On Thu, Mar 20, 2014 at 03:21:05AM +0100, Martin Bähr wrote: > Excerpts from Michael K. Johnson's message of 2014-03-20 02:44:41 +0100: > > On Wed, Mar 19, 2014 at 02:51:05PM +0100, Martin Bähr wrote: > > > Excerpts from Michael K. Johnson's message of 2014-03-19 11:40:22 +0100: > > > > 1. rpmbuild adds functionality to invoke cvc and converts the > > > > resulting build output to a standalone RPM. > > > > 2. conary controls the build process, creating a spec file that > > > > hands control during the build process back to conary > > > > > > > The second possibility could make RPM packaging > > > > a lot easier, but does not create a standalone RPM. > > > > > > what does it create then? > > > from a fedora dev perspective, if it doesn't create an rpm how is it rpm > > > packaging? > > Sorry, I meant doesn't build a standalone SRPM. So it couldn't > > integrate into normal RPM build processes. > > ah, ok, so both methods do create standalone rpms. > but then i am unclear what goes in.
My point is that conary controlling RPM, which matches what you described, wouldn't fit into their existing build system. That would make it not an evolutionary path. Using cvc as an intervening supervisor to make packaging easier would conceivably fit into existing build systems. An evolutionary path is the only way that could possibly be interesting to them. "Scrap your whole build infrastructure and try something new" is not going to get much interest. _______________________________________________ Foresight-devel mailing list [email protected] https://lists.foresightlinux.org/mailman/listinfo/foresight-devel
