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

[arch-dev-public] Time to go forward with arch=any?
by Firmicus 20 Jul '09

20 Jul '09

I've just prepared pre-release packages of TeXLive 2009. They should be ready for release (upstream) within the next 4-8 weeks. Before that, I'd *really* love to benefit from arch=any. That would save considerable space on gerolde, as well as bandwidth, and would also speed up my task with building/uploading. Any prospect of making the transition to the new dbscripts soon? We could use the testing repo first to check that everything runs as expected. F

5 11

[arch-dev-public] [signoff] [extra] mutt-1.5.20-1
by Thayer Williams 20 Jul '09

20 Jul '09

A few bug fixes and a number of enhancements. See http://www.mutt.org/doc/devel/UPDATING for details. It looks like there are some optional deps listed as mandatory, but I didn't go mucking with them as I sort of borrowed this package *ahem*. Signoff please.

1 0

[arch-dev-public] [signoff] dhclient-3.1.2p1
by Daniel Isenmann 19 Jul '09

19 Jul '09

Please signoff both architectures. The new version fixes CVE-2009-0692 (https://www.isc.org/node/468) I have cleaned up the PKGBUILD and removes our patches. So please test it carefully and report any issue to the bugtracker or the ML. On works here on my PC, but it needs definitely more testing. Daniel

1 0

[arch-dev-public] Need some advices for new dhclient
by Daniel Isenmann 19 Jul '09

19 Jul '09

Hi, dhclient is out-of-date and the last update to the package was from September last year. In order to fix this bug: FS#15515 - [dhclient] DHCP Stack Overflow in 'dhclient' script_write_params() (http://bugs.archlinux.org/task/15515) I get some troubles. At the moment our package is at 3.0.6, but ISC doesn't provide a patch for the above stack overflow, instead they adivse you to upgrade to 4.1.0p1, 4.0.1p1, or 3.1.2p1. I have tested 3.1.2p1 and our own huge patch (http://repos.archlinux.org/viewvc.cgi/dhclient/repos/extra-x86_64/arch-nm.p…) doesn't apply anymore. Nevertheless I have absolute no clue what this patch is doing completely. I have seen that it adds an additional option on command line "-x", but I don't know what it's doing. The other two patches from us are more or less needless with the new versions. This (http://repos.archlinux.org/viewvc.cgi/dhclient/repos/extra-x86_64/fix-broke…) was fixed in the new version from ISC and the other patch changes: - ifconfig $interface inet 0 down + ifconfig $interface inet 0 in the scripts and I really don't know why. The patch is called "fix-mac80211-wireless". Nevertheless of the patch problems above, I have tried the newest version 4.1.0p1. It's clear that our patches don't apply anymore, but I think they aren't necessary anymore. Like in 3.1.2p1 the last two patches are needless and the "arch-nm.patch" doesn't apply and I have not enough knowledge about the dhclient source code here. As I said, I have test the 4.1.0p1 and it compiles fine, BUT only if I disable DHCPv6. I haven't tried the package yet, but I'm not sure if we should deliver it without DHCPv6. Otherwise I don't know how many servers uses DHCPv6 at the moment or will use it. Should we provide the 4.1.0p1 package, but without DHCPv6 support? I have looked at some other distros, they have enabled DHCPv6 and it seems that it compiles. But I don't know if they have the latest toolchain. Your advice please...need some input here. Daniel

2 3

[arch-dev-public] Multiple checksum support
by Jan de Groot 19 Jul '09

19 Jul '09

Recently the ftp-release-list for ftp.gnome.org has switched over to sha256sums, which means that I can't copy and paste checksums from the mailinglist anymore. I see that makepkg supports sha256 without problems, but when I enter sha256sums=('somesum') into a PKGBUILD without modifying makepkg.conf to use them, makepkg warns about missing checksums. Shouldn't makepkg check for any supported checksum type by default? If sha256sums is defined, it should check them. If sha1sums is defined, it should check them too, etc. This makes more sense than depending only on the INTEGRITY_CHECK variable set in makepkg.conf.

3 2

[arch-dev-public] mesa 7.5 is in testing
by Jan de Groot 19 Jul '09

19 Jul '09

I've put mesa 7.5, xorg-server 1.6.2, xf86-video-intel 2.7.99.902 and a rebuild/update of all drivers containing DRI drivers in testing. The nouveau package was modified to use the latest version of libdrm instead of building from git, as the libdrm version is actually git master at this moment. I haven't included Radeon Gallium3D support yet, but we might put this into testing in a later stage. The libdrm package contains all the required libraries and changes for this. Please give these packages a test, so we can move them to extra soon.

3 2

[arch-dev-public] [signoff] usbutils-0.84-1
by Allan McRae 19 Jul '09

19 Jul '09

Upstream update. Signoff both arches. Allan

3 2

[arch-dev-public] [signoff] shadow-4.1.4.1-1
by Allan McRae 19 Jul '09

19 Jul '09

Upstream update. Fixes FS#15217. Sign-off both arches. Allan

3 2

[arch-dev-public] [signoff] vc/* -> tty* transition
by Allan McRae 18 Jul '09

18 Jul '09

Signoffs needed for the following packages: filesystem 2009.07-1 initscripts 2009.07-2 (what I tagged as -1 did not shutdown correctly) syslog-ng 3.0.3-2 udev 141-4 Allan

4 6

Re: [arch-dev-public] [arch-general] [signoff] vc/* -> tty* transition
by Allan McRae 18 Jul '09

18 Jul '09

Gerardo Exequiel Pozzi wrote: > Allan McRae wrote: > >> Signoffs needed for the following packages: >> >> filesystem 2009.07-1 >> initscripts 2009.07-2 (what I tagged as -1 did not shutdown correctly) >> syslog-ng 3.0.3-2 >> udev 141-4 >> >> >> Allan >> >> >> >> >> > For udev missing the last patch for PKGBUILD [#1] in my last-1 comment [#2] > This was deliberate - I did not want to deal with needing a -Sf for udev when these four packages are kind of tied together with this update. > And for initscript the patch for fix the rtc path [#3] (oops!) > This is not too much of an issue so I will hold of fixing for a bit in case other issues occur. Will push an update before all this moves out of [testing]. Allan

1 0