On Mon, Sep 14, 2009 at 05:28:54PM -0700, Jesse Keating wrote: > > Is there a way we could make the depsolving done during composes non-fatal > > if it hits something like this ? As long as I didn't try an encrypted > > install, things probably would have worked no? > > That seems like just kicking the can down the road instead of actually > fixing the problem. > > Since you can use the images from the last good tree with any day's repo > makes the non-existence of images less of a major problem. Yes there > was a new anaconda, and a new kernel, so it's not ideal, but I still > think that fixing the original problem is better than trying to work > around it by accepting broken deps. This was exactly the reason I proposed this idea. We lose a days testing of the kernel & installer far too often because of some inconseqential dependancy breakage. "oh well, try tomorrow" gets old real fast during times when we have different depfail every day. I agree we shouldn't _have_ dependancy failures, but as we don't have anything in place to prevent them, I think it's like asking for ponies at this point. Dave -- fedora-test-list mailing list fedora-test-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-test-list
