Research
Security News
Malicious npm Package Targets Solana Developers and Hijacks Funds
A malicious npm package targets Solana developers, rerouting funds in 2% of transactions to a hardcoded address.
The main aim of this package is to define the class DeepImmutableProxy
that prevents a user-defined class to be mutated to any level of depth.
DeepImmutableProxy
takes an object as an argument and acts as a proxy to that object.
Attributes and methods can be accessed normally unless they try to mutate any object that is accessible from the wrapped object.
This package is written in pure python. Simply install it using pip
pip install immutable-proxy
>>> from immutable import DeepImmutableProxy, ConstantAttributeError
We define a simple class:
>>> class Example():
... def __init__(self, x):
... self.x = x
... def __repr__(self):
... return f'{type(self).__qualname__}({self.x!r})'
>>> example = Example(42)
>>> example.x = 0
>>> print(example)
Example(0)
If we wrap an object with DeepImmutableProxy
, then this object cannot be mutated.
>>> proxy = DeepImmutableProxy(Example(42))
>>> try:
... proxy.x = 0
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Example]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... del proxy.x
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Example]' object is immutable. Cannot delete attribute 'x' after initialization.
>>> print(proxy)
Example(42)
The following example demonstrates a more complicated class with properties and methods that set attributes.
>>> class Inner:
... def __init__(self, x, y):
... self.x = x
... self.y = y
... def __repr__(self):
... return f'{type(self).__qualname__}({self.x!r}, {self.y!r})'
... def set_x(self, x):
... "Demonstrates a way to set attribute 'x'."
... self.x = x
... def get_x(self) -> int:
... return self.x
... @property
... def y(self):
... "Demonstrates a property to handle attribute 'y'."
... return self.__dict__['y']
... @y.setter
... def y(self, value):
... self.__dict__['y'] = value
The objects of this class work as expected.
>>> inner = Inner(1, 2)
>>> inner.x, inner.y = -1, -2
>>> print(inner)
Inner(-1, -2)
>>> inner.set_x(3)
>>> print(inner)
Inner(3, -2)
Now we wrap an object of this class with DeepImmutableProxy
.
This object cannot be mutated.
>>> inner_proxy = DeepImmutableProxy(Inner(1, 2))
>>> try:
... inner_proxy.x = -1
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... inner_proxy.y = -2
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'y' after initialization.
>>> try:
... inner_proxy.set_x(3)
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... del inner_proxy.x
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot delete attribute 'x' after initialization.
We define another class whose attributes are of type Inner
.
It also contains an attribute referring to another object that refers back to an object of this class. The initial object cannot be modified indirectly trough this back reference.
>>> class Outer:
... def __init__(self, a: Inner, b: Inner) -> None:
... self.a = a
... self.b = b
... self.delegate = Modifier(self)
... def __repr__(self):
... return f'{type(self).__qualname__}({self.a!r}, {self.b!r})'
... def set_a(self, a):
... 'Sets a.'
... self.a = a
... def get_a(self):
... return self.a
... @property
... def b(self):
... return self.__dict__['b']
... @b.setter
... def b(self, value) -> None:
... self.__dict__['b'] = value
>>> class Modifier():
... def __init__(self, obj):
... self.back = obj
... def reset(self, x):
... self.back.x = x
>>> outer_proxy = DeepImmutableProxy(Outer(Inner(1, 2), Inner(3, 4)))
The object just created is wrapped with DeepImmutableProxy
.
This prevents mutation of attributes at all levels of depth.
>>> try:
... outer_proxy.a = Inner(-1, -2)
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Outer]' object is immutable. Cannot change attribute 'a' after initialization.
>>> try:
... outer_proxy.b = Inner(-3, -4)
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Outer]' object is immutable. Cannot change attribute 'b' after initialization.
>>> try:
... outer_proxy.a.x = -1
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... outer_proxy.a.y = -2
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'y' after initialization.
>>> try:
... outer_proxy.b.x = -3
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... outer_proxy.b.y = -4
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'y' after initialization.
>>> try:
... outer_proxy.set_a(Inner(-1, -2))
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Outer]' object is immutable. Cannot change attribute 'a' after initialization.
>>> try:
... outer_proxy.a.set_x(-1)
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
>>> try:
... del outer_proxy.a
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Outer]' object is immutable. Cannot delete attribute 'a' after initialization.
We recall that instances of Outer
have an attribute that references an object of another class that has a reference to the first. We show that DeepImmutableProxy
still prevents mutation.
>>> try:
... outer_proxy.delegate.back.x = 5
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Outer]' object is immutable. Cannot change attribute 'x' after initialization.
Next we present an example where inheritance is involved and super()
is used.
>>> class Derived(Outer):
... def set_a(self, a):
... modified = Inner(2*a.x, 2*a.y)
... super().set_a(modified)
>>> derived = Derived(Inner(1, 2), Inner(3, 4))
>>> derived.set_a(Inner(10, 20))
>>> print(derived)
Derived(Inner(20, 40), Inner(3, 4))
>>> derived_proxy = DeepImmutableProxy(Derived(Inner(1, 2), Inner(3, 4)))
>>> try:
... derived_proxy.set_a(Inner(10, 20))
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Derived]' object is immutable. Cannot change attribute 'a' after initialization.
The class DeepImmutableProxy
works with instances of user defined classes.
For classes that are not written in Python there might be problems, as the introspection methods used by DeepImmutableProxy
may not work.
This problem is solved with the class methods DeepImmutableProxy.register_immutable
, which can be used to tell DeepImmutableProxy
that certain class is already immutable (hence, wrapping it with DeepImmutableProxy
has no effect), and DeepImmutableProxy.register_proxy
, that allows a user to register a subclass of DeepImmutableProxy
that defines an immutable proxy specific for the given type.
For example, types int
, float
, str
, bytes
, and some others, have already been registered to be immutable:
>>> no_proxy = DeepImmutableProxy(3)
>>> type(no_proxy) == int
True
The class method DeepImmutableProxy.get_registered_immutable
returns a list of the currently registered immutable classes:
>>> [cls.__name__ for cls in DeepImmutableProxy.get_registered_immutable()] # doctest: +ELLIPSIS
['bytes', 'int', 'float', 'complex', 'str', 'NoneType', ...]
Some containers like tuples are immutable only if their content is immutable.
Thus, when applying DeepImmutableProxy
to a container like a tuple, we need a specific wrapper.
This wrapper is already defined in this package.
For example, the second element of the next tuple is mutable, so the resulting object is mutable.
However, when wrapped with DeepImmutableProxy
it really becomes immutable.
This is achieved by returning a special wrapper for tuples.
The usual DeepImmutableProxy
class does not work because attribute access for built-in types does not follow the standard Python rules.
>>> mutable_tuple_proxy = DeepImmutableProxy(('immutable', ['mutable']))
>>> mutable_tuple_proxy
TupleImmutableProxy(('immutable', ['mutable']))
The first item of the tuple is a str
, which is immutable, so it is returned as is.
>>> mutable_tuple_proxy[0]
'immutable'
The second item is mutable, so it is wrapped and cannot be modified either.
>>> mutable_tuple_proxy[1]
ListImmutableProxy(['mutable'])
This second item is of type list, so it also needs a special wrapper. Methods that do not mutate the content can be used normally.
>>> len(mutable_tuple_proxy[1]), mutable_tuple_proxy[1].index('mutable')
(1, 0)
However, methods that do mutate content are forbidden.
>>> try:
... mutable_tuple_proxy[1].append('another')
... except AttributeError as error:
... print(error)
Object of type "ListImmutableProxy" does not have attribute "append".
DeepImmutableProxy
objects include a hook for __copy__
and __deepcopy__
that copies the underlying object. The copy is mutable.
>>> from copy import deepcopy
>>> copied = deepcopy(outer_proxy)
>>> print(copied)
Outer(Inner(1, 2), Inner(3, 4))
>>> copied.a.x = -1
>>> try:
... outer_proxy.a.x = -1
... except ConstantAttributeError as error:
... print(error)
'DeepImmutableProxy[Inner]' object is immutable. Cannot change attribute 'x' after initialization.
FAQs
Provides a proxy to an object that prevents mutation at any depth
We found that immutable-proxy demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Research
Security News
A malicious npm package targets Solana developers, rerouting funds in 2% of transactions to a hardcoded address.
Security News
Research
Socket researchers have discovered malicious npm packages targeting crypto developers, stealing credentials and wallet data using spyware delivered through typosquats of popular cryptographic libraries.
Security News
Socket's package search now displays weekly downloads for npm packages, helping developers quickly assess popularity and make more informed decisions.