Issue #27781: Change file system encoding on Windows to UTF-8 (PEP 529)

3 files changed, 73 insertions, 5 deletions

diff --git a/Python/bltinmodule.c b/Python/bltinmodule.c
index be145609dd..252c0a7b89 100644
--- a/Python/bltinmodule.c
+++ b/Python/bltinmodule.c
@@ -21,16 +21,18 @@
    Don't forget to modify PyUnicode_DecodeFSDefault() if you touch any of the
    values for Py_FileSystemDefaultEncoding!
 */
-#ifdef HAVE_MBCS
-const char *Py_FileSystemDefaultEncoding = "mbcs";
+#if defined(__APPLE__)
+const char *Py_FileSystemDefaultEncoding = "utf-8";
 int Py_HasFileSystemDefaultEncoding = 1;
-#elif defined(__APPLE__)
+#elif defined(MS_WINDOWS)
+/* may be changed by initfsencoding(), but should never be free()d */
 const char *Py_FileSystemDefaultEncoding = "utf-8";
 int Py_HasFileSystemDefaultEncoding = 1;
 #else
 const char *Py_FileSystemDefaultEncoding = NULL; /* set by initfsencoding() */
 int Py_HasFileSystemDefaultEncoding = 0;
 #endif
+const char *Py_FileSystemDefaultEncodeErrors = "surrogateescape";
 
 _Py_IDENTIFIER(__builtins__);
 _Py_IDENTIFIER(__dict__);
diff --git a/Python/pylifecycle.c b/Python/pylifecycle.c
index 1896888916..3f3b614a47 100644
--- a/Python/pylifecycle.c
+++ b/Python/pylifecycle.c
@@ -90,6 +90,9 @@ int Py_NoUserSiteDirectory = 0; /* for -s and site.py */
 int Py_UnbufferedStdioFlag = 0; /* Unbuffered binary std{in,out,err} */
 int Py_HashRandomizationFlag = 0; /* for -R and PYTHONHASHSEED */
 int Py_IsolatedFlag = 0; /* for -I, isolate from user's env */
+#ifdef MS_WINDOWS
+int Py_LegacyWindowsFSEncodingFlag = 0; /* Uses mbcs instead of utf-8 */
+#endif
 
 PyThreadState *_Py_Finalizing = NULL;
 
@@ -321,6 +324,10 @@ _Py_InitializeEx_Private(int install_sigs, int install_importlib)
        check its value further. */
     if ((p = Py_GETENV("PYTHONHASHSEED")) && *p != '\0')
         Py_HashRandomizationFlag = add_flag(Py_HashRandomizationFlag, p);
+#ifdef MS_WINDOWS
+    if ((p = Py_GETENV("PYTHONLEGACYWINDOWSFSENCODING")) && *p != '\0')
+        Py_LegacyWindowsFSEncodingFlag = add_flag(Py_LegacyWindowsFSEncodingFlag, p);
+#endif
 
     _PyRandom_Init();
 
@@ -958,6 +965,18 @@ initfsencoding(PyInterpreterState *interp)
 {
     PyObject *codec;
 
+#ifdef MS_WINDOWS
+    if (Py_LegacyWindowsFSEncodingFlag)
+    {
+        Py_FileSystemDefaultEncoding = "mbcs";
+        Py_FileSystemDefaultEncodeErrors = "replace";
+    }
+    else
+    {
+        Py_FileSystemDefaultEncoding = "utf-8";
+        Py_FileSystemDefaultEncodeErrors = "surrogatepass";
+    }
+#else
     if (Py_FileSystemDefaultEncoding == NULL)
     {
         Py_FileSystemDefaultEncoding = get_locale_encoding();
@@ -968,6 +987,7 @@ initfsencoding(PyInterpreterState *interp)
         interp->fscodec_initialized = 1;
         return 0;
     }
+#endif
 
     /* the encoding is mbcs, utf-8 or ascii */
     codec = _PyCodec_Lookup(Py_FileSystemDefaultEncoding);
diff --git a/Python/sysmodule.c b/Python/sysmodule.c
index a54f266030..0fe76b7a74 100644
--- a/Python/sysmodule.c
+++ b/Python/sysmodule.c
@@ -311,6 +311,23 @@ operating system filenames."
 );
 
 static PyObject *
+sys_getfilesystemencodeerrors(PyObject *self)
+{
+    if (Py_FileSystemDefaultEncodeErrors)
+        return PyUnicode_FromString(Py_FileSystemDefaultEncodeErrors);
+    PyErr_SetString(PyExc_RuntimeError,
+        "filesystem encoding is not initialized");
+    return NULL;
+}
+
+PyDoc_STRVAR(getfilesystemencodeerrors_doc,
+    "getfilesystemencodeerrors() -> string\n\
+\n\
+Return the error mode used to convert Unicode filenames in\n\
+operating system filenames."
+);
+
+static PyObject *
 sys_intern(PyObject *self, PyObject *args)
 {
     PyObject *s;
@@ -866,6 +883,24 @@ sys_getwindowsversion(PyObject *self)
 
 #pragma warning(pop)
 
+PyDoc_STRVAR(enablelegacywindowsfsencoding_doc,
+"_enablelegacywindowsfsencoding()\n\
+\n\
+Changes the default filesystem encoding to mbcs:replace for consistency\n\
+with earlier versions of Python. See PEP 529 for more information.\n\
+\n\
+This is equivalent to defining the PYTHONLEGACYWINDOWSFSENCODING \n\
+environment variable before launching Python."
+);
+
+static PyObject *
+sys_enablelegacywindowsfsencoding(PyObject *self)
+{
+    Py_FileSystemDefaultEncoding = "mbcs";
+    Py_FileSystemDefaultEncodeErrors = "replace";
+    Py_RETURN_NONE;
+}
+
 #endif /* MS_WINDOWS */
 
 #ifdef HAVE_DLOPEN
@@ -1225,6 +1260,8 @@ static PyMethodDef sys_methods[] = {
 #endif
     {"getfilesystemencoding", (PyCFunction)sys_getfilesystemencoding,
      METH_NOARGS, getfilesystemencoding_doc},
+    { "getfilesystemencodeerrors", (PyCFunction)sys_getfilesystemencodeerrors,
+     METH_NOARGS, getfilesystemencodeerrors_doc },
 #ifdef Py_TRACE_REFS
     {"getobjects",      _Py_GetObjects, METH_VARARGS},
 #endif
@@ -1240,6 +1277,8 @@ static PyMethodDef sys_methods[] = {
 #ifdef MS_WINDOWS
     {"getwindowsversion", (PyCFunction)sys_getwindowsversion, METH_NOARGS,
      getwindowsversion_doc},
+    {"_enablelegacywindowsfsencoding", (PyCFunction)sys_enablelegacywindowsfsencoding,
+     METH_NOARGS, enablelegacywindowsfsencoding_doc },
 #endif /* MS_WINDOWS */
     {"intern",          sys_intern,     METH_VARARGS, intern_doc},
     {"is_finalizing",   sys_is_finalizing, METH_NOARGS, is_finalizing_doc},
@@ -1456,14 +1495,21 @@ version -- the version of this interpreter as a string\n\
 version_info -- version information as a named tuple\n\
 "
 )
-#ifdef MS_WINDOWS
+#ifdef MS_COREDLL
 /* concatenating string here */
 PyDoc_STR(
 "dllhandle -- [Windows only] integer handle of the Python DLL\n\
 winver -- [Windows only] version number of the Python DLL\n\
 "
 )
-#endif /* MS_WINDOWS */
+#endif /* MS_COREDLL */
+#ifdef MS_WINDOWS
+/* concatenating string here */
+PyDoc_STR(
+"_enablelegacywindowsfsencoding -- [Windows only] \n\
+"
+)
+#endif
 PyDoc_STR(
 "__stdin__ -- the original stdin; don't touch!\n\
 __stdout__ -- the original stdout; don't touch!\n\