[NEW] Better apostrophes in HTML output

[Rob Beezer]

We aim to make your source as simple as possible to create, provided you do the
hard work of structuring it so there are no ambiguities.

So for possessives and contractions, use the plain old apostrophe you get from
your keyboard, as always. For your HTML output, we will now upgrade it to a
stylish curly version (U+2019 for the cognesceti) for you. You will be able to
tell the difference if you look closely, but it takes some looking since we use
a sans-serif font.

Note that we are not changing your program code, your Sage code, or your LaTeX
math, or other non-literary bits - you've told us where those are and we won't
manipulate that text in ways that wreck its meaning.

There is presently one gotcha leftover from some Runestone integration. A
*horizontal* Parsons problem assumes its blocks are literal code, so an
apostrophe here *will* get mangled into a Unicode version and wreck any
automated testing of correct answers. The solution is to markup the blocks the
PreTeXt way - use "c" for code and leave literary text for "natural language"
problems. But now, the markup for "c" bleeds into the blocks, a situation that
needs to be corrected. You can follow progress on this at:

https://github.com/RunestoneInteractive/RunestoneComponents/issues/1395

Note also, that if you cut/paste from Google Docs or Word, your apostrophes may
have aleady been upgraded for you the minute you let go of the keypress and
could migrate to your PreTeXt source. We advise against this, and do not
consider this legitimate PreTeXt source. But is seems as a practical matter
there is no real harm (yet).

Rob