[New-bugs-announce] [issue35463] mock uses incorrect signature for partial and partialmethod with autospec
Karthikeyan Singaravelan
report at bugs.python.org
Tue Dec 11 09:47:32 EST 2018
New submission from Karthikeyan Singaravelan <tir.karthi at gmail.com>:
This is a bug report for https://bugs.python.org/issue17185#msg331149 that I was asked to raise as a separate issue.
1. When we call create_autospec it calls _get_signature_object that gets the signature for the given parameter. With functools.partial it returns a partial object and hence while getting the signature it returns the signature for the constructor of partial instead of the underlying function passed to functools.partial. I think a check needs to be added to make sure not to use func.__init__ when it's a partial object.
2. When we call create_autospect on a class that has a partialmethod the self parameter is not skipped in the signature and thus it creates a signature with self causing error. The fix would be to handle partialmethod also in _must_skip that determines whether to skip self or not.
Sample reproducer :
from functools import partial, partialmethod
from unittest.mock import create_autospec
import inspect
def foo(a, b):
pass
p = partial(foo, 1)
m = create_autospec(p)
m(1, 2, 3) # passes since signature is set as (*args, **kwargs) the signature of functools.partial constructor. This should throw TypeError under autospec
class A:
def f(self, a, b):
print(a, b)
g = partialmethod(f, 1)
m = create_autospec(A)
m().g(1, 2) # passes since signature is set as (self, b) and self is not skipped in _must_skip thus self=1, b=2. This should throw TypeError under autospec since the valid call is m().g(2)
----------
components: Library (Lib)
messages: 331631
nosy: cjw296, mariocj89, michael.foord, pablogsal, xtreak
priority: normal
severity: normal
status: open
title: mock uses incorrect signature for partial and partialmethod with autospec
type: behavior
versions: Python 3.7, Python 3.8
_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue35463>
_______________________________________
More information about the New-bugs-announce
mailing list