On Wed, Dec 20, 2017 at 05:44:31PM +0000, Holger Levsen wrote: > On Wed, Dec 20, 2017 at 05:13:26PM +0100, Andreas Tille wrote: > > I can confirm that it also affects arch:all packages. But why shouldn't > > it be possible to detect this automatically also in this case? > > because a build on any architecture will make the arch:all package > appear and then you cannot know whether it's a known missing feature > that the package doesnt build on $arch
Sure you can know. You can calculate from UDD on what architecture you can create the arch:all package and on what you can't. > (or a new problem) or you would > need to introduce state tracking and then you still wouldn't know > whether a build failure on $arch is a new problem or an old problem, > because maybe it's a temporary problem... Since you can also query UDD whether a missing package has was there (at least on a previous release) you can know this as well. May be I should write an according query and than close all those bugs ... Kind regards Andreas. -- http://fam-tille.de