Trait paxakos::applicable::ApplicableTo[−][src]

pub trait ApplicableTo<S: State> {
    type Projection: Projection<OutcomeOf<S>>;
    fn into_log_entry(self) -> Arc<LogEntryOf<S>>;
}

Expand description

Describes values that may be applied to type S.

For any given State implementation there is usually a wide range of possible operations. These are encoded by its corresponding LogEntry type, which commonly contains (or is) an enum value whose variants correspond to the state’s operations. Different operations usually have different outcome types, which is why State::Outcome is also commonly an enum.

This presents an ergonomics challenge. Imagine enum MyLogEntry { A, B } and enum MyOutcome { A(i64), B(bool) }. When appending a MyLogEntry::A we’d like to get back an i64 rather than a MyOutcome. This can be achieved as follows.

struct A;

impl paxakos::applicable::ApplicableTo<MyState> for A {
    type Projection = ProjectionA;

    fn into_log_entry(self) -> Arc<MyLogEntry> {
        Arc::new(MyLogEntry::A)
    }
}

struct ProjectionA;

impl paxakos::applicable::Projection<MyOutcome> for ProjectionA {
    type Projected = i64;

    fn project(val: MyOutcome) -> Self::Projected {
        match val {
            MyOutcome::A(i) => i,
            _ => panic!("unexpected: {:?}", val)
        }
    }
}