Separate FILE_ENCODING from TERMINAL_ENCODING

gitlint-core/gitlint/cli.py

@@ -68,7 +68,7 @@ def log_system_info():
     LOG.debug("Git version: %s", git_version())
     LOG.debug("Gitlint version: %s", gitlint.__version__)
     LOG.debug("GITLINT_USE_SH_LIB: %s", os.environ.get("GITLINT_USE_SH_LIB", "[NOT SET]"))
-    LOG.debug("DEFAULT_ENCODING: %s", gitlint.utils.DEFAULT_ENCODING)
+    LOG.debug("DEFAULT_ENCODING: %s", gitlint.utils.TERMINAL_ENCODING)
 
 
 def build_config(
@@ -264,7 +264,7 @@ def __init__(self, config, config_builder, commit_hash, refspec, msg_filename, g
 @click.option("--ignore", envvar="GITLINT_IGNORE", default="", help="Ignore rules (comma-separated by id or name).")
 @click.option("--contrib", envvar="GITLINT_CONTRIB", default="",
               help="Contrib rules to enable (comma-separated by id or name).")
-@click.option("--msg-filename", type=click.File(encoding=gitlint.utils.DEFAULT_ENCODING),
+@click.option("--msg-filename", type=click.File(encoding=gitlint.utils.FILE_ENCODING),
               help="Path to a file containing a commit-msg.")
 @click.option("--ignore-stdin", envvar="GITLINT_IGNORE_STDIN", is_flag=True,
               help="Ignore any stdin data. Useful for running in CI server.")

gitlint-core/gitlint/config.py

@@ -13,7 +13,7 @@
 )
 from gitlint.contrib import rules as contrib_rules
 from gitlint.exception import GitlintError
-from gitlint.utils import DEFAULT_ENCODING
+from gitlint.utils import FILE_ENCODING
 
 
 def handle_option_error(func):
@@ -468,7 +468,7 @@ def set_from_config_file(self, filename):
         try:
             parser = ConfigParser()
 
-            with open(filename, encoding=DEFAULT_ENCODING) as config_file:
+            with open(filename, encoding=FILE_ENCODING) as config_file:
                 parser.read_file(config_file, filename)
 
             for section_name in parser.sections():

gitlint-core/gitlint/hooks.py

@@ -4,7 +4,7 @@
 
 from gitlint.exception import GitlintError
 from gitlint.git import git_hooks_dir
-from gitlint.utils import DEFAULT_ENCODING
+from gitlint.utils import FILE_ENCODING
 
 COMMIT_MSG_HOOK_SRC_PATH = os.path.join(os.path.dirname(os.path.realpath(__file__)), "files", "commit-msg")
 COMMIT_MSG_HOOK_DST_PATH = "commit-msg"
@@ -52,7 +52,7 @@ def uninstall_commit_msg_hook(lint_config):
         if not os.path.exists(dest_path):
             raise GitHookInstallerError(f"There is no commit-msg hook present in {dest_path}.")
 
-        with open(dest_path, encoding=DEFAULT_ENCODING) as fp:
+        with open(dest_path, encoding=FILE_ENCODING) as fp:
             lines = fp.readlines()
             if len(lines) < 2 or lines[1] != GITLINT_HOOK_IDENTIFIER:
                 msg = (

gitlint-core/gitlint/shell.py

@@ -6,7 +6,7 @@
 
 import subprocess
 
-from gitlint.utils import DEFAULT_ENCODING, USE_SH_LIB
+from gitlint.utils import TERMINAL_ENCODING, USE_SH_LIB
 
 
 def shell(cmd):
@@ -64,7 +64,7 @@ def _exec(*args, **kwargs):
             raise CommandNotFound from e
 
         exit_code = p.returncode
-        stdout = result[0].decode(DEFAULT_ENCODING)
+        stdout = result[0].decode(TERMINAL_ENCODING)
         stderr = result[1]  # 'sh' does not decode the stderr bytes to unicode
         full_cmd = "" if args is None else " ".join(args)
 

gitlint-core/gitlint/tests/base.py

@@ -13,7 +13,7 @@
 from gitlint.deprecation import LOG as DEPRECATION_LOG
 from gitlint.deprecation import Deprecation
 from gitlint.git import GitChangedFileStats, GitContext
-from gitlint.utils import DEFAULT_ENCODING, LOG_FORMAT
+from gitlint.utils import FILE_ENCODING, LOG_FORMAT
 
 EXPECTED_REGEX_STYLE_SEARCH_DEPRECATION_WARNING = (
     "WARNING: gitlint.deprecated.regex_style_search {0} - {1}: gitlint will be switching from using "
@@ -95,7 +95,7 @@ def get_sample_path(filename=""):
     def get_sample(filename=""):
         """Read and return the contents of a file in gitlint/tests/samples"""
         sample_path = BaseTestCase.get_sample_path(filename)
-        return Path(sample_path).read_text(encoding=DEFAULT_ENCODING)
+        return Path(sample_path).read_text(encoding=FILE_ENCODING)
 
     @staticmethod
     def patch_input(side_effect):
@@ -109,7 +109,7 @@ def get_expected(filename="", variable_dict=None):
         """Utility method to read an expected file from gitlint/tests/expected and return it as a string.
         Optionally replace template variables specified by variable_dict."""
         expected_path = os.path.join(BaseTestCase.EXPECTED_DIR, filename)
-        expected = Path(expected_path).read_text(encoding=DEFAULT_ENCODING)
+        expected = Path(expected_path).read_text(encoding=FILE_ENCODING)
 
         if variable_dict:
             expected = expected.format(**variable_dict)

gitlint-core/gitlint/tests/cli/test_cli.py

@@ -9,7 +9,7 @@
 from gitlint import __version__, cli
 from gitlint.shell import CommandNotFound
 from gitlint.tests.base import BaseTestCase
-from gitlint.utils import DEFAULT_ENCODING
+from gitlint.utils import FILE_ENCODING, TERMINAL_ENCODING
 
 
 class CLITests(BaseTestCase):
@@ -39,7 +39,7 @@ def get_system_info_dict():
             "gitlint_version": __version__,
             "GITLINT_USE_SH_LIB": BaseTestCase.GITLINT_USE_SH_LIB,
             "target": os.path.realpath(os.getcwd()),
-            "DEFAULT_ENCODING": DEFAULT_ENCODING,
+            "DEFAULT_ENCODING": TERMINAL_ENCODING,
         }
 
     def test_version(self):
@@ -315,7 +315,7 @@ def test_lint_staged_msg_filename(self, sh, _):
 
         with self.tempdir() as tmpdir:
             msg_filename = os.path.join(tmpdir, "msg")
-            with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+            with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                 f.write("WIP: msg-filename tïtle\n")
 
             with patch("gitlint.display.stderr", new=StringIO()) as stderr:
@@ -370,7 +370,7 @@ def test_msg_filename(self, _):
 
         with self.tempdir() as tmpdir:
             msg_filename = os.path.join(tmpdir, "msg")
-            with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+            with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                 f.write("Commït title\n")
 
             with patch("gitlint.display.stderr", new=StringIO()) as stderr:

gitlint-core/gitlint/tests/cli/test_cli_hooks.py

@@ -6,7 +6,7 @@
 from gitlint import cli, config, hooks
 from gitlint.shell import ErrorReturnCode
 from gitlint.tests.base import BaseTestCase
-from gitlint.utils import DEFAULT_ENCODING
+from gitlint.utils import FILE_ENCODING
 
 
 class CLIHookTests(BaseTestCase):
@@ -102,7 +102,7 @@ def test_run_hook_no_tty(self):
 
         with self.tempdir() as tmpdir:
             msg_filename = os.path.join(tmpdir, "hür")
-            with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+            with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                 f.write("WIP: tïtle\n")
 
             with patch("gitlint.display.stderr", new=StringIO()) as stderr:
@@ -130,7 +130,7 @@ def test_run_hook_edit(self, shell):
 
             with self.patch_input(["e", "e", "n"]), self.tempdir() as tmpdir:
                 msg_filename = os.path.realpath(os.path.join(tmpdir, "hür"))
-                with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+                with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                     f.write(commit_messages[i] + "\n")
 
                 with patch("gitlint.display.stderr", new=StringIO()) as stderr:
@@ -158,7 +158,7 @@ def test_run_hook_no(self):
 
         with self.patch_input(["n"]), self.tempdir() as tmpdir:
             msg_filename = os.path.join(tmpdir, "hür")
-            with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+            with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                 f.write("WIP: höok no\n")
 
             with patch("gitlint.display.stderr", new=StringIO()) as stderr:
@@ -175,7 +175,7 @@ def test_run_hook_yes(self):
         """Test for run-hook subcommand, answering 'y(es)' after commit-hook"""
         with self.patch_input(["y"]), self.tempdir() as tmpdir:
             msg_filename = os.path.join(tmpdir, "hür")
-            with open(msg_filename, "w", encoding=DEFAULT_ENCODING) as f:
+            with open(msg_filename, "w", encoding=FILE_ENCODING) as f:
                 f.write("WIP: höok yes\n")
 
             with patch("gitlint.display.stderr", new=StringIO()) as stderr:

gitlint-core/gitlint/tests/test_utils.py

@@ -27,7 +27,7 @@ def test_use_sh_library(self, patched_env):
         self.assertEqual(utils.use_sh_library(), False)
 
     @patch("gitlint.utils.locale")
-    def test_default_encoding_non_windows(self, mocked_locale):
+    def test_terminal_encoding_non_windows(self, mocked_locale):
         utils.PLATFORM_IS_WINDOWS = False
         mocked_locale.getpreferredencoding.return_value = "foöbar"
         self.assertEqual(utils.getpreferredencoding(), "foöbar")
@@ -37,7 +37,7 @@ def test_default_encoding_non_windows(self, mocked_locale):
         self.assertEqual(utils.getpreferredencoding(), "UTF-8")
 
     @patch("os.environ")
-    def test_default_encoding_windows(self, patched_env):
+    def test_terminal_encoding_windows(self, patched_env):
         utils.PLATFORM_IS_WINDOWS = True
         # Mock out os.environ
         mock_env = {}

gitlint-core/gitlint/utils.py

@@ -38,39 +38,50 @@ def use_sh_library():
 USE_SH_LIB = use_sh_library()
 
 ########################################################################################################################
-# DEFAULT_ENCODING
+# TERMINAL_ENCODING
+# Encoding used for terminal encoding/decoding.
 
 
 def getpreferredencoding():
     """Modified version of local.getpreferredencoding() that takes into account LC_ALL, LC_CTYPE, LANG env vars
     on windows and falls back to UTF-8."""
     fallback_encoding = "UTF-8"
-    default_encoding = locale.getpreferredencoding() or fallback_encoding
+    preferred_encoding = locale.getpreferredencoding() or fallback_encoding
 
     # On Windows, we mimic git/linux by trying to read the LC_ALL, LC_CTYPE, LANG env vars manually
     # (on Linux/MacOS the `getpreferredencoding()` call will take care of this).
     # We fallback to UTF-8
     if PLATFORM_IS_WINDOWS:
-        default_encoding = fallback_encoding
+        preferred_encoding = fallback_encoding
         for env_var in ["LC_ALL", "LC_CTYPE", "LANG"]:
             encoding = os.environ.get(env_var, False)
             if encoding:
                 # Support dotted (C.UTF-8) and non-dotted (C or UTF-8) charsets:
                 # If encoding contains a dot: split and use second part, otherwise use everything
                 dot_index = encoding.find(".")
-                default_encoding = encoding[dot_index + 1 :] if dot_index != -1 else encoding
+                preferred_encoding = encoding[dot_index + 1 :] if dot_index != -1 else encoding
                 break
 
         # We've determined what encoding the user *wants*, let's now check if it's actually a valid encoding on the
         # system. If not, fallback to UTF-8.
         # This scenario is fairly common on Windows where git sets LC_CTYPE=C when invoking the commit-msg hook, which
         # is not a valid encoding in Python on Windows.
         try:
-            codecs.lookup(default_encoding)
+            codecs.lookup(preferred_encoding)
         except LookupError:
-            default_encoding = fallback_encoding
+            preferred_encoding = fallback_encoding
 
-    return default_encoding
+    return preferred_encoding
 
 
-DEFAULT_ENCODING = getpreferredencoding()
+TERMINAL_ENCODING = getpreferredencoding()
+
+########################################################################################################################
+# FILE_ENCODING
+# Gitlint assumes UTF-8 encoding for all file operations:
+#  - reading/writing its own hook and config files
+#  - reading/writing git commit messages
+# Git does have i18n.commitEncoding and i18n.logOutputEncoding options which we might want to take into account,
+# but that's not supported today.
+
+FILE_ENCODING = "UTF-8"

qa/test_gitlint.py

@@ -2,7 +2,7 @@
 
 from qa.base import BaseTestCase
 from qa.shell import echo, git, gitlint
-from qa.utils import DEFAULT_ENCODING
+from qa.utils import FILE_ENCODING
 
 
 class IntegrationTests(BaseTestCase):
@@ -58,7 +58,7 @@ def test_fixup_commit(self):
         self.assertEqualStdout(output, expected)
 
         # Make a small modification to the commit and commit it using fixup commit
-        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=DEFAULT_ENCODING) as fh:
+        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=FILE_ENCODING) as fh:
             fh.write("Appending söme stuff\n")
 
         git("add", test_filename, _cwd=self.tmp_git_repo)
@@ -87,7 +87,7 @@ def test_fixup_amend_commit(self):
         self.assertEqualStdout(output, expected)
 
         # Make a small modification to the commit and commit it using fixup=amend commit
-        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=DEFAULT_ENCODING) as fh:
+        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=FILE_ENCODING) as fh:
             fh.write("Appending söme stuff\n")
 
         git("add", test_filename, _cwd=self.tmp_git_repo)
@@ -133,7 +133,7 @@ def test_squash_commit(self):
         self.assertEqualStdout(output, expected)
 
         # Make a small modification to the commit and commit it using squash commit
-        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=DEFAULT_ENCODING) as fh:
+        with open(os.path.join(self.tmp_git_repo, test_filename), "a", encoding=FILE_ENCODING) as fh:
             # Wanted to write a unicode string, but that's obnoxious if you want to do it across Python 2 and 3.
             # https://stackoverflow.com/questions/22392377/
             # error-writing-a-file-with-file-write-in-python-unicodeencodeerror