set.__or__, __and__, etc create subclass types, but ignore __new__

[JonObermark]

BPO	15879
Nosy	@amauryfa, @benjaminp, @bitdancer
Superseder	bpo-1721812: A subclass of set doesn't always have init called.

^{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 = None
closed_at = <Date 2012-09-07.16:44:39.508>
created_at = <Date 2012-09-07.15:44:41.370>
labels = []
title = 'set.__or__, __and__, etc create subclass types, but ignore __new__'
updated_at = <Date 2012-09-07.17:26:30.617>
user = 'https://bugs.python.org/JonObermark'

bugs.python.org fields:

activity = <Date 2012-09-07.17:26:30.617>
actor = 'Jon.Obermark'
assignee = 'none'
closed = True
closed_date = <Date 2012-09-07.16:44:39.508>
closer = 'amaury.forgeotdarc'
components = []
creation = <Date 2012-09-07.15:44:41.370>
creator = 'Jon.Obermark'
dependencies = []
files = []
hgrepos = []
issue_num = 15879
keywords = []
message_count = 4.0
messages = ['169986', '169989', '169991', '169997']
nosy_count = 4.0
nosy_names = ['amaury.forgeotdarc', 'benjamin.peterson', 'r.david.murray', 'Jon.Obermark']
pr_nums = []
priority = 'normal'
resolution = 'out of date'
stage = None
status = 'closed'
superseder = '1721812'
type = None
url = 'https://bugs.python.org/issue15879'
versions = []

[JonObermark]

If they are not going to call the metaclass or the class new, then they should return set objects instead of subclass objects, so that it is clear what is going on.

As it is, the results of set operations receive some subclass information but not all. So they are not really obeying the notion of a subclass: the results are neither set objects, nor properly-constructed objects of the set subclass.

e.g.
class Fooset(Set):
def __new__(cls, s = []):
print 'New called'
self = super(Fooset, cls).__new__(cls)
self.update(s)
if isinstance(s, Fooset):
self.foo = s.foo
else:
self.foo = 'default'
return self

x = Fooset([1,2,5])
y = x|x

The object y reports being of the type Fooset, but has not been constructed by the type that makes Fooset objects.

[bitdancer]

It is probably true that this is a bug, but subclasses of mutable classes do not normally override __new__. (I'm curious what your use case is for doing so.)

[amauryfa]

This was already fixed (in the 3.x releases) by bpo-1721812.

[JonObermark]

The closing author is correct, the use case is adequately covered by __init__, and that has been fixed in 3. It makes sense this will not be backported.