Type subclassing: bug or feature

[Bengt Richter]

On 13 Jun 2002 16:08:40 -0400, aahz at pythoncraft.com (Aahz) wrote:

>Consider the following code:
>
>class MyStr(str):
>    def contains(self, value):
>        return self.find(value) >= 0
>
>s = MyStr("hello, world!")
>s = s.capitalize()
>if s.contains('Hello'):
>    print "Found it!"
>
>It fails with an AttributeError when it calls s.contains(), because
>s.capitalize() returned a str instead of a MyStr.  Anyone want to take a
>whack at defending this as the correct behavior?
I don't think you could generally coerce base class method results
to subclass instances, so I think it has to be explicit. Maybe it
could be done with a method list for a more compact way (using a metaclass?)
Anyway:

 >>> class MyStr(str):
 ...     def contains(self, value):
 ...         return self.find(value) >= 0
 ...     def capitalize(self):
 ...         return MyStr(str.capitalize(self))
 ...
 >>> s = MyStr("hello, world!")
 >>> s = s.capitalize()
 >>> if s.contains('Hello'):
 ...     print "Found it!"
 ...
 Found it!


Regards,
Bengt Richter