Timed out statements raises poor exception content

[gmiejski]

Versions:

aiopg==0.16.0
psycopg2==2.7.7

When statement gets cancelled after setting a timeout, only a coroutine CancelledError is raised without any context. I would expect some more information why the query has been cancelled:

Reproduce:

import asyncio

from aiopg.sa import create_engine


def create_eng():
    engine = create_engine(
        "postgresql://postgres:postgres@localhost:5432/postgres?sslmode=disable",
        options="-c statement_timeout={}".format(1000),
    )
    loop = asyncio.get_event_loop()
    engine = loop.run_until_complete(engine)
    return engine


async def main(engine):
    async with engine.acquire() as conn:
        async with conn.execute("SELECT * FROM pg_sleep(5)") as result:
            rows = await result.fetchall()
            print(rows)


asyncio.get_event_loop().run_until_complete((main(create_eng())))

Expected output:

any message saying that execute statement timed out

Actual output:

Traceback (most recent call last):
  File "aiopg_timeout_minimal_example.py", line 23, in <module>
      asyncio.get_event_loop().run_until_complete((main(create_eng())))
  File "/lib/python3.7/asyncio/base_events.py", line 573, in run_until_complete
      return future.result()
concurrent.futures._base.CancelledError

[aio-libs-bot]

GitMate.io thinks possibly related issues are #65 (Closing connection object throws exception), #67 (Queries start failing after some time), #9 (Publish psycopg2 exceptions in aiopg module), #78 (SQL statements logging), and #110 (Exception in Connection.del when database connection has failed.).

[vir-mir]

@gmiejski Hi.
unfortunately CanceledError does not support extended traceback. You can read it here and here.

But I think. We can add loop.call_exception_handler which can be overridden with loop.set_exception_handler and with the help of the standard python logger, you can send errors in sentry... the here

[Pliner]

Resolved by #537