Struct TaskExecutor 

pub struct TaskExecutor { /* private fields */ }

Available on crate feature tasks only.

Parallel task executor with timeout and retry support.

Works directly with Arc<TaskManager> — since TaskManager uses DashMap internally, no external RwLock is needed for concurrent access.

Implementations

impl TaskExecutor

pub fn new( task_manager: Arc<TaskManager>, config: TaskExecutorConfig, ) -> TaskExecutor

Create a new task executor

pub fn with_execute_fn( self, f: Arc<dyn Fn(TaskContext) -> Pin<Box<dyn Future<Output = Result<String, ReactError>> + Send>> + Sync + Send>, ) -> TaskExecutor

Set a custom execution function

pub fn with_checkpoint_store( self, store: Arc<dyn CheckpointStore>, ) -> TaskExecutor

Set checkpoint store for periodic saving during execute_all

pub fn is_completed(&self) -> bool

Check if all tasks are completed

pub fn get_progress(&self) -> (usize, usize)

Get progress statistics

pub async fn execute_ready_tasks( &self, ) -> Result<Vec<TaskExecutionResult>, ReactError>

Execute all ready tasks in parallel

Returns the number of tasks executed

pub fn cancel_task(&self, task_id: &str) -> bool

Cancel a specific task

Marks the task as cancelled in the manager AND cancels the in-flight execution via CancellationToken, so a running spawned task is aborted.

pub fn cancel_all(&self)

Cancel all tasks

pub async fn execute_all(&self) -> Result<Vec<TaskExecutionResult>, ReactError>

Run the full execution loop until all tasks complete or a deadlock is detected.

This method repeatedly calls execute_ready_tasks and uses wake_dependents after each batch to discover newly-ready tasks, eliminating polling.

Returns all execution results accumulated across batches.