Cyclic reference

[Gribouillis]

Again, when you run the code from the command line as

python test1.py

The code is compiled and ran in a module which name is not 'test1' but '__main__' because it is the main script. Then when the code calls change(), this function runs import test1. This causes the file to be read again because the filename is test1.py, the code is ran again in a new module which name is 'test1'. So there are now two modules with the same code and there are two 'x' variables.

The 'import test1' statement triggers a call to the new function change() from module test1. This call sets test1.x to the value 2 and prints x. Then the 'import test1' is completed and the first call to __main__.change() can complete, but it doesn't change __main__.x which value is still 1, so the last print statement prints 1.

Here is a modified code to show the different steps

#test1.py 
x=1
def change(): 
    print("Entering change() in module", __name__)
    import test1 
    test1.x=2 
    print("Exiting change() in module", __name__)
change()
print("Printing x in module", __name__)
print(x)

Output:
λ python3 test1.py
Entering change() in module __main__
Entering change() in module test1
Exiting change() in module test1
Printing x in module test1
2
Exiting change() in module __main__
Printing x in module __main__
1