fix no auto reconnect after close/connect in TCPclient

[peufeu2]

Greetings,

AsyncModbusTcpClient.close() followed by connect() to reuse the object left it unable to auto-reconnect because delay_ms was set to 0 in close() but not reset to default setting in connect().

Added canceling of auto-reconnect background task in close().

Added check to avoid launching more than one auto-reconnect task in case user calls connect() too many times.

fixes #1198

[janiversen]

Please also solve the CI errors.

[peufeu2]

The stuff lint was complaining about should be OK
The other pytest failure comes from test_client.py:

    client.delay_ms = 5000
    await client.connect()

    run_coroutine(client._reconnect())  # pylint: disable=protected-access
    mock_sleep.assert_called_once_with(5)

connect() now overwrites delay_ms with the default in params (that solved the bug breaking reconnects after close()). So it overwrites the value set to "5000" and fails this test.

I'm not sure what to do here, because it's up to you to decide if client.delay_ms is an API that should be writable by the user or not. Although I think it makes little sense to have it writable because the reconnect functions in the client modify the value to increase the interval between reconnects, and it is reset to the default value from params after a successful connection. So if the user writes something to it, it'll be short lived anyway...

[janiversen]

delay_ms is not part of the API, but in many of the tests are allowed to use internal values.

Please correct the tests, maybe the tests should simply use the timeout parameter.

Please remember some of the tests are very old, so changing them is quite ok even removing some might be ok. When doing the final review, I will be extra careful when looking at the test changes.

[peufeu2]

Done, looks good!

[janiversen]

LGTM. Thanks

[janiversen]

Seems there are a problem with python 3.11, I have restarted the job to see if it is a real error.