Fix event loop timer triggering and async task completion#32
Merged
Conversation
After dispatching timer or FD events, the worker now sends task_ready to
itself to trigger _run_once processing. Also modified process_ready_tasks
to call _run_once when there are pending events (not just new coroutines).
Without this fix, asyncio.sleep and other async operations would never
complete because the event loop wasn't being driven after timer events.
Changes:
- py_event_worker: Send task_ready after handling {timeout, ...} and
{select, ...} messages to trigger event loop processing
- py_event_loop.c: Call _run_once when pending_count > 0 in addition
to when coros_scheduled > 0
Timer events weren't triggering task completion because: - dispatch_timer didn't notify worker to process ready tasks - Python strings became binaries, but await expected atoms - ErlangRef conversion happened before resource type checks Changes: - Add global worker PID storage for timer dispatch notification - Add nif_set_shared_worker NIF and py_nif:set_shared_worker/1 - Add ErlangAtomObject type and erlang.atom() function - Modify _run_and_send to use atoms for message keys - Move enif_is_ref check after resource type checks - Add _get_ready_count method for event loop processing - Fix lazy loop creation to use ErlangEventLoop directly All py_async_task_SUITE (26) and py_SUITE (50) tests pass.
- Skip _run_once when Python loop is already running to prevent reentrancy issues when erlang.run() is used inside py:exec - Don't destroy global loop capsule on close() since it's shared between multiple Python ErlangEventLoop instances - Clean up events_module in early return path Fixes py_async_e2e_SUITE hangs and crashes.
dispatch_timer may be called on a loop different from the one the global worker manages. The worker already sends task_ready to itself after handling the timer timeout, so the global send was redundant and incorrect.
This error is expected when the loop is being destroyed while task_ready messages are still in the worker's mailbox.
- Add _has_loop_ref() to prevent concurrent loops while allowing sequential replacement (checks is_running() not just exists) - Add _clear_loop_ref() called on loop close for proper cleanup - Add global_loop_capsule_destructor to fix resource leak - Rename atom() to _atom() in C, add Python wrapper with cache and configurable limit (ERLANG_PYTHON_MAX_ATOMS, default 10000) - Use enif_make_existing_atom() first to avoid duplicate atoms - Fix venv .pth file processing for Python 3.14 subinterpreters by embedding site-packages path directly in exec code
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
asyncio.sleep()never completingerlang.atom()for explicit atom creation in Pythonerlang.run()used insidepy:exec