> Even if this looks more like an ad-hoc solution, I find it easier to > grasp and review than the trap-based diff. Works fine here and seems to > solve miod's problem. > > ok jca@
I concur. I'm slightly annoyed by the asymmetry between c_fc_depth++ vs c_fc_reset(), but I think we've had enough bikeshedding on this patch. Put it in. ok
