-
Notifications
You must be signed in to change notification settings - Fork 184
pySCG: Adding explanation of the 'is' operator to CWE-595 #997
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Changes from all commits
8c18703
d0e7050
551425b
a307c58
cd70645
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,14 +1,70 @@ | ||
# CWE-595: Comparison of Object References Instead of Object Contents | ||
|
||
In Python, the `==` operator is implemented by the `__eq__` method on an object [[python.org data model 2023](https://docs.python.org/3/reference/datamodel.html?highlight=__eq__#object.__eq__)]. For built-in types like `int` and `str`, the comparison is implemented in the interpreter. The main issue comes when implementing custom classes, where the default implementation compares object references using the `is` operator. The `is` operator compares the identities of the objects, equivalent to `id(obj1) == id(obj2)`. The `id` function is built into Python, and in the CPython interpreter, the standard implementation, it returns the object's memory address [[de Langen 2023](https://realpython.com/python-is-identity-vs-equality/)]. | ||
Prevent unexpected results by knowing the differences between comparison operators such as `==` and `is`. | ||
|
||
Python falls back to comparing objects' `id()` if the `__eq__` implementation is missing for a custom class. In Python, the `==` operator is implemented by the `__eq__` method on an object [[python.org data model 2023](https://docs.python.org/3/reference/datamodel.html?highlight=__eq__#object.__eq__)]. For built-in types like `int` and `str`, the comparison is implemented in the interpreter. The main issue comes when implementing custom classes, where the default implementation compares object references using the `is` operator. The `is` operator compares the identities of the objects, equivalent to `id(obj1) == id(obj2)`. The `id` function is built into Python, and in the CPython interpreter, the standard implementation, it returns the object's memory address [[de Langen 2023](https://realpython.com/python-is-identity-vs-equality/)]. | ||
|
||
You want to implement the `__eq__` method on a class if you believe you ever want to compare it to another object or find it in a list of objects. Actually, it is so common that the `dataclasses.dataclass` decorator by default implements it for you [[dataclasses — Data Classes — Python 3.11.4 documentation](https://docs.python.org/3/library/dataclasses.html#dataclasses.dataclass)]. | ||
|
||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Be aware of Python's memory optimization for strings and numbers as demonstrated in # SPDX-FileCopyrightText: OpenSSF project contributors
# SPDX-License-Identifier: MIT
""" Code Example """
print("-" * 10 + "Memory optimization with strings" + 10 * "-")
a = "foobar"
b = "foobar"
c = ''.join(["foo", "bar"])
print(f"a is b: {a} is {b}?", a is b)
print(f"a is c: {a} is {c}?", a is c)
print(f"a == c: {a} == {c}?", a == c)
print(f"size? len(a)={len(a)} len(b)={len(b)} len(c)={len(c)}")
print("-" * 10 + "Memory optimization with numbers" + 10 * "-")
a = b = 256
print (f"{a} is {b}?", a is b)
a = b = 257
print (f"{a} is {b}?", a is b)
print("-" * 10 + "Memory optimization with numbers in a loop" + 10 * "-")
a = b = 255
while(a is b):
a += 1
b += 1
print (f"{a} is {b}?", a is b) The a is b: foobar is foobar? True
a is c: foobar is foobar? False
a == c: foobar == foobar? True
size? len(a)=6 len(b)=6 len(c)=6
----------Memory optimisation with numbers----------
256 is 256? True
257 is 257? True
----------Memory optimisation with numbers in a loop----------
256 is 256? True
257 is 257? False The first print statement illustrates Python's memory optimization for strings. The example in the middle creates same object number as The last print statements in example02.py: a = 256
b = 256
print(a is b)
a = 257
b = 257
print(a is b) Using an interactive Python shell will print |
||
Be aware of Python's memory optimization for strings and numbers as demonstrated in `example01.py` code. | ||
Python tries to avoid allocating more memory for the same string. The process of reusing already existing strings is a Python optimization technique known as **String interning** [[sys — System-specific parameters and functions — Python 3.11.4 documentation](https://docs.python.org/3/library/sys.html#sys.intern)] According to the documentation, "CPython keeps an array of integer objects for all integers between `-5` and `256`. When you create an `int` in that range you actually just get back a reference to the existing object." [[Integer objects — Python 3.11.4 documentation](https://docs.python.org/3/c-api/long.html#c.PyLong_FromLong)] | ||
|
||
_[example01.py:](example01.py)_ | ||
|
||
```py | ||
""" Code Example """ | ||
|
||
print("-" * 10 + "Memory optimization with strings" + 10 * "-") | ||
a = "foobar" | ||
b = "foobar" | ||
c = ''.join(["foo", "bar"]) | ||
print(f"a is b: {a} is {b}?", a is b) | ||
print(f"a is c: {a} is {c}?", a is c) | ||
print(f"a == c: {a} == {c}?", a == c) | ||
print(f"size? len(a)={len(a)} len(b)={len(b)} len(c)={len(c)}") | ||
|
||
print("-" * 10 + "Memory optimization with numbers" + 10 * "-") | ||
a = b = 256 | ||
print (f"{a} is {b}?", a is b) | ||
a = b = 257 | ||
print (f"{a} is {b}?", a is b) | ||
|
||
print("-" * 10 + "Memory optimization with numbers in a loop" + 10 * "-") | ||
a = b = 255 | ||
while(a is b): | ||
a += 1 | ||
b += 1 | ||
print (f"{a} is {b}?", a is b) | ||
``` | ||
|
||
**Output of example01.py:** | ||
|
||
```bash | ||
----------Memory optimization with strings---------- | ||
a is b: foobar is foobar? True | ||
a is c: foobar is foobar? False | ||
a == c: foobar == foobar? True | ||
size? len(a)=6 len(b)=6 len(c)=6 | ||
----------Memory optimization with numbers---------- | ||
256 is 256? True | ||
257 is 257? True | ||
----------Memory optimization with numbers in a loop---------- | ||
256 is 256? True | ||
257 is 257? False | ||
``` | ||
|
||
The first set of print statements illustrates string interning. While `a` and `b` reuse the same object, `c` is created by joining two new strings, which results in an object with a different `id()`. The variables in the middle example both point to the same number object, which is why comparing them after `a = b = 257` still returns `True` even though `257` falls outside of the cached range. However, when assigning values in a loop, Python needs to allocate new objects for numbers greater than `256` and thus will create two separate objects as soon as it hits `257`. The way caching and interning works may differ between running a Python script from a file and using REPL, which may produce different results when running `example01.py` in Python's interactive mode. | ||
|
||
## Non-Compliant Code Example | ||
|
||
The non-compliant code shows how the default comparison operator compares object references rather than the object values. Furthermore, it displays how this causes issues when comparing lists of objects, although it applies to other types of collections as well. Finally, it shows how the `in` operator also depends on the behavior of the `__eq__` method and, therefore, also returns a non-desirable result. | ||
The `noncompliant01.py` code demonstrates potentially unexpected outcomes when using different comparisons. | ||
|
||
[*noncompliant01.py:*](noncompliant01.py) | ||
* The `==` operator using `__eq__`, checks value equality for most build-in types, checks for reference equality if the `__eq__` is missing in a custom class. So 12 == 12 is True and Integer(12) == Integer(12) is False. | ||
* The `==` comparing lists of objects, that also applies to other types of collections. | ||
* The `in` operator also depends on the behavior of the `__eq__` method | ||
* The `is` operator that checks the references point to the same object regardless of the stored value. | ||
|
||
_[noncompliant01.py:](noncompliant01.py)_ | ||
|
||
```py | ||
""" Non-compliant Code Example """ | ||
|
@@ -27,41 +83,59 @@ print(Integer(12) == Integer(12)) | |
print([Integer(12)] == [Integer(12)]) | ||
# And this is equally this will always be False as well | ||
print(Integer(12) in [Integer(10), Integer(12)]) | ||
# The 'is' will return True only if both references point to the same object | ||
a = Integer(12) | ||
b = a | ||
# Here, a and b point to the same Integer, so 'is' returns True | ||
print(a is b) | ||
|
||
b = Integer(12) | ||
# Even though b still points to an Integer of the same value, it is a new object, so 'is' returns False | ||
print(a is b) | ||
|
||
``` | ||
|
||
## Compliant Solution | ||
|
||
In this compliant solution the `__eq__` method is implemented and all the comparisons now correctly compares the object values, rather than the object reference. | ||
In this compliant solution, the `__eq__` method is implemented and the comparisons that not use `is` now correctly compare the object values, rather than the object reference. The `is` operator does not call `__eq__`, hence the last print will still display `False`. | ||
|
||
[*compliant01.py:*](compliant01.py) | ||
_[compliant01.py:](compliant01.py)_ | ||
|
||
```py | ||
""" Compliant Code Example """ | ||
|
||
|
||
class Integer: | ||
def __init__(self, value): | ||
self.value = value | ||
|
||
def __eq__(self, other): | ||
if isinstance(other, type(self)): | ||
return self.value == other.value | ||
if isinstance(other, int): | ||
return self.value == other | ||
return False | ||
|
||
|
||
##################### | ||
# exploiting above code example | ||
##################### | ||
# All these scenarios will now show True | ||
print(Integer(12) == Integer(12)) | ||
print([Integer(12)] == [Integer(12)]) | ||
print(Integer(12) in [Integer(10), Integer(12)]) | ||
|
||
# By adding the handling for int we also support | ||
print(Integer(12) == 12) | ||
# The 'is' will return True only if both references point to the same object | ||
a = Integer(12) | ||
b = a | ||
# Here, a and b point to the same Integer, so 'is' returns True | ||
print(a is b) | ||
|
||
b = Integer(12) | ||
# Since the 'is' operator does not call __eq__, print below will still return False | ||
print(a is b) | ||
|
||
``` | ||
|
||
|
@@ -86,3 +160,5 @@ print(Integer(12) == 12) | |
|[[python.org data model 2023](https://docs.python.org/3/reference/datamodel.html?highlight=__eq__#object.__eq__)]|[3. Data model — Python 3.11.3 documentation](https://docs.python.org/3/reference/datamodel.html?highlight=__eq__#object.__eq__)| | ||
|[[de Langen 2023](https://realpython.com/python-is-identity-vs-equality/)]|[Python '!=' Is Not 'is not': Comparing Objects in Python – Real Python](https://realpython.com/python-is-identity-vs-equality/)| | ||
|[[dataclasses — Data Classes — Python 3.11.4 documentation](https://docs.python.org/3/library/dataclasses.html#dataclasses.dataclass)]|[9. Classes — Python 3.11.3 documentation](https://docs.python.org/3/tutorial/classes.html)| | ||
|[[sys — System-specific parameters and functions — Python 3.11.4 documentation](https://docs.python.org/3/library/sys.html#sys.intern)]|[sys — System-specific parameters and functions — Python 3.11.3 documentation](https://docs.python.org/3/library/sys.html#sys.intern)| | ||
|[[Integer objects — Python 3.11.4 documentation](https://docs.python.org/3/c-api/long.html#c.PyLong_FromLong)]|[Integer objects — Python 3.11.4 documentation](https://docs.python.org/3/c-api/long.html#c.PyLong_FromLong)| |
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,27 +1,36 @@ | ||
# SPDX-FileCopyrightText: OpenSSF project contributors | ||
# SPDX-License-Identifier: MIT | ||
""" Compliant Code Example """ | ||
|
||
|
||
class Integer: | ||
def __init__(self, value): | ||
self.value = value | ||
|
||
def __eq__(self, other): | ||
if isinstance(other, type(self)): | ||
return self.value == other.value | ||
if isinstance(other, int): | ||
return self.value == other | ||
return False | ||
|
||
|
||
##################### | ||
# exploiting above code example | ||
##################### | ||
# All these scenarios will now show True | ||
print(Integer(12) == Integer(12)) | ||
print([Integer(12)] == [Integer(12)]) | ||
print(Integer(12) in [Integer(10), Integer(12)]) | ||
|
||
# By adding the handling for int we also support | ||
print(Integer(12) == 12) | ||
# The 'is' will return True only if both references point to the same object | ||
a = Integer(12) | ||
b = a | ||
# Here, a and b point to the same Integer, so 'is' returns True | ||
print(a is b) | ||
|
||
b = Integer(12) | ||
# Since the 'is' operator does not call __eq__, print below will still return False | ||
print(a is b) |
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,25 @@ | ||
# SPDX-FileCopyrightText: OpenSSF project contributors | ||
# SPDX-License-Identifier: MIT | ||
""" Code Example """ | ||
|
||
print("-" * 10 + "Memory optimization with strings" + 10 * "-") | ||
a = "foobar" | ||
b = "foobar" | ||
c = ''.join(["foo", "bar"]) | ||
print(f"a is b: {a} is {b}?", a is b) | ||
print(f"a is c: {a} is {c}?", a is c) | ||
print(f"a == c: {a} == {c}?", a == c) | ||
print(f"size? len(a)={len(a)} len(b)={len(b)} len(c)={len(c)}") | ||
|
||
print("-" * 10 + "Memory optimization with numbers" + 10 * "-") | ||
a = b = 256 | ||
print (f"{a} is {b}?", a is b) | ||
a = b = 257 | ||
print (f"{a} is {b}?", a is b) | ||
|
||
print("-" * 10 + "Memory optimization with numbers in a loop" + 10 * "-") | ||
a = b = 255 | ||
while(a is b): | ||
a += 1 | ||
b += 1 | ||
print (f"{a} is {b}?", a is b) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Suggstion for first sentence:
Prevent unexpected results by knowing what comparisment operators do such as
==
andis
actually do.After the existing first sentence we shall also explain:
Python falls back to comparing objects
id()
if the `eq implementation is missing for a custom class.