buffer: add fast api for isUtf8 and isAscii

gurgunday · richardlau · commit 6d14279688c9 · 2026-07-06T11:58:00.000Z
Signed-off-by: Gürgün Dayıoğlu <hey@gurgun.day> PR-URL: #64169 Reviewed-By: Daniel Lemire <daniel@lemire.me> Reviewed-By: Matteo Collina <matteo.collina@gmail.com> Reviewed-By: Anna Henningsen <anna@addaleax.net>
diff --git a/src/node_buffer.cc b/src/node_buffer.cc
@@ -1360,37 +1360,93 @@ void FastSwap64(Local<Value> receiver,
 
 static CFunction fast_swap64(CFunction::Make(FastSwap64));
 
+struct ValidationResult {
+  bool is_valid;
+  bool was_detached;
+};
+
+static ValidationResult ValidateUtf8(Local<Value> value) {
+  ArrayBufferViewContents<char> abv(value);
+  bool was_detached = abv.WasDetached();
+  return {!was_detached && simdutf::validate_utf8(abv.data(), abv.length()),
+          was_detached};
+}
+
 static void IsUtf8(const FunctionCallbackInfo<Value>& args) {
   Environment* env = Environment::GetCurrent(args);
   CHECK_EQ(args.Length(), 1);
   CHECK(args[0]->IsTypedArray() || args[0]->IsArrayBuffer() ||
         args[0]->IsSharedArrayBuffer());
-  ArrayBufferViewContents<char> abv(args[0]);
 
-  if (abv.WasDetached()) {
+  const ValidationResult result = ValidateUtf8(args[0]);
+  if (result.was_detached) {
     return node::THROW_ERR_INVALID_STATE(
         env, "Cannot validate on a detached buffer");
   }
 
-  args.GetReturnValue().Set(simdutf::validate_utf8(abv.data(), abv.length()));
+  args.GetReturnValue().Set(result.is_valid);
+}
+
+static bool FastIsUtf8(Local<Value> receiver,
+                       Local<Value> value,
+                       // NOLINTNEXTLINE(runtime/references)
+                       FastApiCallbackOptions& options) {
+  TRACK_V8_FAST_API_CALL("buffer.isUtf8");
+  HandleScope scope(options.isolate);
+
+  const ValidationResult result = ValidateUtf8(value);
+  if (result.was_detached) {
+    node::THROW_ERR_INVALID_STATE(options.isolate,
+                                  "Cannot validate on a detached buffer");
+    return false;
+  }
+  return result.is_valid;
+}
+
+static CFunction fast_is_utf8(CFunction::Make(FastIsUtf8));
+
+static ValidationResult ValidateAscii(Local<Value> value) {
+  ArrayBufferViewContents<char> abv(value);
+  bool was_detached = abv.WasDetached();
+  return {
+      !was_detached &&
+          !simdutf::validate_ascii_with_errors(abv.data(), abv.length()).error,
+      was_detached};
 }
 
 static void IsAscii(const FunctionCallbackInfo<Value>& args) {
   Environment* env = Environment::GetCurrent(args);
   CHECK_EQ(args.Length(), 1);
   CHECK(args[0]->IsTypedArray() || args[0]->IsArrayBuffer() ||
         args[0]->IsSharedArrayBuffer());
-  ArrayBufferViewContents<char> abv(args[0]);
 
-  if (abv.WasDetached()) {
+  const ValidationResult result = ValidateAscii(args[0]);
+  if (result.was_detached) {
     return node::THROW_ERR_INVALID_STATE(
         env, "Cannot validate on a detached buffer");
   }
 
-  args.GetReturnValue().Set(
-      !simdutf::validate_ascii_with_errors(abv.data(), abv.length()).error);
+  args.GetReturnValue().Set(result.is_valid);
 }
 
+static bool FastIsAscii(Local<Value> receiver,
+                        Local<Value> value,
+                        // NOLINTNEXTLINE(runtime/references)
+                        FastApiCallbackOptions& options) {
+  TRACK_V8_FAST_API_CALL("buffer.isAscii");
+  HandleScope scope(options.isolate);
+
+  const ValidationResult result = ValidateAscii(value);
+  if (result.was_detached) {
+    node::THROW_ERR_INVALID_STATE(options.isolate,
+                                  "Cannot validate on a detached buffer");
+    return false;
+  }
+  return result.is_valid;
+}
+
+static CFunction fast_is_ascii(CFunction::Make(FastIsAscii));
+
 void SetBufferPrototype(const FunctionCallbackInfo<Value>& args) {
   Realm* realm = Realm::GetCurrent(args);
 
@@ -1762,8 +1818,9 @@ void Initialize(Local<Object> target,
   SetFastMethod(context, target, "swap32", Swap32, &fast_swap32);
   SetFastMethod(context, target, "swap64", Swap64, &fast_swap64);
 
-  SetMethodNoSideEffect(context, target, "isUtf8", IsUtf8);
-  SetMethodNoSideEffect(context, target, "isAscii", IsAscii);
+  SetFastMethodNoSideEffect(context, target, "isUtf8", IsUtf8, &fast_is_utf8);
+  SetFastMethodNoSideEffect(
+      context, target, "isAscii", IsAscii, &fast_is_ascii);
 
   target
       ->Set(context,
@@ -1836,7 +1893,9 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(fast_swap64);
 
   registry->Register(IsUtf8);
+  registry->Register(fast_is_utf8);
   registry->Register(IsAscii);
+  registry->Register(fast_is_ascii);
 
   registry->Register(StringSlice<ASCII>);
   registry->Register(StringSlice<BASE64>);
diff --git a/test/parallel/test-buffer-isutf8-isascii-fast.js b/test/parallel/test-buffer-isutf8-isascii-fast.js
@@ -0,0 +1,34 @@
+// Flags: --expose-internals --no-warnings --allow-natives-syntax
+'use strict';
+
+const common = require('../common');
+const assert = require('assert');
+const { Buffer, isAscii, isUtf8 } = require('buffer');
+
+const ascii = Buffer.from('hello');
+const utf8 = Buffer.from('hello \xc4\x9f');
+
+function testFastIsAscii() {
+  assert.strictEqual(isAscii(ascii), true);
+}
+
+function testFastIsUtf8() {
+  assert.strictEqual(isUtf8(utf8), true);
+}
+
+eval('%PrepareFunctionForOptimization(isAscii)');
+testFastIsAscii();
+eval('%OptimizeFunctionOnNextCall(isAscii)');
+testFastIsAscii();
+
+eval('%PrepareFunctionForOptimization(isUtf8)');
+testFastIsUtf8();
+eval('%OptimizeFunctionOnNextCall(isUtf8)');
+testFastIsUtf8();
+
+if (common.isDebug) {
+  const { internalBinding } = require('internal/test/binding');
+  const { getV8FastApiCallCount } = internalBinding('debug');
+  assert.strictEqual(getV8FastApiCallCount('buffer.isAscii'), 1);
+  assert.strictEqual(getV8FastApiCallCount('buffer.isUtf8'), 1);
+}
