On Fri, Oct 10, 2008 at 2:24 AM, Andy Green <[EMAIL PROTECTED]> wrote:
Hello Andy. > Werner Almesberger wrote: >> Andy Green wrote: >>> It seems some other folks are interested to take care about making sure >>> we are clean for config variations, I'll leave this to them. >> >> Sounds like a good plan. Since Nelson should soon become the one who >> will find config conflicts long before anyone else would, I think it > > Being reasonably config-clean is important, but it's basically the > number of possible config combinations is immense and sitting there > studying them all doesn't really move us or Nelson on. I'm sure we'll come up with something fast. I know I should start doing "real work" soon :-) > Only a small subset of the combinations are actually realistic, I think > Nelson's plan of having a whitelist of configs is probably the most > workable. It also has the advantage that he can set it up and move on > to more interesting (for him and us) things, we can just tack configs on > the list as we find new reasonable cases for them. Well, we could do something like: # Permute CONFIG_OPTION_A=y And run a small preprocessor to select the permutations we actually care about. We can use a hash to avoid duplicates. But let's start with the white list. I'll check the project Sven sent. Regards, Nelson.-
