2 unstable releases
0.2.0 | Aug 10, 2024 |
---|---|
0.1.0 | Aug 9, 2024 |
#628 in Game dev
7KB
65 lines
bufferbuffer
This is an implementation of the Double Buffer design pattern from 'Game Programming Patterns' by Robert Nystrom.
In a simulation, you often have to do a lot of processing to prepare the next "frame", but if you're iterating through the current-state data while mutating it, things can slip. The Double Buffer design pattern solves this by keeping two copies of the simulation state (or any variable): the "current" (or previous) state which is immutable, and the "next" (or future) state which is being prepared. When a turn of the simulation is completed, you simply switch the buffers.
Unlike other implementations on crates.io, this one wraps both buffers in std::cell::RefCell
so that
it is possible to borrow one buffer as mutable at the same time the other is borrowed as immutable --
a typical use case is to iterate over objects in the world (current state) and write updated versions of
them to the next state.
Usage
For a simple variable:
let mut my_double_buf: DoubleBuffer<i32> = DoubleBuffer::new(0,0);
*my_double_buf.next() += 10;
my_double_buf.switch();
assert_eq!(*my_double_buf.current(), 10);
For a vector:
let mut my_double_buf: DoubleBuffer<Vec<i32>> = DoubleBuffer::new( vec!(2,4,6), Vec::new());
for number in my_double_buf.current().iter() {
my_double_buf.next().push(*number + 1);
}
my_double_buf.switch();
assert_eq!(*my_double_buf.current(), vec!(3,5,7));