bpo-38324: Fix test__locale.py Windows failures (GH-20529)

Use wide-char _W_* fields of lconv structure on Windows
Remove "ps_AF" from test__locale.known_numerics on Windows

diff --git a/Lib/test/test__locale.py b/Lib/test/test__locale.py
index cda0ee91b700f73889d1d1c4a163b8bc3f278133..59a00bad7d98a3d9fce6188d59c6f7684110b901 100644 (file)
--- a/Lib/test/test__locale.py
+++ b/Lib/test/test__locale.py
@@ -72,6 +72,10 @@ known_numerics = {
     'ps_AF': ('\u066b', '\u066c'),
 }
 
+if sys.platform == 'win32':
+    # ps_AF doesn't work on Windows: see bpo-38324 (msg361830)
+    del known_numerics['ps_AF']
+
 class _LocaleTests(unittest.TestCase):
 
     def setUp(self):

diff --git a/Misc/NEWS.d/next/Windows/2020-05-30-02-46-43.bpo-38324.476M-5.rst b/Misc/NEWS.d/next/Windows/2020-05-30-02-46-43.bpo-38324.476M-5.rst
new file mode 100644 (file)
index 0000000..c45aa13
--- /dev/null
+++ b/Misc/NEWS.d/next/Windows/2020-05-30-02-46-43.bpo-38324.476M-5.rst
@@ -0,0 +1 @@
+Avoid Unicode errors when accessing certain locale data on Windows.

diff --git a/Modules/_localemodule.c b/Modules/_localemodule.c
index 0fe2e08b41f9f6d683dc9a816918deb8f06ecd02..9c7ce876e4059f481f6f52d7b2fe07d48f40f7a6 100644 (file)
--- a/Modules/_localemodule.c
+++ b/Modules/_localemodule.c
@@ -155,6 +155,7 @@ locale_is_ascii(const char *str)
 static int
 locale_decode_monetary(PyObject *dict, struct lconv *lc)
 {
+#ifndef MS_WINDOWS
     int change_locale;
     change_locale = (!locale_is_ascii(lc->int_curr_symbol)
                      || !locale_is_ascii(lc->currency_symbol)
@@ -190,12 +191,18 @@ locale_decode_monetary(PyObject *dict, struct lconv *lc)
         }
     }
 
+#define GET_LOCALE_STRING(ATTR) PyUnicode_DecodeLocale(lc->ATTR, NULL)
+#else  /* MS_WINDOWS */
+/* Use _W_* fields of Windows struct lconv */
+#define GET_LOCALE_STRING(ATTR) PyUnicode_FromWideChar(lc->_W_ ## ATTR, -1)
+#endif /* MS_WINDOWS */
+
     int res = -1;
 
 #define RESULT_STRING(ATTR) \
     do { \
         PyObject *obj; \
-        obj = PyUnicode_DecodeLocale(lc->ATTR, NULL); \
+        obj = GET_LOCALE_STRING(ATTR); \
         if (obj == NULL) { \
             goto done; \
         } \
@@ -211,14 +218,17 @@ locale_decode_monetary(PyObject *dict, struct lconv *lc)
     RESULT_STRING(mon_decimal_point);
     RESULT_STRING(mon_thousands_sep);
 #undef RESULT_STRING
+#undef GET_LOCALE_STRING
 
     res = 0;
 
 done:
+#ifndef MS_WINDOWS
     if (loc != NULL) {
         setlocale(LC_CTYPE, oldloc);
     }
     PyMem_Free(oldloc);
+#endif
     return res;
 }
 
@@ -258,9 +268,15 @@ _locale_localeconv_impl(PyObject *module)
         Py_DECREF(obj); \
     } while (0)
 
+#ifdef MS_WINDOWS
+/* Use _W_* fields of Windows struct lconv */
+#define GET_LOCALE_STRING(ATTR) PyUnicode_FromWideChar(lc->_W_ ## ATTR, -1)
+#else
+#define GET_LOCALE_STRING(ATTR) PyUnicode_DecodeLocale(lc->ATTR, NULL)
+#endif
 #define RESULT_STRING(s)\
     do { \
-        x = PyUnicode_DecodeLocale(lc->s, NULL); \
+        x = GET_LOCALE_STRING(s); \
         RESULT(#s, x); \
     } while (0)
 
@@ -289,8 +305,10 @@ _locale_localeconv_impl(PyObject *module)
     RESULT_INT(n_sign_posn);
 
     /* Numeric information: LC_NUMERIC encoding */
-    PyObject *decimal_point, *thousands_sep;
+    PyObject *decimal_point = NULL, *thousands_sep = NULL;
     if (_Py_GetLocaleconvNumeric(lc, &decimal_point, &thousands_sep) < 0) {
+        Py_XDECREF(decimal_point);
+        Py_XDECREF(thousands_sep);
         goto failed;
     }
 
@@ -319,6 +337,7 @@ _locale_localeconv_impl(PyObject *module)
 #undef RESULT
 #undef RESULT_STRING
 #undef RESULT_INT
+#undef GET_LOCALE_STRING
 }
 
 #if defined(HAVE_WCSCOLL)

diff --git a/Python/fileutils.c b/Python/fileutils.c
index b79067f2b5d38b8047168cab51f4d066b9086b7f..e125ba46c21baee67c3a03e7aa52d7110e4cce8d 100644 (file)
--- a/Python/fileutils.c
+++ b/Python/fileutils.c
@@ -2047,6 +2047,7 @@ _Py_GetLocaleconvNumeric(struct lconv *lc,
     assert(decimal_point != NULL);
     assert(thousands_sep != NULL);
 
+#ifndef MS_WINDOWS
     int change_locale = 0;
     if ((strlen(lc->decimal_point) > 1 || ((unsigned char)lc->decimal_point[0]) > 127)) {
         change_locale = 1;
@@ -2085,14 +2086,20 @@ _Py_GetLocaleconvNumeric(struct lconv *lc,
         }
     }
 
+#define GET_LOCALE_STRING(ATTR) PyUnicode_DecodeLocale(lc->ATTR, NULL)
+#else /* MS_WINDOWS */
+/* Use _W_* fields of Windows strcut lconv */
+#define GET_LOCALE_STRING(ATTR) PyUnicode_FromWideChar(lc->_W_ ## ATTR, -1)
+#endif /* MS_WINDOWS */
+
     int res = -1;
 
-    *decimal_point = PyUnicode_DecodeLocale(lc->decimal_point, NULL);
+    *decimal_point = GET_LOCALE_STRING(decimal_point);
     if (*decimal_point == NULL) {
         goto done;
     }
 
-    *thousands_sep = PyUnicode_DecodeLocale(lc->thousands_sep, NULL);
+    *thousands_sep = GET_LOCALE_STRING(thousands_sep);
     if (*thousands_sep == NULL) {
         goto done;
     }
@@ -2100,11 +2107,15 @@ _Py_GetLocaleconvNumeric(struct lconv *lc,
     res = 0;
 
 done:
+#ifndef MS_WINDOWS
     if (loc != NULL) {
         setlocale(LC_CTYPE, oldloc);
     }
     PyMem_Free(oldloc);
+#endif
     return res;
+
+#undef GET_LOCALE_STRING
 }
 
 /* Our selection logic for which function to use is as follows: