How to avoid node spiralling out when rotating node around point in a circle via translate()/move()

Elf_Ears

When I rotate an object around a point via translate() using get_node("INSERT_NODE_HERE").translate(Vector2(-1,0).rotated(get_node("INSERT_NODE_HERE").get_global_pos().angle_to_point("INSERT_POINT_HERE"))*speed*delta) the node rotates around the point in a circle, but it also begins to spiral out from the center, is there any way to fix this?

Here is a gif of my problem imgur.com/zKoq232

EDIT: This also happens with move() with KinematicBody2D

PaulMurray

Just a guess but is translate(Vector2(-1,0) being applied every frame, causing the location to spiral out?

Elf_Ears

@PaulMurray said: Just a guess but is translate(Vector2(-1,0) being applied every frame, causing the location to spiral out? yes, it's in _fixed_process(delta), is there some reason that this causes the problem? will test with KinematicBody2D to see if it also happens with that...

Elf_Ears

can confirm KinematicBody2D has this problem with move() (will add to OP)

joelns

Well, looking at your code, I suppose you are rotating a object by moving it, every frame, a little amount in a direction perpendicular to the line connecting that object to a point. This won't result in a perfect circular movement because, according to the pythagorean theorem, the new distance to the point will be slightly bigger than the previous one: [(new distance)²=(old distance)² + (dX)²] I think way to fix that is just making the object follow circular path centered on the point

Elf_Ears

@joelns said: Well, looking at your code, I suppose you are rotating a object by moving it, every frame, a little amount in a direction perpendicular to the line connecting that object to a point. This won't result in a perfect circular movement because, according to the pythagorean theorem, the new distance to the point will be slightly bigger than the previous one: [(new distance)²=(old distance)² + (dX)²] I think way to fix that is just making the object follow circular path centered on the point

Yeah, I'd say that would probably be the best hassle free way to do this, thanks!