Struct Executor 

pub struct Executor { /* private fields */ }

Top-level executor. One per process is the typical case.

Implementations

impl Executor

pub fn builder() -> ExecutorBuilder

Start a new builder.

pub fn channel<T: Payload>( &mut self, name: &str, ) -> Result<Arc<Channel<T>>, ExecutorError>

Open or create a pub/sub channel bound to this executor’s node.

pub fn service<Req, Resp>( &mut self, name: &str, ) -> Result<Arc<Service<Req, Resp>>, ExecutorError>

where Req: Payload, Resp: Payload,

Open or create a request/response service bound to this executor’s node.

pub fn add( &mut self, item: impl ExecutableItem, ) -> Result<TaskId, ExecutorError>

Add an item to the executor with an auto-generated id.

pub fn add_with_id( &mut self, id: impl Into<TaskId>, item: impl ExecutableItem, ) -> Result<TaskId, ExecutorError>

Add an item with a user-supplied id.

The item’s ExecutableItem::task_id override takes precedence over the caller-supplied id, which itself takes precedence over the auto-generated id assigned by Executor::add.

pub fn add_with_fault_handler<I, H>( &mut self, main: I, handler: H, ) -> Result<TaskId, ExecutorError>

where I: ExecutableItem, H: ExecutableItem,

Register an item plus a fault-handler item.

The main item is registered through the canonical add path. The handler’s declare_triggers is called (so handlers that internally rely on the declarer being invoked observe the call) but its returned trigger list is ignored — the handler dispatches on the main item’s triggers while the task is in Faulted state and runs in place of the main item’s execute(). The pre-built handler dispatch closure is stashed on the same task entry as the main item’s job, satisfying REQ_0072.

Errors

Propagates any error from registering the main item via add, or from the handler’s declare_triggers call.

Panics

Panics if the task entry just inserted by add cannot be located in self.tasks — this is unreachable by construction and indicates a logic bug.

pub fn clear_task_fault( &self, task: TaskId, ) -> Result<FaultState, ExecutorError>

Clear a per-task fault. Returns the previous FaultState. Fires Observer::on_task_clear if the state changed from Faulted to Running. REQ_0070.

Errors

• ExecutorError::TaskNotFound if task is unknown.
• ExecutorError::TaskNotFaulted if task is already Running.

pub fn clear_executor_fault(&self) -> Result<ExecutorFaultState, ExecutorError>

Clear the executor-wide fault and cascade-clear every task whose state is Faulted{ExecutorFaulted}. Tasks whose state is Faulted{BudgetExceeded} are NOT cleared (their own contract breach is independent). Fires Observer::on_executor_clear and one Observer::on_task_clear per cascade-cleared task. REQ_0071.

Errors

• ExecutorError::ExecutorNotFaulted if the executor is Running.

pub fn overrun_count(&self, task: TaskId) -> Result<u64, ExecutorError>

Return the per-task overrun counter — number of times the task’s execute() exceeded its budget over the executor’s lifetime. Monotonic; not reset by clear_task_fault. REQ_0102.

Errors

• ExecutorError::TaskNotFound if task is unknown.

pub fn task_fault_state( &self, task: TaskId, ) -> Result<FaultState, ExecutorError>

Return a snapshot of the per-task FaultState. REQ_0073 (pull path).

Errors

• ExecutorError::TaskNotFound if task is unknown.

pub fn executor_fault_state(&self) -> ExecutorFaultState

Return a snapshot of the executor-wide ExecutorFaultState. REQ_0073 (pull path).

pub fn add_chain<I, C>(&mut self, items: C) -> Result<TaskId, ExecutorError>

where I: ExecutableItem, C: IntoIterator<Item = I>,

Add a sequential chain of items. Only the head item’s declare_triggers is consulted; non-head triggers are ignored with a tracing warn.

pub fn add_chain_with_id<I, C>( &mut self, id: impl Into<TaskId>, items: C, ) -> Result<TaskId, ExecutorError>

where I: ExecutableItem, C: IntoIterator<Item = I>,

Like Executor::add_chain but with a user-supplied id.

pub fn stoppable(&self) -> Stoppable

Returns a Stoppable handle that is waker-aware from the moment the executor is built. Clone before calling run() — any clone taken at any time will wake the WaitSet when stop() is called.

pub const fn iceoryx_node(&self) -> &Node<Service>

Borrow the underlying iceoryx2 node (escape hatch for power users).

pub fn add_graph(&mut self) -> ExecutorGraphBuilder<'_>

Begin building a graph. Call .build() on the returned builder to register the graph as a task.

impl Executor

pub fn run(&mut self) -> Result<(), ExecutorError>

Run the executor until Stoppable::stop is called or a task signals stop via crate::Context::stop_executor.

Errors

Returns the first ExecutorError surfaced during dispatch:

• ExecutorError::Item if any item returns Err or panics.
• ExecutorError::Iceoryx2 if a WaitSet operation fails.
• ExecutorError::AlreadyRunning if the executor is already running.

If multiple items error in the same dispatch iteration, only the first is preserved; subsequent errors are discarded silently. To observe every error, attach an Observer and read errors via Observer::on_app_error.

pub fn run_for(&mut self, max: Duration) -> Result<(), ExecutorError>

Run for at most max wall-clock duration, then return.

Errors

Returns the first ExecutorError surfaced during dispatch:

• ExecutorError::Item if any item returns Err or panics.
• ExecutorError::Iceoryx2 if a WaitSet operation fails.
• ExecutorError::AlreadyRunning if the executor is already running.

If multiple items error in the same dispatch iteration, only the first is preserved; subsequent errors are discarded silently. To observe every error, attach an Observer and read errors via Observer::on_app_error.

pub fn run_n(&mut self, n: usize) -> Result<(), ExecutorError>

Run until n full barrier-cycles (WaitSet wakeups) have completed.

Errors

Returns the first ExecutorError surfaced during dispatch:

• ExecutorError::Item if any item returns Err or panics.
• ExecutorError::Iceoryx2 if a WaitSet operation fails.
• ExecutorError::AlreadyRunning if the executor is already running.

If multiple items error in the same dispatch iteration, only the first is preserved; subsequent errors are discarded silently. To observe every error, attach an Observer and read errors via Observer::on_app_error.

pub fn run_until<F: FnMut() -> bool>( &mut self, predicate: F, ) -> Result<(), ExecutorError>

Run until predicate() returns true. Checked after each WaitSet wakeup.

Errors

Returns the first ExecutorError surfaced during dispatch:

• ExecutorError::Item if any item returns Err or panics.
• ExecutorError::Iceoryx2 if a WaitSet operation fails.
• ExecutorError::AlreadyRunning if the executor is already running.

If multiple items error in the same dispatch iteration, only the first is preserved; subsequent errors are discarded silently. To observe every error, attach an Observer and read errors via Observer::on_app_error.

Trait Implementations

impl Send for Executor