Vladimir Kotal

I don't follow. The original comment says `OSError` [which is exception](https://docs.python.org/3/library/exceptions.html#OSError) however distinct from `MMQTTException`.

Here's one idea for testing: it might be worthwhile to use a good MQTT server implementation (such as Mosquitto). For a functional PUBLISH test it might look like this: 1....

Another idea which I started to implement in PR #187 is to fake the state and produce/capture socket data and verify it against the data sent by known good MQTT...

> I'm very sorry to say, but I have never experienced such a crappy library as this one. Random crashes and hangs everywhere. To be fair, the library got more...

Code to reproduce the issue with (assumes the server behaves according to the MQTT spec and disconnects the client after the keep alive timeout): ```python #!/usr/bin/env python3 import logging import...

@schaefer01 I assume you meant issue #175 ?

The negative errno value was reported on https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/issues/101#issuecomment-1281650450 and supposedly fixed in CP 8.0.0 beta specifically for Pico W (https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/issues/101#issuecomment-1281728775).

Let's start from the end: the `user_data` parameter is handy when you want to avoid global state. For example, there can be a variable that is defined only in main()...

> However, in [c334c81](https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/commit/c334c8157dddc06f67b85b9c299b49d0b29f3031) we started to utilize `recv_into` rather than `recv`, if it was available. Before that change, `recv_into()` was already used (in that if branch); my change merely...

> I might've gotten the exact commit mixed up, it could've been [d60ab23](https://github.com/adafruit/Adafruit_CircuitPython_MiniMQTT/commit/d60ab23c375f86a74c7ddaa55018f20c9d5c1563). I tested my way back and forth until I found which version it broke in. I think...