bpo-36775: _PyCoreConfig only uses wchar_t* (GH-13062)

_PyCoreConfig: Change filesystem_encoding, filesystem_errors,
stdio_encoding and stdio_errors fields type from char* to wchar_t*.

Changes:

* PyInterpreterState: replace fscodec_initialized (int) with fs_codec
  structure.
* Add get_error_handler_wide() and unicode_encode_utf8() helper
  functions.
* Add error_handler parameter to unicode_encode_locale()
  and unicode_decode_locale().
* Remove _PyCoreConfig_SetString().
* Rename _PyCoreConfig_SetWideString() to _PyCoreConfig_SetString().
* Rename _PyCoreConfig_SetWideStringFromString()
  to _PyCoreConfig_DecodeLocale().

Include/cpython/coreconfig.h

@@ -207,8 +207,8 @@ typedef struct {
 
        See Py_FileSystemDefaultEncoding and Py_FileSystemDefaultEncodeErrors.
        */
-    char *filesystem_encoding;
-    char *filesystem_errors;
+    wchar_t *filesystem_encoding;
+    wchar_t *filesystem_errors;
 
     wchar_t *pycache_prefix;  /* PYTHONPYCACHEPREFIX, -X pycache_prefix=PATH */
     wchar_t *program_name;    /* Program name, see also Py_GetProgramName() */
@@ -334,13 +334,13 @@ typedef struct {
        Value set from PYTHONIOENCODING environment variable and
        Py_SetStandardStreamEncoding() function.
        See also 'stdio_errors' attribute. */
-    char *stdio_encoding;
+    wchar_t *stdio_encoding;
 
     /* Error handler of sys.stdin and sys.stdout.
        Value set from PYTHONIOENCODING environment variable and
        Py_SetStandardStreamEncoding() function.
        See also 'stdio_encoding' attribute. */
-    char *stdio_errors;
+    wchar_t *stdio_errors;
 
 #ifdef MS_WINDOWS
     /* If greater than zero, use io.FileIO instead of WindowsConsoleIO for sys

Include/internal/pycore_coreconfig.h

@@ -106,12 +106,9 @@ PyAPI_FUNC(_PyInitError) _PyCoreConfig_Copy(
     _PyCoreConfig *config,
     const _PyCoreConfig *config2);
 PyAPI_FUNC(_PyInitError) _PyCoreConfig_SetString(
-    char **config_str,
-    const char *str);
-PyAPI_FUNC(_PyInitError) _PyCoreConfig_SetWideString(
     wchar_t **config_str,
     const wchar_t *str);
-PyAPI_FUNC(_PyInitError) _PyCoreConfig_SetWideStringFromString(
+PyAPI_FUNC(_PyInitError) _PyCoreConfig_DecodeLocale(
     wchar_t **config_str,
     const char *str);
 PyAPI_FUNC(_PyInitError) _PyCoreConfig_InitPathConfig(_PyCoreConfig *config);

Include/internal/pycore_pylifecycle.h

@@ -21,6 +21,9 @@ extern int _Py_SetFileSystemEncoding(
     const char *errors);
 extern void _Py_ClearFileSystemEncoding(void);
 extern _PyInitError _PyUnicode_InitEncodings(PyInterpreterState *interp);
+#ifdef MS_WINDOWS
+extern int _PyUnicode_EnableLegacyWindowsFSEncoding(void);
+#endif
 
 PyAPI_FUNC(void) _Py_ClearStandardStreamEncoding(void);
 

Include/internal/pycore_pystate.h

@@ -56,7 +56,14 @@ struct _is {
     PyObject *codec_search_cache;
     PyObject *codec_error_registry;
     int codecs_initialized;
-    int fscodec_initialized;
+
+    /* fs_codec.encoding is initialized to NULL.
+       Later, it is set to a non-NULL string by _PyUnicode_InitEncodings(). */
+    struct {
+        char *encoding;   /* Filesystem encoding (encoded to UTF-8) */
+        char *errors;     /* Filesystem errors (encoded to UTF-8) */
+        _Py_error_handler error_handler;
+    } fs_codec;
 
     _PyCoreConfig core_config;
 #ifdef HAVE_DLOPEN

Objects/stringlib/codecs.h

@@ -260,6 +260,7 @@ Py_LOCAL_INLINE(PyObject *)
 STRINGLIB(utf8_encoder)(PyObject *unicode,
                         STRINGLIB_CHAR *data,
                         Py_ssize_t size,
+                        _Py_error_handler error_handler,
                         const char *errors)
 {
     Py_ssize_t i;                /* index into data of next input character */
@@ -268,7 +269,6 @@ STRINGLIB(utf8_encoder)(PyObject *unicode,
     PyObject *error_handler_obj = NULL;
     PyObject *exc = NULL;
     PyObject *rep = NULL;
-    _Py_error_handler error_handler = _Py_ERROR_UNKNOWN;
 #endif
 #if STRINGLIB_SIZEOF_CHAR == 1
     const Py_ssize_t max_char_size = 2;

Programs/_testembed.c

@@ -488,8 +488,8 @@ static int test_init_from_config(void)
        Force it to 0 through the config. */
     config.legacy_windows_stdio = 0;
 #endif
-    config.stdio_encoding = "iso8859-1";
-    config.stdio_errors = "replace";
+    config.stdio_encoding = L"iso8859-1";
+    config.stdio_errors = L"replace";
 
     putenv("PYTHONNOUSERSITE=");
     Py_NoUserSiteDirectory = 0;

Python/preconfig.c

@@ -14,7 +14,10 @@
 /* --- File system encoding/errors -------------------------------- */
 
 /* The filesystem encoding is chosen by config_init_fs_encoding(),
-   see also initfsencoding(). */
+   see also initfsencoding().
+
+   Py_FileSystemDefaultEncoding and Py_FileSystemDefaultEncodeErrors
+   are encoded to UTF-8. */
 const char *Py_FileSystemDefaultEncoding = NULL;
 int Py_HasFileSystemDefaultEncoding = 0;
 const char *Py_FileSystemDefaultEncodeErrors = NULL;

Python/pylifecycle.c

@@ -1668,7 +1668,7 @@ is_valid_fd(int fd)
 static PyObject*
 create_stdio(const _PyCoreConfig *config, PyObject* io,
     int fd, int write_mode, const char* name,
-    const char* encoding, const char* errors)
+    const wchar_t* encoding, const wchar_t* errors)
 {
     PyObject *buf = NULL, *stream = NULL, *text = NULL, *raw = NULL, *res;
     const char* mode;
@@ -1718,7 +1718,7 @@ create_stdio(const _PyCoreConfig *config, PyObject* io,
 #ifdef MS_WINDOWS
     /* Windows console IO is always UTF-8 encoded */
     if (PyWindowsConsoleIO_Check(raw))
-        encoding = "utf-8";
+        encoding = L"utf-8";
 #endif
 
     text = PyUnicode_FromString(name);
@@ -1754,10 +1754,25 @@ create_stdio(const _PyCoreConfig *config, PyObject* io,
     newline = "\n";
 #endif
 
-    stream = _PyObject_CallMethodId(io, &PyId_TextIOWrapper, "OsssOO",
-                                    buf, encoding, errors,
+    PyObject *encoding_str = PyUnicode_FromWideChar(encoding, -1);
+    if (encoding_str == NULL) {
+        Py_CLEAR(buf);
+        goto error;
+    }
+
+    PyObject *errors_str = PyUnicode_FromWideChar(errors, -1);
+    if (errors_str == NULL) {
+        Py_CLEAR(buf);
+        Py_CLEAR(encoding_str);
+        goto error;
+    }
+
+    stream = _PyObject_CallMethodId(io, &PyId_TextIOWrapper, "OOOsOO",
+                                    buf, encoding_str, errors_str,
                                     newline, line_buffering, write_through);
     Py_CLEAR(buf);
+    Py_CLEAR(encoding_str);
+    Py_CLEAR(errors_str);
     if (stream == NULL)
         goto error;
 
@@ -1874,7 +1889,7 @@ init_sys_streams(PyInterpreterState *interp)
     fd = fileno(stderr);
     std = create_stdio(config, iomod, fd, 1, "<stderr>",
                        config->stdio_encoding,
-                       "backslashreplace");
+                       L"backslashreplace");
     if (std == NULL)
         goto error;
 

Python/sysmodule.c

@@ -424,7 +424,7 @@ sys_getfilesystemencoding_impl(PyObject *module)
 {
     PyInterpreterState *interp = _PyInterpreterState_GET_UNSAFE();
     const _PyCoreConfig *config = &interp->core_config;
-    return PyUnicode_FromString(config->filesystem_encoding);
+    return PyUnicode_FromWideChar(config->filesystem_encoding, -1);
 }
 
 /*[clinic input]
@@ -439,7 +439,7 @@ sys_getfilesystemencodeerrors_impl(PyObject *module)
 {
     PyInterpreterState *interp = _PyInterpreterState_GET_UNSAFE();
     const _PyCoreConfig *config = &interp->core_config;
-    return PyUnicode_FromString(config->filesystem_errors);
+    return PyUnicode_FromWideChar(config->filesystem_errors, -1);
 }
 
 /*[clinic input]
@@ -1211,30 +1211,9 @@ static PyObject *
 sys__enablelegacywindowsfsencoding_impl(PyObject *module)
 /*[clinic end generated code: output=f5c3855b45e24fe9 input=2bfa931a20704492]*/
 {
-    PyInterpreterState *interp = _PyInterpreterState_GET_UNSAFE();
-    _PyCoreConfig *config = &interp->core_config;
-
-    /* Set the filesystem encoding to mbcs/replace (PEP 529) */
-    char *encoding = _PyMem_RawStrdup("mbcs");
-    char *errors = _PyMem_RawStrdup("replace");
-    if (encoding == NULL || errors == NULL) {
-        PyMem_Free(encoding);
-        PyMem_Free(errors);
-        PyErr_NoMemory();
-        return NULL;
-    }
-
-    PyMem_RawFree(config->filesystem_encoding);
-    config->filesystem_encoding = encoding;
-    PyMem_RawFree(config->filesystem_errors);
-    config->filesystem_errors = errors;
-
-    if (_Py_SetFileSystemEncoding(config->filesystem_encoding,
-                                  config->filesystem_errors) < 0) {
-        PyErr_NoMemory();
+    if (_PyUnicode_EnableLegacyWindowsFSEncoding() < 0) {
         return NULL;
     }
-
     Py_RETURN_NONE;
 }