Rod Whitby wrote: > Am I the only one who thinks that Openmoko should be concentrating on > getting a *single* kernel configuration stable and working flawlessly > under all circumstances before worrying about sophisticated tools for > automatically testing random kernel configuration variations?
Perhaps there's some confusion about the goals of this little sub-project: it's about making sure that our configuration dependencies are sane. One aspect is that we need to do this to keep our code from causing trouble when submitting into mainline. Another aspect is, as the comparably large number of problems that popped up in recent weeks confirm - we also have to put some more work into allowing people to make changes to the kernel configuration without suffering bad surprises. > Then perhaps you can move closer to the goal of a single openmoko kernel > that *all* distros use, I gave on up on that concept long before anyone even dreamt of Openmoko. Sure, the One True Configuration is a control freak's wet dream, but it just won't happen if your project is successful. Many people choose Open Source for the freedom it gives them. And some of them are going to use it. Now you can waste a lot of energy on convincing them that their choices are wrong, or you can just make sure their ostensibly valid choices won't cause your code to blow up. While the pop corn lasts, it's always fun to watch someone take the first choice. What we can - and should - do is to provide guidance. But that doesn't mean that there won't be equally valid but different points of view. There are also points where one can only agree on disagreeing, but I would hope we can push those as far into the fringe as possible. > And who knows, once that's done, maybe you can use that as a stable base > for a single openmoko distribution instead of the 8 or 10 that there are > currently ... I think the possibility of having many distributions is actually a very good thing. But it would be nice if they were for different audiences, not just trying to cover up each other's potholes or craters. E.g., if someone made an "embedded controller" distribution, without any GUI, I think nobody would complain if that one had a different kernel configuration than a "general-purpose phone" distribution. I understand that there's a lot of frustration about Openmoko's failure to meet customers' expectations about our products so far. But we're, keenly aware of where the problems are, and we're working on solving them. And sometimes, the most direct way leads you in a wide circle, and not right into the lake, no matter what your navigation system may tell you :-) - Werner
