[Tutor] What protocol to follow when need to pick either one from __getattr__ and __getattribute__ ?

[Steven D'Aprano]

On Tue, Apr 23, 2019 at 08:22:56PM +0530, Arup Rakshit wrote:
> I read today 2 methods regarding the customizing the attribute 
> access:__getattr__ and __getattribute__ from 
> https://docs.python.org/3/reference/datamodel.html#special-method-names. 
> What I understood about them is that __getattr__ is called when the 
> requested attribute is not found, and an AttributeError is raised. But 
> later is called everytime unconditionally.


When you overload __getattribute__, your class will be slow because 
**every** attribute lookup has to go through your method, instead of 
only the lookups which fail. And unless you are very careful, you will 
break things:



py> class X(object):
...     def __init__(self, x):
...             self.x = x
...     def __getattribute__(self, name):
...             if name is 'x': return self.x
...
py> obj = X(999)
py> obj.x
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<stdin>", line 5, in __getattribute__
  [ previous line repeats 332 times ]
RecursionError: maximum recursion depth exceeded while calling a Python 
object



Normally, overriding __getattribute__ is considered a very advanced and 
unusual thing to do.


-- 
Steven