On Mon, Jun 14, 2010 at 09:43:40AM +0300, Panu Matilainen wrote:
If that known-to-build version fails to build with 4.8.1 then sure,
we appear to have an rpm regression no matter how unlikely it should
seems to me given the changes in 4.8.0 -> 4.8.1. But until proven
otherwise, I'd say the kernel.spec cleanup efforts have broken the
noarch case. It's not exactly hard to break a spec of that
complexity :)
Looking at rpm, I was inclined to agree, so I poked more. Looks like
davej's changes last week broke it somehow... I'll puzzle it out.