The uwsgi documentation states that adding uwsgidecorators.mulefunc decorator to any function will offload function call to mule.
The default mule(without a brain) does not execute actual function and just show messages like this:
*** mule 1 received a 97 bytes message ***
However digging into the code shows that there is a uwsgi.mule_msg_hook function which is for handling offloaded calls(BTW nothing in documentation about mule_msg_hook).
And by giving a simple brain to mule like this:
while True: uwsgi.mule_msg_hook(uwsgi.mule_get_msg())
Will solve the problem.
But there is one place in source code of uwsgi(https://github.com/unbit/uwsgi/search?utf8=%E2%9C%93&q=mule_msg_hook file plugins/python/python_plugin.c ) where mule_msg_hook is called. So I’m guessing that there exists another way to handle mulefuncs without using non-documented funcitons, is it?