I'm still confused: the main purpose of 2to3 is to take Python 2 code and turn it into Python 3 code. The docs[1] are fairly unambiguous here:

> 2to3 is a Python program that reads Python 2.x source code and applies a series of fixers to transform it into valid Python 3.x code.

So expecting 2to3 *not* to parse code as though it's Python 2 code seem odd.

[1] https://docs.python.org/3.7/library/2to3.html