I have a base class B and a derived class D which inherits from B. I also have a D2 class which inherits from D. D is used as a base class for most of my generated classes. I have a "special" class which inherits from D2 because I need to override a couple of its methods. Anything based on D2 will inherit Encode and Decode from B.
class D(b.B): def __init__(self, data, timestamp = None): b.B.__init__(self, data) self.timestamp = timestamp class D2( D ): def __init__(self, data, timestamp = None): DS.__init__(self, data, timestamp) def Decode(self): pass def Encode(self): pass The generated classes come from this loop: for m in mdefs: mdef = mdefs[m] name = mdef['name'] + fbase if mdef.has_key('baseclass'): base_class_seq = mdef['baseclass'] else: base_class_seq = DEFAULT_BASE_CLASS nclass = new.classobj( name, base_class_seq, globals() ) base_class_seq is either going to be (D,) or (D2,) After I get through the class generation, I want to say globals()['IFRAMED2'].Decode = dynDecode globals()['IFRAMED2'].Encode = dynEncode for the one class that inherited from D2. If I do that, I find that *all* classes that call Encode or Decode are all calling dyn{En,De}code (which is exactly what I *don't* want). So I looked at the id values after class generation. print 'IFRAMED2.Encode = ', IFRAMED2.Encode, id(IFRAMED2.Encode) print 'SNRMD.Encode = ', SNRMD.Encode, id(SNRMD.Encode) print 'IFRAMED2 = ', IFRAMED2, id(IFRAMED2) print 'SNRMD = ', SNRMD, id(SNRMD) IFRAMED2.Decode = dynDecode IFRAMED2.Encode = dynEncode print 'IFRAMED2.Encode = ', IFRAMED2.Encode, id(IFRAMED2.Encode) print 'SNRMD.Encode = ', SNRMD.Encode, id(SNRMD.Encode) Here's the output: IFRAMED2.Encode = <unbound method IFRAMED2.Encode> 1100623660 SNRMD.Encode = <unbound method SNRMD.Encode> 1100623660 IFRAMED2 = d.IFRAMED2 1076299644 SNRMD = d.SNRMD 1103202364 IFRAMED2.Encode = <unbound method IFRAMED2.dynEncode> 1100623660 SNRMD.Encode = <unbound method SNRMD.dynEncode> 1100623660 So it looks like the IFRAMED2 class which inherits from D2 is starting out with the same id value for Encode as SNRMD which inherits from D, even though D2 defines its own Encode method. Is it me of is it the interpreter doing something wrong? -- Time flies like the wind. Fruit flies like a banana. Stranger things have .0. happened but none stranger than this. Does your driver's license say Organ ..0 Donor?Black holes are where God divided by zero. Listen to me! We are all- 000 individuals! What if this weren't a hypothetical question? steveo at syslang.net -- http://mail.python.org/mailman/listinfo/python-list