Le 14/11/2010 00:41, Matt Madia a écrit :
I think it would also need documenting what's not preventing it, ie, starting to think about what will be in R2. That would include multiuser support, switching to gcc4 as the main compiler (or llvm), and so on.Is anyone willing to have an earnest push for R1 (Final)? Basically, documenting the actual issues that are preventing R1 Beta's and for R1 (Final) to be released. (whether it be bugs, missing implementations/features, documentation, infrastructure, other tools ....)
So far it looks fine, but from my experience of open source development, this kind of discussions tend to end up in big unreadable wiki pages and ML threads and no action taken. Take care of it. I think it's time to start really triaging tickets in trac, creating the milestones (beta1 and R2) and moving tickets where they belong. It's still possible to move things around if other choices are made later, so don't wait forever, all you do can be undone.This has been started at http://dev.haiku-os.org/wiki/R1/PlansOfConquest This page could be used as a collaboration point, much like PackageManagerIdeas
I think R1should provide at least an upgrade path to R2. The ful-featured package manager with handling of 3rd party apps and other similar stuff can wait a little more. However, we need the 'upgrade path' solution in a beta version, to be able to test it between 2 releases (last beta and R1 Release Candidate / Golden Master) to make sure it works in real life.... Speaking of which, Package Management is the 'elephant in the room' as far as missing features. What can be done about this and who is willing to do it (even if finances are required?
The package manager grown so big no one is willing to start.Overall, the current 'plan of conquest' looks mostly fine to me. Maybe actually creating the milestones in trac will make some questions rise up, but then we can still change the plan a bit.
-- Adrien.