> You'd expect the attempt to grow the stack to be made only *after* > keep_going hits zero.
Only if you thought you knew better than the compiler :) > I'd rather not have the compiler presume > certainty of a 10GB stack allocation, especially not if it's actually > pretty unlikely. It could still be done, even if we inline. There is nothing that prevents us from adding space to the stack allocation only at that point, it's just not coded in gcc to do that.
