Note: I have since figured out inheritance in Python, so this is all just a rant that is no longer relevant, really.
So, I’ve been trying to figure out inheritance in Python. In Lisp (and even C++ and Java), derived objects will have the cons of their parent called. For example, look at the following C++ snippet:
#include <iostream>
using namespace std;
class B {
public:
B() {
cout << __PRETTY_FUNCTION__ << endl;
}
};
class A : public B {
public:
A() {
cout << __PRETTY_FUNCTION__ << endl;
}
};
int
main(int argc, const char **argv)
{
A *foo = new A();
delete foo;
return 0;
}
This outputs the following:
B::B() A::A()
As you can tell, the parent object’s constructor is called and then your derived object’s constructor is called. But, in this Python code, the parent object’s constructor is not:
class B:
def __init__(self):
print "B.__init__(self)"
class A(B):
def __init__(self):
print "A.__init__(self)"
A()
This outputs:
A.__init__(self)
Instead you have to do:
class B:
def __init__(self):
print "B.__init__(self)"
class A(B):
def __init__(self):
B.__init__(self)
print "A.__init__(self)"
A()
To get the right output:
B.__init__(self) A.__init__(self)
This seems broken to me. But maybe I’m not understanding something. The way I got bit by this was by trying to design a new replacement to debian/rules which used subclassing. (More on this later.)
Oh, and I intend to try to keep my blog updated more often.