buffer: speed up Buffer.isEncoding() method

JacksonTian · JacksonTian · commit e12a2aec4a0b · 2016-02-17T17:18:31.000+08:00
Use automata to avoid toLowerCase(), faster, but more dirty.
diff --git a/benchmark/buffers/buffer-is-encoding.js b/benchmark/buffers/buffer-is-encoding.js
@@ -0,0 +1,51 @@
+'use strict';
+
+const common = require('../common.js');
+const v8 = require('v8');
+
+const bench = common.createBenchmark(main, {
+  encoding: [
+    'hex',
+    'utf8',
+    'utf-8',
+    'ascii',
+    'binary',
+    'base64',
+    'ucs2',
+    'ucs-2',
+    'utf16le',
+    'utf-16le',
+    'HEX',
+    'UTF8',
+    'UTF-8',
+    'ASCII',
+    'BINARY',
+    'BASE64',
+    'UCS2',
+    'UCS-2',
+    'UTF16LE',
+    'UTF-16LE',
+    'utf9',
+    'utf-7',
+    'utf17le',
+    'utf-17le',
+    'Unicode-FTW',
+    'new gnu gun'
+  ]
+});
+
+function main(conf) {
+  var encoding = conf.encoding;
+  // Force optimization before starting the benchmark
+  Buffer.isEncoding(encoding);
+  v8.setFlagsFromString('--allow_natives_syntax');
+  eval('%OptimizeFunctionOnNextCall(Buffer.isEncoding)');
+  Buffer.isEncoding(encoding);
+
+  bench.start();
+  var n = 1024 * 1024;
+  for (let i = 0; i < n; i++) {
+    Buffer.isEncoding(encoding);
+  }
+  bench.end(n);
+}
diff --git a/lib/buffer.js b/lib/buffer.js
@@ -183,32 +183,67 @@ Buffer.compare = function compare(a, b) {
   return binding.compare(a, b);
 };
 
+var make = function (list) {
+  var tpl = '';
+  var append = function (line) {
+    tpl += line + '\n';
+  };
 
-Buffer.isEncoding = function(encoding) {
-  var loweredCase = false;
-  for (;;) {
-    switch (encoding) {
-      case 'hex':
-      case 'utf8':
-      case 'utf-8':
-      case 'ascii':
-      case 'binary':
-      case 'base64':
-      case 'ucs2':
-      case 'ucs-2':
-      case 'utf16le':
-      case 'utf-16le':
-        return true;
+  var map = {};
+  list.forEach(function (item) {
+    map[item.length] = item.length;
+  });
 
-      default:
-        if (loweredCase)
-          return false;
-        encoding = ('' + encoding).toLowerCase();
-        loweredCase = true;
+  var kind = Object.keys(map).map(function (key) {
+    return map[key];
+  });
+
+  append(`switch (encoding.length) {`);
+  for (var i = 0; i < kind.length; i++) {
+    append(`  case ${kind[i]}:`);
+    var collections = list.filter(function (item) {
+      return item.length === kind[i];
+    });
+
+    for (var k = 0; k < kind[i]; k++) {
+      append(`    var code${k} = encoding.charCodeAt(${k});`);
     }
+
+    for (var j = 0; j < collections.length; j++) {
+      var word = collections[j];
+      append(`    // ${word}`);
+      var conditions = [];
+      for (var k = 0; k < word.length; k++) {
+        var code = word.charCodeAt(k);
+        if (code >= 97 && code <= 122) {
+          var cond = '(code' + k + ' === ' + code +
+            ' || code' + k + ' === ' + (code - 32) + ')';
+          conditions.push(cond);
+        } else {
+          conditions.push('code' + k + ' === ' + code);
+        }
+      }
+
+      append(`    if (` + conditions.join(' && ') + `) {`);
+      append(`      return true;`);
+      append(`    }`);
+      append(``);
+    }
+
+    append(`    break;`);
   }
+  append(`}`);
+
+  append(`return false;`);
+  return new Function('encoding', tpl);
 };
 
+var cases = [
+  'hex', 'utf8', 'utf-8', 'ascii', 'binary', 'base64',
+  'ucs2', 'ucs-2', 'utf16le', 'utf-16le'
+];
+
+Buffer.isEncoding = make(cases);
 
 Buffer.concat = function(list, length) {
   if (!Array.isArray(list))
diff --git a/test/parallel/test-buffer.js b/test/parallel/test-buffer.js
@@ -998,14 +998,17 @@ Buffer(Buffer(0), 0, 0);
   'ucs-2',
   'utf16le',
   'utf-16le' ].forEach(function(enc) {
-    assert.equal(Buffer.isEncoding(enc), true);
+    assert.equal(Buffer.isEncoding(enc), true, `${enc} should be an encoding`);
   });
 
 [ 'utf9',
   'utf-7',
+  'utf17le',
+  'utf-17le',
   'Unicode-FTW',
   'new gnu gun'  ].forEach(function(enc) {
-    assert.equal(Buffer.isEncoding(enc), false);
+    assert.equal(Buffer.isEncoding(enc), false,
+      `${enc} should not an encoding`);
   });
 
 
