Link checker: Check for double forward slashes

[jeff-matthews]

This PR is a:

• New topic
• Content update
• Content fix or rewrite
• Bug fix or improvement

Summary

When this pull request is merged, it will add a custom test called DoubleSlashCheck to extend the html-proofer library, since the library doesn't natively check for this scenario. See gjtorikian/html-proofer#386.

URLs with double slashes will cause 404s on our website because our hosting solution doesn't automatically strip extra forward slashes from URLs like normal web servers.

This PR also fixes existing instances of double forward slashes in URLs.

[dshevtsov]

I cannot reproduce the test.
Debugging shows that class DoubleSlashCheck does not participate in the link checker run.
What is the scenario and expected behavior for failing test?

[jeff-matthews]

The scenario is a writer adding an extra forward slash when entering a URL in an <a> element (e.g., after a Jekyll variable). The LinkChecker class will not identify this as a failure, so we end up with 404s on the website despite the LinkCheck class. That's why we need a custom test/class.

The expected behavior of a failing test is any occurrence of a double forward slash in an <a> element in the compiled HTML (excluding double slashes around the protocol like https://).

For example:

Fail
<a href="https://devdocs.magento.com//path/to/file.html">Fail</a>

Pass
<a href="https://devdocs.magento.com/path/to/file.html">Pass</a>

The new class actually runs with the other checks (e.g., ImageCheck, ScriptCheck, LinkCheck, DoubleSlashCheck).

[dshevtsov]

The new class doesn't run on my environment, I still see only ImageCheck, ScriptCheck, LinkCheck.
My debugging tool shows that there are no entry points to the new class when the link checker is running.

[jeff-matthews]

Have you tried running rake test?

[dshevtsov]

ah, now I see it
so, it is not available for the link checking hook

[jeff-matthews]

Oops! That wasn't my intention. I assumed that the hook would invoke the new class. Let me know how we can change or optimize the new class. I was just looking for the quickest way to get this working. I'm open to suggestions.

[dshevtsov]

Working on it

[jeff-matthews]

running tests