Struct TaskRuntime 

pub struct TaskRuntime { /* private fields */ }

Bridges task storage with runtime execution state.

Owns both:

• A TaskStorage backend (durable, serializable)
• A TaskExecutor for running task work and managing cancellation

Lives in turul-mcp-server (not in turul-mcp-task-storage) because it combines backend-agnostic storage with executor-specific runtime primitives.

Implementations

impl TaskRuntime

pub fn new( storage: Arc<dyn TaskStorage>, executor: Arc<dyn TaskExecutor>, ) -> Self

Create a new task runtime with the given storage backend and executor.

pub fn with_default_executor(storage: Arc<dyn TaskStorage>) -> Self

Create a new task runtime with the given storage and the default TokioTaskExecutor.

pub fn with_recovery_timeout(self, timeout_ms: u64) -> Self

Create with custom recovery timeout.

pub fn in_memory() -> Self

Create a new task runtime with in-memory storage and the default TokioTaskExecutor.

pub fn storage(&self) -> &dyn TaskStorage

Get a reference to the underlying storage.

pub fn storage_arc(&self) -> Arc<dyn TaskStorage>

Get a shared reference to the storage Arc.

pub fn executor(&self) -> &dyn TaskExecutor

Get a reference to the executor.

pub async fn register_task( &self, task: TaskRecord, ) -> Result<TaskRecord, TaskStorageError>

Register a new task in storage. Returns the created record.

Does NOT start execution — call executor().start_task() separately when the work is ready to run.

pub async fn update_status( &self, task_id: &str, new_status: TaskStatus, status_message: Option<String>, ) -> Result<TaskRecord, TaskStorageError>

Update a task’s status in storage.

pub async fn complete_task( &self, task_id: &str, outcome: TaskOutcome, status: TaskStatus, status_message: Option<String>, ) -> Result<(), TaskStorageError>

Store a task’s result and update status atomically.

pub async fn cancel_task( &self, task_id: &str, ) -> Result<TaskRecord, TaskStorageError>

Cancel a task: delegate to executor AND update storage status.

pub async fn await_terminal(&self, task_id: &str) -> Option<TaskStatus>

Wait until a task reaches terminal status via the executor.

Returns None if the task is not tracked by the executor (already completed or not in-flight).

pub async fn get_task( &self, task_id: &str, ) -> Result<Option<TaskRecord>, TaskStorageError>

Get a task by ID from storage.

pub async fn get_task_result( &self, task_id: &str, ) -> Result<Option<TaskOutcome>, TaskStorageError>

Get a task’s stored result.

pub async fn list_tasks( &self, cursor: Option<&str>, limit: Option<u32>, ) -> Result<TaskListPage, TaskStorageError>

List tasks with pagination.

pub async fn list_tasks_for_session( &self, session_id: &str, cursor: Option<&str>, limit: Option<u32>, ) -> Result<TaskListPage, TaskStorageError>

List tasks for a specific session.

pub async fn recover_stuck_tasks(&self) -> Result<Vec<String>, TaskStorageError>

Recover stuck tasks on startup. Called during server initialization.

pub async fn maintenance(&self) -> Result<(), TaskStorageError>

Run periodic maintenance (TTL expiry, cleanup).