perf: improve sort algorithm

Author: tsctx

benchmarks/headers.mjs

@@ -0,0 +1,57 @@
+import { bench, group, run } from 'mitata'
+import { Headers } from '../lib/fetch/headers.js'
+
+const characters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789'
+const charactersLength = characters.length
+
+function generateAsciiString (length) {
+  let result = ''
+  for (let i = 0; i < length; ++i) {
+    result += characters[Math.floor(Math.random() * charactersLength)]
+  }
+  return result
+}
+
+const settings = {
+  'fast-path (tiny array)': 4,
+  'fast-path (small array)': 8,
+  'fast-path (middle array)': 16,
+  'fast-path': 32,
+  'slow-path': 64
+}
+
+for (const [name, length] of Object.entries(settings)) {
+  const headers = new Headers(
+    Array.from(Array(length), () => [generateAsciiString(12), ''])
+  )
+
+  const headersSorted = new Headers(headers)
+
+  const kHeadersList = Reflect.ownKeys(headers).find(
+    (c) => String(c) === 'Symbol(headers list)'
+  )
+
+  const headersList = headers[kHeadersList]
+
+  const headersListSorted = headersSorted[kHeadersList]
+
+  const kHeadersSortedMap = Reflect.ownKeys(headersList).find(
+    (c) => String(c) === 'Symbol(headers map sorted)'
+  )
+
+  group(`length ${length} #${name}`, () => {
+    bench('Headers@@iterator', () => {
+      // prevention of memoization of results
+      headersList[kHeadersSortedMap] = null
+      return [...headers]
+    })
+
+    bench('Headers@@iterator (sorted)', () => {
+      // prevention of memoization of results
+      headersListSorted[kHeadersSortedMap] = null
+      return [...headersSorted]
+    })
+  })
+}
+
+await run()

benchmarks/sort.mjs

@@ -0,0 +1,59 @@
+import { bench, group, run } from 'mitata'
+import { sort, binaryInsertionSort, heapSort, introSort } from '../lib/fetch/sort.js'
+// import { sort as timSort } from './tim-sort.mjs'
+
+function compare (a, b) {
+  return a < b ? -1 : 1
+}
+
+const characters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789'
+const charactersLength = characters.length
+
+function generateAsciiString (length) {
+  let result = ''
+  for (let i = 0; i < length; ++i) {
+    result += characters[Math.floor(Math.random() * charactersLength)]
+  }
+  return result
+}
+
+const settings = {
+  tiny: 32,
+  small: 64,
+  middle: 128,
+  large: 512
+}
+
+for (const [name, length] of Object.entries(settings)) {
+  group(`sort (${name})`, () => {
+    const array = Array.from(new Array(length), () => generateAsciiString(12))
+    // sort(array, compare)
+    bench('Array#sort', () => array.slice().sort(compare))
+    bench('sort (mixed sort)', () => sort(array.slice(), compare))
+    // bench('tim sort', () => timSort(array.slice(), compare, 0, array.length))
+
+    // sort(array, start, end, compare)
+    bench('intro sort', () => introSort(array.slice(), 0, array.length, compare))
+    bench('heap sort', () => heapSort(array.slice(), 0, array.length, compare))
+
+    // Do not run them in large arrays as they are slow.
+    if (array.length <= 1000) bench('binary insertion sort', () => binaryInsertionSort(array.slice(), 0, array.length, compare))
+  })
+
+  group(`sort sortedArray (${name})`, () => {
+    const array = Array.from(new Array(length), () => generateAsciiString(12)).sort(compare)
+    // sort(array, compare)
+    bench('Array#sort', () => array.sort(compare))
+    bench('sort (mixed sort)', () => sort(array, compare))
+    // bench('tim sort', () => timSort(array, compare, 0, array.length))
+
+    // sort(array, start, end, compare)
+    bench('intro sort', () => introSort(array, 0, array.length, compare))
+    bench('heap sort', () => heapSort(array, 0, array.length, compare))
+
+    // Do not run them in large arrays as they are slow.
+    if (array.length <= 1000) bench('binary insertion sort', () => binaryInsertionSort(array, 0, array.length, compare))
+  })
+}
+
+await run()

lib/fetch/headers.js

@@ -12,6 +12,7 @@ const {
 } = require('./util')
 const { webidl } = require('./webidl')
 const assert = require('node:assert')
+const { sort } = require('./sort')
 
 const kHeadersMap = Symbol('headers map')
 const kHeadersSortedMap = Symbol('headers map sorted')
@@ -120,6 +121,10 @@ function appendHeader (headers, name, value) {
   //    privileged no-CORS request headers from headers
 }
 
+function compareHeaderName (a, b) {
+  return a[0] < b[0] ? -1 : 1
+}
+
 class HeadersList {
   /** @type {[string, string][]|null} */
   cookies = null
@@ -237,7 +242,7 @@ class HeadersList {
 
   * [Symbol.iterator] () {
     // use the lowercased name
-    for (const [name, { value }] of this[kHeadersMap]) {
+    for (const { 0: name, 1: { value } } of this[kHeadersMap]) {
       yield [name, value]
     }
   }
@@ -253,6 +258,79 @@ class HeadersList {
 
     return headers
   }
+
+  // https://fetch.spec.whatwg.org/#convert-header-names-to-a-sorted-lowercase-set
+  toSortedArray () {
+    const size = this[kHeadersMap].size
+    const array = new Array(size)
+    // In most cases, you will use the fast-path.
+    // fast-path: Use binary insertion sort for small arrays.
+    if (size <= 32) {
+      if (size === 0) {
+        // If empty, it is an empty array. To avoid the first index assignment.
+        return array
+      }
+      // Improve performance by unrolling loop and avoiding double-loop.
+      // Double-loop-less version of the binary insertion sort.
+      const iterator = this[kHeadersMap][Symbol.iterator]()
+      const firstValue = iterator.next().value
+      // set [name, value] to first index.
+      array[0] = [firstValue[0], firstValue[1].value]
+      // https://fetch.spec.whatwg.org/#concept-header-list-sort-and-combine
+      // 3.2.2. Assert: value is non-null.
+      assert(firstValue[1].value !== null)
+      for (
+        let i = 1, j = 0, right = 0, left = 0, pivot = 0, x, value;
+        i < size;
+        ++i
+      ) {
+        // get next value
+        value = iterator.next().value
+        // set [name, value] to current index.
+        x = array[i] = [value[0], value[1].value]
+        // https://fetch.spec.whatwg.org/#concept-header-list-sort-and-combine
+        // 3.2.2. Assert: value is non-null.
+        assert(x[1] !== null)
+        left = 0
+        right = i
+        // binary search
+        while (left < right) {
+          // middle index
+          pivot = left + ((right - left) >> 1)
+          // compare header name
+          if (array[pivot][0] <= x[0]) {
+            left = pivot + 1
+          } else {
+            right = pivot
+          }
+        }
+        if (i !== pivot) {
+          j = i
+          while (j > left) {
+            array[j] = array[--j]
+          }
+          array[left] = x
+        }
+      }
+      /* c8 ignore next 4 */
+      if (!iterator.next().done) {
+        // This is for debugging and will never be called.
+        throw new TypeError('Unreachable')
+      }
+      return array
+    } else {
+      // This case would be a rare occurrence.
+      // slow-path: fallback
+      let i = 0
+      for (const { 0: name, 1: { value } } of this[kHeadersMap]) {
+        array[i++] = [name, value]
+        // https://fetch.spec.whatwg.org/#concept-header-list-sort-and-combine
+        // 3.2.2. Assert: value is non-null.
+        assert(value !== null)
+      }
+      return sort(array, compareHeaderName)
+    }
+  }
 }
 
 // https://fetch.spec.whatwg.org/#headers-class
@@ -454,27 +532,19 @@ class Headers {
 
     // 2. Let names be the result of convert header names to a sorted-lowercase
     //    set with all the names of the headers in list.
-    const names = [...this[kHeadersList]]
-    const namesLength = names.length
-    if (namesLength <= 16) {
-      // Note: Use insertion sort for small arrays.
-      for (let i = 1, value, j = 0; i < namesLength; ++i) {
-        value = names[i]
-        for (j = i - 1; j >= 0; --j) {
-          if (names[j][0] <= value[0]) break
-          names[j + 1] = names[j]
-        }
-        names[j + 1] = value
-      }
-    } else {
-      names.sort((a, b) => a[0] < b[0] ? -1 : 1)
-    }
+    const names = this[kHeadersList].toSortedArray()
 
     const cookies = this[kHeadersList].cookies
 
+    // fast-path
+    if (cookies === null) {
+      // Note: The non-null assertion of value has already been done by `HeadersList#toSortedArray`
+      return (this[kHeadersList][kHeadersSortedMap] = names)
+    }
+
     // 3. For each name of names:
-    for (let i = 0; i < namesLength; ++i) {
-      const [name, value] = names[i]
+    for (let i = 0; i < names.length; ++i) {
+      const { 0: name, 1: value } = names[i]
       // 1. If name is `set-cookie`, then:
       if (name === 'set-cookie') {
         // 1. Let values be a list of all values of headers in list whose name
@@ -491,17 +561,15 @@ class Headers {
         // 1. Let value be the result of getting name from list.
 
         // 2. Assert: value is non-null.
-        assert(value !== null)
+        // Note: This operation was done by `HeadersList#toSortedArray`.
 
         // 3. Append (name, value) to headers.
         headers.push([name, value])
       }
     }
 
-    this[kHeadersList][kHeadersSortedMap] = headers
-
     // 4. Return headers.
-    return headers
+    return (this[kHeadersList][kHeadersSortedMap] = headers)
   }
 
   [Symbol.for('nodejs.util.inspect.custom')] () {
@@ -546,6 +614,8 @@ webidl.converters.HeadersInit = function (V) {
 
 module.exports = {
   fill,
+  // for test.
+  compareHeaderName,
   Headers,
   HeadersList
 }