hash() yields different results for different platforms

[Tim Peters]

[Grant Edwards]
>> ...
>> The low 32 bits match, so perhaps you should just use that
>> portion of the returned hash?
>>
>> >>> hex(12416037344)
>> '0x2E40DB1E0L'
>> >>> hex(-468864544 & 0xffffffffffffffff)
>> '0xFFFFFFFFE40DB1E0L'
>>
>> >>> hex(12416037344 & 0xffffffff)
>> '0xE40DB1E0L'
>> >>> hex(-468864544 & 0xffffffff)
>> '0xE40DB1E0L'

[Qiangning Hong]
> Is this relationship (same low 32 bits) guaranteed?

No.  Nothing about hashes is guaranteed, except that when x and y are
of a hashable type, and x == y, then hash(x) == hash(y) too.

> Will it change in the future version?

That's possible, but not planned.  Note that the guts of string
hashing in CPython today is implemented via

	while (--len >= 0)
		x = (1000003*x) ^ *p++;

where x is C type "long", and the C language doesn't even define what
that does (behavior when signed multiplication overflows isn't defined
in C).