Re: [arch-general] kernel26-2.6.32.6-1 updates breaks Compiz?
On Fri, Jan 29, 2010 at 11:23 AM, Ng Oon-Ee ngoo...@gmail.com wrote: On Fri, Jan 29, 2010 at 10:38 AM, David C. Rankin drankina...@suddenlinkmail.com wrote: Guys, After updates today on my x86_64 box from testing, including kernel26-2.6.32.6-1, Compiz is broken. It attempts to start, but then all windows are left without any decorations and keyboard input is disabled on kde4. The first time I tried, the plasma panel died leaving me with no way to get kwin back other than finding a black spot on the desktop and logging out. Next attempt, plasma panel stayed with me and changing back and forth from kwin to compiz and back is easily done with fusion-icon. (I put a shortcut on my desktop to 'kwin --replace ' just in case) Dunno what is happening here. All was OK before updates today. pacman.log is at: www.3111skyline.com/download/Archlinux/pacman.log Looks like it is another jpg/png library issue: 20:35 alchemy:~ fusion-icon * Detected Session: kde * Searching for installed applications... * No GLX_EXT_texture_from_pixmap with direct rendering context ... present with indirect rendering, exporting: LIBGL_ALWAYS_INDIRECT=1 * Using the GTK Interface * KWin is already running * Setting window manager to Compiz ... executing: compiz --replace --sm-disable --ignore-desktop-hints --indirect-rendering ccp --indirect-rendering WARNING: Application calling GLX 1.3 function glXCreatePixmap when GLX 1.3 is not supported! This is an application bug! compiz: symbol lookup error: /usr/lib/compiz/libpng.so: undefined symbol: png_check_sig I experienced the same here at the point when the libpng rebuild started (and compiz was first updated). I almost immediately downgraded due to work requiring openoffice though, and haven't gotten round to enabling testing again. Update, I've found that disabling the 'png' plugin under the 'Image Loading' section fixes the issue, at least compiz loads up and runs like it used to. Something must have gone wrong with the libpng-related rebuild. Bug report time?
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Attila vodoo0...@sonnenkinder.org wrote: At Donnerstag, 28. Januar 2010 10:22 Joerg Schilling wrote: I don't find the most of your sugestions in man 7 capabilities. file_dac_read Permission to open any device file = cap_dac_readsearch ?? Most likely CAP_DAC_OVERRIDE sys_devices Permission to send anc SCSI command Nothing found. Most likely at least CAP_SYS_RAWIO I am nowever not sur whether this is sufficient. proc_lock_memory Lock into memory = cap_ipc_lock Looks correct. proc_priocntl Increase priority Nothing found. Most likely CAP_SYS_NICE net_privaddrAllow ports 1024, needed for RSCSI cap_net_bind_service Looks correct. Is it really such a problem to stay with chmod 4710? As long as there is no support code in Linux distros to set capabilities without making the target program suid root anyway, I see no other possibility than to stay with chown root cdrecord cdda2wav readcd chmod 4711 cdrecord cdda2wav readcd Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
On Fri, Jan 29, 2010 at 8:39 AM, Joerg Schilling joerg.schill...@fokus.fraunhofer.de wrote: As long as there is no support code in Linux distros to set capabilities without making the target program suid root anyway, Don't be afraid, Arch Linux has support for that :) BTW, congratulations and thanks for your software. I use cdrtools and it is a nice piece of very high quality software.
Re: [arch-general] kernel26-2.6.32.6-1 updates breaks Compiz?
On 01/29/2010 10:33 AM, Ng Oon-Ee wrote: On Fri, Jan 29, 2010 at 11:23 AM, Ng Oon-Eengoo...@gmail.com wrote: On Fri, Jan 29, 2010 at 10:38 AM, David C. Rankin drankina...@suddenlinkmail.com wrote: Guys, After updates today on my x86_64 box from testing, including kernel26-2.6.32.6-1, Compiz is broken. It attempts to start, but then all windows are left without any decorations and keyboard input is disabled on kde4. The first time I tried, the plasma panel died leaving me with no way to get kwin back other than finding a black spot on the desktop and logging out. Next attempt, plasma panel stayed with me and changing back and forth from kwin to compiz and back is easily done with fusion-icon. (I put a shortcut on my desktop to 'kwin --replace' just in case) Dunno what is happening here. All was OK before updates today. pacman.log is at: www.3111skyline.com/download/Archlinux/pacman.log Looks like it is another jpg/png library issue: 20:35 alchemy:~ fusion-icon * Detected Session: kde * Searching for installed applications... * No GLX_EXT_texture_from_pixmap with direct rendering context ... present with indirect rendering, exporting: LIBGL_ALWAYS_INDIRECT=1 * Using the GTK Interface * KWin is already running * Setting window manager to Compiz ... executing: compiz --replace --sm-disable --ignore-desktop-hints --indirect-rendering ccp --indirect-rendering WARNING: Application calling GLX 1.3 function glXCreatePixmap when GLX 1.3 is not supported! This is an application bug! compiz: symbol lookup error: /usr/lib/compiz/libpng.so: undefined symbol: png_check_sig I experienced the same here at the point when the libpng rebuild started (and compiz was first updated). I almost immediately downgraded due to work requiring openoffice though, and haven't gotten round to enabling testing again. Update, I've found that disabling the 'png' plugin under the 'Image Loading' section fixes the issue, at least compiz loads up and runs like it used to. Something must have gone wrong with the libpng-related rebuild. Bug report time? do LD_DEBUG=files fusion-icon log 21 post somewhere the log and then post here the link to see it -- Ionut
Re: [arch-general] kernel26-lts won't install
On 01/29/2010 12:25 AM, Gaurish Sharma wrote: Hi, the i686 Arch does have udev-compat 150. So I am guess you are building for x86_64 Arch.udev-compat 150 is still in [Testing][1] Either wait for it or Enable [Testing] Repository [1] http://www.archlinux.org/packages/testing/x86_64/udev-compat/ Regards, Gaurish Sharma www.gaurishsharma.com Nope. I am on i686. Leave it, its a bug. -- Nilesh Govindarajan Site Server Adminstrator www.itech7.com
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Attila vodoo0...@sonnenkinder.org wrote: At Donnerstag, 28. Januar 2010 08:35 Gerardo Exequiel Pozzi wrote: Hi, don't need all root privileges/capabilities. Only cap_sys_admin, cap_sys_rawio for some special SCSI commands and cap_sys_resource for incresing resource limits. setcap cap_sys_admin,cap_sys_rawio,cap_sys_resource+ep /usr/bin/cdrecord If i do this than i cannot open '/dev/sg1' and therefore i will stay with my way instead it is obsolete. Aha, now I see that it seems that Arch really may have the needed features. Did you try to also add CAP_DAC_OVERRIDE? Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
ludovic coues cou...@gmail.com wrote: Dev don't care about technical basis. They are ok with the fact that 13 releases per year is better than only one each single year. During the past 4 years, the average was 176.5 releases per year ;-) Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Am Freitag, 29. Januar 2010 11:58:12 schrieb Joerg Schilling: Paulo Matias mat...@archlinux-br.org wrote: On Fri, Jan 29, 2010 at 8:39 AM, Joerg Schilling joerg.schill...@fokus.fraunhofer.de wrote: As long as there is no support code in Linux distros to set capabilities without making the target program suid root anyway, Don't be afraid, Arch Linux has support for that :) How? Is there support for mandatory ACLs? Jörg Finally some interesting discussion came out of this. I am not an expert on linux capability support, but Thomas has posted two blog entries about this in Arch: http://archlinux.me/brain0/2009/07/28/using-posix-capabilities-in-linux- part-one/ and http://archlinux.me/brain0/2010/01/05/using-posix-capabilities- in-linux-part-two/ In general this should work fine. The only problem is that bsdtar did not support storing those information (don't know if future versions support this) so one has to use install scripts to adjust the permissions after install. Pierre -- Pierre Schmitz, https://users.archlinux.de/~pierre
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
joerg.schill...@fokus.fraunhofer.de (Joerg Schilling) wrote: ludovic coues cou...@gmail.com wrote: Dev don't care about technical basis. They are ok with the fact that 13 releases per year is better than only one each single year. During the past 4 years, the average was 176.5 releases per year ;-) Sorry for the typo: should be 17.5 Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Pierre Schmitz pie...@archlinux.de wrote: Finally some interesting discussion came out of this. I am not an expert on linux capability support, but Thomas has posted two blog entries about this in Arch: http://archlinux.me/brain0/2009/07/28/using-posix-capabilities-in-linux- part-one/ and http://archlinux.me/brain0/2010/01/05/using-posix-capabilities- in-linux-part-two/ I'll have a loot at it. In general this should work fine. The only problem is that bsdtar did not support storing those information (don't know if future versions support this) so one has to use install scripts to adjust the permissions after install. I am not sure whether this is the best solution. I recommend to use star as star is the oldest free tar implementation and as it supports ACLs since 10 years already. Adding more meta data is relatively simple. e~A
Re: [arch-general] [aur-general] PKGBUILD for python-numpy, orphaned and out-of-date
On Fri, Jan 29, 2010 at 9:15 AM, Puneeth Chaganti puncha...@gmail.com wrote: On Fri, Jan 29, 2010 at 5:41 PM, Robson Peixoto robsonpeix...@gmail.com wrote: Please, add gcc-fortran in makedepends Sorry, missed it. thanks punchagan Oh I used to maintain numpy when it was in [community] (dsa used to maintain it first, then me, then him) now, this is an important package, if some dev want to maintain it, I could help (or I must say we?) but definetively this package belongs to [extra]) -- Angel Velásquez angvp @ irc.freenode.net Arch Linux Trusted User Linux Counter: #359909 http://www.angvp.com
Re: [arch-general] Netcfg after resume from suspend
On Thu, Jan 28, 2010 at 8:42 PM, Ng Oon-Ee ngoo...@gmail.com wrote: Hi all, I'm using netcfg and recently have been noticing the following. When I resume from suspend I call 'netcfg home' which tries to connect to my home's wifi network. ESSID and WEP key are correctly set. It exist with a 'could not obtain IP address error, this is the dmesg:- Registered led device: iwl-phy0::radio Registered led device: iwl-phy0::assoc Registered led device: iwl-phy0::RX Registered led device: iwl-phy0::TX ADDRCONF(NETDEV_UP): wlan0: link is not ready wlan0: direct probe to AP 00:14:7c:be:39:ac (try 1) wlan0: direct probe responded wlan0: authenticate with AP 00:14:7c:be:39:ac (try 1) wlan0: authenticated wlan0: associate with AP 00:14:7c:be:39:ac (try 1) wlan0: RX AssocResp from 00:14:7c:be:39:ac (capab=0x471 status=0 aid=1) wlan0: associated ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready wlan0: no IPv6 routers present wlan0: deauthenticating from 00:14:7c:be:39:ac by local choice (reason=3) The weird thing is that when I run 'iwconfig wlan0 essid MYESSID key MYKEY', then run 'netcfg home' again it works no problem. Suggestions for what I can check out are welcome. This happens both on stock kernel and my custom -ice kernel, am up-to-date (without testing for now, due to my openoffice variant not being done yet). Perhaps you need to change DHCP_TIMEOUT in the network profile to be longer?
Re: [arch-general] Netcfg after resume from suspend
On Fri, 2010-01-29 at 10:36 -0500, Alexander Lam wrote: On Thu, Jan 28, 2010 at 8:42 PM, Ng Oon-Ee ngoo...@gmail.com wrote: Hi all, I'm using netcfg and recently have been noticing the following. When I resume from suspend I call 'netcfg home' which tries to connect to my home's wifi network. ESSID and WEP key are correctly set. It exist with a 'could not obtain IP address error, this is the dmesg:- Registered led device: iwl-phy0::radio Registered led device: iwl-phy0::assoc Registered led device: iwl-phy0::RX Registered led device: iwl-phy0::TX ADDRCONF(NETDEV_UP): wlan0: link is not ready wlan0: direct probe to AP 00:14:7c:be:39:ac (try 1) wlan0: direct probe responded wlan0: authenticate with AP 00:14:7c:be:39:ac (try 1) wlan0: authenticated wlan0: associate with AP 00:14:7c:be:39:ac (try 1) wlan0: RX AssocResp from 00:14:7c:be:39:ac (capab=0x471 status=0 aid=1) wlan0: associated ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready wlan0: no IPv6 routers present wlan0: deauthenticating from 00:14:7c:be:39:ac by local choice (reason=3) The weird thing is that when I run 'iwconfig wlan0 essid MYESSID key MYKEY', then run 'netcfg home' again it works no problem. Suggestions for what I can check out are welcome. This happens both on stock kernel and my custom -ice kernel, am up-to-date (without testing for now, due to my openoffice variant not being done yet). Perhaps you need to change DHCP_TIMEOUT in the network profile to be longer? Thanks, I should have mentioned that this isn't a timeout issue, extending it to several minutes did not help, while once I run iwconfig prior to netcfg it connects in under 10 seconds.
Re: [arch-general] Netcfg after resume from suspend
Am 29.01.2010 02:42, schrieb Ng Oon-Ee: Hi all, I'm using netcfg and recently have been noticing the following. When I resume from suspend I call 'netcfg home' which tries to connect to my home's wifi network. ESSID and WEP key are correctly set. It exist with a 'could not obtain IP address error, this is the dmesg:- Registered led device: iwl-phy0::radio Registered led device: iwl-phy0::assoc Registered led device: iwl-phy0::RX Registered led device: iwl-phy0::TX ADDRCONF(NETDEV_UP): wlan0: link is not ready wlan0: direct probe to AP 00:14:7c:be:39:ac (try 1) wlan0: direct probe responded wlan0: authenticate with AP 00:14:7c:be:39:ac (try 1) wlan0: authenticated wlan0: associate with AP 00:14:7c:be:39:ac (try 1) wlan0: RX AssocResp from 00:14:7c:be:39:ac (capab=0x471 status=0 aid=1) wlan0: associated ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready wlan0: no IPv6 routers present wlan0: deauthenticating from 00:14:7c:be:39:ac by local choice (reason=3) The weird thing is that when I run 'iwconfig wlan0 essid MYESSID key MYKEY', then run 'netcfg home' again it works no problem. Suggestions for what I can check out are welcome. This happens both on stock kernel and my custom -ice kernel, am up-to-date (without testing for now, due to my openoffice variant not being done yet). You should try the testing version of netcfg instead: http://mirrors.kernel.org/archlinux/testing/os/any/netcfg-2.5.0rc2-1-any.pkg.tar.gz This has two advantages: 1) wpa_supplicant is not only used for WPA, buy also for Open and WEP networks. This increases reliability drastically! (Note that using the deprecated iwconfig-method the connection will not be reestablished by the driver once it is lost - upstream linux-wireless developers recommend always using wpa_supplicant) 2) If you also install wpa_actiond (http://mirrors.kernel.org/archlinux/testing/os/i686/wpa_actiond-1.0-1-i686.pkg.tar.gz, http://mirrors.kernel.org/archlinux/testing/os/x86_64/wpa_actiond-1.0-1-x86_64.pkg.tar.gz), you get the net-auto-wireless mode, which not only automatically connects to the right wireless network whenever it is in range, but also deals with temporary connection losses gracefully and transparently (there is a minor bug where it sometimes won't get DHCP, but that can be worked around and is easy to fix for the next version (WPA_OPTS=-W)). It will also connect to a new network when the old one is out of range and the new one comes in range (assuming the networks are configured as profiles). I guess this will make your problem(s) go away. I don't know why James still hasn't released netcfg 2.5, but I guess he has little time lately. signature.asc Description: OpenPGP digital signature
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
On Fri, 2010-01-29 at 14:35 +0100, Joerg Schilling wrote: I am not sure whether this is the best solution. I recommend to use star as star is the oldest free tar implementation and as it supports ACLs since 10 years already. Adding more meta data is relatively simple. Implementing it in star has no use, as our package manager doesn't use star but libarchive, the library that bsdtar is based on.
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Am Freitag, 29. Januar 2010 17:58:42 schrieb Jan de Groot: On Fri, 2010-01-29 at 14:35 +0100, Joerg Schilling wrote: I am not sure whether this is the best solution. I recommend to use star as star is the oldest free tar implementation and as it supports ACLs since 10 years already. Adding more meta data is relatively simple. Implementing it in star has no use, as our package manager doesn't use star but libarchive, the library that bsdtar is based on. This might be related: http://code.google.com/p/libarchive/wiki/TarPosix1eACLs -- Pierre Schmitz, https://users.archlinux.de/~pierre
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Pierre Schmitz pie...@archlinux.de wrote: Am Freitag, 29. Januar 2010 17:58:42 schrieb Jan de Groot: On Fri, 2010-01-29 at 14:35 +0100, Joerg Schilling wrote: I am not sure whether this is the best solution. I recommend to use star as star is the oldest free tar implementation and as it supports ACLs since 10 years already. Adding more meta data is relatively simple. Implementing it in star has no use, as our package manager doesn't use star but libarchive, the library that bsdtar is based on. This might be related: http://code.google.com/p/libarchive/wiki/TarPosix1eACLs This does just describe what I defined 10 years ago ;-) Some notes: The POSIX.1e ACL draft was withdrawn in 1999. The current ACL standard is from NTFS and part of NVFv4 and ZFS Mandatory ACLs are something completely different. They are similar to extended file attributes - just that the kernel interprets them. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
joerg.schill...@fokus.fraunhofer.de (Joerg Schilling) wrote: This might be related: http://code.google.com/p/libarchive/wiki/TarPosix1eACLs This does just describe what I defined 10 years ago ;-) I forgot, the complete documentation is here: http://cdrecord.berlios.de/private/man/star/star.4.html Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
[arch-general] ndesk-dbus 0.6.0 bugfix
Bug #377672 https://bugs.launchpad.net/ndesk-dbus/+bug/377672 I stumbled upon this while trying to suspend my laptop using Gnome-DO's Gnome-session plugin. It may also cause problems with other C# apps. There's a quick fix for this in the first post of the launchpad's bug report : --- ndesk-dbus-0.6.0.orig/src/TypeImplementer.cs2007-10-11 20:01:11.0 + +++ ndesk-dbus-0.6.0/src/TypeImplementer.cs2010-01-29 18:11:32.0 + @@ -34,7 +34,7 @@ namespace NDesk.DBus InitHack (); -TypeBuilder typeB = modB.DefineType (declType.Name + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); +TypeBuilder typeB = modB.DefineType (declType.FullName + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); Implement (typeB, declType); Might want to add this quick fix to the package in extra.
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Le Fri, 29 Jan 2010 17:58:42 +0100, Jan de Groot j...@jgc.homeip.net a écrit : Implementing it in star has no use, as our package manager doesn't use star but libarchive, the library that bsdtar is based on. Technically the PKGBUILD could makedepends('star') and use it to extract the source with ACL support... -- catwell
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Le Fri, 29 Jan 2010 17:36:58 +, Pierre Chapuis catw...@archlinux.us a écrit : Le Fri, 29 Jan 2010 17:58:42 +0100, Jan de Groot j...@jgc.homeip.net a écrit : Implementing it in star has no use, as our package manager doesn't use star but libarchive, the library that bsdtar is based on. Technically the PKGBUILD could makedepends('star') and use it to extract the source with ACL support... Ignore this, I just understood how stupid that is. The point is not to support ACLs for the source tarball but for the package itself. -- catwell
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
At Freitag, 29. Januar 2010 14:10 Pierre Schmitz wrote: Finally some interesting discussion came out of this. I am not an expert on linux capability support, but Thomas has posted two blog entries about this in Arch: http://archlinux.me/brain0/2009/07/28/using-posix-capabilities-in-linux- part-one/ and http://archlinux.me/brain0/2010/01/05/using-posix-capabilities- in-linux-part-two/ Nice informations, thanks. In general this should work fine. The only problem is that bsdtar did not support storing those information (don't know if future versions support this) so one has to use install scripts to adjust the permissions after install. I must say that using install scripts is from my view for permissions or setcap even the better way because than i don't need to be root to create a package. See you, Attila
Re: [arch-general] ndesk-dbus 0.6.0 bugfix
On Fri, Jan 29, 2010 at 3:34 PM, Andrea Fagiani andfagi...@gmail.com wrote: Bug #377672 https://bugs.launchpad.net/ndesk-dbus/+bug/377672 I stumbled upon this while trying to suspend my laptop using Gnome-DO's Gnome-session plugin. It may also cause problems with other C# apps. There's a quick fix for this in the first post of the launchpad's bug report : --- ndesk-dbus-0.6.0.orig/src/TypeImplementer.cs 2007-10-11 20:01:11.0 + +++ ndesk-dbus-0.6.0/src/TypeImplementer.cs 2010-01-29 18:11:32.0 + @@ -34,7 +34,7 @@ namespace NDesk.DBus InitHack (); - TypeBuilder typeB = modB.DefineType (declType.Name + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); + TypeBuilder typeB = modB.DefineType (declType.FullName + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); Implement (typeB, declType); Might want to add this quick fix to the package in extra. Thanks for the patch Andrea, I am posting this mail on the opened bug on the Arch Linux Bug Tracker [1]. Thanks for your contribution, next time try to see if the bug is reported on the Arch Linux Bug Tracker [2] in order to have the things better organized [1] http://bugs.archlinux.org/task/17900 [2] http://bugs.archlinux.org -- Angel Velásquez angvp @ irc.freenode.net Arch Linux Trusted User Linux Counter: #359909 http://www.angvp.com
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Attila vodoo0...@sonnenkinder.org wrote: At Freitag, 29. Januar 2010 11:39 Joerg Schilling wrote: Thanks for your nice informations and with this line for setcap cap_dac_override,cap_sys_rawio,cap_ipc_lock,cap_sys_nice,cap_net_bind_service+ep a cdrecord --scanbus works as normal user without a problem. BTW: cdrecord in this mode is unsafe as it does not manages the fine grained privileges but would need to give up cap_dac_override before it tries to open other files. If you don't give up cap_dac_override, cdecord can read _any_ local file. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Am Freitag, 29. Januar 2010 18:47:55 schrieb Xavier Chantry: When I looked into that a few months ago, it stored just fine when creating the archive. But it did not restore them when extracting. This got fixed in trunk, so it will probably be in the next major release (2.8 ?). http://code.google.com/p/libarchive/source/detail?r=1590# xps-m1530:~ bsdtar --version bsdtar 2.7.902a - libarchive 2.7.902a 2.7 release does not work, at least on my system. The development version is required. This is good new; afaik 2.8 should be released soon. -- Pierre Schmitz, https://users.archlinux.de/~pierre
Re: [arch-general] ndesk-dbus 0.6.0 bugfix
On 01/29/2010 05:54 PM, Angel Velásquez wrote: On Fri, Jan 29, 2010 at 3:34 PM, Andrea Fagianiandfagi...@gmail.com wrote: Bug #377672https://bugs.launchpad.net/ndesk-dbus/+bug/377672 I stumbled upon this while trying to suspend my laptop using Gnome-DO's Gnome-session plugin. It may also cause problems with other C# apps. There's a quick fix for this in the first post of the launchpad's bug report : --- ndesk-dbus-0.6.0.orig/src/TypeImplementer.cs2007-10-11 20:01:11.0 + +++ ndesk-dbus-0.6.0/src/TypeImplementer.cs2010-01-29 18:11:32.0 + @@ -34,7 +34,7 @@ namespace NDesk.DBus InitHack (); -TypeBuilder typeB = modB.DefineType (declType.Name + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); +TypeBuilder typeB = modB.DefineType (declType.FullName + Proxy, TypeAttributes.Class | TypeAttributes.Public, typeof (BusObject)); Implement (typeB, declType); Might want to add this quick fix to the package in extra. Thanks for the patch Andrea, I am posting this mail on the opened bug on the Arch Linux Bug Tracker [1]. Thanks for your contribution, next time try to see if the bug is reported on the Arch Linux Bug Tracker [2] in order to have the things better organized [1] http://bugs.archlinux.org/task/17900 [2] http://bugs.archlinux.org Thanks Angel, I'll keep that in mind ;-) .
[arch-general] Laptop reboots instead of resume from suspend
I'm using an Inspiron 1720, GeForce 8600M GT, Arch i686. If I update to any kernel after 2.6.31.6-1 then the laptop reboots from suspend or hibernate. According to pm-suspend.log the suspend was fine and there are n problems in the other logs. It just reboots. Anyone else seen this, or any suggestions for a fix? -- Dave.
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
At Freitag, 29. Januar 2010 18:55 Joerg Schilling wrote: BTW: cdrecord in this mode is unsafe as it does not manages the fine grained privileges but would need to give up cap_dac_override before it tries to open other files. If you don't give up cap_dac_override, cdecord can read any local file. Thanks for the warning that is unsafe but how longer we discuss about doing the same with capabilities how more i like to stay with the gold old solution.-) See you, Attila
Re: [arch-general] Laptop reboots instead of resume from suspend
On 01/29/2010 09:46 PM, Dave Morgan wrote: I'm using an Inspiron 1720, GeForce 8600M GT, Arch i686. If I update to any kernel after 2.6.31.6-1 then the laptop reboots from suspend or hibernate. According to pm-suspend.log the suspend was fine and there are n problems in the other logs. It just reboots. Anyone else seen this, or any suggestions for a fix? I've had the same problem with hibernation on my laptop, and solved by adding the option : /resume=/dev/SWAPPARTITION/ to the linux line in my grub.cfg (that would be kernel line in menu.lst if you're using the old grub). I didn't have problems with suspend though. Might also want to try adding the /`resume/` hook to your mkinitcpio.conf and rebuild your initrd. Andrea
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
So, are we going to switch to cdrtools? Best regards, Luís Moreira.
[arch-general] Little problems with arch
I have a little dell 10 with archlinux. Unfortunately the audio and 3d are not really working, I tried hard some time ago but with no luck, anyway other problems are more annoying. - time getting crazy Every time I reboot this little machine the date is set somehow randomly. Any idea of what could that be? Maybe the battery that should keep it up is not working (I think is an OS issue though)?? - screensaver I use xfce but I installed quite a lot of stuff from gnome, but I never asked for the stupid gnome-screensaver to start at every boot, and I don't find how to disable it. My daemons are only those: --8---cut here---start-8--- DAEMONS=(syslog-ng network netfs crond acpid sshd hal bluetooth alsa rsyncd cpufreqd wicd cups fam) --8---cut here---end---8--- - network devices with random name I had some troubles with the wireless card but then it worked smoothly with the proprietary broadcom-wl driver, the only problems is that sometimes eth0 = lan, eth1 = wireless and sometimes is the opposite and I don't understand why. This also happens with normal reboots. - no packages to update This is quite strange, sometimes I do a pacman -Suy and I loop up to date, but if I change the first mirror just editing mirrorlist I'm not up to date anymore and it starts updating. What does that mean? Not all the mirrors are on sync? Could it be related somehow to my problem with the date (shouldn't be though)?
Re: [arch-general] Little problems with arch
On 01/29/2010 11:59 PM, Andrea Crotti wrote: I have a little dell 10 with archlinux. Unfortunately the audio and 3d are not really working, I tried hard some time ago but with no luck, anyway other problems are more annoying. - time getting crazy Every time I reboot this little machine the date is set somehow randomly. Any idea of what could that be? Maybe the battery that should keep it up is not working (I think is an OS issue though)?? - screensaver I use xfce but I installed quite a lot of stuff from gnome, but I never asked for the stupid gnome-screensaver to start at every boot, and I don't find how to disable it. My daemons are only those: --8---cut here---start-8--- DAEMONS=(syslog-ng network netfs crond acpid sshd hal bluetooth alsa rsyncd cpufreqd wicd cups fam) --8---cut here---end---8--- - network devices with random name I had some troubles with the wireless card but then it worked smoothly with the proprietary broadcom-wl driver, the only problems is that sometimes eth0 = lan, eth1 = wireless and sometimes is the opposite and I don't understand why. This also happens with normal reboots. - no packages to update This is quite strange, sometimes I do a pacman -Suy and I loop up to date, but if I change the first mirror just editing mirrorlist I'm not up to date anymore and it starts updating. What does that mean? Not all the mirrors are on sync? Could it be related somehow to my problem with the date (shouldn't be though)? Hi Andrea, 1) time issue check your /etc/rc.conf and make sure you have HARDWARECLOCK=localtime and the correct TIMEZONE setting. Also, if that is already ok, try deleting //var/lib/hwclock/adjtime/ , this fixed it for me. 2) about gnome-screensaver, I don't really know I've never had such problems with it, I'd try running /gnome-screensaver-preferences/ and disabling it from there though. 3) network intefaces issue if that keeps up you could try adding udev rules for your devices: create a new rules file, e.g. //etc/udev/rules.d/66-nic.rules/ , a possible set of rules would look like KERNEL==eth*, SYSFS{address}==aa:bb:cc:dd:ee:ff, NAME=eth0 KERNEL==eth*, SYSFS{address}==ff:ee:dd:cc:bb:aa, NAME=eth1 where the address fields should match that of your network intefaces (you can use /ifconfig /to narrow that down). That should pretty much do the trick. 4) mirrors Yes, some mirrors may be out of sync, take a look at the Mirror Status https://www.archlinux.de/?page=MirrorStatus ;-) Andrea
Re: [arch-general] kernel26-2.6.32.6-1 updates breaks Compiz?
On Fri, 2010-01-29 at 13:19 +0200, Ionut Biru wrote: On 01/29/2010 10:33 AM, Ng Oon-Ee wrote: On Fri, Jan 29, 2010 at 11:23 AM, Ng Oon-Eengoo...@gmail.com wrote: Update, I've found that disabling the 'png' plugin under the 'Image Loading' section fixes the issue, at least compiz loads up and runs like it used to. Something must have gone wrong with the libpng-related rebuild. Bug report time? do LD_DEBUG=files fusion-icon log 21 post somewhere the log and then post here the link to see it Here's the log from a failed run - http://pastebin.com/m6a6cb270 Here's the log from a successful run (with the png plugin turned off) - http://pastebin.com/m62782b96 Both pastebins only available for a month. Interesting line (in my unexperienced opinion) in the first log is line 1914:- 7720: /usr/lib/compiz/libpng.so: error: symbol lookup error: undefined symbol: png_check_sig (fatal) compiz: symbol lookup error: /usr/lib/compiz/libpng.so: undefined symbol: png_check_sig
Re: [arch-general] Netcfg after resume from suspend
On Fri, 2010-01-29 at 17:00 +0100, Thomas Bächler wrote: Am 29.01.2010 02:42, schrieb Ng Oon-Ee: Hi all, I'm using netcfg and recently have been noticing the following. When I resume from suspend I call 'netcfg home' which tries to connect to my home's wifi network. ESSID and WEP key are correctly set. It exist with a 'could not obtain IP address error, this is the dmesg:- Registered led device: iwl-phy0::radio Registered led device: iwl-phy0::assoc Registered led device: iwl-phy0::RX Registered led device: iwl-phy0::TX ADDRCONF(NETDEV_UP): wlan0: link is not ready wlan0: direct probe to AP 00:14:7c:be:39:ac (try 1) wlan0: direct probe responded wlan0: authenticate with AP 00:14:7c:be:39:ac (try 1) wlan0: authenticated wlan0: associate with AP 00:14:7c:be:39:ac (try 1) wlan0: RX AssocResp from 00:14:7c:be:39:ac (capab=0x471 status=0 aid=1) wlan0: associated ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready wlan0: no IPv6 routers present wlan0: deauthenticating from 00:14:7c:be:39:ac by local choice (reason=3) The weird thing is that when I run 'iwconfig wlan0 essid MYESSID key MYKEY', then run 'netcfg home' again it works no problem. Suggestions for what I can check out are welcome. This happens both on stock kernel and my custom -ice kernel, am up-to-date (without testing for now, due to my openoffice variant not being done yet). You should try the testing version of netcfg instead: http://mirrors.kernel.org/archlinux/testing/os/any/netcfg-2.5.0rc2-1-any.pkg.tar.gz This has two advantages: 1) wpa_supplicant is not only used for WPA, buy also for Open and WEP networks. This increases reliability drastically! (Note that using the deprecated iwconfig-method the connection will not be reestablished by the driver once it is lost - upstream linux-wireless developers recommend always using wpa_supplicant) 2) If you also install wpa_actiond (http://mirrors.kernel.org/archlinux/testing/os/i686/wpa_actiond-1.0-1-i686.pkg.tar.gz, http://mirrors.kernel.org/archlinux/testing/os/x86_64/wpa_actiond-1.0-1-x86_64.pkg.tar.gz), you get the net-auto-wireless mode, which not only automatically connects to the right wireless network whenever it is in range, but also deals with temporary connection losses gracefully and transparently (there is a minor bug where it sometimes won't get DHCP, but that can be worked around and is easy to fix for the next version (WPA_OPTS=-W)). It will also connect to a new network when the old one is out of range and the new one comes in range (assuming the networks are configured as profiles). I guess this will make your problem(s) go away. I don't know why James still hasn't released netcfg 2.5, but I guess he has little time lately. Thanks, it does seem to work (just one try so far). I've seen the netcfg testing thread in the forums, is that the authoritative place to go for info on this latest version, since the wiki doesn't seem updated for it yet (which of course makes perfect sense).
Re: [arch-general] An old, tiresome discussion: cdrtools vs cdrkit
Steve Holmes steve.holme...@gmail.com wrote: -BEGIN PGP SIGNED MESSAGE- Hash: RIPEMD160 I don't know much about the licenses differences and all that crap but I experienced a problem with cdrecord several years ago where it would not work with my CD burner. I kept getting wiere I/O errors or some such. When I asked around,some people told me about wodim and when I went out and installed wodim, I've been able to burn CDs and DVDs flawlessly ever since. My time with wodim has transpired over Slackware, Debian, and now Arch. I don't know today if cdrecord would still cause me those errors or not but for me, the drkit has been doing me just fine. As you do not give any facts, this is obviously nonsense. I know of not a single case where cdrecord fails but wodim succeeds. Wodim is nothing than an onl version of cdrecord with bugs added by it't creators that never have been in the original. If you would give evidence, it would be easy to prove that your alleged problem is not related to cdrecord. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily