On Fri, 2011-02-18 at 07:09 -0800, John Reiser wrote: > On 02/18/2011 04:22 AM, Panu Matilainen wrote: > > Traditionally yum has used (rpm-header, path) tuple as the "key" for > > callback to locate the package to-be-installed, but this has the > > downside of keeping the full rpm headers of all installed packages > > to be loaded throughout the entire transaction, wasting (tens of) > > megabytes of memory when it only really needs is nevra info to locate > > the relevant yum package object. > > Not only that, it wastes 2.5 minutes during a fresh install, > extracting the header from each .rpm into a .hdr file. > For a DVD the constant seeking is hard on the drive. > https://bugzilla.redhat.com/show_bug.cgi?id=522838 Anaconda is still downloading the pkgs one at a time right and then installing, right? If so then it will still need the hdrs do to the file-conflicts check. Not sure how the above patches get around that. -sv _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
