17
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
IMO 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 :)