Extend class.__init__ to subclass

[clapperboy]

Hello, 
I'm trying to extend the Base.__init__ to the Derivative.__init__
I've read that by calling it inside the Derivative.__init__ should do the trick.
But I can't get it to work.

class Test():
    def __init__(self, name, default = "default"):
        self.name = name
        self.default = default
 
    def print_init(self):
        print self.name, self.default
 
class Sub(Test):
    def __init__(self, subdefault = "subdefault"):
        Test.__init__(self)
        self.subdefault = subdefault
 
    def print_init(self):
        print self.name, self.default, self.subdefault
 
 
t = Test('hello')
s = Sub('hey', default = "grrr")
t.print_init()
s.print_init()

Now I can completely overwrite the __init__ in Sub to get it to work, but it seems to defeat the purpose of making a subclass.

class Sub(Test):
    def __init__(self, name, default = "default", subdefault = "subdefault"):
        self.name = name
        self.default = default
        self.subdefault = subdefault
 
    def print_init(self):
        print self.name, self.default, self.subdefault

 So is there a way to extend the __init__ of a class in its subclass?

Thanks 

[joojaa]

Well im not really sure a board on maya usage is the best most conductive place to general python usage questions. You can use super to get the superclass. However since python has a multi inheritance so super may become very ugly.  Tough coders usually frown uppon using direct calls to superclass anyway. Mainly because theres no need to do so.

http://en.wikipedia.org/wiki/Call\_super

even if you dont call super you call it implicitly still as bad. So usually you could avoid doing it this way and refactor it away, by making a not so private call that wraps the functionality of the init (it is more discoverable and easier to modularize).

But yes your problem is pretty clear:

def \_\_init\_\_(self, name, default = "default")


def \_\_init\_\_(self, subdefault = "subdefault"):

dont match. so theres no stuff ror the other init to pass the names it must specify something similliar remeber your not feeding the same things. Easiest fix is :

class Sub(Test):
    def __init__(self, name, default = "default",  subdefault = "subdefault"):
        Test.__init__(self, name, default)
        self.subdefault = subdefault
 
    def print_init(self):
        print self.name, self.default, self.subdefault

See subclassing does not do partial lines if you override the function input you override ALL OF IT. No magic there.

[joojaa]

Shit i messed up again. sorry your post got put in devnull by mistake

Well, it is for maya, I just strip down my code to the bare essential for my question.
I did post  on python-forum, but you we're faster.

Well yeah i can appreciate that but still its a bit creepy to be a advisory in both the core aangauage and maya usage at the same time. Because they ar so completely different things. Mayas system is so different form normal python dev that you dont really need to know this kind of stuff.

Id figure most user comes from mel, so it could help them as well.

Well this is actally my main concern. People did all this ranting before that mel is brain dead etc. and then lo behold they get python and they are happy. And i go "WTF, seriously get a grip". Sure mel was retarded, but teh python interface of maya is at least equaly retarded. Mind you the api stuff is a bit better but still not good enough.

Truly get this people want things to get done. But im so sick and fed up with people who and go run around in the dark and then bump to the wall. Now in this analogy mel is a frigging padded room, its a elongated square so once you find the corners you know where everything is. But python on the other hand is a frigging labyrinth which starts with padded walls only to eventually turn into a corridor with razor sharp 10 meter long spikes all over the place.

What i mean with this ia that the constraints of where maya works didnt change. Your now living in a sort of illusion where it feels like your in charge but you never are more than in mel. On the other hand some features do make it nicer. But still. A better language with a crowd that didnt know where to begin with in mel dont know any better in python. But they sure as hell run further in the maze of their own construction.

Besides i want to use Lisp or Haskel

PS: And in all seriousness with all the firgging rpblems with mel its got one of the coolest EVER features. The mayas echo service. I mean itd so frigging great id want EVERY program in the WORLD to emit a echo service like that. Seriously. But thenagain they should have chosen TCL