Remove handling of mbcs encoding

This has been phased out in Python 3.6.

https://peps.python.org/pep-0529/

Author: snejus

test/test_library.py

@@ -19,7 +19,6 @@
 import re
 import shutil
 import stat
-import sys
 import time
 import unicodedata
 import unittest
@@ -423,19 +422,6 @@ def test_unicode_normalized_nfc_on_linux(self):
             dest = self.i.destination(relative_to_libdir=True)
         assert as_string(dest) == unicodedata.normalize("NFC", instr)
 
-    def test_non_mbcs_characters_on_windows(self):
-        oldfunc = sys.getfilesystemencoding
-        sys.getfilesystemencoding = lambda: "mbcs"
-        try:
-            self.i.title = "h\u0259d"
-            self.lib.path_formats = [("default", "$title")]
-            p = self.i.destination()
-            assert b"?" not in p
-            # We use UTF-8 to encode Windows paths now.
-            assert "h\u0259d".encode() in p
-        finally:
-            sys.getfilesystemencoding = oldfunc
-
     def test_unicode_extension_in_fragment(self):
         self.lib.path_formats = [("default", "foo")]
         self.i.path = util.bytestring_path("bar.caf\xe9")

test/test_util.py

@@ -156,13 +156,8 @@ def test_syspath_posix_unchanged(self):
         assert path == outpath
 
     def _windows_bytestring_path(self, path):
-        old_gfse = sys.getfilesystemencoding
-        sys.getfilesystemencoding = lambda: "mbcs"
-        try:
-            with _common.platform_windows():
-                return util.bytestring_path(path)
-        finally:
-            sys.getfilesystemencoding = old_gfse
+        with _common.platform_windows():
+            return util.bytestring_path(path)
 
     def test_bytestring_path_windows_encodes_utf8(self):
         path = "caf\xe9"