On 11/01/2010 10:37 PM, Jason Reardon wrote:

On Mon, Nov 1, 2010 at 10:29 PM, Allan McRae wrote:

...

On 02/11/10 12:19, Jason Reardon wrote:

...

On Mon, Nov 1, 2010 at 5:27 PM, Gaetan Bisson wrote:

[2010-11-01 19:26:31 +1000] Allan McRae:

...

...

db-5.1.19-2 heimdal-1.3.3-2 iproute2-2.3.25-2 libsasl-2.1.23-5 pam-1.1.1-2 perl-5.12.1-3

signoff i686

-- Gaetan

I installed db from testing. Now when I run `pacman -Qu --dbpath /var/lib/pacman` I have one update as 'db 5.1.19-2' (which is currently installed from testing). However, when I run `pacman -Su`, it wants to install 'db 4.8.26-2'. When updating to the testing version, there's a message: 'Consider running db-upgrade on Berkeley DB databases.' I'm not at all familiar with the db package, so I'm not too sure what to do with this, if this is what's causing the problem. Seemed kind of strange anyway, and worth mentioning.

Testing is all or none. You do not just install a single package from [testing] and then disable it unless you know what you are doing.

Allan

I'm somewhat of a mailing list lurker. I've read before that one thing Arch is lacking in are normal users testing packages. So, I wrote a script to install packages from testing when someone asked for a sign-off, so that I could give my input if needed. But, just for clarification, if I wish to contribute in this manner, I need to run a machine solely from the testing repos?

You can get away with excluding groups of packages from testing, but I don't think there's a good way to automate the process. I don't think there is any real strong desire to do so either. (If there was, a solution might be to have groups in testing which define packages in a related rebuild project like texlive recently, xorg-server upgrade, lib* upgrade...)

2010/11/1 Ng Oon-Ee

...

On 11/01/2010 10:37 PM, Jason Reardon wrote:

...

I'm somewhat of a mailing list lurker. I've read before that one thing Arch is lacking in are normal users testing packages. So, I wrote a script to install packages from testing when someone asked for a sign-off, so

On Mon, 2010-11-01 at 22:42 -0400, Matthew Monaco wrote: that I

...

...

could give my input if needed. But, just for clarification, if I wish to contribute in this manner, I need to run a machine solely from the testing repos?

You can get away with excluding groups of packages from testing, but I don't think there's a good way to automate the process. I don't think there is any real strong desire to do so either. (If there was, a solution might be to have groups in testing which define packages in a related rebuild project like texlive recently, xorg-server upgrade, lib* upgrade...)

Additional dev work for no real purpose. As a user with only one main machine, I run [testing] and don't experience much problems. Some basic knowledge on downgrading is of course needed.

Just use [testing] in totality and maybe script a way to reverse any particular update (for example, 30 packages updated, only reverse those) if there's problems.

Ok, cool. Thanks for the info. -- Jason