url: spec-compliant URLSearchParams parser

The entire `URLSearchParams` class is now fully spec-compliant.

PR-URL: https://github.com/nodejs/node/pull/12507
Fixes: https://github.com/nodejs/node/issues/10821
Reviewed-By: James M Snell <jasnell@gmail.com>

benchmark/url/legacy-vs-whatwg-url-searchparams-parse.js

@@ -7,7 +7,7 @@ const inputs = require('../fixtures/url-inputs.js').searchParams;
 const bench = common.createBenchmark(main, {
   type: Object.keys(inputs),
   method: ['legacy', 'whatwg'],
-  n: [1e5]
+  n: [1e6]
 });
 
 function useLegacy(n, input) {

lib/internal/url.js

@@ -1,7 +1,11 @@
 'use strict';
 
 const util = require('util');
-const { hexTable, StorageObject } = require('internal/querystring');
+const {
+  hexTable,
+  isHexTable,
+  StorageObject
+} = require('internal/querystring');
 const binding = process.binding('url');
 const context = Symbol('context');
 const cannotBeBase = Symbol('cannot-be-base');
@@ -575,23 +579,106 @@ function initSearchParams(url, init) {
     url[searchParams] = [];
     return;
   }
-  url[searchParams] = getParamsFromObject(querystring.parse(init));
+  url[searchParams] = parseParams(init);
 }
 
-function getParamsFromObject(obj) {
+// application/x-www-form-urlencoded parser
-  const keys = Object.keys(obj);
+// Ref: https://url.spec.whatwg.org/#concept-urlencoded-parser
-  const values = [];
+function parseParams(qs) {
-  for (var i = 0; i < keys.length; i++) {
+  const out = [];
-    const name = keys[i];
+  var pairStart = 0;
-    const value = obj[name];
+  var lastPos = 0;
-    if (Array.isArray(value)) {
+  var seenSep = false;
-      for (const item of value)
+  var buf = '';
-        values.push(name, item);
+  var encoded = false;
-    } else {
+  var encodeCheck = 0;
-      values.push(name, value);
+  var i;
+  for (i = 0; i < qs.length; ++i) {
+    const code = qs.charCodeAt(i);
+
+    // Try matching key/value pair separator
+    if (code === 38/*&*/) {
+      if (pairStart === i) {
+        // We saw an empty substring between pair separators
+        lastPos = pairStart = i + 1;
+        continue;
+      }
+
+      if (lastPos < i)
+        buf += qs.slice(lastPos, i);
+      if (encoded)
+        buf = querystring.unescape(buf);
+      out.push(buf);
+
+      // If `buf` is the key, add an empty value.
+      if (!seenSep)
+        out.push('');
+
+      seenSep = false;
+      buf = '';
+      encoded = false;
+      encodeCheck = 0;
+      lastPos = pairStart = i + 1;
+      continue;
+    }
+
+    // Try matching key/value separator (e.g. '=') if we haven't already
+    if (!seenSep && code === 61/*=*/) {
+      // Key/value separator match!
+      if (lastPos < i)
+        buf += qs.slice(lastPos, i);
+      if (encoded)
+        buf = querystring.unescape(buf);
+      out.push(buf);
+
+      seenSep = true;
+      buf = '';
+      encoded = false;
+      encodeCheck = 0;
+      lastPos = i + 1;
+      continue;
+    }
+
+    // Handle + and percent decoding.
+    if (code === 43/*+*/) {
+      if (lastPos < i)
+        buf += qs.slice(lastPos, i);
+      buf += ' ';
+      lastPos = i + 1;
+    } else if (!encoded) {
+      // Try to match an (valid) encoded byte (once) to minimize unnecessary
+      // calls to string decoding functions
+      if (code === 37/*%*/) {
+        encodeCheck = 1;
+      } else if (encodeCheck > 0) {
+        // eslint-disable-next-line no-extra-boolean-cast
+        if (!!isHexTable[code]) {
+          if (++encodeCheck === 3)
+            encoded = true;
+        } else {
+          encodeCheck = 0;
+        }
+      }
     }
   }
-  return values;
+
+  // Deal with any leftover key or value data
+
+  // There is a trailing &. No more processing is needed.
+  if (pairStart === i)
+    return out;
+
+  if (lastPos < i)
+    buf += qs.slice(lastPos, i);
+  if (encoded)
+    buf = querystring.unescape(buf);
+  out.push(buf);
+
+  // If `buf` is the key, add an empty value.
+  if (!seenSep)
+    out.push('');
+
+  return out;
 }
 
 // Adapted from querystring's implementation.

test/fixtures/url-searchparams.js

@@ -0,0 +1,68 @@
+module.exports = [
+  ['', '', []],
+  [
+    'foo=918854443121279438895193',
+    'foo=918854443121279438895193',
+    [['foo', '918854443121279438895193']]
+  ],
+  ['foo=bar', 'foo=bar', [['foo', 'bar']]],
+  ['foo=bar&foo=quux', 'foo=bar&foo=quux', [['foo', 'bar'], ['foo', 'quux']]],
+  ['foo=1&bar=2', 'foo=1&bar=2', [['foo', '1'], ['bar', '2']]],
+  [
+    "my%20weird%20field=q1!2%22'w%245%267%2Fz8)%3F",
+    'my+weird+field=q1%212%22%27w%245%267%2Fz8%29%3F',
+    [['my weird field', 'q1!2"\'w$5&7/z8)?']]
+  ],
+  ['foo%3Dbaz=bar', 'foo%3Dbaz=bar', [['foo=baz', 'bar']]],
+  ['foo=baz=bar', 'foo=baz%3Dbar', [['foo', 'baz=bar']]],
+  [
+    'str=foo&arr=1&somenull&arr=2&undef=&arr=3',
+    'str=foo&arr=1&somenull=&arr=2&undef=&arr=3',
+    [
+      ['str', 'foo'],
+      ['arr', '1'],
+      ['somenull', ''],
+      ['arr', '2'],
+      ['undef', ''],
+      ['arr', '3']
+    ]
+  ],
+  [' foo = bar ', '+foo+=+bar+', [[' foo ', ' bar ']]],
+  ['foo=%zx', 'foo=%25zx', [['foo', '%zx']]],
+  ['foo=%EF%BF%BD', 'foo=%EF%BF%BD', [['foo', '\ufffd']]],
+  // See: https://github.com/joyent/node/issues/3058
+  ['foo&bar=baz', 'foo=&bar=baz', [['foo', ''], ['bar', 'baz']]],
+  ['a=b&c&d=e', 'a=b&c=&d=e', [['a', 'b'], ['c', ''], ['d', 'e']]],
+  ['a=b&c=&d=e', 'a=b&c=&d=e', [['a', 'b'], ['c', ''], ['d', 'e']]],
+  ['a=b&=c&d=e', 'a=b&=c&d=e', [['a', 'b'], ['', 'c'], ['d', 'e']]],
+  ['a=b&=&d=e', 'a=b&=&d=e', [['a', 'b'], ['', ''], ['d', 'e']]],
+  ['&&foo=bar&&', 'foo=bar', [['foo', 'bar']]],
+  ['&', '', []],
+  ['&&&&', '', []],
+  ['&=&', '=', [['', '']]],
+  ['&=&=', '=&=', [['', ''], ['', '']]],
+  ['=', '=', [['', '']]],
+  ['+', '+=', [[' ', '']]],
+  ['+=', '+=', [[' ', '']]],
+  ['=+', '=+', [['', ' ']]],
+  ['+=&', '+=', [[' ', '']]],
+  ['a&&b', 'a=&b=', [['a', ''], ['b', '']]],
+  ['a=a&&b=b', 'a=a&b=b', [['a', 'a'], ['b', 'b']]],
+  ['&a', 'a=', [['a', '']]],
+  ['&=', '=', [['', '']]],
+  ['a&a&', 'a=&a=', [['a', ''], ['a', '']]],
+  ['a&a&a&', 'a=&a=&a=', [['a', ''], ['a', ''], ['a', '']]],
+  ['a&a&a&a&', 'a=&a=&a=&a=', [['a', ''], ['a', ''], ['a', ''], ['a', '']]],
+  ['a=&a=value&a=', 'a=&a=value&a=', [['a', ''], ['a', 'value'], ['a', '']]],
+  ['foo%20bar=baz%20quux', 'foo+bar=baz+quux', [['foo bar', 'baz quux']]],
+  ['+foo=+bar', '+foo=+bar', [[' foo', ' bar']]],
+  [
+    // fake percent encoding
+    'foo=%©ar&baz=%A©uux&xyzzy=%©ud',
+    'foo=%25%C2%A9ar&baz=%25A%C2%A9uux&xyzzy=%25%C2%A9ud',
+    [['foo', '%©ar'], ['baz', '%A©uux'], ['xyzzy', '%©ud']]
+  ],
+  // always preserve order of key-value pairs
+  ['a=1&b=2&a=3', 'a=1&b=2&a=3', [['a', '1'], ['b', '2'], ['a', '3']]],
+  ['?a', '%3Fa=', [['?a', '']]]
+];

test/parallel/test-whatwg-url-searchparams.js

@@ -1,8 +1,9 @@
 'use strict';
 
-require('../common');
+const common = require('../common');
 const assert = require('assert');
-const URL = require('url').URL;
+const path = require('path');
+const { URL, URLSearchParams } = require('url');
 
 // Tests below are not from WPT.
 const serialized = 'a=a&a=1&a=true&a=undefined&a=null&a=%EF%BF%BD' +
@@ -77,3 +78,27 @@ assert.throws(() => sp.forEach(1),
 
 m.search = '?a=a&b=b';
 assert.strictEqual(sp.toString(), 'a=a&b=b');
+
+const tests = require(path.join(common.fixturesDir, 'url-searchparams.js'));
+
+for (const [input, expected, parsed] of tests) {
+  if (input[0] !== '?') {
+    const sp = new URLSearchParams(input);
+    assert.strictEqual(String(sp), expected);
+    assert.deepStrictEqual(Array.from(sp), parsed);
+
+    m.search = input;
+    assert.strictEqual(String(m.searchParams), expected);
+    assert.deepStrictEqual(Array.from(m.searchParams), parsed);
+  }
+
+  {
+    const sp = new URLSearchParams(`?${input}`);
+    assert.strictEqual(String(sp), expected);
+    assert.deepStrictEqual(Array.from(sp), parsed);
+
+    m.search = `?${input}`;
+    assert.strictEqual(String(m.searchParams), expected);
+    assert.deepStrictEqual(Array.from(m.searchParams), parsed);
+  }
+}