Re: Selective Updates

2010-05-04 Thread Bernd Stramm
On Tue, 04 May 2010 21:08:23 -0400 Matt McCutchen wrote: > On Tue, 2010-05-04 at 17:44 -0400, Bernd Stramm wrote: > > Too costly with the current tools, I have no doubt. The method you > > describe doesn't look manageable, you're right. > > > > Perhaps I should go research packaging systems and

Re: Selective Updates

2010-05-04 Thread Matt McCutchen
On Tue, 2010-05-04 at 17:44 -0400, Bernd Stramm wrote: > Too costly with the current tools, I have no doubt. The method you > describe doesn't look manageable, you're right. > > Perhaps I should go research packaging systems and strategies, there > may be demand for more customizable systems. This

Re: Selective Updates

2010-05-04 Thread Matt McCutchen
On Wed, 2010-05-05 at 00:16 +0200, Kevin Kofler wrote: > Nor can you really stop us from closing all bugs filed > against the conservative official repo as WONTFIX with a comment of "fixed > in kde-redhat stable, please use that". If that practice became a problem, it would be escalated through

Re: Selective Updates

2010-05-04 Thread Kevin Kofler
Jesse Keating wrote: > On Tue, 2010-05-04 at 17:44 -0400, Orcan Ogetbil wrote: >> How about a de-centralized approach: main repo gets only critical >> fixes. Releng will be responsible only for this repo. >> Other than this, every major SIG gets their own update repo. They can >> choose to (but ar

Re: Selective Updates

2010-05-04 Thread Jesse Keating
nges without also accepting unrelated changes. > >> > Packages depend on each other, the same package receives changes of > >> > varying nature etc. So I am afraid what you're asking for is > >> > technically not possible, sorry. > >> > > >> &g

Re: Selective Updates

2010-05-04 Thread Jesse Keating
On Tue, 2010-05-04 at 17:44 -0400, Bernd Stramm wrote: > > I was thinking about filtering the updates suggested by the current > system. That is more or less what someone has to do manually today if > they want to experiment with advanced versions of a particular > package. That assumes that the

Re: Selective Updates

2010-05-04 Thread Orcan Ogetbil
and in general unsolvable with a single >> > repo. It is just impossible, with a single repository, to be able to >> > select one type of changes without also accepting unrelated changes. >> > Packages depend on each other, the same package receives changes of >> > vary

Re: Selective Updates

2010-05-04 Thread Bernd Stramm
> problem to solve in practice, and in general unsolvable with a > > > single repo. It is just impossible, with a single repository, to > > > be able to select one type of changes without also accepting > > > unrelated changes. Packages depend on each other, the same &

Re: Selective Updates

2010-05-04 Thread Jesse Keating
ble to > > select one type of changes without also accepting unrelated changes. > > Packages depend on each other, the same package receives changes of > > varying nature etc. So I am afraid what you're asking for is > > technically not possible, sorry. > > >

Selective Updates

2010-05-04 Thread Bernd Stramm
package receives changes of > varying nature etc. So I am afraid what you're asking for is > technically not possible, sorry. > Yes I'm asking for selective updates, I know that. That is a hard problem, I know. But it is not impossible. If the current package management tools d