In this part we will be adding health pick ups, ammo pick ups, targets we can destroy, add support for joypads, and add the ability to change weapons with the scroll wheel.
You are assumed to have finished Part 3 before moving on to this part of the tutorial.
The finished project from Part 3 will be the starting project for part 4
Let’s get started!
Adding joypad input¶
In Godot any game controller is referred to as a joypad. This includes: Console controllers, Joysticks (like for flight simulators), Wheels (like for driving simulators), VR Controllers, and more.
First we need to change a few things in our project’s input map. Open up the project settings and select the
Input Map tab.
Now we need to add some joypad buttons to our various actions. Click the plus icon and select
Feel free to use whatever button layout you want. Make sure that the device selected is set to
0. In the finished project, we will be using the following:
Device 0, Button 4 (L, L1)
Device 0, Button 0 (PS Cross, XBox A, Nintendo B)
Device 0, Button 0 (PS Square, XBox X, Nintendo Y)
Device 0, Button 12 (D-Pad Up)
Device 0, Button 15 (D-Pad Right)
Device 0, Button 14 (D-Pad Right)
Device 0, Button 1 (PS Circle, XBox B, Nintendo A).
These are already set up for you if you downloaded the starter assets
Once you are happy with the input, close the project settings and save.
Now let’s open up
Player.gd and add joypad input.
First, we need to define a few new global variables. Add the following global variables to
# You may need to adjust depending on the sensitivity of your joypad var JOYPAD_SENSITIVITY = 2 const JOYPAD_DEADZONE = 0.15
Let’s go over what each of these do:
JOYPAD_SENSITIVITY: This is how fast our joypad joysticks will move our camera.
JOYPAD_DEADZONE: The dead zone for the joypad. You may need to adjust depending on your joypad.
Many joypads jitter around a certain point. To counter this, we ignore any movement in a with a radius of JOYPAD_DEADZONE. If we did not ignore said movement, the camera will jitter.
Also, we are defining
JOYPAD_SENSITIVITY as a variable instead of a constant because we’ll later be changing it.
Now we are ready to start handling joypad input!
process_input add the following code, just before
input_movement_vector = input_movement_vector.normalized():
# Add joypad input, if there is a joypad if Input.get_connected_joypads().size() > 0: var joypad_vec = Vector2(0, 0) if OS.get_name() == "Windows": joypad_vec = Vector2(Input.get_joy_axis(0, 0), -Input.get_joy_axis(0, 1)) elif OS.get_name() == "X11": joypad_vec = Vector2(Input.get_joy_axis(0, 1), Input.get_joy_axis(0, 2)) elif OS.get_name() == "OSX": joypad_vec = Vector2(Input.get_joy_axis(0, 1), Input.get_joy_axis(0, 2)) if joypad_vec.length() < JOYPAD_DEADZONE: joypad_vec = Vector2(0, 0) else: joypad_vec = joypad_vec.normalized() * ((joypad_vec.length() - JOYPAD_DEADZONE) / (1 - JOYPAD_DEADZONE)) input_movement_vector += joypad_vec
Let’s go over what we’re doing.
First we check to see if there is a connected joypad.
If there is a joypad connected, we then get its left stick axes for right/left and up/down. Because a wired Xbox 360 controller has different joystick axis mapping based on OS, we use different axes based on the OS.
This tutorial assumes you are using a XBox 360 wired controller. Also, I do not (currently) have access to a Mac computer, so the joystick axes may need changing. If they do, please open a GitHub issue on the Godot documentation repository!
Next we check to see if the joypad vector length is within the
If it is, we set
joypad_vec to an empty Vector2. If it is not, we use a scaled Radial Dead zone for precise dead zone calculating.
You can find a great article explaining all about how to handle joypad/controller dead zones here: http://www.third-helix.com/2013/04/12/doing-thumbstick-dead-zones-right.html
We’re using a translated version of the scaled radial dead zone code provided in that article. The article is a great read, and I highly suggest giving it a look!
Finally, we add
Remember how we normalize
input_movement_vector? This is why! If we did not normalize
input_movement_vector players could
move faster if they are pushing in the same direction with both their keyboard and their joypad!
Make a new function called
process_view_input and add the following:
func process_view_input(delta): if Input.get_mouse_mode() != Input.MOUSE_MODE_CAPTURED: return # NOTE: Until some bugs relating to captured mouses are fixed, we cannot put the mouse view # rotation code here. Once the bug(s) are fixed, code for mouse view rotation code will go here! # ---------------------------------- # Joypad rotation var joypad_vec = Vector2() if Input.get_connected_joypads().size() > 0: if OS.get_name() == "Windows": joypad_vec = Vector2(Input.get_joy_axis(0, 2), Input.get_joy_axis(0, 3)) elif OS.get_name() == "X11": joypad_vec = Vector2(Input.get_joy_axis(0, 3), Input.get_joy_axis(0, 4)) elif OS.get_name() == "OSX": joypad_vec = Vector2(Input.get_joy_axis(0, 3), Input.get_joy_axis(0, 4)) if joypad_vec.length() < JOYPAD_DEADZONE: joypad_vec = Vector2(0, 0) else: joypad_vec = joypad_vec.normalized() * ((joypad_vec.length() - JOYPAD_DEADZONE) / (1 - JOYPAD_DEADZONE)) rotation_helper.rotate_x(deg2rad(joypad_vec.y * JOYPAD_SENSITIVITY)) rotate_y(deg2rad(joypad_vec.x * JOYPAD_SENSITIVITY * -1)) var camera_rot = rotation_helper.rotation_degrees camera_rot.x = clamp(camera_rot.x, -70, 70) rotation_helper.rotation_degrees = camera_rot # ----------------------------------
Let’s go over what’s happening:
First we check the mouse mode. If the mouse mode is not
MOUSE_MODE_CAPTURED, we want to return, which will skip the code below.
Next we define a new Vector2 called
joypad_vec. This will hold the right joystick position. Based on the OS, we set its values so
it is mapped to the proper axes for the right joystick.
As stated above, I do not (currently) has access to a Mac computer, so the joystick axes may need changing. If they do, please open a GitHub issue on the Godot documentation repository!
We then account for the joypad’s dead zone, exactly like in
Then we rotate
rotation_helper and our KinematicBody using
Notice how the code that handles rotating ourselves and
rotation_helper is exactly the same as the
_input. All we’ve done is change the values to use
Due to few mouse related bugs on Windows, we cannot put mouse rotation in
process_view as well.
Once these bugs are fixed, this will likely be updated to place the mouse rotation here as well.
Finally, we clamp the camera’s rotation so we cannot look upside down.
The last thing you need to do is add
process_view_input is added to
_physics_process, you should be able to play using a joypad!
I decided not to use the joypad triggers for firing because we’d then have to do some more axis managing, and because I prefer to use a shoulder button to fire.
If you want to use the triggers for firing, you will need to change how firing works in
process_input. You need to get the proper axis value for the trigger,
and check if it’s over a certain value, say
0.8 for example. If it is, you add the same code as when the
fire action was pressed.
Adding mouse scroll wheel input¶
Let’s add one input related feature before we start working on the pick ups and target. Let’s add the ability to change weapons using the scroll wheel on the mouse.
Player.gd and add the following global variables:
var mouse_scroll_value = 0 const MOUSE_SENSITIVITY_SCROLL_WHEEL = 0.08
Let’s go over what each of these new variables will be doing:
mouse_scroll_value: The value of the mouse scroll wheel.
MOUSE_SENSITIVITY_SCROLL_WHEEL: How much a single scroll action increases mouse_scroll_value
Now let’s add the following to
if event is InputEventMouseButton and Input.get_mouse_mode() == Input.MOUSE_MODE_CAPTURED: if event.button_index == BUTTON_WHEEL_UP or event.button_index == BUTTON_WHEEL_DOWN: if event.button_index == BUTTON_WHEEL_UP: mouse_scroll_value += MOUSE_SENSITIVITY_SCROLL_WHEEL elif event.button_index == BUTTON_WHEEL_DOWN: mouse_scroll_value -= MOUSE_SENSITIVITY_SCROLL_WHEEL mouse_scroll_value = clamp(mouse_scroll_value, 0, WEAPON_NUMBER_TO_NAME.size()-1) if changing_weapon == false: if reloading_weapon == false: var round_mouse_scroll_value = int(round(mouse_scroll_value)) if WEAPON_NUMBER_TO_NAME[round_mouse_scroll_value] != current_weapon_name: changing_weapon_name = WEAPON_NUMBER_TO_NAME[round_mouse_scroll_value] changing_weapon = true mouse_scroll_value = round_mouse_scroll_value
Let’s go over what’s happening here:
First we check if the event is a
InputEventMouseButton event and that our mouse mode is
Then we check to see if the button index is either a
If the event’s index is indeed a button wheel index, we then check to see if it is a
Based on whether it is up or down we add/remove
Next we clamp mouse scroll value to assure it is inside the range of our weapons.
We then check to see if we are changing weapons or reloading. If we are doing neither, we round
mouse_scroll_value and cast it to a
We are casting
mouse_scroll_value to a
int so we can use it as a key in our dictionary. If we left it as a float,
we would get an error when we try to run the project.
Next we check to see if the weapon name at
round_mouse_scroll_value is not equal to the current weapon name using
If the weapon is different than our current weapon, we assign
changing_weapon to true so we will change weapons in
process_changing_weapon, and set
The reason we are setting
mouse_scroll_value to the rounded scroll value is because we do not want the player to keep their
mouse scroll wheel just in between values, giving them the ability to switch almost extremely fast. By assigning
round_mouse_scroll_value, we assure that each weapon takes exactly the same amount of scrolling to change.
One more thing we need to change is in
process_input. In the code for changing weapons, add the following right after the line
changing_weapon = true:
mouse_scroll_value = weapon_change_number
Now our scroll value we be changed with the keyboard input. If we did not change this, our scroll value will be out of sync. If the scroll wheel is out of sync, scrolling forwards or backwards would not transition to the next/last weapon, but rather the next/last weapon the scroll wheel changed to.
Now you can change weapons using the scroll wheel! Go give it a whirl!
Adding the health pick ups¶
Now that our player has health and ammo, we ideally need a way to replenish those resources.
Holder if it’s not already expanded. Notice how we have two Spatial nodes, one called
Health_Kit and another called
This is because we’re actually going to be making two sizes of health pick ups, one small and one large/normal.
have a single MeshInstance as their children.
Health_Pickup_Trigger. This is an Area node we’re going to use to check if the player has walked close enough to pick up
the health kit. If you expand it you’ll find two collision shapes, one for each size. We will be using a different collision shape size based on the size of the
health pick up, so the smaller health pick up has a trigger collision shape closer to it’s size.
The last thing to note is how we have a AnimationPlayer node so the health kit spins around slowly and bobs up and down.
Health_Pickup and add a new script called
Health_Pickup.gd. Add the following:
extends Spatial export (int, "full size", "small") var kit_size = 0 setget kit_size_change # 0 = full size pickup, 1 = small pickup const HEALTH_AMOUNTS = [70, 30] const RESPAWN_TIME = 20 var respawn_timer = 0 var is_ready = false func _ready(): $Holder/Health_Pickup_Trigger.connect("body_entered", self, "trigger_body_entered") is_ready = true kit_size_change_values(0, false) kit_size_change_values(1, false) kit_size_change_values(kit_size, true) func _physics_process(delta): if respawn_timer > 0: respawn_timer -= delta if respawn_timer <= 0: kit_size_change_values(kit_size, true) func kit_size_change(value): if is_ready: kit_size_change_values(kit_size, false) kit_size = value kit_size_change_values(kit_size, true) else: kit_size = value func kit_size_change_values(size, enable): if size == 0: $Holder/Health_Pickup_Trigger/Shape_Kit.disabled = !enable $Holder/Health_Kit.visible = enable elif size == 1: $Holder/Health_Pickup_Trigger/Shape_Kit_Small.disabled = !enable $Holder/Health_Kit_Small.visible = enable func trigger_body_entered(body): if body.has_method("add_health"): body.add_health(HEALTH_AMOUNTS[kit_size]) respawn_timer = RESPAWN_TIME kit_size_change_values(kit_size, false)
Let’s go over what this script is doing, starting with its global variables:
kit_size: The size of the health pick up. Notice how we’re using a
setgetfunction to tell if it’s changed.
HEALTH_AMMOUNTS: The amount of health each pick up in each size contains.
RESPAWN_TIME: The amount of time, in seconds, it takes for the health pick up to respawn
respawn_timer: A variable used to track how long the health pick up has been waiting to respawn.
is_ready: A variable to track whether the
_readyfunction has been called or not.
setget functions are called before
_ready, we need to ignore the
first kit_size_change call, because we cannot access child nodes until
_ready is called. If we did not ignore the
setget call, we would get several errors in the debugger.
Also, notice how we’re using a exported variable. This is so we can change the size of the health pick up in the editor, for each pick up. This makes it where we do not have to make two scenes for the two sizes, since we can easily change sizes in the editor using the exported variable.
See GDScript and scroll down to the Exports section for a list of of export hints you can use.
Let’s look at
First we connect the
body_entered signal from our
Health_Pickup_Trigger to the
trigger_body_entered function. This makes is where any
body that enters the Area triggers the
Next we set
true so we can use our
Then we hide all of the possible kits and their collision shapes using
kit_size_change_values. The first argument is the size of the kit, while the second argument
is whether to enable or disable the collision shape and mesh at that size.
Then we make only the kit size we selected visible, calling
kit_size_change_values and passing in
true, so the size at
kit_size is enabled.
Next let’s look at
The first thing we do is check to see if
true, we then make whatever kit is currently assigned to
kit_size disabled using
kit_size_change_values, passing in
Then we assign
kit_size to the new value passed in,
value. Then we call
kit_size_change_values passing in
kit_size again, but this time
with the second argument as
true so we enable it. Because we changed
kit_size to the passed in value, this will make whatever kit size we passed in visible.
is_ready is not
true, we simply assign
kit_size to the passed in
Now let’s look at
The first thing we do is check to see which size we’re using. Based on which size we’re wanting to enable/disable, we want to get different nodes.
We get the collision shape for the node corresponding to
size and disable it based on the
enabled passed in argument/variable.
Why are we using
!enable instead of
enable? This is so when we say we want to enable the node, we can pass in
true, but since
CollisionShape uses disabled instead of enabled, we need to flip it. By flipping it, we can enable the collision shape
and make the mesh visible when
true is passed in.
We then get the correct Spatial node holding the mesh and set its visibility to
This function may be a little confusing, try to think of it like this: We’re enabling/disabling the proper nodes for
enabled. This is so we cannot pick up
health for a size that is not visible, and so only the mesh for the proper size will be visible.
Finally, let’s look at
The first thing we do is see whether or not the body that just entered has a method/function called
add_health. If it does, we then
add_health and pass in the health provided by the current kit size.
Then we set
RESPAWN_TIME so we have to wait before we can get health again. Finally, call
false so the kit at
kit_size is invisible until we’ve waited long enough to respawn.
The last thing we need to do before we can use this health pick up is add a few things to our player.
Player.gd and add the following global variable:
const MAX_HEALTH = 150
MAX_HEALTH: The maximum amount of health a player can have.
Now we need to add the
add_health function to our player. Add the following to
func add_health(additional_health): health += additional_health health = clamp(health, 0, MAX_HEALTH)
Let’s quickly go over what this does.
We first add
additional_health to our current health. We then clamp the health so that it cannot exceed a value higher than
MAX_HEALTH, nor a value lower
With that done, now we can collect health! Go place a few
Health_Pickup scenes around and give it a try. You can change the size of the health pick up in the editor
Health_Pickup instanced scene is selected, from a convenient drop down.
Adding the ammo pick ups¶
While adding health is good and all, we can’t reap the rewards from it since nothing can (currently) damage us. Let’s add some ammo pick ups next!
Ammo_Pickup.tscn. Notice how it’s structured exactly the same as
Health_Pickup.tscn, but with the meshes and trigger collision shapes changed slightly to adjust
for the difference in mesh sizes.
Ammo_Pickup and add a new script called
Ammo_Pickup.gd. Add the following:
extends Spatial export (int, "full size", "small") var kit_size = 0 setget kit_size_change # 0 = full size pickup, 1 = small pickup const AMMO_AMOUNTS = [4, 1] const RESPAWN_TIME = 20 var respawn_timer = 0 var is_ready = false func _ready(): $Holder/Ammo_Pickup_Trigger.connect("body_entered", self, "trigger_body_entered") is_ready = true kit_size_change_values(0, false) kit_size_change_values(1, false) kit_size_change_values(kit_size, true) func _physics_process(delta): if respawn_timer > 0: respawn_timer -= delta if respawn_timer <= 0: kit_size_change_values(kit_size, true) func kit_size_change(value): if is_ready: kit_size_change_values(kit_size, false) kit_size = value kit_size_change_values(kit_size, true) else: kit_size = value func kit_size_change_values(size, enable): if size == 0: $Holder/Ammo_Pickup_Trigger/Shape_Kit.disabled = !enable $Holder/Ammo_Kit.visible = enable elif size == 1: $Holder/Ammo_Pickup_Trigger/Shape_Kit_Small.disabled = !enable $Holder/Ammo_Kit_Small.visible = enable func trigger_body_entered(body): if body.has_method("add_ammo"): body.add_ammo(AMMO_AMOUNTS[kit_size]) respawn_timer = RESPAWN_TIME kit_size_change_values(kit_size, false)
You may have noticed this code looks almost exactly the same as the health pick up. That’s because it largely is the same! Only a few things have been changed, and that’s what we’re going to go over.
First, notice how we have
AMMO_AMOUNTS instead of
AMMO_AMOUNTS will be how many ammo clips/magazines we add to the current weapon.
HEALTH_AMMOUNTS which was how many health points, we instead add an entire clip for the current weapon, instead of the raw ammo amount)
The only other thing to notice is in
trigger_body_entered we’re checking and calling a function called
Other than those two small changes, everything else is exactly the same as the health pickup!
All we need to do make the ammo pick ups work is add a new function to our player. Open
Player.gd and add the following function:
func add_ammo(additional_ammo): if (current_weapon_name != "UNARMED"): if (weapons[current_weapon_name].CAN_REFILL == true): weapons[current_weapon_name].spare_ammo += weapons[current_weapon_name].AMMO_IN_MAG * additional_ammo
Let’s go over what this function does.
The first thing we check is to see whether we’re using
UNARMED or not. Because
UNARMED does not have a node/script, we want to make sure we’re not using
UNARMED before trying to get the node/script attached to
Next we check to see if the current weapon can be refilled. If the current weapon can, we add a full clip/magazine worth of ammo to the weapon by
multiplying the current weapon’s
AMMO_IN_MAG variable times however much ammo clips we’re adding (
With that done, you should now be able to get additional ammo! Go place some ammo pick ups in one/both/all of the scenes and give it a try!
Notice how we’re not limiting the amount of ammo you can carry. To limit the amount of ammo each weapon can carry, you need to add a additional variable to
each weapon’s script, and then clamp the weapon’s
spare_ammo variable after adding ammo in
Adding breakable targets¶
Before we end this part, let’s add some targets.
Target.tscn and take a look at the scenes in the scene tree.
First, notice how we’re not using a RigidBody node, but rather a StaticBody node instead. The reason behind this is our non-broken targets will not be moving anywhere, using a RigidBody would be more hassle then its worth, since all it has to do is stay still.
The other thing to note is we have a node called
Broken_Target_Holder. This node is going to hold a spawned/instanced scene called
Broken_Target.tscn. Open up
Notice how the target is broken up into five pieces, each a RigidBody node. We’re going to spawn/instance this scene when the target takes too much damage and needs to be destroyed. Then we’re going to hide the non-broken target, so it looks like the target shattered rather than a shattered target was spawned/instanced.
While you still have
Broken_Target.tscn open, attach
RigidBody_hit_test.gd to all of the RigidBody nodes. This will make
it where we can shoot at the broken pieces and they will react to the bullets.
Alright, now switch back to
Target.tscn, select the
Target StaticBody node and created a new script called
Add the following code to
extends StaticBody const TARGET_HEALTH = 40 var current_health = 40 var broken_target_holder # The collision shape for the target. # NOTE: this is for the whole target, not the pieces of the target var target_collision_shape const TARGET_RESPAWN_TIME = 14 var target_respawn_timer = 0 export (PackedScene) var destroyed_target func _ready(): broken_target_holder = get_parent().get_node("Broken_Target_Holder") target_collision_shape = $Collision_Shape func _physics_process(delta): if target_respawn_timer > 0: target_respawn_timer -= delta if target_respawn_timer <= 0: for child in broken_target_holder.get_children(): child.queue_free() target_collision_shape.disabled = false visible = true current_health = TARGET_HEALTH func bullet_hit(damage, bullet_hit_pos): current_health -= damage if current_health <= 0: var clone = destroyed_target.instance() broken_target_holder.add_child(clone) for rigid in clone.get_children(): if rigid is RigidBody: var center_in_rigid_space = broken_target_holder.global_transform.origin - rigid.global_transform.origin var direction = (rigid.transform.origin - center_in_rigid_space).normalized() # Apply the impulse with some additional force (I find 12 works nicely) rigid.apply_impulse(center_in_rigid_space, direction * 12 * damage) target_respawn_timer = TARGET_RESPAWN_TIME target_collision_shape.disabled = true visible = false
Let’s go over what this script does, starting with the global variables:
TARGET_HEALTH: The amount of damage needed to break a fully healed target.
current_health: The amount of health this target currently has.
broken_target_holder: A variable to hold the
Broken_Target_Holdernode so we can use it easily.
target_collision_shape: A variable to hold the CollisionShape for the non-broken target.
TARGET_RESPAWN_TIME: The length of time, in seconds, it takes for a target to respawn.
target_respawn_timer: A variable to track how long a target has been broken.
destroyed_target: A PackedScene to hold the broken target scene.
Notice how we’re using an exported variable (a PackedScene) to get the broken target scene instead of
preload. By using an exported variable, we can chose the scene from the editor, and when/if we need to use a different scene,
it’s as easy as selecting a different scene in the editor, we don’t need to go to the code to change the scene we’re using.
Let’s look at
The first thing we do is get the broken target holder and assign it to
broken_target_holder. Notice how we’re using
get_parent().get_node() here, instead
$. If you want to use
$, then you’d need to change
At the time of when this was written, I did not realize you can use
$"../NodeName" to get the parent nodes using
$, which is why
is used instead.
Next we get the collision shape and assign it to
target_collision_shape. The reason we need to collision shape is because even when the mesh is invisible, the
collision shape will still exist in the physics world. This makes it where the player can interact with a non-broken target even though it’s invisible, which is
not what we want. To get around this, we will disable/enable the collision shape as we make the mesh visible/invisible.
Next let’s look at
We’re only going to be using
_physics_process for respawning, and so the first thing we do is check to see if
target_respawn_timer is more than
If it is, we then remove
delta from it.
Then we check to see if
0 or less. The reason behind this is since we just removed
target_respawn_timer, if it’s
0 or less then we’ve just got here, effectively allowing us to do whatever we need to do when the timer is finished.
In this case, we want to respawn our target.
The first thing we do is remove all children in the broken target holder. We do this by iterating over all of the children in
broken_target_holder and free them.
Next we enable our collision shape by setting its
disabled boolean to
Then we make ourselves, and all of our children nodes, visible.
Finally, we reset
Finally, let’s look at
The first the we do is remove however much damage the bullet does from our health.
Next we check to see if we’re at
0 health or lower. If we are, then we’ve just died and need to spawn a broken target.
We first instance a new destroyed target scene, and assign it to a new variable,
Next we add
clone as a child of our broken target holder.
For an added bonus, we want to make all of the target pieces explode outwards. Do to this, we iterate over all of the children in
For each child, we first check to see if it’s a RigidBody node. If it is, we then calculate the center position of the target relative to the child node. Then we figure out which direction we are relative to the center. Using those calculated variables, we push the child from the calculated center, in the direction away from the center, using the damage of the bullet as the force.
We multiply the damage by
12 so it has a more dramatic effect. You can change this to a higher or lower value depending on how explosive you want
your targets to shatter.
Next we set our respawn timer for our non-broken target. We set it to
TARGET_RESPAWN_TIME, so it takes
TARGET_RESPAWN_TIME many seconds to respawn.
Then we disable the non-broken target’s collision shape, and set our visibility to
Make sure to set the exported
destroyed_target value for
Target.tscn in the editor! Otherwise the targets will not be destroyed
and you will get an error!
With that done, go place some
Target.tscn instances around in one/both/all of the levels. You should find they explode into five pieces after they’ve taken enough
damage. After a little while, they’ll respawn into a whole target again.
Now you can use a joypad, change weapons with the mouse’s scroll wheel, replenish your health and ammo, and break targets with your weapons.
In the next part, Part 5, we’re going to add grenades to our player, give our player the ability to grab and throw objects, and add turrets!
If you ever get lost, be sure to read over the code again!
You can download the finished project for this part here: