Rebuilding RPMs results in bad update behavior

[Jeff Spaleta]

On 6/12/07, Mike Kearey <mkearey at redhat.com> wrote:
> Also exactly what 'optimizations' are you doing? Simply rebuilding is a
> waste of time.

Yes, knowing what differences are there during the local rebuild
process. Since the distribution builders themselves are 64bit machines
its not even clear to me that there's a significant difference between
the system he's rebuilding on and the actual build systems.

Is he perhaps hitting some sort of oddity with prelink and or
readahead memory caching and misdiagnosing some sort of slowdown due
to library loading?  I'm only guessing, but without a concrete
procedure to attempt to reproduce and a benchmark metric to record
this is just noise with no way forward.

-jef