Why defining a default list for a class method's parameter makes it shared across instances?
Does that mean that this list param, in the following example it is props, got declared as a class attribute instead of an instance attribute?
And if so, does that mean that all the class method's params got defined as class attributes? and if so, why setting the name by appending a text string does not produce the same issue?
Example
Given the following example
class Node():
def __init__(self, name="", props=[]):
self.name = name
self.props = props
def __str__(self):
return f"{self.name:<10} : props [{self._joined_props()}]"
def AddProp(self, ptext):
self.props.append(ptext)
def AppendName(self, nodename):
self.name += nodename
def _joined_props(self):
return ", ".join(self.props)
node1 = Node()
node1.AppendName("node1")
node1.AddProp("p1")
node2 = Node()
node2.AppendName("node2")
node2.AddProp("p2")
print(node1)
print(node2)
outputs
node1 : props [p1, p2]
node2 : props [p1, p2]
Changing the __init__() to be
def __init__(self, name="", props=[]):
self.name = name
# self.props = props # old
self.props = [] # new
outputs the expected result
node1 : props [p1]
node2 : props [p2]
For a more realistic example, follow the Node class in the tree data structure demo defined @ weshouman\tut-py-tree:tree.py#L61
