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

Re: [arch-dev-public] [prerelease] util-linux-2.20-rc2
by Tom Gundersen 23 Aug '11

23 Aug '11

Hi guys, I put up the next prerelease of util-linux in case someone wants to play around: <https://dev.archlinux.org/~tomegun/util-linux-2.20rc2-1-i686.pkg.tar.xz> <https://dev.archlinux.org/~tomegun/util-linux-2.20rc2-1-x86_64.pkg.tar.xz> Same warnings apply as to rc1: - mountpoint will conflict with sysvinit (I'll propose moving this from sysvinit to util-linux when we have a final release); and - libmount is enabled even though it will probably not be in the final version. Cheers, Tom On Sat, Jul 30, 2011 at 12:43 AM, Tom Gundersen <teg(a)jklm.no> wrote: > Hi guys, > > A new and very exciting release of util-linux is imminent. Even though > we only have an -rc1 so far, I thought it might be worthwhile to put > it out there so people can try it out (in case we want anything fixed > before the final release): > > <https://dev.archlinux.org/~tomegun/util-linux-2.20rc1-1-i686.pkg.tar.xz> > <https://dev.archlinux.org/~tomegun/util-linux-2.20rc1-1-x86_64.pkg.tar.xz> > > I enabled libmount support in mount, but this will probably be > reverted before the final release unless we have good reason to > believe it to be safe. > > Major changes: > > - agetty gets most of the features from mingetty. > - dmesg gets lots of new filtering capabilities > - mountpoint moves here from sysvinit (we probably have to manually > remove it from the sysvinit package). > > We still patch hwclock to move the adjustment file from /etc to /var. > I'm not a big fan of this, and hope to have a backwards compatible > patch to sort things out before 2.21 is released. > > Cheers, > > Tom > > > Util-linux 2.20 Release Notes > ============================= > > Release highlights > ------------------ > > agetty(8): > - mingetty features have been merged to agetty > > chrt(1), taskset(1): > - supports new command line option "--all-tasks" to set or retrieve the > scheduling attributes of all the tasks (threads) for a given PID > > dmesg(1): > - supports new command line options: --clear, --console-on, --console-off, > --ctime, --decode, --facility=<list>, --level=<list>, --show-delta, > --notime, --kernel and --userspace > > fdisk(8): > - improved dialogs to be more user-friendly > > findmnt(8), partx(8), lsblk(8) > - support new command line option "--pairs" to enable key="value" > output format > > findmnt(8): > - supports new command line options "--poll" and "--timeout" to monitor > /proc/self/mountinfo changes > > ionice(1): > - supports human-readable scheduling class names, for example: > ionice -c best-effort $PID > > kill(1): > - supports new command line option "-q <sigval>" to use sigqueue(2) > - supports real-time signals in formats RT<n>, RTMIN+<n> and RTMAX-<n> > > libmount: > - the API officially stable > > lsblk(8): > - supports new columns: > TYPE - device type > RQ-SIZE - queue request size > STATE - device state (e.g. running, suspended) > DALIGN - discard alignment offset > DISC-GRAN - discard granularity > DISC-MAX - discard max bytes > DISC-ZERO - discard zeroes data > - supports new command line option "-D" to print device discard topology > > lscpu(8): > - improved support for s390 boxes > > mkfs.minix: > - supports minix version 3 > > mountpoint(1): > - this NEW COMMAND is libmount based re-implementation of mountpoint(1) from > sysvinit suite > > simpleinit: > - this set of deprecated utils has been REMOVED > > Wall(1): > - support new command line option "--timeout" to specify write timeout to > terminals in seconds. > > It's not supported to link with external (e.g. from e2fsprogs) libblkid and > libuuid any more. >

1 0

[arch-dev-public] PHP 5.3.7 and Arch Linux
by Pierre Schmitz 23 Aug '11

23 Aug '11

Hi all, as it is all over the press now that one should not update to PHP 5.3.7 due to a crypt/md5 bug* I feel the need to advice every Arch user to update to PHP 5.3.7-3 and not to keep 5.3.6. That version has the simple patch applied while 5.3.6 has known security issues. I expect 5.3.8 to be announced today though; but who knows? :-) Greetings, Pierre *) https://bugs.php.net/bug.php?id=55439 -- Pierre Schmitz, https://users.archlinux.de/~pierre

1 0

[arch-dev-public] Archweb login cookies are now secure-only
by Dan McGee 23 Aug '11

23 Aug '11

I switched cookies in the developer interface to be secure-only as of this morning. This means you won't see any of the developer stuff if you are viewing the site over plain HTTP once your existing cookie expires. You will be redirected to HTTPS if you are required to log in, but devs/TUs may not see the links they are used to if you are not viewing over a secure connection. -Dan

1 0

[arch-dev-public] [signoff] isl-0.07-1, cloog-0.16.3-1, gcc-4.6.1-4
by Allan McRae 23 Aug '11

23 Aug '11

This is going to hit [testing] as soon as gcc-multilib gets a rebuild so the signoff thread is slightly early... isl-0.07-1 - upstream update, soname bump cloog-0.16.3-1 - upstream compatibility update for isl bump - has its own soname bump... gcc-4.6.1-4 - new upstream snapshot from 4.6 branch (there are some nice bug fixes on the way to the 4.6.2 release) - libstdc++ man page update - disable libssp building - it is not needed on glibc systems The only other package requiring rebuild for the isl/cloog soname bumps is gcc-multilib. @heftig: I am leaving that to you. Can you move isl/cloog/gcc from [staging] to [testing] once you are done? The only package require rebuilt for removing libssp is avahi and I have put that in [testing] now. Allan

2 1

[arch-dev-public] [signoff] fakeroot-1.17-1
by Allan McRae 22 Aug '11

22 Aug '11

Minor upstream bug fix release. Signoff both, Allan

3 2

[arch-dev-public] [isgnoff] linux-firmware-20110822-1
by Tobias Powalowski 22 Aug '11

22 Aug '11

Hi, bumped to new intel wireless firmware. greetings tpowa -- Tobias Powalowski Archlinux Developer & Package Maintainer (tpowa) http://www.archlinux.org tpowa(a)archlinux.org

1 0

[arch-dev-public] [signoff] lvm2/device-mapper 2.02.88-1
by Eric Bélanger 22 Aug '11

22 Aug '11

Hi, lvm2/device-mapper 2.02.88-1 are in testing for upstream update. Please test and signoff. Eric

2 2

Re: [arch-dev-public] [arch-general] gnutls 3.0.0 pushed to testing
by Allan McRae 21 Aug '11

21 Aug '11

On 21/08/11 18:50, Andreas Radke wrote: > I'm running into similar issue running now valgrind tests when building > gnutls. The gnutls upstream developer says it's probably a bug in > glibc headers breaking memcpy. > See also this Debian report mentioned in the gnutls bug: > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=619963 That bug report seems to be only with gcc-4.6 and (e)glibc < 2.13 so does not seem relevant to us. Other than that, I have no idea what the context of this message as you did not quote anything, so can not look into the issue further... Does it get fixed with gnutls-3.0.1 which I note was released 12 hours ago? Allan

1 0

[arch-dev-public] New CFLAGS/LDFLAGS plus complete toolchain rebuild
by Allan McRae 21 Aug '11

21 Aug '11

This has been discussed a couple of times previously on the mailing lists and there were no objections so I have finally gotten around to adding some hardening options to our CFLAGS/LDFLAGS. With pacman-3.5.4-4 the defaults in makepkg.conf become: CFLAGS="-march=i686 -mtune=generic -O2 -pipe -fstack-protector --param=ssp-buffer-size=4 -D_FORTIFY_SOURCE=2" LDFLAGS="-Wl,-O1,--sort-common,--as-needed,-z,relro,--hash-style=gnu" As discussed previously, the addition of -Wl,-O1,--sort-common to LDFLAGS is not hardening... but these are safe options and they do appear to more than counter the slight overhead that stack smashing protection adds. These are all fairly standard flags being used to build the major distros these days (other distros patch their toolchain to make these the default), so there should be few issues. Probably the only thing to watch out for is to disable them when building bootloaders. The toolchain and all its (real) dependencies has been rebuilt with these flags and the necessary adjustments made to the packages. See notes below: All toolchain dependencies (just rebuilds): cloog-0.16.2-2 gmp-5.0.2-3 isl-0.06-2 libmpc-0.9-2 mpfr-3.0.1.p4-2 ppl-0.11.2-2 zlib-1.2.5-4 Toolchain components: linux-api-headers-3.0.1-1 (upstream update) binutils-2.21.1-2 gcc{,-libs}-4.6.1-3 (do not build libssp with hardening flags) glibc-2.14-5 (do not build libraries with hardening flags) I intend to leave this in [testing] for a couple of weeks to make sure there are no issues. I have been running this locally for about a week and am fairly sure I have the kinks worked out now... I will call for the sign-off later. Allan

7 10

[arch-dev-public] [signoff] usbutils-004-1
by Tobias Powalowski 20 Aug '11

20 Aug '11

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

3 2