#20246: Use `with strict_equality(True)` to work around hashing of p-adics
-------------------------------------+-------------------------------------
Reporter: saraedum | Owner:
Type: enhancement | Status: needs_work
Priority: major | Milestone: sage-7.2
Component: padics | Resolution:
Keywords: days71 | Merged in:
Authors: Julian Rüth | Reviewers:
Report Upstream: N/A | Work issues:
Branch: | Commit:
u/saraedum/use__with_strict_equality_true___to_work_around_hashing_of_p_adics|
f4c5a9db729ccf911517039fdced282588210133
Dependencies: #16342, #16339 | Stopgaps:
-------------------------------------+-------------------------------------
Comment (by nbruin):
OK, finally I've found why making objects only hashable part of the time
is a no-go: weakref callback and deletion finalizers can lead to arbitrary
code execution. So once you've put an object into a dictionary as a key,
you have to make sure it remains hashable, because key lookup can happen
at arbitrary times:
{{{
class T(object):
def __init__(self,V):
self.V=V
def __repr__(self):
return "<T(%s)>"%V
def __hash__(self):
if hashflag:
return hash(self.V)
else:
raise TypeError("T is not hashable at the moment")
def __eq__(self,other):
return isinstance(other,T) and self.V == other.V
def __ne__(self,other):
return not(self==other)
}}}
You get this:
{{{
sage: v=T(0) #this just needs to be a weakreffable object
sage: k=T(1)
sage: D=weakref.WeakValueDictionary()
sage: hashflag=True
sage: D[k]=v
sage: hashflag=False
sage: del v
Exception TypeError: TypeError('T is not hashable at the moment',) in
<function remove at 0x7f384da9a0c8> ignored
}}}
(and yes, particularly the cache for `UniqueRepresetation` is a
`WeakValueDictionary`. That particular one is a slightly safer one than
what comes bundled with python by default, so this particular issue isn't
triggered there, but I think it illustrates that in python it's illegal to
have a hashable object that becomes unhashable afterwards).
So it's not even a question design preferences: the design is not
according to the behaviour that python objects are expected to adhere to,
so it's a no-go unless we decide that sage isn't python anymore.
--
Ticket URL: <http://trac.sagemath.org/ticket/20246#comment:27>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.
