When running "yum --skip-broken update" and getting this message (currently in Rawhide), should it always be considered a bug and reported? Normally, I can get some packages to update by attempting to update just a subset of them (say one at a time). Should I always try to update as many packages as possible (to reduce the transaction size) before reporting this?
On Sun, 2011-01-09 at 21:03 +0000, Andre Robatino wrote:
When running "yum --skip-broken update" and getting this message (currently in Rawhide), should it always be considered a bug and reported?
Depends what you mean. If you mean 'is it a bug in yum', no. There are some types of dependency issues that can't really sensibly be resolved by ignoring some updates. It's clearly a bug in *something*, though. You should treat each case individually, figure out what the ultimate root of the problem is, and file a bug against the appropriate package.
On Sun, 2011-01-09 at 21:59 +0000, Adam Williamson wrote:
On Sun, 2011-01-09 at 21:03 +0000, Andre Robatino wrote:
When running "yum --skip-broken update" and getting this message (currently in Rawhide), should it always be considered a bug and reported?
Depends what you mean. If you mean 'is it a bug in yum', no. There are some types of dependency issues that can't really sensibly be resolved by ignoring some updates. It's clearly a bug in *something*, though. You should treat each case individually, figure out what the ultimate root of the problem is, and file a bug against the appropriate package.
well some of them are a bug in yum. :) And, in theory, yes we should be able to walk it back.
there's a couple of rawhide bugs which are ugly to trace out right now.
-sv
seth vidal <skvidal <at> fedoraproject.org> writes:
Depends what you mean. If you mean 'is it a bug in yum', no. There are some types of dependency issues that can't really sensibly be resolved by ignoring some updates. It's clearly a bug in *something*, though. You should treat each case individually, figure out what the ultimate root of the problem is, and file a bug against the appropriate package.
well some of them are a bug in yum. :) And, in theory, yes we should be able to walk it back.
there's a couple of rawhide bugs which are ugly to trace out right now.
https://bugzilla.redhat.com/show_bug.cgi?id=668301
It was reassigned to shotwell after I guessed the wrong component (exiv2).
On Sun, 2011-01-09 at 18:14 -0500, seth vidal wrote:
On Sun, 2011-01-09 at 21:59 +0000, Adam Williamson wrote:
On Sun, 2011-01-09 at 21:03 +0000, Andre Robatino wrote:
When running "yum --skip-broken update" and getting this message (currently in Rawhide), should it always be considered a bug and reported?
Depends what you mean. If you mean 'is it a bug in yum', no. There are some types of dependency issues that can't really sensibly be resolved by ignoring some updates. It's clearly a bug in *something*, though. You should treat each case individually, figure out what the ultimate root of the problem is, and file a bug against the appropriate package.
well some of them are a bug in yum. :) And, in theory, yes we should be able to walk it back.
right, my answer was 'no it's not *always* a bug in yum', not 'no it's never a bug in yum'. sorry if that wasn't clear =)