Python: Understanding yield assignment in generator

Here's an infinite looper which I've added row number prints to to facilitate tracing of the program execution.

def infinite_looper(objects):
    count = 0
    print("row 35")
    while True:
        print("row 37")
        if count >= len(objects):
            count = 0
        print("Row 40")
        message = yield objects[count]
        print("row 42")
        print("The message is "+str(message))
        print("row 44")
        if message != None:
            count = 0 if message < 0 else message
            print("row 47, count = "+str(count))
        else:
            count += 1
            print("row 50")
        print("Row 51")

x = infinite_looper("abcdefghijkl")

print("executing next 1st time")
print(next(x))

print("executing next 2nd time")
print(next(x))

print("executing send 1st time")
print(x.send(10))

The output is:

executing next 1st time
row 35
row 37
Row 40
a
executing next 2nd time
row 42
The message is None
row 44
row 50
Row 51
row 37
Row 40
b
executing send 1st time
row 42
The message is 10
row 44
row 47, count = 10
Row 51
row 37
Row 40
k

What I don't understand is what happens before "executing send 1st time" is printed. b has just been output by the program, presumably through the message = yield objects[count] row in infinite_looper. But then, the value of message is changed to 10 from None even though message = yield objects[count] has already been executed! My only theory is that the yield keyword works in such a way that the execution "stays" on its row after it's been executed, and a send statement to the looper can make the same row (in this case message = yield objects[count]) execute again. Otherwise, we would have:

executing send 1st time
row 42
The message is **None**

Is this a correct theory? More details on how this works?