Server method calls

[migel]

Hi,
I modified the server-methods.py example to print a timestamp whenever multiple_async is called:

@uamethod
async def multiply_async(parent, x, y):
    sleep_time = x * y
    print(f"{time.time()} Sleeping asynchronously for {x * y} seconds")
    await asyncio.sleep(sleep_time)

on the client side, I called made two calls in parallel:

        async with Client(url=url, timeout=10) as client:
            obj = await client.nodes.root.get_child(["0:Objects", "2:MyObject"])

            t0 = time.time()
            res = await asyncio.gather(obj.call_method("2:multiply_async", 2, 1), obj.call_method("2:multiply_async", 1, 1))
            dt = time.time() - t0
            _logger.info("method result is: %r dt: %f", res, dt)

One call is for a 2 seconds wait and the other is for a 1 second wait. However, since they are called in parallel I expected the overall wait to be around 2 seconds but the log shows 3 seconds:

INFO:__main__:method result is: [None, None] dt: 3.005395

Looking at the server side, it looks like it serializes the calls:

1687280357.4277189 Sleeping asynchronously for 2 seconds
1687280359.433525 Sleeping asynchronously for 1 seconds

Since multiply_async has an async sleep, why aren't the calls processed in parallel?