joedad

For anyone interested in the game mechanics used I just tonight updated to correct a glitch in the game and at the same time make it more realistic. I was 'dropping' mines at about half the depth of the ocean. At that depth the player could actually motor past them in the water column. Not very realistic. The reason for this was because of the ocean floor contours which were pre-made and just loaded as 3 different scenes the game level generator chooses from.

This turned out to not be too difficult to fix; it involved using a linear function on the one ocean floor which present problems. The mines are dropped just above the ocean floor at a height that changes based on the global x-position of the mine to be placed. Then game physics does the rest.

The other two ocean profiles are flat enough that a simple adjustment to the proper height above the floor fixed those.

I had a similar problem with the DISSUB - I was dropping it too low and it was spawning going through the tilemap ocean floor!

Final gd script code:

## deeptexture node script; defines uniforms for use in the material shader
extends ColorRect

## Currently entire scene needs to be a child of GameLevel.tscn
onready var my_parent = get_parent().get_parent().get_parent().get_node("GameLevel")
onready var player = get_parent().get_parent().get_node("RescueSub")
onready var global_posit = player.global_position    ## current screen depth calculated from Player y position + 450
onready var ocean_depth = my_parent.depth_get()  ## Max ocean depth from stage generator
onready var depth_factor

func _ready():
	pass

func _physics_process(_delta):
	ocean_depth = float(my_parent.depth_get())
	global_posit = player.global_position
	depth_factor = float(global_posit.y/ocean_depth)
	self.material.set_shader_param("depth_factor", depth_factor)

@joedad said: Okay - I must have missed a tutorial note. Why do you have to select local to scene in the inspector when I trying to pass the uniform code to the shader? I made a couple of changes... to my script which is shown in the two pictures; I changed "self.get_material().set_shader_param()" to "self.material.set_shader_param()".

2nd - I simply clicked "local to scene" as true. Then it started updating my parameters real time! If anyone has a quick answer to why the need for local to scene so I understand better I would appreciate it.

EDIT: The local to scene had nothing to do with it. I tried to use "get_material()" and apparently that did not work. I changed back to self.material. and it works.

Unfortunately, can't get my shader to actually darken the screen....

UPDATE: Shader is working! Easily and very well I might add. So my trouble was not passing the uniform to the shader code and it was not finding the "correct shader material". The trouble turned out to be trying to change variables within the shader incorrectly. I simply changed my method of updating the shader to be based on the global position and ocean depth instead of trying to get each screen line to have a different depth factor. I feed the Depth Factor from the gd_script into the shader. Turns out this is a LOT simpler and looks more realistic in my opinion. Final Shader code below....

shader_type canvas_item;

uniform vec4 deep_color : hint_color = vec4(0.0, 0.0, 0.0, 1.0);
uniform float depth_factor;

void fragment() {
	vec3 c = textureLod(SCREEN_TEXTURE, SCREEN_UV, 0.0).rgb;
	c.rgb = mix(c.rgb, deep_color.rgb, depth_factor);
	COLOR.rgb = c;
}

Okay - I must have missed a tutorial note. Why do you have to select local to scene in the inspector when I trying to pass the uniform code to the shader? I made a couple of changes... to my script which is shown in the two pictures; I changed "self.get_material().set_shader_param()" to "self.material.set_shader_param()".

2nd - I simply clicked "local to scene" as true. Then it started updating my parameters real time! If anyone has a quick answer to why the need for local to scene so I understand better I would appreciate it.

EDIT: The local to scene had nothing to do with it. I tried to use "get_material()" and apparently that did not work. I changed back to self.material. and it works.

Unfortunately, can't get my shader to actually darken the screen....

And a copy of the shader code:

shader_type canvas_item;

uniform vec2 global_posit;
uniform float ocean_depth;
uniform vec4 deep_color : hint_color = vec4(0.0, 0.0, 0.0, 1.0);
varying float depth_factor; 
//	self.material.set_shader_param("ocean_depth", ocean_depth)
//	self.material.set_shader_param("global_posit", global_posit)

void fragment() {
	vec3 c = texture(SCREEN_TEXTURE, SCREEN_UV).rgb;
	depth_factor = (FRAGCOORD.y)/ocean_depth;
	c.rgb = mix(c.rgb, deep_color.rgb, depth_factor);
	COLOR.rgb = c.rgb;
}

More modifications and game logic fix to my submarine rescue game. See an updated test run here.

Next things to fix: 1. User option to transfer rescuees to the surface ship. 2. User option to return for the other survivors. 3. Slow down the rescues - it is instantaneous right now. 4. Correct my audio bus problem. 5. Insert game session anomalies - storms, broken air compressors, injured survivors requiring more time etc..

I made some modifications and got the hose working the way I think looks much better. No bouncing jitter during game play. A little adaptive coding for the torque on the reel keeps the extra hose coiled up but the rescue submarine can counteract the torque so that the sub can go to the bottom. Seems to work well.

Video sample of the game showing the hose reel working.

I have made use of others' rigid body designs for rope for some time now but cannot find a good method to add links to the rope or chain. For a game I am designing (2D platformer) I need to be able to lengthen a hose as the player vehicle moves away from a starting position. The code for the most part is working - as the Player moves away a comparison is done in code and when the hose is stretched too far a segment is duplicated and attached to the previous segment.

Now the trouble and question: I modeled my code after how I would duplicate the segments in the editor, then move the segment and attach node_b. That seemed to work but each successive segment seems to be attached at one corner where the entire length of rope (initially 20 segments) are all aligned on center. I am uncertain why the offset.

Key Part of the Code: "link_players()", "calc_length()", and "add_segments()".

func link_players():
	
	#prevents the rope from freaking out, when relocated
	yield(get_tree(), "idle_frame")
	
	pjoint = get_parent().get_node("RescueSub/Player_pinjoint")
	ppos = pjoint.global_position 
	pjoint.node_a = $Segments.get_child(19).get_path()
	
	player2 = get_parent().get_node("VOO")
	p2pos = player2.get_node("Position2D").global_position
	var p2seg = $Segments.get_child(0)
	var p2joint = p2seg.get_node("PinJoint2D")
	p2joint.node_b = player2.get_path()

	linked = true

func calc_length():
	hose_length = num_segs*9.7

func add_segments():
	for chld in $Segments.get_children():
		chld.mode = 1
		chld.set_physics_process(false)

	player = get_parent().get_node("RescueSub")
	var init_vel = player.velocity
	player.set_process_input(false)
	player.set_physics_process(false)
	player.velocity = Vector2.ZERO
	
	var segs_list = $Segments.get_children()
	var last_seg = segs_list[num_segs-1]
	var seg_dup = last_seg.duplicate(15)
	print("did we pause")
	seg_dup.rotate(180)
	seg_dup.set_physics_process(false)
	$Segments.add_child(seg_dup, true)

	yield(get_tree(), "idle_frame")
	var dupjoint = seg_dup.get_node("PinJoint2D")
	dupjoint.node_b = last_seg.get_path()
	num_segs = $Segments.get_child_count()

	ppos = pjoint.global_position 
	seg_dup.get_node("North_Pole").global_position = ppos
	pjoint.node_a = seg_dup.get_path()
	
	player.velocity = init_vel
	player.set_process_input(true)
	player.set_physics_process(true)
	
	for chld in $Segments.get_children():
		chld.mode = 0
		chld.set_physics_process(true)