I am working on the input and it feels like its getting out of hand. I wanted to check in with people and see how people structure their input handling.
Currently I have 1 function _unhandled_unput(event) and inside there I have a ton of elif statements trying to handle every possible situation and event. Its manageable at the moment but I only have like 4 events so its going to get very out of hand if I continue.
I need to have 100s of these events based on whats selected and what mouse/keyboard buttons are being pressed and I need some way to resuse the actions.
spoiler
func _unhandled_input(event):
if event is InputEventMouseButton and event.pressed:
if event.button_index == MOUSE_BUTTON_RIGHT:
clear_selection()
gui.queue_redraw()
get_viewport().set_input_as_handled()
return
if selected_item == "colonist": #broken
if event is InputEventKey:
if event.OS.get_keycode_string() == "r":
for colonist in selected_group:
colonist.set_state("DRAFT")
get_viewport().set_input_as_handled()
gui.queue_redraw()
#nothing selected dragbox to select things and single click to select things - does not work at the moment
elif selected_type == "" or selected_type == "basic":
if is_dragging and event is InputEventMouseMotion:
drag_end = camera.get_global_mouse_position()
cam_drag_end= get_viewport().get_mouse_position()
get_selection(drag_start, drag_end)
gui.queue_redraw()
get_viewport().set_input_as_handled()
return
elif event is InputEventMouseButton and not event.pressed:
is_dragging = false
gui.queue_redraw()
drag_start = null
drag_end = null
get_viewport().set_input_as_handled()
return
elif event is InputEventMouseButton and event.pressed:
if event.button_index == MOUSE_BUTTON_LEFT:
selected_type = "basic"
is_dragging = true
drag_start = camera.get_global_mouse_position()
cam_drag_start = get_viewport().get_mouse_position()
gui.queue_redraw()
get_viewport().set_input_as_handled()
return
#command flow for dragging a selection box
elif selected_type == "command":
if selected_item == "structure_dict_growing":
if is_dragging and event is InputEventMouseMotion:
var grid_pos = tilemap.local_to_map(camera.get_global_mouse_position())
var local_pos = tilemap.map_to_local(grid_pos)
drag_end = local_pos + Vector2(32, 32)
cam_drag_end = get_viewport().get_mouse_position()
gui.queue_redraw()
get_viewport().set_input_as_handled()
return
elif event is InputEventMouseButton and not event.pressed:
if event.button_index == MOUSE_BUTTON_LEFT:
is_dragging = false
gui.queue_redraw()
get_viewport().set_input_as_handled()
MessageBus.rpc_id(1, "request_zone_growing", selected_item ,drag_start, drag_end, multiplayer.get_unique_id())
drag_start = null
drag_end = null
return
elif event is InputEventMouseButton and event.pressed:
if event.button_index == MOUSE_BUTTON_LEFT:
is_dragging = true
#to snap to grid
var grid_pos = tilemap.local_to_map(camera.get_global_mouse_position())
var local_pos = tilemap.map_to_local(grid_pos)
drag_start = local_pos - Vector2(32, 32)
cam_drag_start = get_viewport().get_mouse_position() #this is broken cbf fixing maybe one day after selection is working
gui.queue_redraw()
get_viewport().set_input_as_handled()
return
elif selected_type == "floor":
if event is InputEventMouseButton and event.pressed:
if event.button_index == MOUSE_BUTTON_LEFT:
var global_mouse_pos = camera.get_global_mouse_position()
var grid_pos = tilemap.local_to_map(global_mouse_pos)
if selected_item == "":
return
MessageBus.rpc_id(1, "request_build_floor", selected_item, grid_pos, multiplayer.get_unique_id())
get_viewport().set_input_as_handled()
return
elif selected_type == "building":
if event is InputEventMouseButton and event.pressed:
if event.button_index == MOUSE_BUTTON_LEFT:
var global_mouse_pos = camera.get_global_mouse_position()
var grid_pos = tilemap.local_to_map(global_mouse_pos)
if selected_item == "":
return
MessageBus.rpc_id(1, "request_build_structure", selected_item, grid_pos, multiplayer.get_unique_id())
get_viewport().set_input_as_handled()
return
You must log in or # to comment.
if you hardcode letters for keys, you can’t handle alternate keyboard layouts. without rebinding, that means some users will have to be contortionists to play.
I love when I have to use ZX for actions on German keyboard layout /s
Between QWERTZ vs QWERTY the Z and Y keys are swapped, which gives YX quite the big spread.
I’m mostly experienced in c++ and python, but I would use a dictionary/map/associative array/whatever name you call it, where you put keypress/event/… as key, and a function with the result you want as value (as a first version, there’s other more specialised ways, but this will be enough for most cases).
If there’s a case/switch statement in gdscript (that’s what you’re using I think?), you could also look into that, though I think it’s likely just syntactic sugar for either what you’re already doing or what I suggested, but it might be easier to read, so there’s that.
recommend using godot’s action system instead of checking for specific button inputs https://docs.godotengine.org/en/stable/getting_started/first_3d_game/02.player_input.html#creating-input-actions since it can more easily allow for controller support and remapping, unless there’s a reason you are checking for specific keys? also is there a reason to using _unhandled_input() over _input()? Otherwise I would also recommend moving some of this logic into separate functions as a start
Good point, this probably also allows players to remap inputs if they want.
The reason for checking keys is it is the only way i know how to do it. This input handling script is what im finding the most trouble with, a lot of the lines are added because I read a 6yr old reddit comment that said to use it and im not sure if I should remove it because then bugs start appearing.
The reason for unhandled input was the first input handling code example I saw and honestly I forgot about input(). I’ll look into the creating input actions but at a glance it doesn’t seem like it could work for my game. This game is pretty much left click right click on a 2d space and depending on what tool is selected.
I really do want to move some of the code into other functions but im not sure how to make a function that handles the input over multiple different states.
For example dragging a box. needs to check for the left click then check for the dragging state and draw on drag then check for a release. I have a lot of tools that will need to implement this so I’d really like to get this logic out of the main function and into a nice reusable block.
code example
elif selected_type == "command": if selected_item == "structure_dict_growing": if is_dragging and event is InputEventMouseMotion: var grid_pos = tilemap.local_to_map(camera.get_global_mouse_position()) var local_pos = tilemap.map_to_local(grid_pos) drag_end = local_pos + Vector2(32, 32) cam_drag_end = get_viewport().get_mouse_position() gui.queue_redraw() get_viewport().set_input_as_handled() return elif event is InputEventMouseButton and not event.pressed: if event.button_index == MOUSE_BUTTON_LEFT: is_dragging = false gui.queue_redraw() get_viewport().set_input_as_handled() MessageBus.rpc_id(1, "request_zone_growing", selected_item ,drag_start, drag_end, multiplayer.get_unique_id()) drag_start = null drag_end = null return elif event is InputEventMouseButton and event.pressed: if event.button_index == MOUSE_BUTTON_LEFT: is_dragging = true #to snap to grid var grid_pos = tilemap.local_to_map(camera.get_global_mouse_position()) var local_pos = tilemap.map_to_local(grid_pos) drag_start = local_pos - Vector2(32, 32) cam_drag_start = get_viewport().get_mouse_position() gui.queue_redraw() get_viewport().set_input_as_handled() returnIMO this is a great case for Finite State Machines. Unfortunately I can’t provide a GDScript example since I write in C#, but I think there are some assets for this in the asset library. The basic idea is that you define states, and valid transitions between those states (e.g. for a drag’n’drop system given the states
idle,dragginganddroppedit makes sense for the user to goidle->dragging,dragging->droppedanddropped->idle, but not fromidle->dropped). Then you can separate the state transition logic from the main handler for each state, which I think is what you want, right?This pattern can also easily be made reusable, either by having no dedicated logic in the state machine (-> everything is handled through events), or by using an adapter interface that components create and pass into the state machine.
Let me know if you have questions, or would like to see an example - I can at least provide pseudo code :)
yea honestly JaN0h4ck’s advice by moving the input check into is probably what you’d want to have a crack into trying otherwise for cleaning up code I would recommend 100% using Godot’s actions so instead of doing an if statement for checking InputEventMouseButton and if its pressed or not and then checking the button index, just use an event.is_action_[pressed/released], or you check if the event is an action type and then call a separate function passing the event in to check the action just to make your _input function a little cleaner by only having the logic for the non button events happening in there
also considering how much youre calling camera.get_global_mouse_position() and whatnot, I feel like you may as well just set a variable at the top of the func for it, having the space for that would also help if you were to look into any other “mouse” position stuff such as touch screen inputs or virtual pointers from a controller
You could have your Drag Box use the Area2D class and listen for it’s events like this: (input_pickable need to be set to true)
class_name DraggableBox extend Area2D var mouse_inside: bool = false func _ready(): mouse_entered.connect(_on_mouse_entered) mouse_exited.connect(_on_mouse_exited) func _process(delta: float): if (not mouse_inside): reuturn if (Input.is_action_pressed("name_for_action_with_mouse_click"): #handle your object movement if (Input.is_action_just_released("name_for_action_with_mouse_click"): #handle letting go of the object func _on_mouse_entered(): mouse_inside = true func _on_mouse_exited(): mouse_inside = falseThis is a solid advice. In the last Unreal Engine project that I was working, this is the approach I went for, entities handle their own input, with a few exceptions of course. Harder to find input handling but at least it’s contained in a class the input actually needs to effect.
inside there I have a ton of elif statements
I don’t even use godot and I can already tell you this is not how you want to do it.
Haha, yeah it feels wrong but im pretty not an experienced programmer so its all I know.
I am thinking if I could figure out how to put each type of command into its own function. The main input handling loop could just be calling a single line in response to an input event and that would be more managable. However im stuck because a lot of the logic requires multiple input events.
All the godot tutorials keep talking about state and after waiting 8 I still have no idea how to use state to control this.
I highly, highly recommend using the built-in system for receiving inputs. From there you can build whatever abstraction you want on top of it.
I just dump a lot of if statements into _process. One thing good about it is that it gets rid of some nesting and shortens it a bit
