handle nbsp (nonbreaking space) in headings

[zebapy]

Our site uses a plugin to prevent widows by adding a   between the last 2 words of most elements (h1, p, etc).

When using anchorjs, this is causing some headings to come out as

<h2 id="where-to&nbsp;start">Where to&nbsp;Start?
<a class="anchorjs-link " aria-label="Anchor" data-anchorjs-icon="" href="#where-to&nbsp;start"></a>
</h2>

Everything is functional but the anchor link suffers in readability after.

[bryanbraun]

@zebapy, I'm on board with this.

It's definitely an edge-case, but a good-enough solution seems pretty low-overhead and straightforward.

As for the approach, he looks nice but I want to keep anchorjs dependency-free. I like the simplicity of the textarea approach. It shouldn't be hard to add it at the beginning of our urlify function, and add a test for it.

The biggest downside of this change is that it would require at least a minor (if not major) version bump, since we could be potentially changing link locations for existing users. I'm ok with that, but if I'm doing a major version bump, I want to include #109 in it as well.

Feel free to submit a PR if you want to take a pass at this. Otherwise, I'll get to it eventually.

[zebapy]

I'm also encountering curly quotes and degree icons not being replaced with nothing. #there’s-a-heading-after-intro-360°-degrees

[bryanbraun]

@zebapy, the quotes and degree icons are a part of the decision to support UTF-8 characters (except for a small set of non-safe characters). I had originally stripped out these kinds of characters, but I changed the approach in order to better support international characters. For more background on that decision, see: #41

The   showing up is different to me, since that's not an instance of a UTF-8 character, and I don't know of anybody who would want to see those in their urls. 😄