As has been discussed earlier on MO,<sup>1,2</sup>
recently an impressive advance was proved concerning
internally illuminating a mirrored polygon.
Here is the result:

> **Corollary 3**. Let $P$ be a rational polygon. 
Then for any $x \in P$ there are at most
finitely many points $y$ for which there is no 
geodesic trajectory between $x$ and $y$.

Here "no geodesic trajectory" means no illumination path,
using angle-of-incidence equals angle-of-reflection from the
polygon edges.
A *rational polygon* is one each of whose angles is a rational
multiple of $\pi$. So, a match lit at $x$ will leave at most
a finite number of points in $P$ dark.
[Examples were known](https://mathworld.wolfram.com/IlluminationProblem.html):

&nbsp; &nbsp; &nbsp;
<img src="https://i.sstatic.net/yxa2r.gif" />

The theorem is proved in this "Everything is illuminated" paper:

> Lelievre, Samuel, Thierry Monteil, and Barak Weiss. 
"Everything is illuminated." 
*Geometry & Topology* 20, no. 3 (2016): 1737-1762.

The authors say they depend on
"recent breakthrough results of
Eskin-Mirzakhani [EM] and Eskin-Mirzakhani-Mohammadi [EMM],"


<hr />

Here finally is my question:

> ***Q***.
Suppose $P$ has all rational vertex angles, except for two
irrational vertex angles. 
(A polygon cannot have just one irrational angle, because the angle sum is $(n-2)\pi$.)
Is it now true that, for any $x$,
the entirety of $P$ is illuminated, i.e., no point is dark?

I am wondering if this is implied by their results. I do not have 
sufficient grasp of their use of
the "Magic Wand Theorem" 
and other translation-surface technology to understand if the answer to ***Q*** is just a
corollary, or a possibly open question.
It makes some sense to me that one irrational angle will suffice to fill
$P$ with light, because the lightray reflections become ergodic.

<hr />

 - <sup>1</sup>[Current state of Straus's illumination
   problem](https://mathoverflow.net/a/252441/6094),
   
 - <sup>2</sup>[Using mirrors to make a non-convex polygon visible from a fixed
   interior point](https://mathoverflow.net/a/42007/6094)