I've found the following open source code in Python:
class Wait:
timeout = 9
def __init__(self, timeout=None):
if timeout is not None:
self.timeout = timeout
...
I'm trying to understand if there are advantages of the code above vs using default argument's value:
class Wait:
def __init__(self, timeout=9):
...
It's possible to change the default value this way:
Wait.timeout = 20
Will mean that, if unset, the default will be 20.
E.g:
>>> class Wait:
... timeout = 9
... def __init__(self, timeout=None):
... if timeout is not None:
... self.timeout = timeout
...
>>> a = Wait()
>>> b = Wait(9)
>>> a.timeout
9
>>> b.timeout
9
>>> Wait.timeout = 20
>>> a.timeout
20
>>> b.timeout
9
This utilises the fact that Python looks for class attributes if it doesn't find an instance attribute.
Semantically, a class attribute is like making the default timeout part of the public interface of the class. Depending on the documentation, an end user may be encouraged to read or possibly change the default.
Using a default parameter value instead strongly suggests that the particular default value is an implementation detail, not to be fiddled with by end users of the class.
_timeout=9.
self.timeout = 9, in the second example construct, you're setting a class variable/attribute.self.x = ..., you are setting an instance attribute.Wait.timeoutthrows anattributeError. In the first, it does not.