The temp-solid solution of saiyadjin works pretty well, but only using solids as obstacles. These are moving obstacles, where the moving should use each other as obstacle, but not themself. So there is heavy CPU-load and I dont think this is the right solution.

I also tried your idea, but no success so far. Maybe you can provide a more detailed event.

Could it be that obstacles can not be set per instance (when not using solids as them) ?

Thank you very much.

I tried this before, but did not pick all object instances before the evaluation.

Is it really needed to pick all before, and why?

Hi,

I just can clear the whole obstacle stack. How to remove single obstacles, to refresh the map with less of them?

Thank you very much!

Thank you.

How did you set only other instances as obstacles? If I add all, they see themself as obstacles too... so they can't find any path, being blocked by their own instance.

Another way would be to remove an obstacle from the obstacle map. But I can't see a way to do that. I might create another thread for this problem.

Would be nice if you could share a capx.

MultipleChoice

This helps a lot, many thanks!

Hi,

maybe this has been asked before, but how can I get the instances (IID, UID or custom variable) of 2 colliding objects of the same type?

[attachment=0:1nzetvxe][/attachment:1nzetvxe]

Thank you very much!

A condition as "Select all instances, but (self.IID)" would help a lot.

[quote:30euxzz6]Each displays it's own IID. When the layout starts up, both objects have IID 0,

Thats the problem. I need to access them through family(iid).value, not family.object(iid).value - I dont need families for this expression object(iid).value.

Where is the sense of a family.IID value when each object types inside are seperated in a different order?

And again, it works for DIFFERENT objects created during runtime. Then the order is done for all family objects... no matter what type of.

I see the problem in creation of new family members... which is also random.

Maybe the Scirra team can get an answer how to fix this without extra instance variables.

All needed is a realtime IID management for family objects which can be accessed through the same id as the objects inside.

Anyway, Ashley told me to fix that soon.... before few months

At the end, its the same sh*t as my problem

The c2 families are useless, for RTS unlimited you need kind of temp objects which hold all values for the childs as a transparent family dad.

This makes the event sheets rock like a hippie concert on valium.

Add a third transparent object "touchable" with the relevant values in an upper layer. Link them (or dont) maybe via Pin behavior to those "real" objects when needed.

Thank you for the reply.

I can't confirm that. For example, Family(IID).width does only work on objects generated during runtime.