On Wed, Nov 4, 2009 at 11:50 AM, Jesse Keating <jkeating(a)redhat.com> wrote:
Until we add a new arch. But that still leaves things like java,
mono,
ruby, etc as problem areas where "noarch" may not actually be
"noarch".
We seem to be using "noarch" in two different senses:
1. Contains no machine code, other architecture-specific bits, or
build-system-specific artifacts (like build timestamps, build machine
names, etc.)
2. Can be built/installed/consumed on any architecture.
Those aren't the same. Since the addition of a new arch can break #2,
how can packagers mean anything other than #1 by "noarch"?
--
Jerry James
http://www.jamezone.org/