fixture: make the logic of both non-English and English characters fi…

…ltering with the same logic (nodejs#2570)

1. We should make sure that either your English anchor name or non-English
anchor name is filtered with the same logic, for non-English characters,
we should also include mid-spaces filtered out.
2. Move the ANCHOR_COMMENTREG out of the scope as the global-based one.
3. Remove the 'g', because it will remember the lastIndex where it began to
search for the last time.

scripts/plugins/anchor-markdown-headings.js

@@ -12,12 +12,19 @@
  * `<!---->` to quote your English anchor name inside, with your
  * own title beside it.
  */
-module.exports = function anchorMarkdownHeadings (text, level, raw, slugger) {
-  // Check whether we've got the comment matches
-  // <!--comment-->, <!--comment --> and even <!-- comment-->
-  // (20 hex = 32 dec = space character)
-  const ANCHOR_COMMENTREG = /<!--\x20?([\w|-]+)\x20?-->/gi
 
+// Check whether we've got the comment matches
+// <!--comment-->, <!--comment --> and even <!-- comment-->
+// (20 hex = 32 dec = space character)
+// Only need to find one that matches the Regex, we cannot use
+// 'g' here because it will continue searching for the rest string,
+// and you have to reset the lastIndex, which isn't a necessary for
+// our current situation.
+// For more, you can see:
+// https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/RegExp/lastIndex#Description
+const ANCHOR_COMMENTREG = /<!--\x20?([\w\x20-]+)\x20?-->/
+
+module.exports = function anchorMarkdownHeadings (text, level, raw, slugger) {
   let anchorTitle = null
 
   // If we've checked the title has a comment symbol,
@@ -27,15 +34,15 @@ module.exports = function anchorMarkdownHeadings (text, level, raw, slugger) {
   if (anchorTitleArray !== null) {
     anchorTitle = anchorTitleArray[1]
   } else {
-    // For others, directly replace all non-English characters
-    // to '-' in the middle, only keep the English characters
     anchorTitle = raw
-      .replace(/(\[([^\]]+)]\([^)]+\))/g, '$2')
-      .replace(/[^\w]+/g, '-')
-      .replace(/-{2,}/g, '-')
-      .replace(/(^-|-$)/g, '')
   }
 
+  anchorTitle = anchorTitle.replace(/(\[([^\]]+)]\([^)]+\))/g, '$2')
+    .replace(/[^\w]+/g, '-')
+    .replace(/[\x20]+/g, '-')
+    .replace(/-{2,}/g, '-')
+    .replace(/(^-|-$)/g, '')
+
   if (!anchorTitle) {
     return `<h${level}>${text}</h${level}>`
   }

tests/scripts/anchor-mardown-headings.test.js

@@ -5,7 +5,7 @@ const test = require('tape')
 const anchorMarkdownHeadings = require('../../scripts/plugins/anchor-markdown-headings')
 
 test('anchorMarkdownHeadings', (t) => {
-  t.plan(7)
+  t.plan(8)
 
   t.test('correctly parses markdown heading without links', (t) => {
     const text = 'Simple title'
@@ -76,6 +76,21 @@ test('anchorMarkdownHeadings', (t) => {
     t.equal(output, expected)
   })
 
+  t.test('correctly parses markdown heading with empty spaces', (t) => {
+    const text = '这是<a href="b">链接</a>的<a href="d">测试！</a>'
+    const level = 2
+    const raw = '<!-- empty  spaces - id -->这是[链接](b)c[测试！](d)'
+    const output = anchorMarkdownHeadings(text, level, raw)
+    const expected = '<h2 id="header-empty-spaces-id">' +
+      '这是<a href="b">链接</a>的<a href="d">测试！</a>' +
+      '<a id="empty-spaces-id" class="anchor" ' +
+      'href="#empty-spaces-id" ' +
+      'aria-labelledby="header-empty-spaces-id"></a></h2>'
+
+    t.plan(1)
+    t.equal(output, expected)
+  })
+
   t.test('does not generate empty anchors', (t) => {
     const text = 'إنضم إلينا'
     const level = 2