Do not raise AttributeError on instance attribute update/deletion if data descriptor with missing __set__/__delete__ method found on its type

[maggyero]

BPO	43639
Nosy	@rhettinger, @maggyero
PRs	bpo-43639: Do not raise AttributeError on instance attribute update/deletion if data descriptor with missing __set__/__delete__ method found on its type #25033

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/rhettinger'
closed_at = <Date 2021-12-29.09:19:39.583>
created_at = <Date 2021-03-27.12:40:32.176>
labels = ['interpreter-core', 'type-bug', '3.10']
title = 'Do not raise AttributeError on instance attribute update/deletion if data descriptor with missing __set__/__delete__ method found on its type'
updated_at = <Date 2021-12-29.09:19:49.485>
user = 'https://github.com/maggyero'

bugs.python.org fields:

activity = <Date 2021-12-29.09:19:49.485>
actor = 'rhettinger'
assignee = 'rhettinger'
closed = True
closed_date = <Date 2021-12-29.09:19:39.583>
closer = 'rhettinger'
components = ['Interpreter Core']
creation = <Date 2021-03-27.12:40:32.176>
creator = 'maggyero'
dependencies = []
files = []
hgrepos = []
issue_num = 43639
keywords = ['patch']
message_count = 2.0
messages = ['389598', '409297']
nosy_count = 2.0
nosy_names = ['rhettinger', 'maggyero']
pr_nums = ['25033']
priority = 'normal'
resolution = 'rejected'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue43639'
versions = ['Python 3.10']

[maggyero]

Currently, the object.__setattr__ and type.__setattr__ methods raise an AttributeError during attribute update on an instance if its type has an attribute which is a data descriptor without a __set__ method. Likewise, the object.__delattr__ and type.__delattr__ methods raise an AttributeError during attribute deletion on an instance if its type has an attribute which is a data descriptor without a __delete__ method.

This should not be the case. When update/deletion is impossible through a data descriptor found on the type, update/deletion should carry its process on the instance, like when there is no data descriptor found on the type. And this is what the object.__getattribute__ and type.__getattribute__ methods already do: they do not raise an AttributeError during attribute lookup on an instance if its type has an attribute which is a data descriptor without a __get__ method. See the discussion on Python Discuss.

Here is a simple program illustrating the differences between attribute lookup by object.__getattribute__ on the one hand (AttributeError is not raised), and attribute update by object.__setattr__ and attribute deletion by object.__delattr__ on the other hand (AttributeError is raised):

class DataDescriptor1:  # missing __get__
    def __set__(self, instance, value): pass
    def __delete__(self, instance): pass

class DataDescriptor2:  # missing __set__
    def __get__(self, instance, owner=None): pass
    def __delete__(self, instance): pass

class DataDescriptor3:  # missing __delete__
    def __get__(self, instance, owner=None): pass
    def __set__(self, instance, value): pass

class A:
    x = DataDescriptor1()
    y = DataDescriptor2()
    z = DataDescriptor3()

a = A()
vars(a).update({'x': 'foo', 'y': 'bar', 'z': 'baz'})

a.x
# actual: returns 'foo'
# expected: returns 'foo'

a.y = 'qux'
# actual: raises AttributeError: __set__
# expected: vars(a)['y'] == 'qux'

del a.z
# actual: raises AttributeError: __delete__
# expected: 'z' not in vars(a)

Here is another simple program illustrating the differences between attribute lookup by type.__getattribute__ on the one hand (AttributeError is not raised), and attribute update by type.__setattr__ and attribute deletion by type.__delattr__ on the other hand (AttributeError is raised):

class DataDescriptor1:  # missing __get__
    def __set__(self, instance, value): pass
    def __delete__(self, instance): pass

class DataDescriptor2:  # missing __set__
    def __get__(self, instance, owner=None): pass
    def __delete__(self, instance): pass

class DataDescriptor3:  # missing __delete__
    def __get__(self, instance, owner=None): pass
    def __set__(self, instance, value): pass

class M(type):
    x = DataDescriptor1()
    y = DataDescriptor2()
    z = DataDescriptor3()

class A(metaclass=M):
    x = 'foo'
    y = 'bar'
    z = 'baz'

A.x
# actual: returns 'foo'
# expected: returns 'foo'

A.y = 'qux'
# actual: raises AttributeError: __set__
# expected: vars(A)['y'] == 'qux'

del A.z
# actual: raises AttributeError: __delete__
# expected: 'z' not in vars(A)

[rhettinger]

I'm going to decline this one.

• It is arguable whether or not this behavior should have been designed in originally. However, changing it now is risky (as noted by Brett and Ethan).

• Personally, I disagree with the notion of allowing a partial pass through. That seems hazardous and error-prone to me. It is easier to reason about data descriptors being all or nothing. I like that AttributeError is raised while still allowing me to add the missing methods if I want to explicitly define some other behavior.

• This has been open for 9 months and no one else stepped forward to champion it.

• For two decades, no one has complained that the current behavior got in the way of what they were trying to do. That provides some evidence that there isn't a real world problem to be solved.