url: reduce deplicated codes in autoEscapeStr

PR-URL: #18613 Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
nodejs · Feb 26, 2018 · def51ba · def51ba
1 parent 8532251
commit def51ba
Show file tree

Hide file tree

Showing 2 changed files with 47 additions and 87 deletions.
diff --git a/benchmark/url/url-parse.js b/benchmark/url/url-parse.js
@@ -0,0 +1,22 @@
+'use strict';
+const common = require('../common.js');
+const url = require('url');
+
+const inputs = {
+  normal: 'http://foo.com/bar',
+  escaped: 'https://foo.bar/{}^`/abcd'
+};
+
+const bench = common.createBenchmark(main, {
+  type: Object.keys(inputs),
+  n: [1e7]
+});
+
+function main({ type, n }) {
+  const input = inputs[type] || '';
+
+  bench.start();
+  for (var i = 0; i < n; i += 1)
+    url.parse(input);
+  bench.end(n);
+}
diff --git a/lib/url.js b/lib/url.js
@@ -439,101 +439,39 @@ function validateHostname(self, rest, hostname) {
   }
 }
 
+// Escaped characters. Use empty strings to fill up unused entries.
+// Using Array is faster than Object/Map
+const escapedCodes = [
+  /*0 - 9*/ '', '', '', '', '', '', '', '', '', '%09',
+  /*10 - 19*/ '%0A', '', '', '%0D', '', '', '', '', '', '',
+  /*20 - 29*/ '', '', '', '', '', '', '', '', '', '',
+  /*30 - 39*/ '', '', '%20', '', '%22', '', '', '', '', '%27',
+  /*40 - 49*/ '', '', '', '', '', '', '', '', '', '',
+  /*50 - 59*/ '', '', '', '', '', '', '', '', '', '',
+  /*60 - 69*/ '%3C', '', '%3E', '', '', '', '', '', '', '',
+  /*70 - 79*/ '', '', '', '', '', '', '', '', '', '',
+  /*80 - 89*/ '', '', '', '', '', '', '', '', '', '',
+  /*90 - 99*/ '', '', '%5C', '', '%5E', '', '%60', '', '', '',
+  /*100 - 109*/ '', '', '', '', '', '', '', '', '', '',
+  /*110 - 119*/ '', '', '', '', '', '', '', '', '', '',
+  /*120 - 125*/ '', '', '', '%7B', '%7C', '%7D'
+];
+
 // Automatically escape all delimiters and unwise characters from RFC 2396.
 // Also escape single quotes in case of an XSS attack.
 // Return the escaped string.
 function autoEscapeStr(rest) {
   var escaped = '';
   var lastEscapedPos = 0;
   for (var i = 0; i < rest.length; ++i) {
-    // Manual switching is faster than using a Map/Object.
     // `escaped` contains substring up to the last escaped character.
-    switch (rest.charCodeAt(i)) {
-      case 9:   // '\t'
-        // Concat if there are ordinary characters in the middle.
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%09';
-        lastEscapedPos = i + 1;
-        break;
-      case 10:  // '\n'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%0A';
-        lastEscapedPos = i + 1;
-        break;
-      case 13:  // '\r'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%0D';
-        lastEscapedPos = i + 1;
-        break;
-      case 32:  // ' '
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%20';
-        lastEscapedPos = i + 1;
-        break;
-      case 34:  // '"'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%22';
-        lastEscapedPos = i + 1;
-        break;
-      case 39:  // '\''
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%27';
-        lastEscapedPos = i + 1;
-        break;
-      case 60:  // '<'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%3C';
-        lastEscapedPos = i + 1;
-        break;
-      case 62:  // '>'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%3E';
-        lastEscapedPos = i + 1;
-        break;
-      case 92:  // '\\'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%5C';
-        lastEscapedPos = i + 1;
-        break;
-      case 94:  // '^'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%5E';
-        lastEscapedPos = i + 1;
-        break;
-      case 96:  // '`'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%60';
-        lastEscapedPos = i + 1;
-        break;
-      case 123: // '{'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%7B';
-        lastEscapedPos = i + 1;
-        break;
-      case 124: // '|'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%7C';
-        lastEscapedPos = i + 1;
-        break;
-      case 125: // '}'
-        if (i > lastEscapedPos)
-          escaped += rest.slice(lastEscapedPos, i);
-        escaped += '%7D';
-        lastEscapedPos = i + 1;
-        break;
+    var escapedChar = escapedCodes[rest.charCodeAt(i)];
+    if (escapedChar) {
+      // Concat if there are ordinary characters in the middle.
+      if (i > lastEscapedPos)
+        escaped += rest.slice(lastEscapedPos, i);
+      escaped += escapedChar;
+      lastEscapedPos = i + 1;
     }
   }
   if (lastEscapedPos === 0)  // Nothing has been escaped.