Use a random seed for the string hash algorithm.

src/objects-inl.h

 // Copyright 2011 the V8 project authors. All rights reserved.
 // Redistribution and use in source and binary forms, with or without
 // modification, are permitted provided that the following conditions are
 // met:
 //
 //     * Redistributions of source code must retain the above copyright
 //       notice, this list of conditions and the following disclaimer.
 //     * Redistributions in binary form must reproduce the above
 //       copyright notice, this list of conditions and the following
 //       disclaimer in the documentation and/or other materials provided
@@ skipping 2044 matching lines @@
 
 // Find entry for key otherwise return kNotFound.
 template<typename Shape, typename Key>
 int HashTable<Shape, Key>::FindEntry(Isolate* isolate, Key key) {
   uint32_t capacity = Capacity();
   uint32_t entry = FirstProbe(Shape::Hash(key), capacity);
   uint32_t count = 1;
   // EnsureCapacity will guarantee the hash table is never full.
   while (true) {
     Object* element = KeyAt(entry);
-    if (element == isolate->heap()->undefined_value()) break;  // Empty entry.
-    if (element != isolate->heap()->the_hole_value() &&
+    // Empty entry.
+    if (element == isolate->heap()->raw_unchecked_undefined_value()) break;
+    if (element != isolate->heap()->raw_unchecked_the_hole_value() &&
         Shape::IsMatch(key, element)) return entry;
     entry = NextProbe(entry, count++, capacity);
   }
   return kNotFound;
 }
 
 
 bool NumberDictionary::requires_slow_elements() {
   Object* max_index_object = get(kMaxNumberKeyIndex);
   if (!max_index_object->IsSmi()) return false;
@@ skipping 2236 matching lines @@
 
 uint32_t String::Hash() {
   // Fast case: has hash code already been computed?
   uint32_t field = hash_field();
   if (IsHashFieldComputed(field)) return field >> kHashShift;
   // Slow case: compute hash code and set it.
   return ComputeAndSetHash();
 }
 
 
-StringHasher::StringHasher(int length)
+StringHasher::StringHasher(int length, uint32_t seed)
   : length_(length),
-    raw_running_hash_(0),
+    raw_running_hash_(seed),
     array_index_(0),
     is_array_index_(0 < length_ && length_ <= String::kMaxArrayIndexSize),
     is_first_char_(true),
-    is_valid_(true) { }
+    is_valid_(true) {
+  ASSERT(FLAG_randomize_string_hashes == (raw_running_hash_ != 0));
+}
 
 
 bool StringHasher::has_trivial_hash() {
   return length_ > String::kMaxHashCalcLength;
 }
 
 
 void StringHasher::AddCharacter(uc32 c) {
   // Use the Jenkins one-at-a-time hash function to update the hash
   // for the given character.
@@ skipping 39 matching lines @@
   result ^= (result >> 11);
   result += (result << 15);
   if (result == 0) {
     result = 27;
   }
   return result;
 }
 
 
 template <typename schar>
-uint32_t HashSequentialString(const schar* chars, int length) {
-  StringHasher hasher(length);
+uint32_t HashSequentialString(const schar* chars, int length, uint32_t seed) {
+  StringHasher hasher(length, seed);
   if (!hasher.has_trivial_hash()) {
     int i;
     for (i = 0; hasher.is_array_index() && (i < length); i++) {
       hasher.AddCharacter(chars[i]);
     }
     for (; i < length; i++) {
       hasher.AddCharacterNoIndex(chars[i]);
     }
   }
   return hasher.GetHashField();
@@ skipping 365 matching lines @@
 #undef WRITE_INT_FIELD
 #undef READ_SHORT_FIELD
 #undef WRITE_SHORT_FIELD
 #undef READ_BYTE_FIELD
 #undef WRITE_BYTE_FIELD
 
 
 } }  // namespace v8::internal
 
 #endif  // V8_OBJECTS_INL_H_