ipython / ipython/ipykernel

Removing `@register_integration("asyncio")` and `def loop_asyncio(kernel):`

Abierto
#1,532 0 comentarios 0 reacciones 0 asignados Ver en GitHub
Lenguaje dominante
Python
Estrellas
734
Forks
412
Merge medio
1 d 5 h
PR fusionados (30 d)
8

Descripción

In current versions of ipykernel, when you execute in jupyter, vscode jupyter,
```
%gui asyncio
```
then after a while, when you execute
```
import asyncio
asyncio.all_tasks()```
```
you will get and see like
```
{.run_in_context() running at D:\test\ipykernel\ipykernel\utils.py:60> wait_for= cb=[Task.task_wakeup()]> cb=[ZMQStream._run_callback.._log_error() at e:\.pixi\envs\default\Lib\site-packages\zmq\eventloop\zmqstream.py:563]>,
cb=[Task.task_wakeup()]>}
```
It keeps increasing task count and it wastes your some cpu resource in the background. It's acually running these below codes, it just enters the function the return repeatedly, by `self.io_loop.call_later(0.001, advance_eventloop)` of `enter_eventloop()` in `kernelbase.py`.
https://github.com/ipython/ipykernel/blob/8816d024bf730e9fdb051b44ab2fc1ee2276cea2/ipykernel/eventloops.py#L430-L473

How it happened:
The `@register_integration("asyncio")` and `def loop_asyncio(kernel):` was at the old ages, to drive the asyncio codes that inputed by users, like `asyncio.ensure_future(...)`, when ipykernel was running under `IOloop` of `tornado<6.1` which using the `selectorloop` or even older type eventloop. It was to make users able to run asyncio codes when the underhood is other things.
The current ipykernel its main loop is already asyncio loop, by `IOloop` of `tornado>6.1`. The user inputed asyncio codes already run in the main loop.

Therefore the related codes `@register_integration("asyncio")` and `def loop_asyncio(kernel):` could be deleted.

Solution:
related issue: https://github.com/ipython/ipykernel/pull/1469 solving the proactor and selector eventloop within asyncio

Guía de contribución

Abrir la guía de contribución

Evaluación

Este issue todavía no se ha evaluado.

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.