Verlet integration for GMS2.3+

(Seems like I can't upload screenshots for some reason.. :c)

Problem 1:

The verlet system works by having a list of vertices and a list of sticks. Every update, we iterate through the list of vertices from the start, forcing the sticks to their original length. Since we're starting at the first vertex, its position will always be correct, but the rope will drag away from the last point as a result. One way to solve that might be to rewrite how we iterate through the list every update. So update the first vertex, then the last, then the second, second from last, etc. until we reach the center. I'm not sure if this will result in other unwanted behaviour though.

I'd manipulate this for loop and see what happens:

verletFunctions line 299:
for (var i = 0; i < stickAmount; i++) {
    currentStick = stickList[| i]; 

A simple and dirty "fix" would be to rewrite the draw function and draw another line from the last vertex to the point where it is supposed to be attached. It won't be perfect but it will look like it's attached.

verletFunctions line 497:
draw_vertex(currentStick.v2.x + lengthdir_x(wHalf, stickDir - 90), currentStick.v2.y + lengthdir_y(wHalf, stickDir - 90));
draw_vertex(currentStick.v2.x + lengthdir_x(wHalf, stickDir + 90), currentStick.v2.y + lengthdir_y(wHalf, stickDir + 90));
// Draw another line here
draw_primitive_end(); 

Problem 2:

When the verlet system updates, the "stiffness" is simply the amount of times we do the update. The more updates per frame, the more correct the simulation will be, but the more performace we draw.

verletFunctions line 294:
repeat (stiffness) {
    #region Update sticks
    #region Update vertices
} 

A stiffness value below 1 would therefore result in errors.

Problem 3:

That will be a difficult one.. I didn't think of a use case like this.

The sticks already have an individual length value. To retract the rope, you probably want to create some function that retracts the rope by one pixel. It would decrease the length of the first stick by 1 as long as the stick's length is greater than 1. A length of 0 would probably create errors. Otherwise delete the first stick and the first vertex, then continue with the next stick until the rope is gone.

New function for the rope group:
function retractRopeByOnePixel() {
    if (<amount of vertices> > 2 AND <amount of sticks> > 1) {
        var currentStick = <first stick>
        if (currentStick.length > 1) {
            currentStick.length --;
        } else {
            <delete first stick>
            <delete first vertex>
        }
    } else {
        return false;
    }
    return true;
} 

Something like that maybe. You'll need to figure out where to get the amount of ropes/sticks and where to delete them. Call the function until it returns false, then delete the rope.

Hope I could help :)

Thank you! I think I added an option to change the 'stiffness' of the verlet system so it doesn't stretch so much. You should be able to manipulate it either in the rope function or the function which sets up the verlet system. But do note that increasing the stiffness draws a lot of performance, as this basically multiplies how many times per frame the whole physics simulation gets updated.

Thank you :) The verlet system is drawn using the VerletSystem() struct's Draw() function. The depth would be equal to the object which calls the Draw() function. By changing either that object's layer, or where you call the Draw() function (depending on your case), you can control the depth at which it is drawn. There is however no functionality to sort different verlet groups within the same verlet system after they are created. Groups within a system will be drawn in the order in which you added them with the VerletGroupCreateRope() or other functions. If you need to change them dynamically, you could either create multiple verlet systems, or write a new function inside the VerletSystem() constructor, which rearanges the "verletGroups" ds_list. Hope this helps!