New submission from Leopold Talirz <leopold.tal...@epfl.ch>:
Disclaimer: this is my first issue on the python bug tracker. Please feel free
to close this issue and redirect this to a specific mailing list, if more
appropriate.
The implementation of `functools._HashedSeq` [1] does not include an
implementation of `__eq__` that takes advantage of the hash:
```
class _HashedSeq(list):
""" This class guarantees that hash() will be called no more than once
per element. This is important because the lru_cache() will hash
the key multiple times on a cache miss.
"""
__slots__ = 'hashvalue'
def __init__(self, tup, hash=hash):
self[:] = tup
self.hashvalue = hash(tup)
def __hash__(self):
return self.hashvalue
```
As far as I can tell, the `_HashedSeq` object is used as a key for looking up
values in the python dictionary that holds the LRU cache, and this lookup
mechanism relies on `__eq__` over `__hash__` (besides shortcuts for objects
with the same id, etc.).
This can cause potentially expensive `__eq__` calls on the arguments of the
cached function and I wonder whether this is intended?
Here is a short example code to demonstrate this:
```
from functools import _HashedSeq
class CompList(list):
"""Hashable list (please forgive)"""
def __eq__(self, other):
print("equality comparison")
return super().__eq__(other)
def __hash__(self):
return hash(tuple(self))
args1=CompList((1,2,3)) # represents function arguments passed to lru_cache
args2=CompList((1,2,3)) # identical content but different object
hs1=_HashedSeq( (args1,))
hs2=_HashedSeq( (args2,))
hs1 == hs2 # True, prints "equality comparison"
d={}
d[hs1] = "cached"
d[hs2] # "cached", prints "equality comparison"
```
Adding the following to the implementation of `_HashedSeq` gets rid of the
calls to `__eq__`:
```
def __eq__(self, other):
return self.hashvalue == other.hashvalue
```
Happy to open a PR for this.
I'm certainly a bit out of my depth here, so apologies if that is all intended
behavior.
[1]
https://github.com/python/cpython/blob/679cb4781ea370c3b3ce40d3334dc404d7e9d92b/Lib/functools.py#L432-L446
----------
components: Library (Lib)
messages: 400858
nosy: leopold.talirz
priority: normal
severity: normal
status: open
title: functools._HashedSeq implements __hash__ but not __eq__
type: behavior
versions: Python 3.10, Python 3.9
_______________________________________
Python tracker <rep...@bugs.python.org>
<https://bugs.python.org/issue45080>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
