Duplicate header names aren't differentiated in IDs

[ChiriVulpes]

Other markdown parsers support having multiple sections with the same name, but this one doesn't.

Input:

# test
# test

Expected Output:

<h1 id="test">test</h1>
<h1 id="test-1">test</h1>

Actual Output:

<h1 id="test">test</h1>
<h1 id="test">test</h1>

Supporting this means subsections can go by the same name. (for instance, in an md file I have, I have a section for specifications of files, and another section for examples of the files in action, and each section needs a subsection of the same name, as they're the same file)
There are probably other use cases as well. Also supporting this would make it not be invalid html =P

[charlieTheBotDev]

I'm using markdown-toc to insert a table of contents in one of my projects and wrote a workaround for this that handles duplicates in the same way.

Input

<!-- toc -->

# test
# test

Output

<!-- toc -->
<ul>
<li><a href="#test">test</a></li>
<li><a href="#test-1">test</a></li>
</ul>
<!-- tocstop -->
<h1 id="test">test</h1>
<h1 id="test-1">test</h1>

ES6 Version

const content = `
<!-- toc -->

# test
# test
`

const marked = require('marked')
const toc = require('markdown-toc')
const renderer = new marked.Renderer()
const headings = []
renderer.heading = (text, level) => {
  const escapedText = text.toLowerCase().replace(/[^\w]+/g, '-')
  const duplicateIndex = headings.map(({ text }) => text).indexOf(escapedText)
  let duplicateText = undefined
  if (duplicateIndex === -1) {
    headings.push({
      text: escapedText,
      count: 0
    })
  } else {
    headings[duplicateIndex].count++
    duplicateText = `${escapedText}-${headings[duplicateIndex].count}`
  }
  return `<h${level} id="${duplicateText || escapedText}">${text}</h${level}>\n`
}
const parsedMarkdown = marked(toc.insert(content), { renderer })

console.log('INPUT\n' + content + '\nEND INPUT')
console.log('OUTPUT\n' + parsedMarkdown + '\nEND OUTPUT')

ES5 Version

var content = '<!-- toc -->\n\n# test\n# test'
var marked = require('marked')
var toc = require('markdown-toc')
var renderer = new marked.Renderer()
var headings = []
renderer.heading = function (text, level) {
  var escapedText = text.toLowerCase().replace(/[^\w]+/g, '-')
  var duplicateIndex = headings.map(function (h) { return h.text }).indexOf(escapedText)
  var duplicateText
  if (duplicateIndex === -1) {
    headings.push({
      text: escapedText,
      count: 0
    })
  } else {
    headings[duplicateIndex].count++
    duplicateText = escapedText + '-' + headings[duplicateIndex].count
  }
  return '<h' + level + ' id="'+ (duplicateText || escapedText) + '">' + text + '</h' + level + '>\n'
}
var parsedMarkdown = marked(toc.insert(content), { renderer: renderer })

console.log('INPUT\n' + content + '\nEND INPUT')
console.log('OUTPUT\n' + parsedMarkdown + '\nEND OUTPUT')

I'm pretty sure this is how all markdown parsers handle duplicates

It also works fine if you remove the <!-- toc --> and don't use toc.insert

[carmacleod]

This is also causing Eclipse Orion bug: https://bugs.eclipse.org/bugs/show_bug.cgi?id=525834

[joshbruce]

See #981