Ignoring KeyboardInterrupt #295

amackillop · 2019-10-27T21:46:40Z

uvloop version: 0.14.0rc1
Python version: 3.7.4
Platform: Manjaro
Can you reproduce the bug with PYTHONASYNCIODEBUG in env?: Yes
Does uvloop behave differently from vanilla asyncio? How?: Vanilla loop respects the KeyboardInterrupt while uvloop ignore it.

Here is a code snippet:

import asyncio
from aiohttp import web
import uvloop

async def start(app: web.Application, host: str, port: int) -> web.AppRunner:
    """Start the server"""
    runner = web.AppRunner(app)
    await runner.setup()
    server = web.TCPSite(runner, host, port)
    await server.start()
    return runner

def main() -> None:
    """Entrypoint"""
    host = "0.0.0.0"
    port = 8000
    app = web.Application()
    loop = asyncio.get_event_loop()
    runner = loop.run_until_complete(start(app, host, port))
    print(f"======== Running on http://{host}:8000 ========\n" "(Press CTRL+C to quit)")
    try:
        loop.run_forever()
    except KeyboardInterrupt:
        loop.run_until_complete(runner.cleanup())


if __name__ == "__main__":
    uvloop.install()
    main()

Execute the file:
python issue.py

Hit Ctrl+C.

The keyboard interrupt will be ignored. If you comment out line 28 uvloop.install() the interrupt is respected. This used to work fine, only noticed that it stopped working today.

1st1 · 2019-10-28T03:35:03Z

Seems that the regression was introduced in 48d376d cc @vladima

1st1 · 2019-10-28T04:04:49Z

Alright, I think I know what causes the bug. Will need to refactor our signals setup/processing code a bit to untangle it. I'll release RC2 tomorrow or the day after with the fix.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.


        Fix KeyboardInterrupt handling logic.

When uvloop is run in the main thread we *always* want to set up a self-pipe and a signal wakeup FD. That's the only way how libuv can be notified that a ^C happened and break away from selecting on sockets. asyncio does not need to do that, as the 'selectors' module it uses is already aware of the way Python implements ^C handling. This translates to a slightly different behavior between asyncio & uvloop: 1. uvloop needs to always call signal.set_wakeup_fd() when run in the main thread; 2. asyncio only needs to call signal.set_wakeup_fd() when a user registers a signal handler. (2) means that if the user had not set up any signals, the signal wakeup FD stays the same between different asyncio runs. This commit fixes uvloop signal implementation to make sure that uvloop behaves the same way as asyncio in regards to signal wakeup FD between the loop runs. It also ensures that uvloop always have a proper self-pipe set up so that ^C is always supported when it is run in the main thread. Issue #295.

1st1 · 2019-10-29T18:43:25Z

Fixed in uvloop 0.14.0rc2. Thank you!

amackillop · 2019-10-29T23:50:14Z

@1st1 I appreciate the quick turnaround. Keep up the great work.

1st1 added the bug label Oct 28, 2019

1st1 mentioned this issue Oct 28, 2019

Fix KeyboardInterrupt handling logic. #296

Merged

1st1 added the fixed in master label Oct 29, 2019

1st1 closed this Oct 29, 2019

MagicStack / uvloop

Ignoring KeyboardInterrupt #295

Ignoring KeyboardInterrupt #295

amackillop commented Oct 27, 2019

1st1 commented Oct 28, 2019

1st1 commented Oct 28, 2019

1st1 commented Oct 29, 2019

amackillop commented Oct 29, 2019

MagicStack / uvloop

Join GitHub today

Ignoring KeyboardInterrupt #295

Ignoring KeyboardInterrupt #295

Comments

amackillop commented Oct 27, 2019

1st1 commented Oct 28, 2019

1st1 commented Oct 28, 2019

1st1 commented Oct 29, 2019

amackillop commented Oct 29, 2019