asyncio eager task inherits the creators exception state

[deadlovelll]

Bug report

Bug description:

Eager tasks created from an except block inherits the creator's exception state, a bare raise re-raises it, and any other exception gets it as __context__.

Repro:

import asyncio


async def bare():
    raise


async def explicit():
    raise ValueError("child's own")


async def main(coro):
    try:
        raise ConnectionError("primary failed")
    except ConnectionError:
        t = asyncio.create_task(coro)
    try:
        await t
    except BaseException as e:
        print(f"  {type(e).__name__}: {e} | __context__: {e.__context__!r}")


for eager in (False, True):
    print("eager_task_factory:", eager)
    for coro_fn in (bare, explicit):
        loop = asyncio.new_event_loop()
        if eager:
            loop.set_task_factory(asyncio.eager_task_factory)
        try:
            loop.run_until_complete(main(coro_fn()))
        finally:
            loop.close()

expected:

eager_task_factory: False
  RuntimeError: No active exception to reraise | __context__: None
  ValueError: child's own | __context__: None
eager_task_factory: True
  RuntimeError: No active exception to reraise | __context__: None
  ValueError: child's own | __context__: None

actually:

eager_task_factory: False
  RuntimeError: No active exception to reraise | __context__: None
  ValueError: child's own | __context__: None
eager_task_factory: True
  ConnectionError: primary failed | __context__: None
  ValueError: child's own | __context__: ConnectionError('primary failed')

proposed fix - run the eager first step under an empty exception state in task_eager_start

    _PyErr_StackItem exc_state = { .exc_value = NULL, .previous_item = NULL };
    _PyErr_StackItem *prev_exc_info = ts->base.exc_info;
    ts->base.exc_info = &exc_state;
    Py_BEGIN_CRITICAL_SECTION(task);
    stepres = task_step_impl(state, task, NULL);
    Py_END_CRITICAL_SECTION();
    ts->base.exc_info = prev_exc_info;

This error also affects python implementation of task, but i think this needs a discussion in a separate issue

Have a fix ready

CPython versions tested on:

CPython main branch

Operating systems tested on:

macOS

Linked PRs

• gh-155575: Fix eager tasks inheriting the creator's exception state #155578

[kumaraditya303]

Can you describe more about the expected behaviour? Is this causing any problem in practise?

[deadlovelll]

Can you describe more about the expected behaviour? Is this causing any problem in practise?

Task shouldnt see the exception state of whoever started it

In practice it produces misleading tracebacks, creating task is a common thing during exception catching i.e we need to send something, and task exception get chained with unrelated one

Example this code:

import asyncio
import logging


async def notify():
    raise ValueError("invalid webhook url")


async def handle():
    try:
        raise ConnectionError("database is down")
    except ConnectionError:
        t = asyncio.create_task(notify())
    try:
        await t
    except Exception:
        logging.exception("failed to notify on-call")


loop = asyncio.new_event_loop()
loop.set_task_factory(asyncio.eager_task_factory)
loop.run_until_complete(handle())

outputs:

ERROR:root:failed to notify on-call
Traceback (most recent call last):
  File "/Users/timofeiivankov/cpython/pp.py", line 11, in handle
    raise ConnectionError("database is down")
ConnectionError: database is down

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/timofeiivankov/cpython/pp.py", line 15, in handle
    await t
  File "/Users/timofeiivankov/cpython/pp.py", line 6, in notify
    raise ValueError("invalid webhook url")
ValueError: invalid webhook url

It gives us a false picture of what caused what because traceback should only be :

ERROR:root:failed to notify on-call
Traceback (most recent call last):
  File "/Users/timofeiivankov/cpython/pp.py", line 15, in handle
    await t
  File "/Users/timofeiivankov/cpython/pp.py", line 6, in notify
    raise ValueError("invalid webhook url")
ValueError: invalid webhook url

[deadlovelll]

And this if we raise explicitly, if we just do raise (that i think is not good, but I can accept that people do that), task fails with exception that was never raised

import asyncio
import logging


async def notify():
    raise


async def handle():
    try:
        raise ConnectionError("database is down")
    except ConnectionError:
        t = asyncio.create_task(notify())
    try:
        await t
    except Exception:
        logging.exception("failed to notify on-call")


loop = asyncio.new_event_loop()
loop.set_task_factory(asyncio.eager_task_factory)
loop.run_until_complete(handle())

ERROR:root:failed to notify on-call
Traceback (most recent call last):
  File "/Users/timofeiivankov/cpython/pp.py", line 15, in handle
    await t
  File "/Users/timofeiivankov/cpython/pp.py", line 11, in handle
    raise ConnectionError("database is down")
ConnectionError: database is down

There should be:

ERROR:root:failed to notify on-call
Traceback (most recent call last):
  File "/Users/timofeiivankov/cpython/pp.py", line 15, in handle
    await t
  File "/Users/timofeiivankov/cpython/pp.py", line 6, in notify
    raise
RuntimeError: No active exception to reraise