Wait for cleanup coroutines before event loop is closed.

[xloem]

BPO	42838
Nosy	@asvetlov, @1st1, @xloem

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2022-03-18.10:24:26.007>
created_at = <Date 2021-01-06.13:56:08.313>
labels = ['type-feature', '3.8', '3.10', 'expert-asyncio']
title = 'Wait for cleanup coroutines before event loop is closed.'
updated_at = <Date 2022-03-18.10:26:23.855>
user = 'https://github.com/xloem'

bugs.python.org fields:

activity = <Date 2022-03-18.10:26:23.855>
actor = 'xloem'
assignee = 'none'
closed = True
closed_date = <Date 2022-03-18.10:24:26.007>
closer = 'xloem'
components = ['asyncio']
creation = <Date 2021-01-06.13:56:08.313>
creator = 'xloem'
dependencies = []
files = []
hgrepos = []
issue_num = 42838
keywords = []
message_count = 5.0
messages = ['384503', '415411', '415430', '415432', '415481']
nosy_count = 3.0
nosy_names = ['asvetlov', 'yselivanov', 'xloem']
pr_nums = []
priority = 'normal'
resolution = 'out of date'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue42838'
versions = ['Python 3.8', 'Python 3.10']

[xloem]

To handle destruction of resources especially during exceptions, it would be nice if there were some way to provide coroutines/tasks that run at the termination or closure of an event loop. There are a lot of api options here. Maybe a simple one would be to have close() call a run_until_closed() function that starts all of these and steps the loop to wait for them, before closing, and provide run_close() functions to queue them for delay.

[asvetlov]

asyncio.run() does the task

[xloem]

I'm sorry, is this closure an error? Could you explain more how to use asyncio.run() to clean up resources when an unhandled exception is thrown? Is this new with a recent python improvement?

[asvetlov]

asyncio.run() present since Python 3.7
After the main coroutine finish, it cancels all background tasks and waits for their termination.

Background tasks can use old good try/finally for resources cleanup.
An additional API is not required IMHO.

[xloem]

hey, I don't have the capacity to stay on this, but thanks for the attention, time, and clear response.

there are of course other situations such as returning a resource to library code or manual loop management, but I don't have the use case present any more.

maybe an aatexit library could patch this in if needed.