How does Timer in Python work, regarding mutlithreading?

Question

If I call

Timer(.1, some_function, [some_arguments]).start()

multiple times, what exactly happens behind the scenes?

The source of our problem is ...

We have a method that's essentially:

def move(target):
 force = calculateForce(target-getCurrentPosition())
 if(force != 0)
   setForce(force)
   Timer(.1, moveCursor, [tx]).start()
 else:
   setForce(0)

After setting the force, we need to check after a certain amount of time whether it should be stopped. (This information is to/from an external physical device that doesn't fire events.)

There's weird issues in how much time this function is taking, and also we're getting "can't start new thread" errors after a certain amount of time.

This leads me to believe that Timer(...) does not reuse threads but creates a new one every time.

Combined with a belief that the library we're using isn't threadsafe, these errors would make some sense....

Answer 1

Right: each call to Timer does start a new thread. Indeed, class threading.Timer is documented as being "a thread". You can confirm this by reading the source code, line 707.

A good alternative is to run a scheduler in a single thread, receiving requests through a Queue.Queue instance (intrinsically threadsafe) and intrinsically serializing them (which may also take care of your "non-thread-safe-library" problems without needing further locking or synchronization, depending how you arrange your overall architecture of course).