🚀 go-pugleaf

>Also, can someone explain to me how the Emacs M-q command works within
>LaTeX mode?  Its behavior makes no sense to me (based on my experience
>with it in regular text mode) and the Emacs documentation says nothing
>on the matter.

I think the emacs documentation says something, although I suppose it is
could be more readable.  The emacs info node Paragraphs says:
"   The precise definition of a paragraph boundary is controlled by the
variables `paragraph-separate' and `paragraph-start'.  "
And every major mode has a different value for these regexps.  So M-q
will work differently in different major modes.  

In Emacs 19.28 in latex mode the regexps are defined by the code in
tex-mode.el, which begins:
  ;; A line containing just $$ is treated as a paragraph separator.
  ;; A line starting with $$ starts a paragraph,
  ;; but does not separate paragraphs if it has more stuff on it.
  (setq paragraph-start "^[ \t]*$\\|^[\f%]\\|^[ \t]*\\$\\$\\|\
^\\\\begin\\>\\|^\\\\label\\>\\|^\\\\end\\>\\|^\\\\\\[\\|^\\\\\\]\\|\
^\\\\chapter\\>\\|^\\\\section\\>\\|\
...

So for instance, M-q at any place will not fill the following paragraph:

\begin{Theorem}
\label{important-stuff}
A well know factoid is that 
$$
1 + 1 = 2.
$$
This is useful in hypereleptic Quantum $C^*$ crystaline cohomology.
\end{Theorem}

However, M-q will fill the paragraph:

It's clear that $1 
\cdot 1 = 1$, and therefore we have the following result.

Formerly M-q would fill any lines that did not begin with a backslash.
The new regexps are supposed to be better suited for actual latex usage.