Issue #10785: Store the filename as Unicode in the Python parser.

9 files changed, 94 insertions, 42 deletions

diff --git a/Include/parsetok.h b/Include/parsetok.h
index 4b7694f..911dfc1 100644
--- a/Include/parsetok.h
+++ b/Include/parsetok.h
@@ -9,7 +9,10 @@ extern "C" {
 
 typedef struct {
     int error;
-    const char *filename;       /* decoded from the filesystem encoding */
+#ifndef PGEN
+    /* The filename is useless for pgen, see comment in tok_state structure */
+    PyObject *filename;
+#endif
     int lineno;
     int offset;
     char *text;                 /* UTF-8-encoded string */
@@ -66,8 +69,10 @@ PyAPI_FUNC(node *) PyParser_ParseStringFlagsFilenameEx(
     perrdetail *err_ret,
     int *flags);
 
-/* Note that he following function is defined in pythonrun.c not parsetok.c. */
+/* Note that the following functions are defined in pythonrun.c,
+   not in parsetok.c */
 PyAPI_FUNC(void) PyParser_SetError(perrdetail *);
+PyAPI_FUNC(void) PyParser_ClearError(perrdetail *);
 
 #ifdef __cplusplus
 }
diff --git a/Makefile.pre.in b/Makefile.pre.in
index c3004ce..6a206c2 100644
--- a/Makefile.pre.in
+++ b/Makefile.pre.in
@@ -238,14 +238,13 @@ POBJS=		\
 		Parser/listnode.o \
 		Parser/node.o \
 		Parser/parser.o \
-		Parser/parsetok.o \
 		Parser/bitset.o \
 		Parser/metagrammar.o \
 		Parser/firstsets.o \
 		Parser/grammar.o \
 		Parser/pgen.o
 
-PARSER_OBJS=	$(POBJS) Parser/myreadline.o Parser/tokenizer.o
+PARSER_OBJS=	$(POBJS) Parser/myreadline.o Parser/parsetok.o Parser/tokenizer.o
 
 PGOBJS=		\
 		Objects/obmalloc.o \
@@ -254,10 +253,12 @@ PGOBJS=		\
 		Python/pyctype.o \
 		Parser/tokenizer_pgen.o \
 		Parser/printgrammar.o \
+		Parser/parsetok_pgen.o \
 		Parser/pgenmain.o
 
 PARSER_HEADERS= \
 		Parser/parser.h \
+		Include/parsetok.h \
 		Parser/tokenizer.h
 
 PGENOBJS=	$(PGENMAIN) $(POBJS) $(PGOBJS)
@@ -593,6 +594,7 @@ Parser/grammar.o:	$(srcdir)/Parser/grammar.c \
 Parser/metagrammar.o:	$(srcdir)/Parser/metagrammar.c
 
 Parser/tokenizer_pgen.o:	$(srcdir)/Parser/tokenizer.c
+Parser/parsetok_pgen.o:	$(srcdir)/Parser/parsetok.c
 
 Parser/pgenmain.o:	$(srcdir)/Include/parsetok.h
 
@@ -700,7 +702,6 @@ PYTHON_HEADERS= \
 		Include/objimpl.h \
 		Include/opcode.h \
 		Include/osdefs.h \
-		Include/parsetok.h \
 		Include/patchlevel.h \
 		Include/pgen.h \
 		Include/pgenheaders.h \
diff --git a/Misc/NEWS b/Misc/NEWS
index d4297d2..30d7c50 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,8 @@ What's New in Python 3.3 Alpha 1?
 Core and Builtins
 -----------------
 
+- Issue #10785: Store the filename as Unicode in the Python parser.
+
 - Issue #11619: _PyImport_LoadDynamicModule() doesn't encode the path to bytes
   on Windows.
 
diff --git a/Modules/parsermodule.c b/Modules/parsermodule.c
index 3cdf135..e5b4e55 100644
--- a/Modules/parsermodule.c
+++ b/Modules/parsermodule.c
@@ -584,6 +584,7 @@ parser_do_parse(PyObject *args, PyObject *kw, char *argspec, int type)
         else
             PyParser_SetError(&err);
     }
+    PyParser_ClearError(&err);
     return (res);
 }
 
diff --git a/Parser/parsetok.c b/Parser/parsetok.c
index 2251cac..eef650a 100644
--- a/Parser/parsetok.c
+++ b/Parser/parsetok.c
@@ -13,7 +13,7 @@
 
 /* Forward */
 static node *parsetok(struct tok_state *, grammar *, int, perrdetail *, int *);
-static void initerr(perrdetail *err_ret, const char* filename);
+static int initerr(perrdetail *err_ret, const char* filename);
 
 /* Parse input coming from a string.  Return error code, print some errors. */
 node *
@@ -48,7 +48,8 @@ PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename,
     struct tok_state *tok;
     int exec_input = start == file_input;
 
-    initerr(err_ret, filename);
+    if (initerr(err_ret, filename) < 0)
+        return NULL;
 
     if (*flags & PyPARSE_IGNORE_COOKIE)
         tok = PyTokenizer_FromUTF8(s, exec_input);
@@ -59,7 +60,10 @@ PyParser_ParseStringFlagsFilenameEx(const char *s, const char *filename,
         return NULL;
     }
 
-    tok->filename = filename ? filename : "<string>";
+#ifndef PGEN
+    Py_INCREF(err_ret->filename);
+    tok->filename = err_ret->filename;
+#endif
     return parsetok(tok, g, start, err_ret, flags);
 }
 
@@ -90,13 +94,17 @@ PyParser_ParseFileFlagsEx(FILE *fp, const char *filename,
 {
     struct tok_state *tok;
 
-    initerr(err_ret, filename);
+    if (initerr(err_ret, filename) < 0)
+        return NULL;
 
     if ((tok = PyTokenizer_FromFile(fp, (char *)enc, ps1, ps2)) == NULL) {
         err_ret->error = E_NOMEM;
         return NULL;
     }
-    tok->filename = filename;
+#ifndef PGEN
+    Py_INCREF(err_ret->filename);
+    tok->filename = err_ret->filename;
+#endif
     return parsetok(tok, g, start, err_ret, flags);
 }
 
@@ -267,14 +275,24 @@ done:
     return n;
 }
 
-static void
+static int
 initerr(perrdetail *err_ret, const char *filename)
 {
     err_ret->error = E_OK;
-    err_ret->filename = filename;
     err_ret->lineno = 0;
     err_ret->offset = 0;
     err_ret->text = NULL;
     err_ret->token = -1;
     err_ret->expected = -1;
+#ifndef PGEN
+    if (filename)
+        err_ret->filename = PyUnicode_DecodeFSDefault(filename);
+    else
+        err_ret->filename = PyUnicode_FromString("<string>");
+    if (err_ret->filename == NULL) {
+        err_ret->error = E_ERROR;
+        return -1;
+    }
+#endif
+    return 0;
 }
diff --git a/Parser/parsetok_pgen.c b/Parser/parsetok_pgen.c
new file mode 100644
index 0000000..97b9288
--- /dev/null
+++ b/Parser/parsetok_pgen.c
@@ -0,0 +1,2 @@
+#define PGEN
+#include "parsetok.c"
diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index 3f6be2f..5edd958 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -128,7 +128,6 @@ tok_new(void)
     tok->prompt = tok->nextprompt = NULL;
     tok->lineno = 0;
     tok->level = 0;
-    tok->filename = NULL;
     tok->altwarning = 1;
     tok->alterror = 1;
     tok->alttabsize = 1;
@@ -140,6 +139,7 @@ tok_new(void)
     tok->encoding = NULL;
     tok->cont_line = 0;
 #ifndef PGEN
+    tok->filename = NULL;
     tok->decoding_readline = NULL;
     tok->decoding_buffer = NULL;
 #endif
@@ -545,7 +545,6 @@ decoding_fgets(char *s, int size, struct tok_state *tok)
 {
     char *line = NULL;
     int badchar = 0;
-    PyObject *filename;
     for (;;) {
         if (tok->decoding_state == STATE_NORMAL) {
             /* We already have a codec associated with
@@ -586,16 +585,12 @@ decoding_fgets(char *s, int size, struct tok_state *tok)
     if (badchar) {
         /* Need to add 1 to the line number, since this line
            has not been counted, yet.  */
-        filename = PyUnicode_DecodeFSDefault(tok->filename);
-        if (filename != NULL) {
-            PyErr_Format(PyExc_SyntaxError,
-                    "Non-UTF-8 code starting with '\\x%.2x' "
-                    "in file %U on line %i, "
-                    "but no encoding declared; "
-                    "see http://python.org/dev/peps/pep-0263/ for details",
-                    badchar, filename, tok->lineno + 1);
-            Py_DECREF(filename);
-        }
+        PyErr_Format(PyExc_SyntaxError,
+                "Non-UTF-8 code starting with '\\x%.2x' "
+                "in file %U on line %i, "
+                "but no encoding declared; "
+                "see http://python.org/dev/peps/pep-0263/ for details",
+                badchar, tok->filename, tok->lineno + 1);
         return error_ret(tok);
     }
 #endif
@@ -853,6 +848,7 @@ PyTokenizer_Free(struct tok_state *tok)
 #ifndef PGEN
     Py_XDECREF(tok->decoding_readline);
     Py_XDECREF(tok->decoding_buffer);
+    Py_XDECREF(tok->filename);
 #endif
     if (tok->fp != NULL && tok->buf != NULL)
         PyMem_FREE(tok->buf);
@@ -1247,8 +1243,13 @@ indenterror(struct tok_state *tok)
         return 1;
     }
     if (tok->altwarning) {
-        PySys_WriteStderr("%s: inconsistent use of tabs and spaces "
+#ifdef PGEN
+        PySys_WriteStderr("inconsistent use of tabs and spaces "
+                          "in indentation\n");
+#else
+        PySys_FormatStderr("%U: inconsistent use of tabs and spaces "
                           "in indentation\n", tok->filename);
+#endif
         tok->altwarning = 0;
     }
     return 0;
@@ -1718,6 +1719,11 @@ PyTokenizer_FindEncoding(int fd)
         fclose(fp);
         return NULL;
     }
+#ifndef PGEN
+    tok->filename = PyUnicode_FromString("<string>");
+    if (tok->filename == NULL)
+        goto error;
+#endif
     while (tok->lineno < 2 && tok->done == E_OK) {
         PyTokenizer_Get(tok, &p_start, &p_end);
     }
@@ -1727,6 +1733,9 @@ PyTokenizer_FindEncoding(int fd)
         if (encoding)
         strcpy(encoding, tok->encoding);
     }
+#ifndef PGEN
+error:
+#endif
     PyTokenizer_Free(tok);
     return encoding;
 }
diff --git a/Parser/tokenizer.h b/Parser/tokenizer.h
index 2be3bf2..3a0d3cb 100644
--- a/Parser/tokenizer.h
+++ b/Parser/tokenizer.h
@@ -40,7 +40,13 @@ 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;   /* encoded to the filesystem encoding */
+#ifndef PGEN
+    /* pgen doesn't have access to Python codecs, it cannot decode the input
+       filename. The bytes filename might be kept, but it is only used by
+       indenterror() and it is not really needed: pgen only compiles one file
+       (Grammar/Grammar). */
+    PyObject *filename;
+#endif
     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 */
diff --git a/Python/pythonrun.c b/Python/pythonrun.c
index 1c36e63..a6787c4 100644
--- a/Python/pythonrun.c
+++ b/Python/pythonrun.c
@@ -62,6 +62,7 @@ static PyObject *run_mod(mod_ty, const char *, PyObject *, PyObject *,
 static PyObject *run_pyc_file(FILE *, const char *, PyObject *, PyObject *,
                               PyCompilerFlags *);
 static void err_input(perrdetail *);
+static void err_free(perrdetail *);
 static void initsigs(void);
 static void call_py_exitfuncs(void);
 static void wait_for_thread_shutdown(void);
@@ -1887,12 +1888,13 @@ PyParser_ASTFromString(const char *s, const char *filename, int start,
         flags->cf_flags |= iflags & PyCF_MASK;
         mod = PyAST_FromNode(n, flags, filename, arena);
         PyNode_Free(n);
-        return mod;
     }
     else {
         err_input(&err);
-        return NULL;
+        mod = NULL;
     }
+    err_free(&err);
+    return mod;
 }
 
 mod_ty
@@ -1917,14 +1919,15 @@ PyParser_ASTFromFile(FILE *fp, const char *filename, const char* enc,
         flags->cf_flags |= iflags & PyCF_MASK;
         mod = PyAST_FromNode(n, flags, filename, arena);
         PyNode_Free(n);
-        return mod;
     }
     else {
         err_input(&err);
         if (errcode)
             *errcode = err.error;
-        return NULL;
+        mod = NULL;
     }
+    err_free(&err);
+    return mod;
 }
 
 /* Simplified interface to parsefile -- return node or set exception */
@@ -1938,6 +1941,7 @@ PyParser_SimpleParseFileFlags(FILE *fp, const char *filename, int start, int fla
                                       start, NULL, NULL, &err, flags);
     if (n == NULL)
         err_input(&err);
+    err_free(&err);
 
     return n;
 }
@@ -1952,6 +1956,7 @@ PyParser_SimpleParseStringFlags(const char *str, int start, int flags)
                                         start, &err, flags);
     if (n == NULL)
         err_input(&err);
+    err_free(&err);
     return n;
 }
 
@@ -1964,6 +1969,7 @@ PyParser_SimpleParseStringFlagsFilename(const char *str, const char *filename,
                             &_PyParser_Grammar, start, &err, flags);
     if (n == NULL)
         err_input(&err);
+    err_free(&err);
     return n;
 }
 
@@ -1977,11 +1983,23 @@ PyParser_SimpleParseStringFilename(const char *str, const char *filename, int st
    even parser modules. */
 
 void
+PyParser_ClearError(perrdetail *err)
+{
+    err_free(err);
+}
+
+void
 PyParser_SetError(perrdetail *err)
 {
     err_input(err);
 }
 
+static void
+err_free(perrdetail *err)
+{
+    Py_CLEAR(err->filename);
+}
+
 /* Set the error appropriate to the given input error code (see errcode.h) */
 
 static void
@@ -1989,7 +2007,6 @@ err_input(perrdetail *err)
 {
     PyObject *v, *w, *errtype, *errtext;
     PyObject *msg_obj = NULL;
-    PyObject *filename;
     char *msg = NULL;
 
     errtype = PyExc_SyntaxError;
@@ -2075,17 +2092,8 @@ err_input(perrdetail *err)
         errtext = PyUnicode_DecodeUTF8(err->text, strlen(err->text),
                                        "replace");
     }
-    if (err->filename != NULL)
-        filename = PyUnicode_DecodeFSDefault(err->filename);
-    else {
-        Py_INCREF(Py_None);
-        filename = Py_None;
-    }
-    if (filename != NULL)
-        v = Py_BuildValue("(NiiN)", filename,
-                          err->lineno, err->offset, errtext);
-    else
-        v = NULL;
+    v = Py_BuildValue("(OiiN)", err->filename,
+                      err->lineno, err->offset, errtext);
     if (v != NULL) {
         if (msg_obj)
             w = Py_BuildValue("(OO)", msg_obj, v);