Previously if an action was both pressed and released on the same tick or frame, `is_action_just_pressed()` would return false, resulting in missed input.
This PR separately the timestamp for pressing and releasing so each can be tested independently.
- lawnjelly
63d208d1b0
The `InputEvent` class currently supports the `pressed` and `released` states, which given the binary nature, is represented by a `bool` field.
This commit introduced the `CANCELED` state, which signals that an ongoing input event has been canceled.
To represent all the states, the `InputEventState` enum is added and the `InputEvent` logic is refactored accordingly.
- m4gr3d
94d6c3dcc6
- Uses all accumulated movements when calculating speed
- Discards old accumulated movements
- Sets last mouse speed to zero when there is no movement
- madmiraal
3d96d7d9fb
Input accumulation was implemented and enabled by default in 3.1, and
I don't recall major complaints around it (or bugs were fixed).
In 3.4, #42220 added input buffering and apparently toggled input
accumulation off by mistake.
This led to multiple bug reports about degraded performance on Windows,
or simply unexpected behavior change (see linked issues in #55037).
Fixes#55037.
- akien-mga
d6bcdd18c3
As _buttons and _axes have both valid string and nullptr.
When iterating over them, if given key exists it will work correctly.
But if given key does not exist, it will end up with
String::operator=(nullptr). As String constructor from nullptr exists, I
use it. - Snowapril - c77b710035
