Pierre Fersing

Done some fix in e96ff7578f7f56a7f3068fbc3e7d0b48e1c64a97 which seems good enough. The last run on master worker without need to retry.

Fixing thread-safety is something we want. If thread-safety is fixed, I don't think such callback are still needed. I would prefer avoiding to add more callback / way to use...

But in your case the thread safety issue is not on paho library side, but on your data structure. You should take a lock before updating the data structure and...

But the user supplied method will only solve simple lock needs. One that only need a single lock which aren't tricky. Or do you have a use-case where the callback...

The publisher client is not running. You never called loop/loop_start or loop_forever. It also seems you never wait for publisher to finish publishing (no disconnect, no even a arbitrary sleep)....

Do you see an use case of reinitialise that is not solved by really re-creating the client ? E.g. instead of: ``` myclient.reinitialise([...]) ``` Do: ``` myclient = paho.Client([...]) ```...

Thanks for your contribution. I've written some test in https://github.com/eclipse/paho.mqtt.python/commit/a2574e61aea8ebb3bb5df3926f4b8771ca0afe8e (pushed on master) which show the problem exist on master. It also show that there is another problem: message are...

Do you have more details on when that happen ? It's Windows with Python 3.10.N ? Are you using SSL connection ? Websocket ? This only happen while debugging is...

Version 1.1 starts to be an old version :( Can you try with latest version ?

I'm also hitting this issue. And I *think* it the cause of the production issue seen in #2480 For me, what happen is: * uWSGI start the worker loop here...