Anchor link checking fails

[MikeMcC399]

markdown-link-check reports a 404 error when checking a correctly formatted link such as

[Good regular link](#link1)
# link1

This is a regression from markdown-link-check@3.11.2.

markdown-link-check version

3.12.1

Operating system

Ubuntu 22.04.4 LTS

Node.js version

v20.11.1 LTS

Steps to reproduce

Execute:

mkdir md-link-test
cd md-link-test
npm install markdown-link-check

Create a Markdown file test.md with the following contents

| Link source                 | Should | 3.8.7    | 3.11.2   | 3.12.1   |
| --------------------------- | ------ | -------- | -------- | -------- |
| [Good regular link](#link1) | Pass   | Pass     | Pass     | **Fail** |
| [Good link in HTML](#link2) | Pass   | Pass     | **Fail** | **Fail** |
| [Bad link](#link3)          | Fail   | **Pass** | Fail     | Fail     |

**Targets**

# link1

<a id='link2'>Link 2</a>

Execute:

npx markdown-link-check test.md

Note the failure: [✖] #link1 → Status: 404

Test with 3.11.2

Execute:

npm install markdown-link-check@3.11.2
npx markdown-link-check test.md

Note the correct success of testing link1: [✓] #link1

Logs from 3.12.1

$ npx markdown-link-check test.md

FILE: test.md
  [✖] #link1
  [✖] #link2
  [✖] #link3

  3 links checked.

  ERROR: 3 dead links found!
  [✖] #link1 → Status: 404
  [✖] #link2 → Status: 404
  [✖] #link3 → Status: 404

Related issue

• Parse HTML in order to find anchors (like #tomato) in the same document #195

[MikeMcC399]

This is a showstopper for updating to markdown-link-check@latest (3.12.1).

[dklimpel]

It is a regression in 2eb21a1

IMO it is dependency markdown-link-extractor.

[MikeMcC399]

@dklimpel

It is a regression in 2eb21a1

IMO it is dependency markdown-link-extractor.

Thank you for locating the source of the regression. I can confirm that the issue also exists in 3.12.0. You would need to add the --ignore workaround to demonstrate this, i.e.:

npx markdown-link-check test.md --ignore nothing

For some reason, the CI test has the hash-links test being skipped, so anchors are not tested:

markdown-link-check/test/markdown-link-check.test.js

Lines 343 to 354 in c8c394c

	it.skip('check hash links', function (done) {
	markdownLinkCheck(fs.readFileSync(path.join(__dirname, 'hash-links.md')).toString(), {}, function (err, result) {
	expect(err).to.be(null);
	expect(result).to.eql([
	{ link: '#foo', statusCode: 200, err: null, status: 'alive' },
	{ link: '#bar', statusCode: 200, err: null, status: 'alive' },
	{ link: '#potato', statusCode: 404, err: null, status: 'dead' },
	{ link: '#tomato', statusCode: 404, err: null, status: 'dead' },
	]);
	done();
	});
	});

[dklimpel]

It was broken.

98ba6bf

:-/

[dklimpel]

Reason is a missing upstream feature: tcort/markdown-link-extractor#24

• added: Reintroduce anchor detection as configurable step markdown-link-extractor#25
• reverted: tcort/markdown-link-extractor@685474f

[AvinZarlez]

I also ran into this issue today on 3.12.1.

[MikeMcC399]

• This issue is no longer reproducible in release v3.12.2, possibly due to PR Check GitHub markdown section links #312