10

I'm running the following code which makes 5 requests via aiohttp:

import aiohttp
import asyncio

def fetch_page(url, idx):
    try:
        url = 'http://google.com'
        response = yield from aiohttp.request('GET', url)

        print(response.status)
    except Exception as e:
        print(e)

def main():
    try:
        url = 'http://google.com'
        urls = [url] * 5

        coros = []
        for idx, url in enumerate(urls):
            coros.append(asyncio.Task(fetch_page(url, idx)))

        yield from asyncio.gather(*coros)
    except Exception as e:
        print(e)

if __name__ == '__main__':
    try:
        loop = asyncio.get_event_loop()
        loop.run_until_complete(main())
    except Exception as e:
        print(e)

Output:

200
200
200
200
200
Exception ignored in: Exception ignored in: Exception ignored in: Exception ignored in: Exception ignored in:

Note: There is no additional information as to what/where the exception is.

What's causing this and are there any tips to debug it?

dano
  • 91,354
  • 19
  • 222
  • 219
okoboko
  • 4,332
  • 8
  • 40
  • 67
  • Note that wrapping whole functions in `try: ... except Exception` is not best practice. What are you running this code *in* - is the *"Exception ignored"* message from your IDE? – jonrsharpe Apr 07 '15 at 22:56
  • I added those to try to catch the error - the "Exception ignored" is actually printed without the `try... except` present. I'm thinking that something in `aiohttp` or `asyncio` is muting out the exceptions. I'm running this in a script executed via terminal in Python 3.4. – okoboko Apr 07 '15 at 22:58
  • Looks like this might be related to [this python bug](https://bugs.python.org/issue22836). I'm not sure why an unraisable exception is occurring in the first place, though; I can't reproduce it on my system. – dano Apr 07 '15 at 23:04
  • @okoboko Also see the comments at the end of [this bug](http://bugs.python.org/issue23548) where this same error is mentioned when exiting an `asyncio` application. Are you using Python 3.4.3? – dano Apr 07 '15 at 23:12
  • Hmm, that's strange indeed. I'm using 3.5.0a1. Glad it appears to be unrelated to my code. – okoboko Apr 07 '15 at 23:24
  • @okoboko Well, according to the comments it probably means there's a task that isn't complete at the end of your program, but that doesn't seem to be the case here. Try adding `response.close()` at the end of `fetch_page`, and `loop.close()` after the `loop.run_until_complete(..)` line, and see if that makes a difference. – dano Apr 07 '15 at 23:25
  • @dano `response.close()` fixed it. Will you add that as an answer so I can accept? – okoboko Apr 07 '15 at 23:28
  • @dano Even I am facing the same issue though I am not using asyncio or aiohttp. It occurs when I am trying to allocate a class instance to a dictionary as a value to an integer key. Any idea how to stop getting the error. – blackmamba Apr 30 '15 at 11:03

2 Answers2

12

I'm not exactly sure why, but it seems that leaving the aiohttp.ClientResponse object open is causing an unraisable exception to be thrown when the interpreter exits. On my system, this results in warnings like this, rather than "Exception ignored in" messages:

sys:1: ResourceWarning: unclosed <socket object at 0x7f44fce557a8>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fce55718>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fcc24a78>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fcc248c8>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fcc24958>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fcc249e8>
sys:1: ResourceWarning: unclosed <socket object at 0x7f44fcc24b08>

In any case, you can fix it by explicitly closing the ClientResponse object at the end of fetch_objects, by calling response.close().

dano
  • 91,354
  • 19
  • 222
  • 219
0

I had a similar problem but using the Context class and its method request, you have to not forget to call shutdown() method on your Context object at the end to avoid the Exception ignored in: message. As the shutdown method is async and I was using python 3.4 the call is done using yield from.

Hamdi Ryad
  • 1
  • 1