Resolve EncodingWarnings

mypy/build.py

@@ -17,6 +17,7 @@
 import errno
 import gc
 import json
+import locale
 import os
 import platform
 import re
@@ -1194,7 +1195,7 @@ def add_catch_all_gitignore(target_dir: str) -> None:
     """
     gitignore = os.path.join(target_dir, ".gitignore")
     try:
-        with open(gitignore, "x") as f:
+        with open(gitignore, "x", encoding=locale.getpreferredencoding(False)) as f:
             print("# Automatically created by mypy", file=f)
             print("*", file=f)
     except FileExistsError:
@@ -1208,7 +1209,7 @@ def exclude_from_backups(target_dir: str) -> None:
     """
     cachedir_tag = os.path.join(target_dir, "CACHEDIR.TAG")
     try:
-        with open(cachedir_tag, "x") as f:
+        with open(cachedir_tag, "x", encoding=locale.getpreferredencoding(False)) as f:
             f.write(
                 """Signature: 8a477f597d28d172789f06886806bc55
 # This file is a cache directory tag automatically created by mypy.
@@ -3600,7 +3601,7 @@ def record_missing_stub_packages(cache_dir: str, missing_stub_packages: set[str]
     """
     fnam = missing_stubs_file(cache_dir)
     if missing_stub_packages:
-        with open(fnam, "w") as f:
+        with open(fnam, "w", encoding=locale.getpreferredencoding(False)) as f:
             for pkg in sorted(missing_stub_packages):
                 f.write(f"{pkg}\n")
     else:

mypy/config_parser.py

@@ -3,6 +3,7 @@
 import argparse
 import configparser
 import glob as fileglob
+import locale
 import os
 import re
 import sys
@@ -236,7 +237,7 @@ def parse_config_file(
                 parser: MutableMapping[str, Any] = destructure_overrides(toml_data)
                 config_types = toml_config_types
             else:
-                config_parser.read(config_file)
+                config_parser.read(config_file, encoding=locale.getpreferredencoding(False))
                 parser = config_parser
                 config_types = ini_config_types
         except (tomllib.TOMLDecodeError, configparser.Error, ConfigTOMLValueError) as err:

mypy/metastore.py

@@ -11,6 +11,7 @@
 from __future__ import annotations
 
 import binascii
+import locale
 import os
 import time
 from abc import abstractmethod
@@ -93,7 +94,9 @@ def read(self, name: str) -> str:
         if not self.cache_dir_prefix:
             raise FileNotFoundError()
 
-        with open(os.path.join(self.cache_dir_prefix, name)) as f:
+        with open(
+            os.path.join(self.cache_dir_prefix, name), encoding=locale.getpreferredencoding(False)
+        ) as f:
             return f.read()
 
     def write(self, name: str, data: str, mtime: float | None = None) -> bool:
@@ -106,7 +109,7 @@ def write(self, name: str, data: str, mtime: float | None = None) -> bool:
         tmp_filename = path + "." + random_string()
         try:
             os.makedirs(os.path.dirname(path), exist_ok=True)
-            with open(tmp_filename, "w") as f:
+            with open(tmp_filename, "w", encoding=locale.getpreferredencoding(False)) as f:
                 f.write(data)
             os.replace(tmp_filename, path)
             if mtime is not None:

mypy/modulefinder.py

@@ -8,6 +8,7 @@
 import ast
 import collections
 import functools
+import locale
 import os
 import re
 import subprocess
@@ -866,7 +867,7 @@ def load_stdlib_py_versions(custom_typeshed_dir: str | None) -> StdlibVersions:
 
     versions_path = os.path.join(stdlib_dir, "VERSIONS")
     assert os.path.isfile(versions_path), (custom_typeshed_dir, versions_path, __file__)
-    with open(versions_path) as f:
+    with open(versions_path, encoding=locale.getpreferredencoding(False)) as f:
         for line in f:
             line = line.split("#")[0].strip()
             if line == "":

mypy/test/testapi.py

@@ -1,7 +1,11 @@
 from __future__ import annotations
 
+import shutil
+import subprocess
 import sys
 from io import StringIO
+from pathlib import Path
+from tempfile import NamedTemporaryFile, mkdtemp
 
 import mypy.api
 from mypy.test.helpers import Suite
@@ -13,12 +17,18 @@ def setUp(self) -> None:
         self.sys_stderr = sys.stderr
         sys.stdout = self.stdout = StringIO()
         sys.stderr = self.stderr = StringIO()
+        with NamedTemporaryFile(delete=False) as tmp:
+            tmp.write(b"x: int = 5\n")
+        self.tmp_path = Path(tmp.name)
+        self.tmp_cache_dir = Path(mkdtemp())
 
     def tearDown(self) -> None:
         sys.stdout = self.sys_stdout
         sys.stderr = self.sys_stderr
         assert self.stdout.getvalue() == ""
         assert self.stderr.getvalue() == ""
+        self.tmp_path.unlink()
+        shutil.rmtree(self.tmp_cache_dir)
 
     def test_capture_bad_opt(self) -> None:
         """stderr should be captured when a bad option is passed."""
@@ -43,3 +53,20 @@ def test_capture_version(self) -> None:
         stdout, _, _ = mypy.api.run(["--version"])
         assert isinstance(stdout, str)
         assert stdout != ""
+
+    def test_default_encoding_warnings(self) -> None:
+        """No EncodingWarnings should be emitted."""
+        for empty_cache in [True, False]:
+            res = subprocess.run(
+                [
+                    sys.executable,
+                    "-c",
+                    "import mypy.api;"
+                    "mypy.api.run("
+                    f"['--cache-dir', '{self.tmp_cache_dir}', '{self.tmp_path}']"
+                    ")",
+                ],
+                capture_output=True,
+                env={"PYTHONWARNDEFAULTENCODING": "1"},
+            )
+            assert b"EncodingWarning" not in res.stderr