Disable the language detection feature of the Nu HTML5 validator

Right now we're getting:

Validating HTML5 validity for [http://127.0.0.1:8080/xwiki/bin/view/Templates/Encyclopedia/Template/WebHome] executed with credentials Admin:admin
Warning at 2:578 This document appears to be written in Lorem ipsum text but the “html” start tag has “lang="en"”. Consider using “lang="zxx"” (or variant) instead.

In practice the doc content is not only Lorem Ipsum but it also contains English, see: https://github.com/xwiki-contrib/application-templates/blob/master/application-templates-ui/src/main/resources/Templates/Encyclopedia/Template/WebHome.xml

On https://github.com/validator/validator/issues/321 it says:

I very recently added a language-detection feature and the intermittent nature of the warnings is due to the fact that it bases the language detection on a random sample of text from the document. So it’s not deterministic. However it is usually accurate as a far as the determination it makes from what it samples. But the feature still needs some refinement and troubleshooting yet.

And

Yes. The lang value can somewhat affect the layout of a document—at least as far as things like hyphenation—so it seems like a good idea to have the correct lang value set even when you’re the stage of only evaluating the layout.

So it seems the reason for validating the lang is mostly for the layout, and that's not interesting for us.

In any case it seems it would lead to too many false positives and thus we're turning it off for the moment.