bpo-40334: Fix error location upon parsing an invalid string literal (GH-19962)

When parsing a string with an invalid escape, the old parser used to
point to the beginning of the invalid string. This commit changes the new
parser to match that behaviour, since it's currently pointing to the
end of the string (or to be more precise, to the beginning of the next
token).

diff --git a/Lib/test/test_cmd_line_script.py b/Lib/test/test_cmd_line_script.py
index 1fc9500738f3522cbb540e520e768a5ea6c094c7..171340581af228b05b75ca0ae2488ffe3bd9cf04 100644 (file)
--- a/Lib/test/test_cmd_line_script.py
+++ b/Lib/test/test_cmd_line_script.py
@@ -648,7 +648,7 @@ class CmdLineTest(unittest.TestCase):
             self.assertEqual(
                 stderr.splitlines()[-3:],
                 [   b'    foo = """\\q"""',
-                    b'                 ^',
+                    b'          ^',
                     b'SyntaxError: invalid escape sequence \\q'
                 ],
             )

diff --git a/Lib/test/test_string_literals.py b/Lib/test/test_string_literals.py
index 5b5477d14d467dd6168d165a9ca6783f4738d96b..9565ee2485afd10f1179e79911eff2b8c160552a 100644 (file)
--- a/Lib/test/test_string_literals.py
+++ b/Lib/test/test_string_literals.py
@@ -118,8 +118,7 @@ class TestLiterals(unittest.TestCase):
             eval("'''\n\\z'''")
         self.assertEqual(len(w), 1)
         self.assertEqual(w[0].filename, '<string>')
-        if use_old_parser():
-            self.assertEqual(w[0].lineno, 1)
+        self.assertEqual(w[0].lineno, 1)
 
         with warnings.catch_warnings(record=True) as w:
             warnings.simplefilter('error', category=DeprecationWarning)
@@ -128,8 +127,8 @@ class TestLiterals(unittest.TestCase):
             exc = cm.exception
         self.assertEqual(w, [])
         self.assertEqual(exc.filename, '<string>')
-        if use_old_parser():
-            self.assertEqual(exc.lineno, 1)
+        self.assertEqual(exc.lineno, 1)
+        self.assertEqual(exc.offset, 1)
 
     def test_eval_str_raw(self):
         self.assertEqual(eval(""" r'x' """), 'x')

diff --git a/Parser/pegen/parse_string.c b/Parser/pegen/parse_string.c
index d96303dc183fa79e579742d127d40f1685c77bf1..ca4b733c153b57b427cb73b2d3756725942d2dde 100644 (file)
--- a/Parser/pegen/parse_string.c
+++ b/Parser/pegen/parse_string.c
@@ -12,7 +12,7 @@
 // file (like "_PyPegen_raise_syntax_error").
 
 static int
-warn_invalid_escape_sequence(Parser *p, unsigned char first_invalid_escape_char)
+warn_invalid_escape_sequence(Parser *p, unsigned char first_invalid_escape_char, Token *t)
 {
     PyObject *msg =
         PyUnicode_FromFormat("invalid escape sequence \\%c", first_invalid_escape_char);
@@ -20,11 +20,16 @@ warn_invalid_escape_sequence(Parser *p, unsigned char first_invalid_escape_char)
         return -1;
     }
     if (PyErr_WarnExplicitObject(PyExc_DeprecationWarning, msg, p->tok->filename,
-                                 p->tok->lineno, NULL, NULL) < 0) {
+                                 t->lineno, NULL, NULL) < 0) {
         if (PyErr_ExceptionMatches(PyExc_DeprecationWarning)) {
             /* Replace the DeprecationWarning exception with a SyntaxError
                to get a more accurate error report */
             PyErr_Clear();
+
+            /* This is needed, in order for the SyntaxError to point to the token t,
+               since _PyPegen_raise_error uses p->tokens[p->fill - 1] for the
+               error location, if p->known_err_token is not set. */
+            p->known_err_token = t;
             RAISE_SYNTAX_ERROR("invalid escape sequence \\%c", first_invalid_escape_char);
         }
         Py_DECREF(msg);
@@ -47,7 +52,7 @@ decode_utf8(const char **sPtr, const char *end)
 }
 
 static PyObject *
-decode_unicode_with_escapes(Parser *parser, const char *s, size_t len)
+decode_unicode_with_escapes(Parser *parser, const char *s, size_t len, Token *t)
 {
     PyObject *v, *u;
     char *buf;
@@ -110,7 +115,7 @@ decode_unicode_with_escapes(Parser *parser, const char *s, size_t len)
     v = _PyUnicode_DecodeUnicodeEscape(s, len, NULL, &first_invalid_escape);
 
     if (v != NULL && first_invalid_escape != NULL) {
-        if (warn_invalid_escape_sequence(parser, *first_invalid_escape) < 0) {
+        if (warn_invalid_escape_sequence(parser, *first_invalid_escape, t) < 0) {
             /* We have not decref u before because first_invalid_escape points
                inside u. */
             Py_XDECREF(u);
@@ -123,7 +128,7 @@ decode_unicode_with_escapes(Parser *parser, const char *s, size_t len)
 }
 
 static PyObject *
-decode_bytes_with_escapes(Parser *p, const char *s, Py_ssize_t len)
+decode_bytes_with_escapes(Parser *p, const char *s, Py_ssize_t len, Token *t)
 {
     const char *first_invalid_escape;
     PyObject *result = _PyBytes_DecodeEscape(s, len, NULL, &first_invalid_escape);
@@ -132,7 +137,7 @@ decode_bytes_with_escapes(Parser *p, const char *s, Py_ssize_t len)
     }
 
     if (first_invalid_escape != NULL) {
-        if (warn_invalid_escape_sequence(p, *first_invalid_escape) < 0) {
+        if (warn_invalid_escape_sequence(p, *first_invalid_escape, t) < 0) {
             Py_DECREF(result);
             return NULL;
         }
@@ -146,9 +151,14 @@ decode_bytes_with_escapes(Parser *p, const char *s, Py_ssize_t len)
    If the string is an f-string, set *fstr and *fstrlen to the unparsed
    string object.  Return 0 if no errors occurred.  */
 int
-_PyPegen_parsestr(Parser *p, const char *s, int *bytesmode, int *rawmode, PyObject **result,
-         const char **fstr, Py_ssize_t *fstrlen)
+_PyPegen_parsestr(Parser *p, int *bytesmode, int *rawmode, PyObject **result,
+                  const char **fstr, Py_ssize_t *fstrlen, Token *t)
 {
+    const char *s = PyBytes_AsString(t->bytes);
+    if (s == NULL) {
+        return -1;
+    }
+
     size_t len;
     int quote = Py_CHARMASK(*s);
     int fmode = 0;
@@ -245,7 +255,7 @@ _PyPegen_parsestr(Parser *p, const char *s, int *bytesmode, int *rawmode, PyObje
             *result = PyBytes_FromStringAndSize(s, len);
         }
         else {
-            *result = decode_bytes_with_escapes(p, s, len);
+            *result = decode_bytes_with_escapes(p, s, len, t);
         }
     }
     else {
@@ -253,7 +263,7 @@ _PyPegen_parsestr(Parser *p, const char *s, int *bytesmode, int *rawmode, PyObje
             *result = PyUnicode_DecodeUTF8Stateful(s, len, NULL, NULL);
         }
         else {
-            *result = decode_unicode_with_escapes(p, s, len);
+            *result = decode_unicode_with_escapes(p, s, len, t);
         }
     }
     return *result == NULL ? -1 : 0;
@@ -637,7 +647,7 @@ exit:
 */
 static int
 fstring_find_literal(Parser *p, const char **str, const char *end, int raw,
-                     PyObject **literal, int recurse_lvl)
+                     PyObject **literal, int recurse_lvl, Token *t)
 {
     /* Get any literal string. It ends when we hit an un-doubled left
        brace (which isn't part of a unicode name escape such as
@@ -660,7 +670,7 @@ fstring_find_literal(Parser *p, const char **str, const char *end, int raw,
                 }
                 break;
             }
-            if (ch == '{' && warn_invalid_escape_sequence(p, ch) < 0) {
+            if (ch == '{' && warn_invalid_escape_sequence(p, ch, t) < 0) {
                 return -1;
             }
         }
@@ -704,7 +714,7 @@ done:
                                                     NULL, NULL);
         else
             *literal = decode_unicode_with_escapes(p, literal_start,
-                                                   s - literal_start);
+                                                   s - literal_start, t);
         if (!*literal)
             return -1;
     }
@@ -1041,7 +1051,7 @@ fstring_find_literal_and_expr(Parser *p, const char **str, const char *end, int
     assert(*literal == NULL && *expression == NULL);
 
     /* Get any literal string. */
-    result = fstring_find_literal(p, str, end, raw, literal, recurse_lvl);
+    result = fstring_find_literal(p, str, end, raw, literal, recurse_lvl, t);
     if (result < 0)
         goto error;
 

diff --git a/Parser/pegen/parse_string.h b/Parser/pegen/parse_string.h
index 4f2aa94fc19b057660063106bad8272a400b82fd..cd85bd57d0a383308eb2d012a3eb84358997adf5 100644 (file)
--- a/Parser/pegen/parse_string.h
+++ b/Parser/pegen/parse_string.h
@@ -34,8 +34,8 @@ typedef struct {
 } FstringParser;
 
 void _PyPegen_FstringParser_Init(FstringParser *);
-int _PyPegen_parsestr(Parser *, const char *, int *, int *, PyObject **,
-             const char **, Py_ssize_t *);
+int _PyPegen_parsestr(Parser *, int *, int *, PyObject **,
+                      const char **, Py_ssize_t *, Token *);
 int _PyPegen_FstringParser_ConcatFstring(Parser *, FstringParser *, const char **,
                                 const char *, int, int, Token *, Token *,
                                 Token *);

diff --git a/Parser/pegen/pegen.c b/Parser/pegen/pegen.c
index c311593af70f58ea1a2d0c4a8c131038b74ddb35..06af53b3597f7438a4d0d99c5a7e60a39222dbc0 100644 (file)
--- a/Parser/pegen/pegen.c
+++ b/Parser/pegen/pegen.c
@@ -383,7 +383,7 @@ _PyPegen_raise_error(Parser *p, PyObject *errtype, int with_col_number, const ch
     PyObject *errstr = NULL;
     PyObject *loc = NULL;
     PyObject *tmp = NULL;
-    Token *t = p->tokens[p->fill - 1];
+    Token *t = p->known_err_token != NULL ? p->known_err_token : p->tokens[p->fill - 1];
     Py_ssize_t col_number = !with_col_number;
     va_list va;
     p->error_indicator = 1;
@@ -1053,6 +1053,7 @@ _PyPegen_Parser_New(struct tok_state *tok, int start_rule, int flags,
     p->starting_col_offset = 0;
     p->flags = flags;
     p->feature_version = feature_version;
+    p->known_err_token = NULL;
 
     return p;
 }
@@ -1972,12 +1973,7 @@ _PyPegen_concatenate_strings(Parser *p, asdl_seq *strings)
         const char *fstr;
         Py_ssize_t fstrlen = -1;
 
-        char *this_str = PyBytes_AsString(t->bytes);
-        if (!this_str) {
-            goto error;
-        }
-
-        if (_PyPegen_parsestr(p, this_str, &this_bytesmode, &this_rawmode, &s, &fstr, &fstrlen) != 0) {
+        if (_PyPegen_parsestr(p, &this_bytesmode, &this_rawmode, &s, &fstr, &fstrlen, t) != 0) {
             goto error;
         }
 

diff --git a/Parser/pegen/pegen.h b/Parser/pegen/pegen.h
index cbe6f197ac742366318c19f098cebdd1101ffc82..ffb18e47e4a9a8e87ceccae80713dcc678e1a2df 100644 (file)
--- a/Parser/pegen/pegen.h
+++ b/Parser/pegen/pegen.h
@@ -71,6 +71,7 @@ typedef struct {
     int flags;
     int feature_version;
     growable_comment_array type_ignore_comments;
+    Token *known_err_token;
 } Parser;
 
 typedef struct {