Last week end, I finally got to the point where I wrote some script to give it existence in the ongoing 3-room demo. There's a thing that went significantly wrong, though: Bilou couldn't notice when the platform is gone.
I've had a similar issue with dumbladors when time came to 'stack' them: when one recovered, those stacked on top of it had somehow to detect they'd no longer have a platform behind them. It would be trivial if the gameobject implementing the platform was destroyed, but here it stays around.
- extend dumblador's extra
gobbit(when 7.1)
to Bilou - trigger an extra "throws up" temporary hitbox when flapping down the platform
- implement it all with brand new "larger animated blocks" feature, not with a sprite, because that platform doesn't truly "move".
Yet, I hesitate going for an additional micro-controller gobbit
(despite it already has C++ code) because that would introduce new events to deal with. I'd rather have the onpath
controller be able to fire an event when the platform-object it is attached to signals it is no longer a platform, the same way the controller can signal the object no longer exists.
I could try having some automated test into 'onpath' that would ensure the current state of the 'platform' object still has the area active and using the proper flags, but that would require extra care if I don't want Bilou to behave weirdly when the platform turns back or does any state transition that don't actually imply losing its 'platform' behaviour.
reality check: using onpath(when w7.2);
(from the worst-possibly-named commit)