Raj B <[EMAIL PROTECTED]> wrote: > Hi > > My question is about how special methods are stored internally in > Python objects. > Consider a new-style class which implements special methods such as > __call__ and __new__ > > class C(type): > def __call__(...): > <body> > > class B: > __metaclass__ = C > <stuff> > > b= B() > > The type of C is 'type', that of B is 'C'. When B is instantiated, > the __call__ method of C is first invoked, since C is the metaclass > for B. > > Internally, when a Python callable object 'obj' is called, the actual > function called seems to be > 'obj->ob_type->tp_call'. > > Does this that somehow the '__call__' method defined in C above is > assigned to the 'tp_call' slot in the object representing the class > C, instead of it just being stored in the dictionary like a normal > attribute? Where and how does this magic happen exactly? I'd > appreciate any level of detail.
Yes, special methods populate the slots in the structures which Python uses to represent types. Objects/typeobject.c in the Python source distribution does the hard work, particularly in function type_new (line 1722 in my current SVN checkout). If you're not comfortable reading C code you may want to try looking at the "Python implemented in Python" project, pypy, or perhaps alternatives such as Jython (in Java) or better IronPython (in C#), but I am not familiar in detail with how they deal with the issue. Alex -- http://mail.python.org/mailman/listinfo/python-list
