Use input override global position if set to calculate diff.

[Boxfort]

Ive made a change to use the input override position to calculate the rope interaction diff if it is set.
As far as I can tell this is a bug?

When the override position is set without this change the CharacterBody2D on that side of the RopeInteraction will be pushed towards the input override position point continuously.

I could also have been using it incorrectly.

[mphe]

Thank you for your contribution.
The behavior is not a bug, it's intended to work that way and documented as such:

GDNative-Ropesim/demo/addons/ropesim/RopeInteraction.gd

Lines 42 to 44 in 52ec1ec

	## (Optional) Use the given node instead of the target node to update the rope point's position.
	## If set, the target node will only be snapped to the rope, but it will not affect it.
	@export var input_node_override: Node2D

In other words, you can use input_node_override when the node that should snap to the rope and the node that should affect the rope are different. target_node will be snapped, but input_node_override won't.
One use-case is shown in the rigidbody_rope_attachment (and rope_with_rigidbody) example scenes. The rigidbody should affect the rope but we can't just snap the rigidbody to the rope. Instead, a PintJoint2D has to be used.

However, it is not unlikely that you might have found a missing feature.
Could you tell me a bit more about what you are trying to accomplish, possibly with an MRP?

[Boxfort]

Hopefully this is a clear enough explanation of what I was attempting:

I have a player scene, with a rope as a child. During gameplay I'm attaching the end of the rope to another object to pull it around. I'd like to have the rope originate from the player at a specific point, without having to offset my sprite.

I first attempted to set the Input Node Override on the Start interaction node to the point at which I wanted to rope to originate from on the player, but doing this caused the player to be pulled from the origin of the CharacterBody, towards the point I had specified.

Implementing the change I made meant that the player CharacterBody could still drag using the rope, but the rope originated from somewhere other than the middle of the CharacterBody.

[mphe]

I see, this is indeed an oversight.
I can think of two ways to implement that feature:

• A) Adding an offset property to RopeInteraction which is applied during _on_before_update and _on_after_update. Also the on_movement_request signal should be changed to include the target position as a third argument, now that the computation is more than just a position delta.
• B) Adding a boolean use_input_override_as_offset (or something similar) property which uses your snapping method when true, otherwise the current one.

B) would be much simpler to implement but I find the input_node_override property already confusing enough. I like A) much better as it's more straightforward usability-wise.
What do you think?

Would you like to update your PR with these changes or should I make a fix?