On Fri, 2011-04-29 at 16:34 -0400, seth vidal wrote: > On Fri, 2011-04-29 at 16:32 -0400, James Antill wrote: > > Huge perf. gain, now takes ~60% of time to get to y/N choice. 45% > > when obsoletes are off. > > > > Somewhat ugly workaround needed to get obsoletes data. > > > > Also fixes bug when obsoletes are off and we install something which > > obsoletes things (they still need to be obsoleted, as rpm will do it). [...] > NACK. > > obsoletes are never really off. > > rpm will honor them if we even just have the pkg in the ts - whether or > not obsoletes=0 or not.
Yeh, that's what the bug fix does (3rd comment). > I don't think we should be making code exceptions for things which > encourage people to set variables which aren't actually DOING what they > think they are doing. All it does is make "yum install zzuf" faster, obsoletes are still done. All the 2nd comment refers to is how we get rawobsoletes in _getUpdates, when we've already got it for the install case (it's a bit hacky, but it seemed better than creating a real shared "obsoletes" object). _______________________________________________ Yum-devel mailing list [email protected] http://lists.baseurl.org/mailman/listinfo/yum-devel
