Yeah we should agree and document it. I think I have leaned towards pushing the final off if it means final will be better.
I guess I am inspired by how the Linux kernel is handling their RCs. If we see a lot of changes between two RCs they do another one until the RC merge requests slow down and stabilize. So in the end we will end up with RCX->final that is very small, because we have merged all fixes.
IMHO having a solid final is more important than hitting a specific date or number of RCs. BUT I think it’s more important that we do the same - so I am open to do it another way.