Issue #9713, #10114: Parser functions (eg. PyParser_ASTFromFile) expects

filenames encoded to the filesystem encoding with surrogateescape error handler
(to support undecodable bytes), instead of UTF-8 in strict mode.

Misc/NEWS

@@ -10,6 +10,10 @@ What's New in Python 3.2 Beta 1?
 Core and Builtins
 -----------------
 
+- Issue #9713, #10114: Parser functions (eg. PyParser_ASTFromFile) expects
+  filenames encoded to the filesystem encoding with surrogateescape error
+  handler (to support undecodable bytes), instead of UTF-8 in strict mode.
+
 - Issue #9997: Don't let the name "top" have special significance in scope
   resolution.
 
@@ -39,7 +43,7 @@ Library
   XML namespace attribute is encountered.
 
 - Issue #2830: Add the ``html.escape()`` function, which quotes all problematic
-  characters by default.  Deprecate ``cgi.escape()``.  
+  characters by default.  Deprecate ``cgi.escape()``.
 
 - Issue #9409: Fix the regex to match all kind of filenames, for interactive
   debugging in doctests.

Parser/tokenizer.h

@@ -40,7 +40,7 @@ struct tok_state {
     int level;          /* () [] {} Parentheses nesting level */
             /* Used to allow free continuations inside them */
     /* Stuff for checking on different tab sizes */
-    const char *filename;       /* For error messages */
+    const char *filename;   /* encoded to the filesystem encoding */
     int altwarning;     /* Issue warning if alternate tabs don't match */
     int alterror;       /* Issue error if alternate tabs don't match */
     int alttabsize;     /* Alternate tab spacing */

Python/ast.c

@@ -102,6 +102,7 @@ static void
 ast_error_finish(const char *filename)
 {
     PyObject *type, *value, *tback, *errstr, *offset, *loc, *tmp;
+    PyObject *filename_obj;
     long lineno;
 
     assert(PyErr_Occurred());
@@ -130,7 +131,11 @@ ast_error_finish(const char *filename)
         Py_INCREF(Py_None);
         loc = Py_None;
     }
-    tmp = Py_BuildValue("(zlOO)", filename, lineno, offset, loc);
+    filename_obj = PyUnicode_DecodeFSDefault(filename);
+    if (filename_obj != NULL)
+        tmp = Py_BuildValue("(NlOO)", filename_obj, lineno, offset, loc);
+    else
+        tmp = NULL;
     Py_DECREF(loc);
     if (!tmp) {
         Py_DECREF(errstr);

Python/bltinmodule.c

@@ -524,6 +524,7 @@ static PyObject *
 builtin_compile(PyObject *self, PyObject *args, PyObject *kwds)
 {
     char *str;
+    PyObject *filename_obj;
     char *filename;
     char *startstr;
     int mode = -1;
@@ -535,12 +536,16 @@ builtin_compile(PyObject *self, PyObject *args, PyObject *kwds)
     static char *kwlist[] = {"source", "filename", "mode", "flags",
                              "dont_inherit", NULL};
     int start[] = {Py_file_input, Py_eval_input, Py_single_input};
+    PyObject *result;
 
-    if (!PyArg_ParseTupleAndKeywords(args, kwds, "Oss|ii:compile",
-                                     kwlist, &cmd, &filename, &startstr,
-                                     &supplied_flags, &dont_inherit))
+    if (!PyArg_ParseTupleAndKeywords(args, kwds, "OO&s|ii:compile",  kwlist,
+                                     &cmd,
+                                     PyUnicode_FSConverter, &filename_obj,
+                                     &startstr, &supplied_flags,
+                                     &dont_inherit))
         return NULL;
 
+    filename = PyBytes_AS_STRING(filename_obj);
     cf.cf_flags = supplied_flags | PyCF_SOURCE_IS_UTF8;
 
     if (supplied_flags &
@@ -548,7 +553,7 @@ builtin_compile(PyObject *self, PyObject *args, PyObject *kwds)
     {
         PyErr_SetString(PyExc_ValueError,
                         "compile(): unrecognised flags");
-        return NULL;
+        goto error;
     }
     /* XXX Warn if (supplied_flags & PyCF_MASK_OBSOLETE) != 0? */
 
@@ -565,14 +570,13 @@ builtin_compile(PyObject *self, PyObject *args, PyObject *kwds)
     else {
         PyErr_SetString(PyExc_ValueError,
                         "compile() arg 3 must be 'exec', 'eval' or 'single'");
-        return NULL;
+        goto error;
     }
 
     is_ast = PyAST_Check(cmd);
     if (is_ast == -1)
-        return NULL;
+        goto error;
     if (is_ast) {
-        PyObject *result;
         if (supplied_flags & PyCF_ONLY_AST) {
             Py_INCREF(cmd);
             result = cmd;
@@ -585,20 +589,27 @@ builtin_compile(PyObject *self, PyObject *args, PyObject *kwds)
             mod = PyAST_obj2mod(cmd, arena, mode);
             if (mod == NULL) {
                 PyArena_Free(arena);
-                return NULL;
+                goto error;
             }
             result = (PyObject*)PyAST_Compile(mod, filename,
                                               &cf, arena);
             PyArena_Free(arena);
         }
-        return result;
+        goto finally;
     }
 
     str = source_as_string(cmd, "compile", "string, bytes, AST or code", &cf);
     if (str == NULL)
-        return NULL;
+        goto error;
 
-    return Py_CompileStringFlags(str, filename, start[mode], &cf);
+    result = Py_CompileStringFlags(str, filename, start[mode], &cf);
+    goto finally;
+
+error:
+    result = NULL;
+finally:
+    Py_DECREF(filename_obj);
+    return result;
 }
 
 PyDoc_STRVAR(compile_doc,

Python/compile.c

@@ -3942,7 +3942,7 @@ makecode(struct compiler *c, struct assembler *a)
     freevars = dict_keys_inorder(c->u->u_freevars, PyTuple_Size(cellvars));
     if (!freevars)
         goto error;
-    filename = PyUnicode_FromString(c->c_filename);
+    filename = PyUnicode_DecodeFSDefault(c->c_filename);
     if (!filename)
         goto error;
 

Python/pythonrun.c

@@ -1213,7 +1213,7 @@ PyRun_SimpleFileExFlags(FILE *fp, const char *filename, int closeit,
     d = PyModule_GetDict(m);
     if (PyDict_GetItemString(d, "__file__") == NULL) {
         PyObject *f;
-        f = PyUnicode_FromString(filename);
+        f = PyUnicode_DecodeFSDefault(filename);
         if (f == NULL)
             return -1;
         if (PyDict_SetItemString(d, "__file__", f) < 0) {
@@ -1968,7 +1968,9 @@ err_input(perrdetail *err)
 {
     PyObject *v, *w, *errtype, *errtext;
     PyObject *msg_obj = NULL;
+    PyObject *filename;
     char *msg = NULL;
+
     errtype = PyExc_SyntaxError;
     switch (err->error) {
     case E_ERROR:
@@ -2052,8 +2054,12 @@ err_input(perrdetail *err)
         errtext = PyUnicode_DecodeUTF8(err->text, strlen(err->text),
                                        "replace");
     }
-    v = Py_BuildValue("(ziiN)", err->filename,
-                      err->lineno, err->offset, errtext);
+    filename = PyUnicode_DecodeFSDefault(err->filename);
+    if (filename != NULL)
+        v = Py_BuildValue("(NiiN)", filename,
+                          err->lineno, err->offset, errtext);
+    else
+        v = NULL;
     if (v != NULL) {
         if (msg_obj)
             w = Py_BuildValue("(OO)", msg_obj, v);

Python/traceback.c

@@ -142,16 +142,19 @@ _Py_FindSourceFile(PyObject *filename, char* namebuf, size_t namelen, PyObject *
     Py_ssize_t npath;
     size_t taillen;
     PyObject *syspath;
-    const char* path;
+    PyObject *path;
     const char* tail;
+    PyObject *filebytes;
     const char* filepath;
     Py_ssize_t len;
+    PyObject* result;
 
-    filepath = _PyUnicode_AsString(filename);
-    if (filepath == NULL) {
+    filebytes = PyUnicode_EncodeFSDefault(filename);
+    if (filebytes == NULL) {
         PyErr_Clear();
         return NULL;
     }
+    filepath = PyBytes_AS_STRING(filebytes);
 
     /* Search tail of filename in sys.path before giving up */
     tail = strrchr(filepath, SEP);
@@ -163,7 +166,7 @@ _Py_FindSourceFile(PyObject *filename, char* namebuf, size_t namelen, PyObject *
 
     syspath = PySys_GetObject("path");
     if (syspath == NULL || !PyList_Check(syspath))
-        return NULL;
+        goto error;
     npath = PyList_Size(syspath);
 
     for (i = 0; i < npath; i++) {
@@ -174,14 +177,18 @@ _Py_FindSourceFile(PyObject *filename, char* namebuf, size_t namelen, PyObject *
         }
         if (!PyUnicode_Check(v))
             continue;
-        path = _PyUnicode_AsStringAndSize(v, &len);
+        path = PyUnicode_EncodeFSDefault(v);
         if (path == NULL) {
             PyErr_Clear();
             continue;
         }
-        if (len + 1 + (Py_ssize_t)taillen >= (Py_ssize_t)namelen - 1)
+        len = PyBytes_GET_SIZE(path);
+        if (len + 1 + (Py_ssize_t)taillen >= (Py_ssize_t)namelen - 1) {
+            Py_DECREF(path);
             continue; /* Too long */
-        strcpy(namebuf, path);
+        }
+        strcpy(namebuf, PyBytes_AS_STRING(path));
+        Py_DECREF(path);
         if (strlen(namebuf) != len)
             continue; /* v contains '\0' */
         if (len > 0 && namebuf[len-1] != SEP)
@@ -189,11 +196,19 @@ _Py_FindSourceFile(PyObject *filename, char* namebuf, size_t namelen, PyObject *
         strcpy(namebuf+len, tail);
 
         binary = PyObject_CallMethod(io, "open", "ss", namebuf, "rb");
-        if (binary != NULL)
-            return binary;
+        if (binary != NULL) {
+            result = binary;
+            goto finally;
+        }
         PyErr_Clear();
     }
-    return NULL;
+    goto error;
+
+error:
+    result = NULL;
+finally:
+    Py_DECREF(filebytes);
+    return result;
 }
 
 int