1 unstable release

0.1.0 Mar 10, 2021

#1070 in Procedural macros

Apache-2.0

56KB
1.5K SLoC

Rerust

NOT YET PUBLISHED

This is the source code for ReRust, a functional reactive programming language (FRP) that provides dataflow programming in Rust.

Usage

To get started, install the Rust toolchain and add following line to your Cargo.toml file:

[dependencies]
rerust = "1.0"

ReRust uses procedural macros to generate native Rust code. Each macro invocation should be placed inside a separate module to avoid name conflicts. Look at examples/chat.rs and examples/diamond.rs to get some inspiration.

Getting started

ReRust code is placed inside a procedural macro to generate structs and functions that can be used from plain Rust. A macro invocation typically looks like this:

mod macrodemo {
	use rerust::rerust;
	rerust! {
        let x = Var::<u32>(1u32);
        let y = x.map(|x: &u32| -> u32 {x * 2});
        let z = x.map(|x: &u32| -> u32 {x * 3});
        let pin t = (y,z).map(|y: &u32, z: &u32| -> u32 {y + z});
	}
}

Afterwards you can use as many instances of your program as you would like. To update your program state, you need to retrieve a Sink which allows updating reactives with new values. If the state has changed between two iterations all registered observers are notified, this is useful for updating GUIs for example. Finally, you need to regularly call Program::run() to poll the sink and update the state. Pushing values to a sink is done via Sink::send_<source_name>(<val>).

    let mut prog = generated::Program::new();
    let mut sink = prog.sink();

	// register observer, takes reference of t's type
    let observer = Rc::new(RefCell::new(observer_cb)) as Rc<_>;
    prog.observe_t(Rc::downgrade(&observer));

	// update x value
    sink.send_x(2);
   
    // initialize program and call observers with initial values
	prog.init();
    for _ in 0..5 {
		// check for new input, udpate state and notify observers
        prog.run();
    }

Available primitives

  • Var/Evt: Source reactives that either preserve state for the next iteration (Variable) or are invalidated after one evaluation (Event)
  • Map: Maps (multiple) reactives to a new output reactive by calling the provided closure on their values. Available for Events and Variables. If at least one Event is present as input, Map will be an Event reactive as well.
  • Fold: Takes at least one Event and any number of Variables as input and is of type Variable. Accumulates a value over time.
  • Changed: Takes exactly one Variable and transforms it into an Event, firing only if the incoming reactive has changed.

Dependencies

~4–6MB
~99K SLoC