python - How to avoid pickling errors when sharing objects between threads? -

i have program in need store global variable file. doing using pickle module.

i have thread(daemon=false, threading module) changes value of global variable. value modified in global scope(the main program).

i dumping value of variable .pkl file every 5 seconds (using thread threading module).

but found following error when dump method executed:

typeerror: can't pickle _thread.lock objects 

why happening? , can fix it?

note: have found similar answers multiprocessing module. need answer threading module.

code:

def save_state():     while true:         global variable          lastsession = open('lastsession.pkl', 'wb')          # error occurs on line         pickle.dump(variable, lastsession)          lastsession.close()                   time.sleep(5)  state_thread = threading.thread(target = save_state) state_thread.setdaemon(false) state_thread.start()  # variable changed outside function , in thread(not state_thread). 

as others have mentioned, cannot pickle "volatile" entities (threads, connections, synchonization primitives etc.) because don't make sense persistent data.

looks you're trying saving session variables later continuation of said session. task, there's nothing can save objects that, well, cannot saved due nature.

the simplest solution ignore them. replacing them "stubs" yield error anytime them makes little sense, missing variable yields error anyway (it won't if masking global variable questionable practice in itself).

alternatively, can set these objects anew when restoring session. such logic necessity task-specific.

finally, e.g. ipython already has session saving/restoration logic, don't need reinvent wheel @ all.