Update link tests

[hadro]

No description provided.

[hadro]

This is to finally update the link tests to a more modern version of html-proofer, 4.4.3.
This required re-rewriting many of the opts for the link checker, and the biggest change is that they've dropped HTML validation altogether, claiming that browsers are very good at figuring out intent and correcting errors, and that this caused needless errors. If we want to enforce HTML validation still, we'll have to do it another way.

[hadro]

What is this test doing that is not covered by either the internal link checker, or the external link checker?

[glenrobson]

Hi Josh,

I'm just looking into this and I've got a few questions:

• How are you updating the version of HTML proofer? I can't see the version in the Gemfile but I can see it in the website version...
• I can't seem to find version 4.4.3 on https://github.com/gjtorikian/html-proofer/releases
• Where do you see the HTML validation being removed?

Thanks

[hadro]

• Yes, I updated the gem file in this branch, with some other necessary tests as well: https://github.com/IIIF/website/tree/update_link_tests
• Interesting, I don't know why it's not in the releases, but it's here: https://github.com/gjtorikian/html-proofer/tree/v4.4.3
  The changes I made could probably just as easily make use of the more recent 5.0 as well
• The removal of HTML validation is in the intro of this upgrade path document: https://github.com/gjtorikian/html-proofer/blob/main/UPGRADING.md

[hadro]

From the upgrade link above: The biggest change is that the HTML parsing check has been removed. HTML parsers like Nokogumbo/Nokogiri, and indeed, many browsers, are fairly generous when it comes to accepting malformed HTML. This feature detected more false issues than actual ones.

[glenrobson]

Possible html5 checker: https://pypi.org/project/html5validator/