This patch addresses two main issues: (1) There exist some non-fatal
errors in some of the hash algorithms. For exmaple, in float_hash and
complex_hash a certain part of the value is not included in the hash
calculation. See Tim's, Guido's, and my discussion of this on
python-dev in May under the title "fix float_hash and complex_hash for
64-bit *nix"
(2) The hash algorithms that use pointers (e.g. func_hash, code_hash)
are universally not correct on Win64 (they assume that sizeof(long) ==
sizeof(void*))
As well, this patch significantly cleans up the hash code. It adds the
two function _Py_HashDouble and _PyHash_VoidPtr that the various
hashing routine are changed to use.
These help maintain the hash function invariant: (a==b) =>
(hash(a)==hash(b))) I have added Lib/test/test_hash.py and
Lib/test/output/test_hash to test this for some cases.
diff --git a/Objects/floatobject.c b/Objects/floatobject.c
index 69b66b7..29ade28 100644
--- a/Objects/floatobject.c
+++ b/Objects/floatobject.c
@@ -59,7 +59,13 @@
#endif
#ifndef LONG_MAX
+#if SIZEOF_LONG == 4
#define LONG_MAX 0X7FFFFFFFL
+#elif SIZEOF_LONG == 8
+#define LONG_MAX 0X7FFFFFFFFFFFFFFFL
+#else
+#error "could not set LONG_MAX"
+#endif
#endif
#ifndef LONG_MIN
@@ -357,12 +363,12 @@
return (i < j) ? -1 : (i > j) ? 1 : 0;
}
+
static long
float_hash(v)
PyFloatObject *v;
{
double intpart, fractpart;
- int expo;
long x;
/* This is designed so that Python numbers with the same
value hash to the same value, otherwise comparisons
@@ -379,7 +385,7 @@
#endif
if (fractpart == 0.0) {
- if (intpart > 0x7fffffffL || -intpart > 0x7fffffffL) {
+ if (intpart > LONG_MAX || -intpart > LONG_MAX) {
/* Convert to long int and use its hash... */
PyObject *w = PyLong_FromDouble(v->ob_fval);
if (w == NULL)
@@ -393,14 +399,9 @@
else {
/* Note -- if you change this code, also change the copy
in complexobject.c */
- long hipart;
- fractpart = frexp(fractpart, &expo);
- fractpart = fractpart * 2147483648.0; /* 2**31 */
- hipart = (long)fractpart; /* Take the top 32 bits */
- fractpart = (fractpart - (double)hipart) * 2147483648.0;
- /* Get the next 32 bits */
- x = hipart + (long)fractpart + (long)intpart + (expo << 15);
- /* Combine everything */
+ x = _Py_HashDouble(v->ob_fval);
+ if (x == -1)
+ return -1;
}
if (x == -1)
x = -2;