metamon/command

Stateful / model-based testing primitives.

A Command(model, real) represents one transition that can be applied to two parallel worlds:

• model: a pure description of the expected state (typically a record or a Dict). The next_model step says how the command updates this expected state.
• real: the system under test (a database connection, a stateful module, an in-memory cache). The run step performs the command on the real world and returns Ok(Nil) on success or Error(reason) when an invariant is violated.

The precondition decides whether the command can fire in the current model state; the runner skips commands whose preconditions are false.

See metamon/stateful for the runner that drives a list of commands against an initial (model, real) pair.

pub type Command(model, real) {
  Command(
    name: String,
    precondition: fn(model) -> Bool,
    next_model: fn(model) -> model,
    run: fn(real) -> Result(Nil, String),
  )
}

pub fn always(
  name name: String,
  next_model next: fn(model) -> model,
  run run: fn(real) -> Result(Nil, String),
) -> Command(model, real)

pub fn name_of(cmd: Command(model, real)) -> String

pub fn new(
  name name: String,
  precondition pre: fn(model) -> Bool,
  next_model next: fn(model) -> model,
  run run: fn(real) -> Result(Nil, String),
) -> Command(model, real)