Rene Gollent wrote:
IMO it's exactly the other way around. If we're going to make a compat breaking change, I'd rather see it done all at once rather than breaking compat twice (I am generally not a fan of the Linux-esque 'break compat whenever we feel like it' approach). Besides, we already know it more or less works on gcc4, so in my opinion you're vastly overstating the complexity of the compiler transition. Especially given that it's already more or less a given that we'll be making compatibility-breaking changes to the API in R2, I'd rather see that done in conjunction with the compiler transition.
It is simply good software engineering practice, you do not do 2 completely unrelated major changes at the same time because then you have a lot more work to figure out if a problem showed up because one change or the other.
Anyway, as I said, I already stated my points. Whatever you guys decide to do I am fine with that but I am convinced that what I suggested would be for the best and the only drawback right now would be that we would probably need a bit more time (and someone willing to work on the GCC 4 port) and that there would be more bugs to fix during the alpha (which we would have to fix at some point anyway).
-Bruno
