bpo-35336: Fix PYTHONCOERCECLOCALE=1 (GH-10806) (GH-10813)

Fix PYTHONCOERCECLOCALE=1 environment variable: only coerce the C locale if the LC_CTYPE locale is "C". (cherry picked from commit 55e4980)

Author: vstinner

Lib/test/test_c_locale_coercion.py

@@ -1,11 +1,12 @@
 # Tests the attempted automatic coercion of the C locale to a UTF-8 locale
 
-import unittest
 import locale
 import os
+import shutil
+import subprocess
 import sys
 import sysconfig
-import shutil
+import unittest
 from collections import namedtuple
 
 import test.support
@@ -25,6 +26,8 @@
 # Set our expectation for the default locale used when none is specified
 EXPECT_COERCION_IN_DEFAULT_LOCALE = True
 
+TARGET_LOCALES = ["C.UTF-8", "C.utf8", "UTF-8"]
+
 # Apply some platform dependent overrides
 if sys.platform.startswith("linux"):
  if test.support.is_android:
@@ -404,6 +407,27 @@ def test_LC_ALL_set_to_C(self):
  expected_warnings=[LEGACY_LOCALE_WARNING],
  coercion_expected=False)
 
+ def test_PYTHONCOERCECLOCALE_set_to_one(self):
+ # skip the test if the LC_CTYPE locale is C or coerced
+ old_loc = locale.setlocale(locale.LC_CTYPE, None)
+ self.addCleanup(locale.setlocale, locale.LC_CTYPE, old_loc)
+ loc = locale.setlocale(locale.LC_CTYPE, "")
+ if loc == "C":
+ self.skipTest("test requires LC_CTYPE locale different than C")
+ if loc in TARGET_LOCALES :
+ self.skipTest("coerced LC_CTYPE locale: %s" % loc)
+
+ # bpo-35336: PYTHONCOERCECLOCALE=1 must not coerce the LC_CTYPE locale
+ # if it's not equal to "C"
+ code = 'import locale; print(locale.setlocale(locale.LC_CTYPE, None))'
+ env = dict(os.environ, PYTHONCOERCECLOCALE='1')
+ cmd = subprocess.run([sys.executable, '-c', code],
+ stdout=subprocess.PIPE,
+ env=env,
+ text=True)
+ self.assertEqual(cmd.stdout.rstrip(), loc)
+
+
 def test_main():
  test.support.run_unittest(
  LocaleConfigurationTests,

Misc/NEWS.d/next/Core and Builtins/2018-11-29-23-59-52.bpo-35336.8LOz4F.rst

@@ -0,0 +1,2 @@
+Fix PYTHONCOERCECLOCALE=1 environment variable: only coerce the C locale
+if the LC_CTYPE locale is "C".

Modules/main.c

@@ -2191,11 +2191,17 @@ pymain_read_conf(_PyMain *pymain, _PyCoreConfig *config, _PyCmdline *cmdline)
 static void
 config_init_locale(_PyCoreConfig *config)
 {
- if (config->coerce_c_locale < 0) {
+ /* Test also if coerce_c_locale equals 1: PYTHONCOERCECLOCALE=1 doesn't
+ imply that the C locale is always coerced. It is only coerced if
+ if the LC_CTYPE locale is "C". */
+ if (config->coerce_c_locale != 0) {
  /* The C locale enables the C locale coercion (PEP 538) */
  if (_Py_LegacyLocaleDetected()) {
  config->coerce_c_locale = 1;
  }
+ else {
+ config->coerce_c_locale = 0;
+ }
  }
 
 #ifndef MS_WINDOWS
@@ -2376,7 +2382,7 @@ _PyCoreConfig_Read(_PyCoreConfig *config)
  }
  }
 
- if (config->utf8_mode < 0 || config->coerce_c_locale < 0) {
+ if (config->coerce_c_locale != 0 || config->utf8_mode < 0) {
  config_init_locale(config);
  }