UTF-8 encoding seems to get lost in PyScope.Execute #2392

Probably doing something wrong here, but no idea what.

This is the Python code (this is a copy-paste of the code variable from the C# code below, as viewed in the Visual Studio watch window, if that matters):

# encoding: UTF-8
import io

with io.open("phi.txt", "r", encoding="utf-8") as f:
    phi = f.read()

if phi != "φ":
    raise RuntimeError(f"Expected 'φ' but got '{phi}'")

This is phi.txt (no newline):

φ

Running this with CPython 3.8 outputs nothing as expected. Running this via pythonnet_py_38_win v2.5.2 using the same Python distribution

using( Py.GIL() )
{
  var scope = Py.CreateScope();
  var script = "/path/to/phi.py";
  var code = File.ReadAllText( script, Encoding.UTF8 );
  scope.Execute( PythonEngine.Compile( code, script, RunFlagType.File ) );
}

results in a PythonException: {"RuntimeError : Expected 'f' but got 'φ'"}. What would be going on here?