March 2011 - Arch-dev-public - lists.archlinux.org

Re: [arch-dev-public] [initscripts] Linux distributions and /run/
by Tom Gundersen 31 Mar '11

31 Mar '11

On Thu, Mar 31, 2011 at 2:22 PM, Andrea Scarpino <andrea(a)archlinux.org> wrote: > On Thursday 31 March 2011 14:15:37 Tom Gundersen wrote: >> I guess the list can be generated by grepping all the PKGBUILDs for >> writes to /var/run and /var/lock. > I guess that instead of grepping PKGBUILDs we've to grep the package files > list. D'oh, very good point! -t

1 0

Re: [arch-dev-public] [initscripts] Linux distributions and /run/
by Tom Gundersen 31 Mar '11

31 Mar '11

On Thu, Mar 31, 2011 at 1:56 PM, Andrea Scarpino <andrea(a)archlinux.org> wrote: > About FS#18157, many of them have been fixed (mysql, mpd, httpd...), but we > need a complete list. I guess the list can be generated by grepping all the PKGBUILDs for writes to /var/run and /var/lock. Maybe it would be worth adding a warning to makepkg if files are written to /var/run, /var/lock, /tmp or /run? Cheers, Tom

3 2

[arch-dev-public] [arch-projects] [initscripts] Linux distributions and /run/
by Tom Gundersen 31 Mar '11

31 Mar '11

On Thu, Mar 31, 2011 at 1:21 PM, Thomas Bächler <thomas(a)archlinux.org> wrote: > This was discussed in a bug report iirc, here is some summary: > http://www.h-online.com/open/news/item/Linux-distributions-to-include-run-d… > > Apparently, many tools (among them udev) will soon require a /run/ > directory that is mounted early (as tmpfs). It is probably a good idea > to add this to the filesystem and initscripts packages. I have just been reading up on this, and I agree that we should add it. At some point in the future we should probably consider symlinking /var/run to /run and /var/lock to /run/lock, but we would have to implement <https://bugs.archlinux.org/task/18157> first, and /run is useful even without the symlinks. My understanding of the situation is that it is ok or /var/run and /run to be different, but please let me know if I'm wrong on this. I guess we'll need to add the mountpoint to the filesystem package (possibly with permissions 0000 to avoid accidents) and mount the tmpfs early in rc.sysinit (I'll write the patch as soon as people agree that this is the right thing to do). Cheers, Tom

1 0

[arch-dev-public] [initscripts] Linux distributions and /run/
by Thomas Bächler 31 Mar '11

31 Mar '11

This was discussed in a bug report iirc, here is some summary: http://www.h-online.com/open/news/item/Linux-distributions-to-include-run-d… Apparently, many tools (among them udev) will soon require a /run/ directory that is mounted early (as tmpfs). It is probably a good idea to add this to the filesystem and initscripts packages.

3 2

[arch-dev-public] Removing network utilities from base group?
by Daniel Isenmann 29 Mar '11

29 Mar '11

Hi, I have a feature request (https://bugs.archlinux.org/task/22481) since January which stated that rp-pppoe should be removed from the base group. Described in the details he stated that every network package should be removed from the base group and should only be in core. I'm not really have a opinion to that topic, so what do you think about it? For me we can leave it as it is and do nothing at this topic, then I will close this feature request as "won't implement". -Daniel

3 2

[arch-dev-public] [signoff] kernel26-lts 2.6.32.36-1
by Andreas Radke 29 Mar '11

29 Mar '11

Upstream update with security fixes. Please signoff. -Andy

1 0

[arch-dev-public] New list: arch-projects
by Thomas Bächler 28 Mar '11

28 Mar '11

A new list has been created: arch-projects. This list is for development discussion, patches and pull requests for initscripts, netcfg, dbscripts, devtools (and maybe other projects, those are the ones I could think of). Anyone is free to subscribe, but please, no user discussions (keep those to the forums or arch-general). Fun fact: There was a mailing list called arch-projects once, and it has been deleted, but the old archives are still there.

2 1

[arch-dev-public] [signoff] usbutils-002-1
by Tobias Powalowski 28 Mar '11

28 Mar '11

Hi guys, latest upstream release, only bugfixes. please signoff both arches greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa(a)archlinux.org

3 2

[arch-dev-public] gcc-4.6.0 toolchain in [testing]
by Allan McRae 27 Mar '11

27 Mar '11

The gcc-4.6 toolchain rebuild has been uploaded to [testing]: gcc-4.6.0-1: - major upstream update. See http://gcc.gnu.org/gcc-4.6/changes.html - new dependency - cloog built with isl backend (replaces cloog-ppl) - remove elfutils dep (no longer needed) - clean-up of configure options - soname bump libobjc.so.2 -> .3 - package gcc-go frontend - please test calling C/C++ code, please report results with and without libffi installed linux-api-headers-2.6.38.1-1: - update to latest version binutils-2.21-6: - new upstream snapshot from 2.21 branch - remove patch included upstream glibc-2.13-5: - patch futex issue rather than disabling it libtool-2.4-3: - gcc version bump rebuild Test suite results can be seen at https://wiki.archlinux.org/index.php/User:Allan/Toolchain . There are a few new failures in the gcc and binutils testsuites. Most appear unimportant, but I will investigate further before I consider moving this to [core]. I expect an update to binutils with quite a few backports soon (which may fix some of the gcc lto issues) and I will probably need to grab an updated gcc snapshot in a couple of weeks anyway... Note that I have not done the small libobjc rebuild yet so if you are using gnustep then you might want to hold off updating for a bit. Allan

1 2

[arch-dev-public] The future of heimdal and nfs-utils
by Tobias Powalowski 27 Mar '11

27 Mar '11

Hi guys, The reason for not updating nfs-utils in [core] is the fact that, nfs-utils and heimdal patches don't work together anymore. I don't use this feature at all. Other distributions don't use heimdal as kerberos implementation, so no luck to find any help there. I asked diego from gentoo who did some work on gentoo and nfs stuff, but he is busy since ages, but agrees that the whole thing is in a real bad state. What to do? Shall we drop the heimdal implementation on nfs? Go away from heimdal and use an other kerberos implementation? As I said before I don't use this features at all. Thanks for your help. greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa(a)archlinux.org

3 6