AW: The rules of reference

[Gerson Kurz]

Jeff Epler wrote

> The code in "a.py" is executed once when it's named on the commandline
> (inside a module called '__main__') and then again when 'import a' is
> executed (inside a module called 'a')

Well, yes. But take again a look at the code of a.py:

num = 42

def print_num():
    global num
    print "a.print_num() called, num is %d" % num

if __name__ == "__main__":
    import b
    print "Step 1: num should be 42"
    print_num()
    num = 60
    print "Step 2: num should be 60"
    print_num()
    print "Step 3: num should *still* be 60, but is probably not:"
    b.do_stuff()

As you can see, the assignment "num = 60" is called *after* "import b". I
know that "import b" does a recursive "import a", but *that is already done*
at the time I do the "num=60" bit. So this is what I don't get.