Struct OperationCompletion 

pub struct OperationCompletion { /* private fields */ }

The completion signal every operation publishes to when it ends.

This is what lets a caller block until an operation finishes instead of re-checking it on a timer. A worker marks its id finished here as the very last thing it does — after the result is stored, the event emitted, and the final status set — so a waiter woken by it is guaranteed to observe all three. In particular, is_finished(id) == true implies LongOperationManager::get_operation_result already has the result (for a Completed operation), which is what makes an event-free blocking wait correct.

Obtain one via LongOperationManager::completion_signal and block on it after releasing any lock guarding the manager — see that method.

Implementations

impl OperationCompletion

pub fn is_finished(&self, id: &str) -> bool

Has id already reached a final status?

pub fn wait_for(&self, id: &str, timeout: Option<Duration>) -> bool

Block until id reaches a final status; true once it has.

There is no lost-wakeup window: the finished-set is inspected while holding the very mutex the condvar is paired with, and a worker must take that same mutex to publish — so an operation that ends between a caller’s check and its wait is already recorded and returns immediately.

timeout bounds one wait, not the whole call, and exists only as a backstop for a signal that can never arrive (e.g. a worker killed before publishing). None waits indefinitely.

Trait Implementations

impl Debug for OperationCompletion

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.