Refactor string internalization.

src/objects-inl.h

Index: src/objects-inl.h
diff --git a/src/objects-inl.h b/src/objects-inl.h
index 7bd8d59801f249183ea5b4e454af576d07cf4734..340aa8b9866c81c97c930ce052315f2cbf6461db 100644
--- a/src/objects-inl.h
+++ b/src/objects-inl.h
@@ -6672,6 +6672,133 @@ void FlexibleBodyDescriptor<start_offset>::IterateBody(HeapObject* obj,
 #undef WRITE_BYTE_FIELD
 
 
+template <typename Char>
+class SequentialStringKey : public HashTableKey {
+ public:
+  explicit SequentialStringKey(Vector<const Char> string, uint32_t seed)
+      : string_(string), hash_field_(0), seed_(seed) { }
+
+  uint32_t Hash() {

[marja, 2014/01/16 16:18:18]

Drive-by coding style nit: don't you need to mark virtual funcs virtual in the
subclasses too?

[ulan, 2014/01/16 16:23:11]

I guess we don't need to mark them as virtual. I saw several other places in the
code where they are not marked.

+    hash_field_ = StringHasher::HashSequentialString<Char>(string_.start(),
+                                                           string_.length(),
+                                                           seed_);
+
+    uint32_t result = hash_field_ >> String::kHashShift;
+    ASSERT(result != 0);  // Ensure that the hash value of 0 is never computed.
+    return result;
+  }
+
+
+  uint32_t HashForObject(Object* other) {
+    return String::cast(other)->Hash();
+  }
+
+  Vector<const Char> string_;
+  uint32_t hash_field_;
+  uint32_t seed_;
+};
+
+
+class OneByteStringKey : public SequentialStringKey<uint8_t> {
+ public:
+  OneByteStringKey(Vector<const uint8_t> str, uint32_t seed)
+      : SequentialStringKey<uint8_t>(str, seed) { }
+
+  bool IsMatch(Object* string) {
+    return String::cast(string)->IsOneByteEqualTo(string_);
+  }
+
+  MaybeObject* AsObject(Heap* heap);
+};
+
+
+class SubStringOneByteStringKey : public HashTableKey {
+ public:
+  explicit SubStringOneByteStringKey(Handle<SeqOneByteString> string,
+                                     int from,
+                                     int length)
+      : string_(string), from_(from), length_(length) { }
+
+  uint32_t Hash() {
+    ASSERT(length_ >= 0);
+    ASSERT(from_ + length_ <= string_->length());
+    uint8_t* chars = string_->GetChars() + from_;
+    hash_field_ = StringHasher::HashSequentialString(
+        chars, length_, string_->GetHeap()->HashSeed());
+    uint32_t result = hash_field_ >> String::kHashShift;
+    ASSERT(result != 0);  // Ensure that the hash value of 0 is never computed.
+    return result;
+  }
+
+
+  uint32_t HashForObject(Object* other) {
+    return String::cast(other)->Hash();
+  }
+
+  bool IsMatch(Object* string) {
+    Vector<const uint8_t> chars(string_->GetChars() + from_, length_);
+    return String::cast(string)->IsOneByteEqualTo(chars);
+  }
+
+  MaybeObject* AsObject(Heap* heap);
+
+ private:
+  Handle<SeqOneByteString> string_;
+  int from_;
+  int length_;
+  uint32_t hash_field_;
+};
+
+
+class TwoByteStringKey : public SequentialStringKey<uc16> {
+ public:
+  explicit TwoByteStringKey(Vector<const uc16> str, uint32_t seed)
+      : SequentialStringKey<uc16>(str, seed) { }
+
+  bool IsMatch(Object* string) {
+    return String::cast(string)->IsTwoByteEqualTo(string_);
+  }
+
+  MaybeObject* AsObject(Heap* heap);
+};
+
+
+// Utf8StringKey carries a vector of chars as key.
+class Utf8StringKey : public HashTableKey {
+ public:
+  explicit Utf8StringKey(Vector<const char> string, uint32_t seed)
+      : string_(string), hash_field_(0), seed_(seed) { }
+
+  bool IsMatch(Object* string) {
+    return String::cast(string)->IsUtf8EqualTo(string_);
+  }
+
+  uint32_t Hash() {
+    if (hash_field_ != 0) return hash_field_ >> String::kHashShift;
+    hash_field_ = StringHasher::ComputeUtf8Hash(string_, seed_, &chars_);
+    uint32_t result = hash_field_ >> String::kHashShift;
+    ASSERT(result != 0);  // Ensure that the hash value of 0 is never computed.
+    return result;
+  }
+
+  uint32_t HashForObject(Object* other) {
+    return String::cast(other)->Hash();
+  }
+
+  MaybeObject* AsObject(Heap* heap) {
+    if (hash_field_ == 0) Hash();
+    return heap->AllocateInternalizedStringFromUtf8(string_,
+                                                    chars_,
+                                                    hash_field_);
+  }
+
+  Vector<const char> string_;
+  uint32_t hash_field_;
+  int chars_;  // Caches the number of characters when computing the hash code.
+  uint32_t seed_;
+};
+
+
 } }  // namespace v8::internal
 
 #endif  // V8_OBJECTS_INL_H_