The Bayou code has been out there for a while now, and our European friends are starting to come back online after exams (welcome back!), so this is an excellent time to discuss the future.
My immediate goal for Bayou is to get it checked in. I know that some of the code is a duplicate of other code in the project (especially the LAR stuff) but I think we can clean those up over time. As it stands, Bayou works fine with the built-in LAR, so we don't need to solve the SELF/ELF right away. For this to happen, we need code review (pick those nits!).
Eventually we do need to solve SELF/ELF if we want to make Bayou work with the boot LAR in v3. The behavior of the 'pbuilder' utility will also depend greatly on the outcome of a LAR redesign. Furthermore, if LAR is redesigned, I think we should consider a generic LAR library, as I have very crudely prototyped in the current Bayou code.
Second goal is to make it work with buildrom. This is a big deal, and I honestly have _zero_ idea how to make it happen. Suggestions welcome.
Now that we have chaining behavior implemented, the eyes of most developers will move to the second part of the 'traditional solution': a setup utility. We need somebody to step up and volunteer to own and drive that particular part of the project. I hope that libpayload and Bayou are mature enough to make it easy, or at least to make it so you don't trip over the same pitfalls that the other payloads seem to hit. And if they aren't mature enough, then know that we have text editors and we aren't afraid to use them.
Thank you for listening. Jordan
how about we do the code review first, I assume you want bayou to be its own top-level svn project?
ron
On 25/06/08 20:12 -0700, ron minnich wrote:
how about we do the code review first, I assume you want bayou to be its own top-level svn project?
It should live with the other payloads.
Jordan
-
Jordan Crouse -
ron minnich