BUG: Timeout via ThreadPoolExecutor does not kill the running thread

[FernandoCelmer]

Description

In action.py lines 147-154, when future.result(timeout=self.timeout) raises TimeoutError, the underlying thread continues running in the background. Python's ThreadPoolExecutor cannot cancel running threads.

Location

dotflow/core/action.py:147-154

if self.timeout:
    with ThreadPoolExecutor(max_workers=1) as executor:
        future = executor.submit(self._call_func, is_async, *args, **kwargs)
        result = future.result(timeout=self.timeout)

Impact

Resource leak — timed-out functions keep running, consuming CPU/memory and potentially causing side effects.

Severity

Significant

[FernandoCelmer]

Closing — the current implementation in engine.py already applies the best mitigation available with Python threads:

future.cancel()
executor.shutdown(wait=False, cancel_futures=True)

The fundamental limitation is that Python cannot kill a running thread. future.cancel() only works if the future hasn't started, and shutdown(cancel_futures=True) cancels queued futures, not threads already executing.

This is a known Python limitation, not a Dotflow bug. The current implementation is the best that can be done with ThreadPoolExecutor.

A separate issue will be opened for a future enhancement using multiprocessing to achieve real timeout enforcement via process.terminate().