Merge pull request RustPython#3951 from jopemachine/float-nan-hash

Fix `float` and `complex`'s incorrect hash value when using with `nan`

Author: youknowone

Lib/test/test_float.py

@@ -582,8 +582,6 @@ def test_hash(self):
         self.assertEqual(hash(float('inf')), sys.hash_info.inf)
         self.assertEqual(hash(float('-inf')), -sys.hash_info.inf)
 
-    # TODO: RUSTPYTHON
-    @unittest.expectedFailure
     def test_hash_nan(self):
         value = float('nan')
         self.assertEqual(hash(value), object.__hash__(value))

common/src/hash.rs

@@ -1,11 +1,7 @@
 use num_bigint::BigInt;
-use num_complex::Complex64;
 use num_traits::ToPrimitive;
 use siphasher::sip::SipHasher24;
-use std::{
-    hash::{BuildHasher, Hash, Hasher},
-    num::Wrapping,
-};
+use std::hash::{BuildHasher, Hash, Hasher};
 
 pub type PyHash = i64;
 pub type PyUHash = u64;
@@ -94,17 +90,14 @@ impl HashSecret {
     }
 }
 
-pub fn hash_float(value: f64) -> PyHash {
+#[inline]
+pub fn hash_float(value: f64) -> Option<PyHash> {
     // cpython _Py_HashDouble
     if !value.is_finite() {
         return if value.is_infinite() {
-            if value > 0.0 {
-                INF
-            } else {
-                -INF
-            }
+            Some(if value > 0.0 { INF } else { -INF })
         } else {
-            NAN
+            None
         };
     }
 
@@ -136,14 +129,7 @@ pub fn hash_float(value: f64) -> PyHash {
     };
     x = ((x << e) & MODULUS) | x >> (BITS32 - e);
 
-    fix_sentinel(x as PyHash * value.signum() as PyHash)
-}
-
-pub fn hash_complex(value: &Complex64) -> PyHash {
-    let re_hash = hash_float(value.re);
-    let im_hash = hash_float(value.im);
-    let Wrapping(ret) = Wrapping(re_hash) + Wrapping(im_hash) * Wrapping(IMAG);
-    fix_sentinel(ret)
+    Some(fix_sentinel(x as PyHash * value.signum() as PyHash))
 }
 
 pub fn hash_iter_unordered<'a, T: 'a, I, F, E>(iter: I, hashf: F) -> Result<PyHash, E>
@@ -192,3 +178,19 @@ pub fn lcg_urandom(mut x: u32, buf: &mut [u8]) {
         *b = ((x >> 16) & 0xff) as u8;
     }
 }
+
+#[inline]
+pub fn hash_object_id_raw(p: usize) -> PyHash {
+    // TODO: Use commented logic when below issue resolved.
+    // Ref: https://github.com/RustPython/RustPython/pull/3951#issuecomment-1193108966
+
+    /* bottom 3 or 4 bits are likely to be 0; rotate y by 4 to avoid
+    excessive hash collisions for dicts and sets */
+    // p.rotate_right(4) as PyHash
+    p as PyHash
+}
+
+#[inline]
+pub fn hash_object_id(p: usize) -> PyHash {
+    fix_sentinel(hash_object_id_raw(p))
+}

vm/src/builtins/complex.rs

@@ -16,6 +16,7 @@ use crate::{
 use num_complex::Complex64;
 use num_traits::Zero;
 use rustpython_common::{float_ops, hash};
+use std::num::Wrapping;
 
 /// Create a complex number from a real part and an optional imaginary part.
 ///
@@ -417,7 +418,16 @@ impl Comparable for PyComplex {
 impl Hashable for PyComplex {
     #[inline]
     fn hash(zelf: &crate::Py<Self>, _vm: &VirtualMachine) -> PyResult<hash::PyHash> {
-        Ok(hash::hash_complex(&zelf.value))
+        let value = zelf.value;
+
+        let re_hash =
+            hash::hash_float(value.re).unwrap_or_else(|| hash::hash_object_id(zelf.get_id()));
+
+        let im_hash =
+            hash::hash_float(value.im).unwrap_or_else(|| hash::hash_object_id(zelf.get_id()));
+
+        let Wrapping(ret) = Wrapping(re_hash) + Wrapping(im_hash) * Wrapping(hash::IMAG);
+        Ok(hash::fix_sentinel(ret))
     }
 }
 

vm/src/builtins/float.rs

@@ -539,7 +539,7 @@ impl Comparable for PyFloat {
 impl Hashable for PyFloat {
     #[inline]
     fn hash(zelf: &crate::Py<Self>, _vm: &VirtualMachine) -> PyResult<hash::PyHash> {
-        Ok(hash::hash_float(zelf.to_f64()))
+        Ok(hash::hash_float(zelf.to_f64()).unwrap_or_else(|| hash::hash_object_id(zelf.get_id())))
     }
 }