src: add encoding_methods with fast api

Author: anonrig

lib/internal/bootstrap/loaders.js

@@ -85,6 +85,7 @@ const internalBindingAllowlist = new SafeSet([
   'constants',
   'contextify',
   'crypto',
+  'encoding_methods',
   'fs',
   'fs_event_wrap',
   'http_parser',

lib/internal/encoding.js

@@ -37,6 +37,7 @@ const {
   customInspectSymbol: inspect,
   kEmptyObject,
   kEnumerableProperty,
+  getUtf8Length,
 } = require('internal/util');
 
 const {
@@ -52,10 +53,13 @@ const {
 
 const {
   encodeInto,
-  encodeUtf8String,
   decodeUTF8,
 } = internalBinding('buffer');
 
+const {
+  encodeUtf8,
+} = internalBinding('encoding_methods');
+
 let Buffer;
 function lazyBuffer() {
   if (Buffer === undefined)
@@ -337,7 +341,12 @@ class TextEncoder {
 
   encode(input = '') {
     validateEncoder(this);
-    return encodeUtf8String(`${input}`);
+    const utf8Length = getUtf8Length(input);
+    const array = new Uint8Array(utf8Length);
+    if (utf8Length > 0) {
+      encodeUtf8(`${input}`, array);
+    }
+    return array;
   }
 
   encodeInto(src, dest) {

lib/internal/util.js

@@ -570,6 +570,13 @@ function isArrayBufferDetached(value) {
   return false;
 }
 
+function getUtf8Length(str) {
+  let length = 0;
+  for (let i = 0; i < str.length; i++)
+    length += str.charCodeAt(i) > 127 ? 2 : 1;
+  return length;
+}
+
 module.exports = {
   assertCrypto,
   cachedResult,
@@ -585,6 +592,7 @@ module.exports = {
   filterOwnProperties,
   getConstructorOf,
   getInternalGlobal,
+  getUtf8Length,
   getSystemErrorMap,
   getSystemErrorName,
   isArrayBufferDetached,

node.gyp

@@ -498,6 +498,7 @@
         'src/node_contextify.cc',
         'src/node_credentials.cc',
         'src/node_dir.cc',
+        'src/node_encoding.cc',
         'src/node_env_var.cc',
         'src/node_errors.cc',
         'src/node_external_reference.cc',

src/node_binding.cc

@@ -43,6 +43,7 @@
   V(contextify)                                                                \
   V(credentials)                                                               \
   V(errors)                                                                    \
+  V(encoding_methods)                                                          \
   V(fs)                                                                        \
   V(fs_dir)                                                                    \
   V(fs_event_wrap)                                                             \

src/node_buffer.cc

@@ -1146,40 +1146,6 @@ void Swap64(const FunctionCallbackInfo<Value>& args) {
   args.GetReturnValue().Set(args[0]);
 }
 
-
-// Encode a single string to a UTF-8 Uint8Array (not Buffer).
-// Used in TextEncoder.prototype.encode.
-static void EncodeUtf8String(const FunctionCallbackInfo<Value>& args) {
-  Environment* env = Environment::GetCurrent(args);
-  Isolate* isolate = env->isolate();
-  CHECK_GE(args.Length(), 1);
-  CHECK(args[0]->IsString());
-
-  Local<String> str = args[0].As<String>();
-  size_t length = str->Utf8Length(isolate);
-
-  Local<ArrayBuffer> ab;
-  {
-    NoArrayBufferZeroFillScope no_zero_fill_scope(env->isolate_data());
-    std::unique_ptr<BackingStore> bs =
-        ArrayBuffer::NewBackingStore(isolate, length);
-
-    CHECK(bs);
-
-    str->WriteUtf8(isolate,
-                   static_cast<char*>(bs->Data()),
-                   -1,  // We are certain that `data` is sufficiently large
-                   nullptr,
-                   String::NO_NULL_TERMINATION | String::REPLACE_INVALID_UTF8);
-
-    ab = ArrayBuffer::New(isolate, std::move(bs));
-  }
-
-  auto array = Uint8Array::New(ab, 0, length);
-  args.GetReturnValue().Set(array);
-}
-
-
 static void EncodeInto(const FunctionCallbackInfo<Value>& args) {
   Environment* env = Environment::GetCurrent(args);
   Isolate* isolate = env->isolate();
@@ -1345,7 +1311,6 @@ void Initialize(Local<Object> target,
   SetMethod(context, target, "swap64", Swap64);
 
   SetMethod(context, target, "encodeInto", EncodeInto);
-  SetMethodNoSideEffect(context, target, "encodeUtf8String", EncodeUtf8String);
 
   target
       ->Set(context,
@@ -1400,7 +1365,6 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(Swap64);
 
   registry->Register(EncodeInto);
-  registry->Register(EncodeUtf8String);
 
   registry->Register(StringSlice<ASCII>);
   registry->Register(StringSlice<BASE64>);

src/node_encoding.cc

@@ -0,0 +1,88 @@
+#include "env-inl.h"
+#include "node.h"
+#include "node_external_reference.h"
+#include "node_internals.h"
+#include "util-inl.h"
+#include "v8-fast-api-calls.h"
+#include "v8.h"
+
+namespace node {
+
+using v8::ArrayBuffer;
+using v8::BackingStore;
+using v8::CFunction;
+using v8::Context;
+using v8::FastApiTypedArray;
+using v8::FastOneByteString;
+using v8::Isolate;
+using v8::Local;
+using v8::Object;
+using v8::String;
+using v8::Uint8Array;
+using v8::Value;
+
+namespace encoding_methods {
+
+void EncodeUtf8(const v8::FunctionCallbackInfo<Value>& args) {
+  Environment* env = Environment::GetCurrent(args);
+  Isolate* isolate = env->isolate();
+  CHECK_GE(args.Length(), 2);
+  CHECK(args[0]->IsString());
+  CHECK(args[1]->IsUint8Array());
+
+  Local<String> str = args[0].As<String>();
+  Local<Uint8Array> out = args[1].As<Uint8Array>();
+  Local<ArrayBuffer> buf = out->Buffer();
+  size_t out_length = out->ByteLength();
+  char* write_result = static_cast<char*>(buf->Data()) + out->ByteOffset();
+
+  str->WriteUtf8(isolate,
+                 write_result,
+                 out_length,
+                 nullptr,
+                 String::NO_NULL_TERMINATION | String::REPLACE_INVALID_UTF8);
+}
+
+void latin1_to_utf8(const char* str, char* c) {
+  for (; *str; ++str) {
+    if (*str & 0x80) {
+      *c++ = *str;
+    } else {
+      *c++ = (char)(0xc0 | (unsigned)*str >> 6);
+      *c++ = (char)(0x80 | (*str & 0x3f));
+    }
+  }
+  *c++ = '\0';
+}
+
+void FastEncodeUtf8(Local<Value> receiver,
+                    const FastOneByteString& source,
+                    const FastApiTypedArray<uint8_t>& output) {
+  uint8_t* storage;
+  auto is_available = output.getStorageIfAligned(&storage);
+  CHECK(is_available);
+  latin1_to_utf8(source.data, reinterpret_cast<char*>(storage));
+}
+
+CFunction fast_encode_utf8_(CFunction::Make(FastEncodeUtf8));
+
+static void Initialize(Local<Object> target,
+                       Local<Value> unused,
+                       Local<Context> context,
+                       void* priv) {
+  SetFastMethod(context, target, "encodeUtf8", EncodeUtf8, &fast_encode_utf8_);
+}
+
+void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
+  registry->Register(EncodeUtf8);
+  registry->Register(FastEncodeUtf8);
+  registry->Register(fast_encode_utf8_.GetTypeInfo());
+}
+
+}  // namespace encoding_methods
+}  // namespace node
+
+NODE_BINDING_CONTEXT_AWARE_INTERNAL(encoding_methods,
+                                    node::encoding_methods::Initialize)
+NODE_BINDING_EXTERNAL_REFERENCE(
+    encoding_methods, node::encoding_methods::RegisterExternalReferences)

src/node_external_reference.h

@@ -10,8 +10,14 @@
 
 namespace node {
 
+// TODO(anonrig): Find a good way of reusing existing types for fast api usages.
 using CFunctionCallback = void (*)(v8::Local<v8::Value> receiver);
 
+using CFunctionCallbackWithInput =
+    void (*)(v8::Local<v8::Value> receiver,
+             const v8::FastOneByteString& input,
+             const v8::FastApiTypedArray<uint8_t>& output);
+
 // This class manages the external references from the V8 heap
 // to the C++ addresses in Node.js.
 class ExternalReferenceRegistry {
@@ -20,6 +26,7 @@ class ExternalReferenceRegistry {
 
 #define ALLOWED_EXTERNAL_REFERENCE_TYPES(V)                                    \
   V(CFunctionCallback)                                                         \
+  V(CFunctionCallbackWithInput)                                                \
   V(const v8::CFunctionInfo*)                                                  \
   V(v8::FunctionCallback)                                                      \
   V(v8::AccessorGetterCallback)                                                \
@@ -67,6 +74,7 @@ class ExternalReferenceRegistry {
   V(credentials)                                                               \
   V(env_var)                                                                   \
   V(errors)                                                                    \
+  V(encoding_methods)                                                          \
   V(fs)                                                                        \
   V(fs_dir)                                                                    \
   V(fs_event_wrap)                                                             \

test/parallel/test-bootstrap-modules.js

@@ -18,6 +18,7 @@ const expectedModules = new Set([
   'Internal Binding contextify',
   'Internal Binding credentials',
   'Internal Binding errors',
+  'Internal Binding encoding_methods',
   'Internal Binding fs_dir',
   'Internal Binding fs_event_wrap',
   'Internal Binding fs',