R0J0hound , actually, nevermind.. they're accurate.

sliding joints have been confusing me for a while. I think it must be something with my events..

R0J0hound , the positions aren't accurate for sliding joints.

R0J0hound , Okay.. I think I found a bug.. Seems like when you move an object that a joint is attached to, the joint anchor position properties don't get updated until the next event cycle. So if you use the joint position expression right after moving the object, it won't be the accurate location.

C2 is really great, and I enjoy working with it. Thanks Ashley.

Update 2.2

Download in first post.

* Fixed: Adding joints in to new objects in "post collide"

* Fixed: Movement actions for new objects in "post collide" were ignored.

Prominent

In "post collide" new objects aren't added to the simulation till right after the physics step. So they won't trigger a "post collide" till at least the next frame.

Awesome! My game is working as expected now. Thanks a lot!

If I come across anything else I'll let you know..

R0J0hound , okay here are some bugs..

first example: object created inside on post collision doesn't get positioned. If you re-enable the joint action, it does for some reason.. weird. Also doesn't get pinned.

[attachment=1:2rut6ghw][/attachment:2rut6ghw]

second example: object created inside on post collision doesn't get pinned..

[attachment=0:2rut6ghw][/attachment:2rut6ghw]

R0J0hound , I have a question about creating objects during the on post collide. When you create the same type of object as the object calling the post collide, and that object is created in the same position, size, as the original object- does that new object instantly call a post collide event before the current event is finished??

I'm having strange results when I do this, and it seems like it gets called more than it should- so objects keep getting created even when I check to see if the position is the same then don't add one, etc, it still does..

I commented out line 4277 and put in if(this.isStatic())return;

Not sure how that will affect other things, but it's allowing me to continue developing stuff for now.

R0J0hound , I get immediate error in my game now when collision happens.

Error: Assertion failed: Body.activateStatic() called on a non-static body.

at assert (http://localhost:50000/rojoChipmunkBeta ... or.js:51:9)

at Body.activateStatic (http://localhost:50000/rojoChipmunkBeta ... .js:4277:2)

at behinstProto.updateCOM (http://localhost:50000/rojoChipmunkBeta ... js:6967:10)

at behinstProto.updateShape (http://localhost:50000/rojoChipmunkBeta ... .js:6714:8)

at behinstProto.changeShapeCheck (http://localhost:50000/rojoChipmunkBeta ... js:7064:10)

at behinstProto.postCreate (http://localhost:50000/rojoChipmunkBeta ... .js:6882:8)

at Runtime.createInstanceFromInit (http://localhost:50000/preview.js:3247:28)

at Runtime.createInstance (http://localhost:50000/preview.js:2885:15)

at SysActs.CreateObject (http://localhost:50000/system.js:1337:33)

at Action.run_system (http://localhost:50000/eveng.js:1743:20)

edit: seems to happen if create object on post collision.

mattb , I just realized you can use query closest to point to find if a point is inside an object..

it's a condition though.. so can't use it like an expression. :\</p>

Prominent you might already have tried these, but for checking internal points there's the Pick overlapping point condition, & for moving them back out the Custom Movement behavior has lots of 'push out solid' options.

mattb , pick overlapping point doesn't work for tilemaps. But I can check what the tile is at a position, and if it is -1 it means it is outside of it. push out solid seems to push them out at closest direction, which may not be accurate. I'm using the normal to push it out so it seems to work.. Sometimes I get it still pushing it out at wrong directions, so maybe I can use push out solid if the normal sends it too far in weird cases...

R0J0hound , do contact points remain after the on post collision event? Since it only fires once, there's no way to reference the contact point afterwards unless you store it in an instance variable- that's what it seems like anyways.

Maybe there could be an expression that checks if a point is inside an object? Seems like a lot of cases the collision points are within the object, and I need a way to move them out, so I think testing if they're inside and then moving via their normal until they're out might work, but there's no expression to test if a point is in an object at the moment.