-tools/modules, but I don't think it is the nicest way. I meant /etc/init.d/multipath-tools-boot (doing the module loading in a udev rule too). We can strip down /usr/share/initramfs-tools/scripts/local-top/multipath then too. But I'd say we go for a minimal fix for lenny and delay these things for afterwards. Thanks again for you help, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489850: multipath-tools-boot: fails to boot from a slow devices
Hi Janusz, On Tue, Jul 29, 2008 at 04:43:02AM -0400, Guido Günther wrote: Thanks a lot for testing! I'll use this for lenny then. It's not as nice as major:minor but the advantage is rather academic - having stable multipath in lenny is more important. I just noticed that We also need to care about change events: ACTION==add|change, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 /dev/%k This is what I intend to commit (since this isn't only useful in the initramfs this way we also have it in the acutal root fs) - just in case you feel like testing a bit more (otherwise I'll git it a whirl tomorrow): This helps slow devices when either /etc/init.d/multipath-tools-boot or the initramfs script are being run although the devices are not ready yet. Thanks: Janusz Dziemidowicz for his suggestions and testing --- debian/multipath.udev |5 + 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/debian/multipath.udev b/debian/multipath.udev index 9d3579f..a689c67 100644 --- a/debian/multipath.udev +++ b/debian/multipath.udev @@ -5,3 +5,8 @@ # socket for uevents RUN+=socket:/org/kernel/dm/multipath_event + +# Coalesce multipath devices before multipathd is running (initramfs, early +# boot) +ACTION==add|change, SUBSYSTEM==block, RUN+=/sbin/multipath -v0 /dev/%k + -- 1.5.6.3 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
Package: udev Version: 0.124-5 Severity: critical Hi, up to version 114 scsi_id accepted: ./scsi_id -u -g -s /block/sdg 1494554010e9b000e Whereas 124 doesn't: /lib/udev/scsi_id -u -g -s /block/sdg /lib/udev/scsi_id: invalid option -- s it needs: /lib/udev/scsi_id -u -g --device=/dev/sdg 1494554010e7c000e This breaks multipath-tools since it calls scsi_id internally. We could probably fix this up for lenny but it would nevertheless break user's udev rules. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493078: dmsetup export uses different syntax than multipath upstream
Package: dmsetup Version: 2:1.02.27-3 Severity: wishlist Hi, the kpartx udev rules have: ENV{DM_TABLE_STATE}!=LIVE, GOTO=kpartx_end which works with the export patch SuSE, Ubuntu and the other distros use. Debian's very own dmesetup patch (in contrast to the originally proposed patch) exports: DM_TABLE_LIVE=1 instead of DM_TABLE_STATE=LIVE This can be fixed easily within multipath-tools but I wonder why we have to divert from other distros in these details? Are there any technical reasons? Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Thu, Jul 31, 2008 at 10:42:30AM +0200, Marco d'Itri wrote: On Jul 31, Guido Günther [EMAIL PROTECTED] wrote: This breaks multipath-tools since it calls scsi_id internally. We could probably fix this up for lenny but it would nevertheless break user's udev rules. Users will have to update their rules too, but I highly doubt that a relevant number of people created custom rules which call scsi_id. I doubt that patching the program to support the old syntax is worth the effort, but feel free to try. I do not consider this a bug. It breaks software currently targeted for lenny - at least multipath-tools (there might be others) so this is indeed a but. I also wonder what a relevant number would be? We didn't announce the change and didn't provide a transition phase. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Thu, Jul 31, 2008 at 11:24:25AM +0200, Marco d'Itri wrote: On Jul 31, Guido Günther [EMAIL PROTECTED] wrote: It breaks software currently targeted for lenny - at least I am not going to revert the change, so the possibile actions on my part are: - merge a lenny-only patch from you, who apparently are the only other maintainer affected - add a Conflict with the multipath-tools versions you will specify - nothing I've already sent a patch upstream for multipath-tools to switch to -d instead of -s. Unfortunately -d/--device doesn't work in e.g. udev 0.114 (currently in lenny): #./scsi_id -u -g -s /block/sdf 1494554010e91000e #./scsi_id -u -g -d /dev/sdf While 0.125 does: # /lib/udev/scsi_id -u -g -d /dev/sdf 1494554010e91000e Thank's to that we don't have an upgrade path: Updating udev first doesn't work since we'd then have the wrong multipath configuration and updating multipath first doesn't work since udev's scsi_id (in e.g. 0.114) can't cope with it's own command line options. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Thu, Jul 31, 2008 at 06:20:44PM +0300, Teodor wrote: On Thu, Jul 31, 2008 at 3:39 PM, Guido Günther [EMAIL PROTECTED] wrote: I've already sent a patch upstream for multipath-tools to switch to -d instead of -s. Unfortunately -d/--device doesn't work in e.g. udev 0.114 (currently in lenny): Thank's to that we don't have an upgrade path: Updating udev first doesn't work since we'd then have the wrong multipath configuration and updating multipath first doesn't work since udev's scsi_id (in e.g. 0.114) can't cope with it's own command line options. I think this can be fixed by depending on udev = 0.125 and hitting both together in lenny. First you'll have to ask for a freeze exception since this seems to be RC. No it can't. This allows for a udev update without updating multipath-tools. This would have worked if scsi_id in the old versions would have supported -d properly. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Thu, Jul 31, 2008 at 06:11:01PM +0200, Marco d'Itri wrote: On Jul 31, Teodor [EMAIL PROTECTED] wrote: I think this can be fixed by depending on udev = 0.125 and hitting both together in lenny. First you'll have to ask for a freeze exception since this seems to be RC. This is planned, and I expect that the freeze exception will be granted. But please let me know in advance if I am supposed to do anything, and what. In order to support the upgrade we need to: udev: * hack in enough support for -s so we can update udev without updating multipath-tools - this can probably be done by just implementing -s as -d and simply doing a: s%^/block/%^/dev/% in set_options if -s was set. I just did this but then noticed: scsi_id.c: In function ‘set_options’: scsi_id.c:465: error: duplicate case value scsi_id.c:419: error: previously used here They reused '-s' for something completely different: to specify the SG version. Now how broken can stuff be? * Deprecate -s/--devpath in NEWS.Debian once this is sorted out: multipath-tools: * move multipath-tools over to using -d (I'm currently discussing this with upstream - the fix is obvious but just to be sure) * depend on udev (= 0.125) - since older versions of udev don't support -d correctly. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Fri, Aug 01, 2008 at 03:35:06AM -0400, Guido Günther wrote: * hack in enough support for -s so we can update udev without updating multipath-tools - this can probably be done by just implementing -s as -d and simply doing a: s%^/block/%^/dev/% in set_options if -s was set. I just did this but then noticed: It turned out that the code for another usage of '-s' is there but it isn't in the options passed to getopt_long, so we can safely keep the compat code. Possible patch attached. -- Guido From 749a45a54e8c12aa2ca2a89b3768b89c3c50ee66 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Fri, 1 Aug 2008 03:16:05 -0400 Subject: [PATCH] convert sysfs paths for block devices into device names adds some backwardscompatibility for programms/udev rules that use -s/--devpath --- extras/scsi_id/scsi_id.c | 27 +-- 1 files changed, 17 insertions(+), 10 deletions(-) diff --git a/extras/scsi_id/scsi_id.c b/extras/scsi_id/scsi_id.c index 5eb95e8..badda7f 100644 --- a/extras/scsi_id/scsi_id.c +++ b/extras/scsi_id/scsi_id.c @@ -33,12 +33,12 @@ static const struct option options[] = { { device, 1, NULL, 'd' }, + { devpath, 1, NULL, 's' }, { config, 1, NULL, 'f' }, { page, 1, NULL, 'p' }, { blacklisted, 0, NULL, 'b' }, { whitelisted, 0, NULL, 'g' }, { replace-whitespace, 0, NULL, 'u' }, - { sg-version, 1, NULL, 's' }, { verbose, 0, NULL, 'v' }, { version, 0, NULL, 'V' }, { export, 0, NULL, 'x' }, @@ -46,7 +46,7 @@ static const struct option options[] = { {} }; -static const char short_options[] = d:f:ghip:uvVx; +static const char short_options[] = d:f:ghip:uvVxs:; static const char dev_short_options[] = bgp:; static int all_good; @@ -374,6 +374,16 @@ static int get_file_options(const char *vendor, const char *model, return retval; } +static void convert_sysfs_path(const char* sysfs, char *devpath) +{ + if (!strncmp(optarg, /block/, 7)) { + snprintf(devpath, MAX_PATH_LEN, /dev/%s, (optarg[7])); +} else { + strncpy(devpath, optarg, MAX_PATH_LEN); + } + devpath[MAX_PATH_LEN-1] = '\0'; +} + static int set_options(int argc, char **argv, const char *short_opts, char *maj_min_dev) { @@ -406,6 +416,11 @@ static int set_options(int argc, char **argv, const char *short_opts, maj_min_dev[MAX_PATH_LEN-1] = '\0'; break; + case 's': + dev_specified = 1; + convert_sysfs_path(optarg, maj_min_dev); + break; + case 'e': use_stderr = 1; break; @@ -447,14 +462,6 @@ static int set_options(int argc, char **argv, const char *short_opts, } break; - case 's': - sg_version = atoi(optarg); - if (sg_version 3 || sg_version 4) { -err(Unknown SG version '%s'\n, optarg); -return -1; - } - break; - case 'u': reformat_serial = 1; break; -- 1.5.6.3
Bug#493216: labeled mounts break with recent udev
Package: udev Version: 0.125 Severity: grave Hi, the removal of the /dev/.static/dev/ hack in 0.124-1 breaks labeled mounts (via LABEL=...) and might leave people with an unbootable system. Reason is that findfs and friends query /etc/blkid.tab to find the device matching the UUID. Since blkid.tab has things like /dev/.static/dev/hda7 (no idea why blkid picked that one in favour of /dev/hda7) this breaks the mount. I see three possible solutions: * put the hack back in (bad) or * put code into e2fsprogs that updates /etc/blkid.tag and let udev conflict on older than that versions of e2fsprogs. or * document that we stop supporting mounts by LABEL=... and people should use /dev/disk/by-label/ Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493216: labeled mounts break with recent udev
Hi Marco, On Fri, Aug 01, 2008 at 02:13:31PM +0200, Marco d'Itri wrote: [..snip..] * put code into e2fsprogs that updates /etc/blkid.tag and let udev conflict on older than that versions of e2fsprogs. This is a bug which needs to be fixed, we generally do not conflict with packages just because they are buggy. I meant: _once_ this is fixed udev should conflict on older than fixed e2fsprogs since otherwise people might update udev and be left with an unbootable system. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493214: git-buildpackage: git-import-dscs do not support epoch
Hi Frederic, On Fri, Aug 01, 2008 at 01:23:24PM +0200, Frederic-Emmanuel PICCA wrote: I tryed to import all the current *.dsc of the lisaac package from the debian repository. In fact version 0.84 and 1:0.13.1 But the order done by import-dscs was in a wrong order. It first import the 0.13.1 version before the 0.84. This is wrong. It is. A fixed version is here: http://honk.sigxcpu.org/projects/git-buildpackage/snapshots/ updated sources are in git. Sorry for the inconvenience, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493216: labeled mounts break with recent udev
On Fri, Aug 01, 2008 at 08:57:43AM -0400, Theodore Tso wrote: [..snip..] All I can think of is that at some point someone accidentally typed the command blkid /dev/.static/dev/sda7 while running as root, and this errant got stuck in your /etc/blkid.tab file. I see (and will fix) the bug which causes blkid to not fix this problem automatically, but this is something which I've never seen show up in peoples /etc/blkid.tab files in the normal course of events. This is not the case. I never used the blkid command before today. After installing the laptop (that was probably sarge back then) - I switch to labeled mounts by using tune2fs and changing the fstab to LABEL=... No blkid was ever involved here. If you run the command the command blkid -g as root the errant /dev/.static entry will go away. cp /dev/null /etc/blkid.tab will also make the problem go away, and I haven't been able to find a way to make blkid put those names into /etc/blkid.tab unless forced via explicit human intervention. Actually blkid -g is what I did to get this fixed but we should automate this in a way. Maybe by checking of blkid.tab has .static/ - if so run blkid -g or something similar. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: scsi_id changed commandline syntax (breaks multipath)
On Fri, Aug 01, 2008 at 02:22:09PM +0200, Marco d'Itri wrote: On Aug 01, Guido Günther [EMAIL PROTECTED] wrote: It turned out that the code for another usage of '-s' is there but it isn't in the options passed to getopt_long, so we can safely keep the compat code. Possible patch attached. WTF? You cannot just remove the code, -s is not used but --sg-version is. You need to change the index from 's' to something else (even non-alphanumeric). You're of course right - thought this got dropped. New version attached, I use '*' as character so we won't clash with any future short options. -- Guido From 221d4561c3c94d9dc105658dafe4aaed279f7d55 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Fri, 1 Aug 2008 03:16:05 -0400 Subject: [PATCH] convert sysfs paths for block devices into device names adds some backwardscompatibility for programms/udev rules that use -s/--devpath --- extras/scsi_id/scsi_id.c | 22 +++--- 1 files changed, 19 insertions(+), 3 deletions(-) diff --git a/extras/scsi_id/scsi_id.c b/extras/scsi_id/scsi_id.c index 5eb95e8..1c03009 100644 --- a/extras/scsi_id/scsi_id.c +++ b/extras/scsi_id/scsi_id.c @@ -33,12 +33,13 @@ static const struct option options[] = { { device, 1, NULL, 'd' }, + { devpath, 1, NULL, 's' }, { config, 1, NULL, 'f' }, { page, 1, NULL, 'p' }, { blacklisted, 0, NULL, 'b' }, { whitelisted, 0, NULL, 'g' }, { replace-whitespace, 0, NULL, 'u' }, - { sg-version, 1, NULL, 's' }, + { sg-version, 1, NULL, '*' }, { verbose, 0, NULL, 'v' }, { version, 0, NULL, 'V' }, { export, 0, NULL, 'x' }, @@ -46,7 +47,7 @@ static const struct option options[] = { {} }; -static const char short_options[] = d:f:ghip:uvVx; +static const char short_options[] = d:f:ghip:uvVxs:; static const char dev_short_options[] = bgp:; static int all_good; @@ -374,6 +375,16 @@ static int get_file_options(const char *vendor, const char *model, return retval; } +static void convert_sysfs_path(const char* sysfs, char *devpath) +{ + if (!strncmp(optarg, /block/, 7)) { + snprintf(devpath, MAX_PATH_LEN, /dev/%s, (optarg[7])); + } else { + strncpy(devpath, optarg, MAX_PATH_LEN); + } + devpath[MAX_PATH_LEN-1] = '\0'; +} + static int set_options(int argc, char **argv, const char *short_opts, char *maj_min_dev) { @@ -406,6 +417,11 @@ static int set_options(int argc, char **argv, const char *short_opts, maj_min_dev[MAX_PATH_LEN-1] = '\0'; break; + case 's': + dev_specified = 1; + convert_sysfs_path(optarg, maj_min_dev); + break; + case 'e': use_stderr = 1; break; @@ -447,7 +463,7 @@ static int set_options(int argc, char **argv, const char *short_opts, } break; - case 's': + case '*': sg_version = atoi(optarg); if (sg_version 3 || sg_version 4) { err(Unknown SG version '%s'\n, optarg); -- 1.5.6.3
Bug#412010: pxe images would be nice
Hi, can't the boot images from the etherboot packages be used? Not having this in qemu is a bit annoying since one can't PXE boot from libvirt/virt-manager. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493592: pygrub fails on lenny's ext3
Package: xen-utils-3.2-1 Version: 3.2.0-3~bpo4+2 Severity: normal Hi, pygrub has problems parsing block devices partitioned with Lenny Beta 2 and ext3 on paravirt guests: # /usr/lib/xen-3.2-1/bin/pygrub /var/lib/xen/images/test2.img Traceback (most recent call last): File /usr/lib/xen-3.2-1/bin/pygrub, line 653, in ? chosencfg = run_grub(file, entry, fs) File /usr/lib/xen-3.2-1/bin/pygrub, line 510, in run_grub g = Grub(file, fs) File /usr/lib/xen-3.2-1/bin/pygrub, line 203, in __init__ self.read_config(file, fs) File /usr/lib/xen-3.2-1/bin/pygrub, line 393, in read_config raise RuntimeError, couldn't find bootloader config file in the image provided. RuntimeError: couldn't find bootloader config file in the image provided. Creating the filesystem with ext2 works and pygrub is able to find the domU kernels. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493758: [Pkg-libvirt-maintainers] Bug#493758: Poor virt-image errors with faulty XML file
On Mon, Aug 04, 2008 at 08:42:50PM +0200, Filip Van Raemdonck wrote: Package: virtinst Version: 0.300.3-4 Severity: minor When specifying a file that isn't actually an XML file, virt-image spits out an unfriendly python backtrace instead of a proper error message: Yes. Error handling in virtinst isn't one of the greatest. I'll recheck with current hg once I touch that stuff again. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#440675: Patches updated for 1.8.8.git.2008.03.24-8
Hi Otavio, Just to make merging easier: This is the git tree as mentioned on debian-boot. http://git.debian.org/git/users/agx/parted1.8.git I additionally attach the patch to this mail so we have it in the BTS. -- Guido From bd2038e2f1efd0528e6a69aaf7fbedabcd531407 Mon Sep 17 00:00:00 2001 From: Guido Guenther [EMAIL PROTECTED] Date: Thu, 7 Aug 2008 12:11:36 +0200 Subject: [PATCH] linux device-mapper map type detection detect the type of the device map and add it to the displayed type information. This is also included in upstream parted. --- include/parted/linux.h |1 + libparted/arch/linux.c | 57 +++- 2 files changed, 57 insertions(+), 1 deletions(-) diff --git a/include/parted/linux.h b/include/parted/linux.h index 05a2f0e..1b0d8f8 100644 --- a/include/parted/linux.h +++ b/include/parted/linux.h @@ -32,6 +32,7 @@ typedef struct _LinuxSpecific LinuxSpecific; struct _LinuxSpecific { int fd; + char* dmtype; /** device map target type */ #if defined(__s390__) || defined(__s390x__) unsigned int real_sector_size; /* IBM internal dasd structure (i guess ;), required. */ diff --git a/libparted/arch/linux.c b/libparted/arch/linux.c index 9876956..14baa69 100644 --- a/libparted/arch/linux.c +++ b/libparted/arch/linux.c @@ -301,6 +301,44 @@ _is_sx8_major (int major) #ifdef ENABLE_DEVICE_MAPPER static int +_dm_maptype (PedDevice* dev) +{ +LinuxSpecific* arch_specific = LINUX_SPECIFIC (dev); +struct dm_task *dmt; +void *next = NULL; +uint64_t start, length; +char *target_type = NULL; +char *params; +int r = -1; +const char* dev_dir = getenv (DM_DEV_DIR); + +if (dev_dir *dev_dir !dm_set_dev_dir(dev_dir)) +return r; + +if (!(dmt = dm_task_create(DM_DEVICE_TABLE))) +return r; + +if (!dm_task_set_name(dmt, dev-path)) +goto bad; + +dm_task_no_open_count(dmt); + +if (!dm_task_run(dmt)) +goto bad; + +next = dm_get_next_target(dmt, next, start, length, + target_type, params); + +arch_specific-dmtype = strdup(target_type); +if (arch_specific-dmtype == NULL) +goto bad; +r = 0; +bad: +dm_task_destroy(dmt); +return r; +} + +static int readFD (int fd, char **buf) { char* p; @@ -488,6 +526,12 @@ _device_probe_type (PedDevice* dev) #ifdef ENABLE_DEVICE_MAPPER } else if (_is_dm_major(dev_major)) { dev-type = PED_DEVICE_DM; +if (_dm_maptype(dev)) { +LinuxSpecific* arch_specific = LINUX_SPECIFIC (dev); + +arch_specific = LINUX_SPECIFIC (dev); +arch_specific-dmtype = strdup(unknown); + } #endif } else if (dev_major == XVD_MAJOR (dev_minor % 0x10 == 0)) { dev-type = PED_DEVICE_XVD; @@ -1105,6 +1149,8 @@ static PedDevice* linux_new (const char* path) { PedDevice* dev; +char* type; +LinuxSpecific* arch_specific; PED_ASSERT (path != NULL, return NULL); @@ -1120,6 +1166,8 @@ linux_new (const char* path) = (LinuxSpecific*) ped_malloc (sizeof (LinuxSpecific)); if (!dev-arch_specific) goto error_free_path; +arch_specific = LINUX_SPECIFIC (dev); +arch_specific-dmtype = NULL; dev-open_count = 0; dev-read_only = 0; @@ -1188,7 +1236,11 @@ linux_new (const char* path) #ifdef ENABLE_DEVICE_MAPPER case PED_DEVICE_DM: -if (!init_generic (dev, _(Linux device-mapper))) +if (arch_specific-dmtype == NULL + || asprintf(type, _(Linux device-mapper (%s)), + arch_specific-dmtype) == -1) +goto error_free_arch_specific; +if (!init_generic (dev, type)) goto error_free_arch_specific; break; #endif @@ -1227,6 +1279,9 @@ linux_destroy (PedDevice* dev) ped_free (dev-arch_specific); ped_free (dev-path); ped_free (dev-model); +#ifdef ENABLE_DEVICE_MAPPER +ped_free (((LinuxSpecific*)dev-arch_specific)-dmtype); +#endif ped_free (dev); } -- 1.5.6.3 From bb51c236d88605d114f062b5f1484be4ff56c37f Mon Sep 17 00:00:00 2001 From: =?utf-8?q?Guido=20G=C3=BCnther?= [EMAIL PROTECTED] Date: Tue, 10 Jun 2008 18:35:41 +0200 Subject: [PATCH] Return the correct partition names for multipath devices --- libparted/arch/linux.c |8 1 files changed, 8 insertions(+), 0 deletions(-) diff --git a/libparted/arch/linux.c b/libparted/arch/linux.c index 14baa69..3d84373 100644 --- a/libparted/arch/linux.c +++ b/libparted/arch/linux.c @@ -2015,6 +2015,7 @@ _device_get_part_path (PedDevice* dev, int num) int path_len =
Bug#440675: Patches updated for 1.8.8.git.2008.03.24-8
On Thu, Aug 07, 2008 at 12:19:46PM +0200, Guido Günther wrote: Hi Otavio, Just to make merging easier: This is the git tree as mentioned on debian-boot. http://git.debian.org/git/users/agx/parted1.8.git I additionally attach the patch to this mail so we have it in the BTS. -- Guido Until the RT has sorted out the current state in unstable/lenny an upload to experimental would be great. This would make sure we at least have the patches in a package _somewhere_ and I could simply pickup the udebs from there. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#490221: Any progress
Hi Sylvain, any progress on bitstring? Anything I can do to help with that? Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494288: dh_pycentral/dh_pysupport should run before dh_installinit
Package: cdbs Version: 0.4.52 Severity: normal This ensures the code for python based daemons is being processed correctly before the daemin is being started. Currently the postinst fragment produced is: # Automatically added by dh_installinit [..snip..] # End automatically added section # Automatically added by dh_pycentral [..snip..] # End automatically added section But it should be the oter way around. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494321: [xserver-xorg-video-intel] Error in I830WaitLpRing(), timeout for 2 seconds - reboot needed
On Fri, Aug 08, 2008 at 04:37:25PM +0200, Julien Cristau wrote: On Fri, Aug 8, 2008 at 16:23:20 +0200, Guido Guenther wrote: Not nice for a upcoming stable release. You're welcome to send a patch. I'd love to, if you point me in the right direction. This was not meant as criticism on your maintainership - I just wanted to point out that there are worrysome issues in the driver. Let me know if I can be of any help to debug this. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494321: [xserver-xorg-video-intel] Error in I830WaitLpRing(), timeout for 2 seconds - reboot needed
On Fri, Aug 08, 2008 at 05:25:06PM +0200, Brice Goglin wrote: Switching to XAA is usually a good workaround for this kind of problem. Option AccelMethod XAA in the Device section. Since this also helps with the ugly scrolling in firefox3 (which I checked actually helps a lot): http://ubuntuforums.org/showthread.php?t=708928page=2 should we make this the default for lenny? I am downgrading severity to important to match many other Xorg crash/lockup reports. I'd fully agree here with the XAA workaround in place. It's a little bit strange to mark these things only as important otherwise - what would be grave then? Anyway, it's your call. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494592: [Pkg-libvirt-maintainers] Bug#494592: virtinst: Unable to boot on CDROM
On Sun, Aug 10, 2008 at 10:40:35PM +0200, Daniel Dehennin wrote: Package: virtinst Version: 0.300.3-4 Severity: normal Hello, I'm trying to manage some kvm systems with libvirtd and all the virt* stuffs. The system can not boot on the debian netinstall CDROM. I found an Ubuntu bug report for that: - https://bugs.launchpad.net/ubuntu/+source/kvm/+bug/220463 Any plan to solve this issue ? Please give details and logs of your problem. Booting systems from CD works fine here. Set up several machines this way. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494592: [Pkg-libvirt-maintainers] Bug#494592: virtinst: Unable to boot on CDROM
On Sun, Aug 10, 2008 at 11:14:07PM +0200, Daniel Dehennin wrote: Le 5457 Septembre 1993, Guido Günther a envoyé: I found an Ubuntu bug report for that: - https://bugs.launchpad.net/ubuntu/+source/kvm/+bug/220463 Any plan to solve this issue ? Please give details and logs of your problem. Booting systems from CD works fine here. Set up several machines this way. I use the command: sudo virt-install --connect qemu:///system --ram 256 -n lenny -f /dev/hati/lenny -w network:default --accelerate --cdrom ~/download/debian/debian-testing-amd64-netinst.iso --os-type=linux --os-variant=debianLenny --hvm The boot on the CDROM gives me: === KVM graphic display === Booting from CD-Rom... ISOLINUX 3.63 Debian-2008-06-10 Copyright (C) 1994-2008 H. Peter Anvin vesamenu.c32 attempted DOS system call boot: === END === I'm not quite sure if it's a virt-install problem or kvm one but the command sudo /usr/bin/kvm -m 256 -smp 1 -name lenny -monitor stdio -boot d \ -drive file=/dev/hati/lenny,if=virtio,index=0 \ -cdrom /home/dad/download/debian/debian-testing-amd64-netinst.iso \ -net nic,vlan=0 -net tap,script=,vlan=0,ifname=vnet0 -serial pty \ -parallel none -k fr Works fine :-/ Looks like a kvm problem but please compare this with the kvm call libvirt is using (either with ps or by looking into libvirt's log file - it has the verbatim kvm call in it). -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494592: [Pkg-libvirt-maintainers] Bug#494592: virtinst: Unable to boot on CDROM
On Mon, Aug 11, 2008 at 12:02:49PM +0200, Daniel Dehennin wrote: [..snip..] It seems that booting on cdrom does not works when using -drive to declare cdrom :-/ So this doesn't work for you: /usr/bin/kvm -S -M pc -m 256 -smp 1 -name lenny -monitor stdio \ -drive file=/home/dad/download/debian/debian-testing-amd64-netinst.iso,if=ide,media=cdrom,index=2,boot=on \ ... but this works: /usr/bin/kvm -S -M pc -m 256 -smp 1 -name lenny -monitor stdio \ ... -boot d -cdrom /home/dad/download/debian/debian-testing-amd64-netinst.iso \ ... Changing this one line fixes the problem? Is it possible to forward the bug report to kvm or do I need to make another? We can forward the report, no problem - let's just get this as detailed as possible first. Thanks for your feedback, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494592: [Pkg-libvirt-maintainers] Bug#494592: Bug#494592: virtinst: Unable to boot on CDROM
On Mon, Aug 11, 2008 at 10:51:06AM -0300, Marco Sinhoreli wrote: Only to report, this issue isn't happen with virtinst/virt-manager and Xen. Thanks for the heads up. It doesn't happen here with KVM 63 either, it think that's broken in KVM in Lenny. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494753: git-buildpackage: git-import-dsc empty ident error
On Mon, Aug 11, 2008 at 06:02:15PM -0400, Adam C Powell IV wrote: My gpg key id is Adam C. Powell, IV [EMAIL PROTECTED], which works fine with most tools, though it trips up lintian, and apparently git-import-dsc. This is not rf822 since the , is unquoted which lets git-import-dsc barf at the moment. I'll Upload a new version soonish. Thanks for the report. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#495013: [Pkg-libvirt-maintainers] Bug#495013: virt-image sets wrong mac address
On Wed, Aug 13, 2008 at 09:38:51PM +0200, Filip Van Raemdonck wrote: $ virt-image -d -p manpage.xml | grep -e 'domain typ' -e 'mac add' domain type='qemu' mac address='00:16:3e:17:4f:dd'/ Yes this is wrong. I'll send a patch upstream to fix at least parts of the issue. I'll hope they'll fix up the rest. Thanks for the report, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#493075: closed by Marco d'Itri [EMAIL PROTECTED] (Bug#493075: fixed in udev 0.125-4)
reopen 493075 thanks On Mon, Aug 04, 2008 at 10:57:08AM +, Debian Bug Tracking System wrote: * Added patch scsi_id_compat: until lenny is released, support again scsi_id -s for the benefit of multipath-tools. Patch contributed by Guido Günther. (Closes: #493075) This is not fixed since the patch is missing from debian/patches/series. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494836: paramiko in etch official
Hi Michel, On Tue, Aug 12, 2008 at 11:19:37AM -0400, Grenier,Michel [CMC] wrote: Package: python-paramiko Version: 1.7.1 Severity: grave Justification: renders package unusable The problem is that within debian etch the paramiko package is completely UNUSABLE as it is now... Once installed, importing into python causes an error and paramiko would simply not work at all. Please provide an _example_ (that is a short python script that fails). We have several packages with reverse dependencies needing paramiko in Debian etch that also work so the package just can't be unusable. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494836: paramiko in etch official
On Fri, Aug 15, 2008 at 07:45:25AM -0400, Grenier,Michel [CMC] wrote: Hi Guido it was simply when trying to use it... so the following line was failing import paramiko Then you are probably not using the paramiko shipped in Debian Etch? Please check if you have other versions on you system (e.g. in /usr/local/lib/python*). I checked on two etch systems and it works there - and it hasto since it would otherwise break packages like xenman. Maybe something is wrong with python-support - try reinstalling the package too. Cheers, -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494836: paramiko in etch official
On Fri, Aug 15, 2008 at 07:58:12AM -0400, Grenier,Michel [CMC] wrote: version is 1.5.2 right !? Check yourself if you have the latest version in etch via: http://packages.debian.org/python-paramiko -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#494836: paramiko in etch official
On Fri, Aug 15, 2008 at 07:53:26AM -0400, Grenier,Michel [CMC] wrote: Did you at least found the typo I described !? Sure. But if it's not affecting anyting badly we won't make an update for Etch. It's fixed in Lenny anyway. -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#495330: should filter out errors from /sbin/multipath
Package: partman-multipath Version: 2 Severity: normal When the scsi_id callout fails it prints it's error message on stdout. This should be filtered out via: grep -v error calling out since it confuses mp device detection (the devices get detected properly but the underlying SCSI devices get also listed). -- Guido -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]