On Tue, 2004-02-24 at 01:57 -0500, seth vidal wrote: > On Tue, 2004-02-24 at 00:55 -0600, David Farning wrote: > > > I am looking at HeaderInfoNevralLoad and trying to determine why you do > > not do a determination for bestarch arch here. It could significantly > > reduce the nerval size. > > > > Will there be reasons downstream where you will want to use a less then > > optimal arch? > > exactarch being set differently. > > -sv OK, the determination of which arch to use cannot be made until we are knowledgeable about the arch of a existing package in the rpmdb? I assume there are cases where a less optimized compile runs better on a given system? Dave
