[perf] Add optimization for ASCII case

Read chunks of 8 bytes until a non-ASCII byte is found.

src/validation.c

@@ -1,5 +1,6 @@
 #define NAPI_VERSION 1
 #include <assert.h>
+#include <string.h>
 #include <node_api.h>
 
 napi_value IsValidUTF8(napi_env env, napi_callback_info info) {
@@ -10,13 +11,13 @@ napi_value IsValidUTF8(napi_env env, napi_callback_info info) {
   status = napi_get_cb_info(env, info, &argc, argv, NULL, NULL);
   assert(status == napi_ok);
 
-  uint8_t *s;
-  size_t length;
+  uint8_t *buf;
+  size_t len;
 
-  status = napi_get_buffer_info(env, argv[0], (void **)&s, &length);
+  status = napi_get_buffer_info(env, argv[0], (void **)&buf, &len);
   assert(status == napi_ok);
 
-  uint8_t *end = s + length;
+  size_t i = 0;
 
   //
   // This code has been taken from utf8_check.c which was developed by
@@ -25,51 +26,71 @@ napi_value IsValidUTF8(napi_env env, napi_callback_info info) {
   // For original code / licensing please refer to
   // https://www.cl.cam.ac.uk/%7Emgk25/ucs/utf8_check.c
   //
-  while (s < end) {
-    if (*s < 0x80) {  // 0xxxxxxx
-      s++;
-    } else if ((s[0] & 0xe0) == 0xc0) {  // 110xxxxx 10xxxxxx
-      if (
-        s + 1 == end ||
-        (s[1] & 0xc0) != 0x80 ||
-        (s[0] & 0xfe) == 0xc0  // overlong
-      ) {
-        break;
-      } else {
-        s += 2;
+  while (i < len) {
+    size_t j = i + 8;
+
+    if (j <= len) {
+      //
+      // Read 8 bytes and check if they are ASCII.
+      //
+      uint64_t chunk;
+      memcpy(&chunk, buf + i, 8);
+
+      if ((chunk & 0x8080808080808080) == 0) {
+        i = j;
+        continue;
       }
-    } else if ((s[0] & 0xf0) == 0xe0) {  // 1110xxxx 10xxxxxx 10xxxxxx
-      if (
-        s + 2 >= end ||
-        (s[1] & 0xc0) != 0x80 ||
-        (s[2] & 0xc0) != 0x80 ||
-        (s[0] == 0xe0 && (s[1] & 0xe0) == 0x80) ||
-        (s[0] == 0xed && (s[1] & 0xe0) == 0xa0)
-      ) {
-        break;
-      } else {
-        s += 3;
+    }
+
+    while ((buf[i] & 0x80) == 0) { // 0xxxxxxx
+      if (++i == len) {
+        goto exit;
       }
-    } else if ((s[0] & 0xf8) == 0xf0) {  // 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+    }
+
+    if ((buf[i] & 0xe0) == 0xc0) {  // 110xxxxx 10xxxxxx
       if (
-        s + 3 >= end ||
-        (s[1] & 0xc0) != 0x80 ||
-        (s[2] & 0xc0) != 0x80 ||
-        (s[3] & 0xc0) != 0x80 ||
-        (s[0] == 0xf0 && (s[1] & 0xf0) == 0x80) ||  // overlong
-        (s[0] == 0xf4 && s[1] > 0x8f) || s[0] > 0xf4  // > U+10FFFF
+        i + 1 == len ||
+        (buf[i + 1] & 0xc0) != 0x80 ||
+        (buf[i] & 0xfe) == 0xc0  // overlong
       ) {
         break;
       } else {
-        s += 4;
+        i += 2;
+      }
+    } else if ((buf[i] & 0xf0) == 0xe0) {  // 1110xxxx 10xxxxxx 10xxxxxx
+      if (
+        i + 2 >= len ||
+        (buf[i + 1] & 0xc0) != 0x80 ||
+        (buf[i + 2] & 0xc0) != 0x80 ||
+        (buf[i] == 0xe0 && (buf[i + 1] & 0xe0) == 0x80) ||  // overlong
+        (buf[i] == 0xed && (buf[i + 1] & 0xe0) == 0xa0)  // surrogate (U+D800 - U+DFFF)
+      ) {
+        break;
+      } else {
+        i += 3;
+      }
+    } else if ((buf[i] & 0xf8) == 0xf0) {  // 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
+      if (
+        i + 3 >= len ||
+        (buf[i + 1] & 0xc0) != 0x80 ||
+        (buf[i + 2] & 0xc0) != 0x80 ||
+        (buf[i + 3] & 0xc0) != 0x80 ||
+        (buf[i] == 0xf0 && (buf[i + 1] & 0xf0) == 0x80) ||  // overlong
+        (buf[i] == 0xf4 && buf[i + 1] > 0x8f) || buf[i] > 0xf4  // > U+10FFFF
+      ) {
+        break;
+      } else {
+        i += 4;
       }
     } else {
       break;
     }
   }
 
+exit:;
   napi_value result;
-  status = napi_get_boolean(env, s == end, &result);
+  status = napi_get_boolean(env, i == len, &result);
   assert(status == napi_ok);
 
   return result;