[Python-Dev] ctypes: is it intentional that id() is the only way to get the address of an object?
On 18 Jan 2019, at 11:57, Antoine Pitrou wrote:
> On Fri, 18 Jan 2019 03:00:54 +0000
> MRAB <python at mrabarnett.plus.com> wrote:
>> On 2019-01-18 00:48, Gregory P. Smith wrote:
>>> I've heard that libraries using ctypes, cffi, or cython code of
>>> sorts in the real world wild today does abuse the unfortunate side
>>> effect of CPython's implementation of id(). I don't have specific
>>> instances of this in mind but trust what I've heard: that it is
>>> id() should never be considered to be the PyObject*.? In as much as
>>> shouldn't assume it is running on top of a specific CPython
>>> If there is a _need_ to get a pointer to a C struct handle
>>> referencing a
>>> CPython C API PyObject, we should make an explicit API for that
>>> than the id() hack.? That way code can be explicit about its need,
>>> code that is just doing a funky form of identity tracking without
>>> is and is not can continue using id() without triggering regressive
>>> behavior on VMs that don't have a CPython compatible PyObject under
>>> hood by default.
>>> [who uses id() anyways?]
>> I use it in some of my code.
>> If I want to cache some objects, I put them in a dict, using the id
>> the key. If I wanted to locate an object in a cache and didn't have
>> id(), I'd have to do a linear search for it.
> Indeed. I've used it for the same purpose in the past
Its useful in all situations where you do topology preserving
transformations, for example pickling (i.e. object serialization) or a
deep copy of some object structures.
In these cases you need a way to record and quickly detect whether
you've handled a specific object before. In Python we can do that with a
dictionary that has object ids as keys. Java provides IdentityHashMap