README.rst

中文

Check function with timeout for tracing or handle it

Feature

• Check a function call is timeout or not.
• The timeout callback and function call are on the same thread.
• Multi-thread support. Nest call support.

Usage

def on_timeout(start_time: float):
    """
    :param start_time:
    """
    traceback.print_stack()
    pdb.set_trace()
    raise Exception("time_out")

# time unit is millisecond
@pymonitor.check_time(10, on_timeout)
def function_1():
    pass

def function_2():
    with pymonitor.check_context(20, on_timeout):
        # do something
        with pymonitor.check_context(10, on_timeout):
            # do something

Implementation Comparison

Here are some comparisons of the other implementations.

• Use signal module and emit a signal
  • Only works in main thread.
  • Not good for nest call becauseof one signal correspond one handler. If you need nest support you need to enter the timeout function continually and call alarm. The cost depend on your accuracy.
  • Support Linux only.
• Start new thread for work and join it with a time, if it had timeout, handle with it (eg. terminate it)
  • Can’t inject the function call.
  • Overhead from threading.
• Use sys.settrace keep tracing for each
  • A huge cost for that.