On 12/08/10 02:24, Pierre Schmitz wrote:
On Wed, 11 Aug 2010 10:56:28 -0500, Aaron Griffin <aaronmgriffin@gmail.com> wrote:
On Wed, Aug 11, 2010 at 10:34 AM, Thomas Bächler <thomas@archlinux.org> wrote:
So, we are adding an extra layer of confusion which there was no need for in the past. What is the problem you are trying to solve?
As I understand it, the point is to help with overly large rebuilds, allowing for multiple developers to work in parallel better. Is this correct, Pierre?
Exactly. In fact we already used this idea somehow when we had a jpng repo to prepare the libjpeg/libpng rebuilds or when we virtually freeze testing for a while and warning people to not use it atm because a rebuild is going on.
My approach is simplify this and make it easier for us; so we can even use it for small rebuilds with just a dozen of packages involved.
Let me underline again that [staging] would be no regular repo that would be used by anyone directly. It mainly meant for collecting rebuilds. Till now we either had a special repo, broke testing or one dev had to do the rebuild completely on his own. Using a common use case for these should make our lives easier.
I think this is worth implementing. No matter how often we post to hold off updating from the [testing] repo while we do a big rebuild, people still do and break their system. A staging repo would avoid that. My only question is, how soon can this be rolled out? I intend the big, big, big python rebuild to start as soon as KDE is pushed to [extra]. Allan