gh-84489: Properly handle trailing spaces in Py_BuildValue() format strings (GH-21158)

The docs state that the space, tab, colon, and comma characters are
ignored in Py_BuildValue() format strings.

Co-authored-by: Serhiy Storchaka <storchaka@gmail.com>

Lib/test/test_capi/test_misc.py

@@ -319,6 +319,7 @@ class CAPITest(unittest.TestCase):
         self.assertEqual(buildvalue('((OO))', 1, 2), ((1, 2),))
 
         self.assertEqual(buildvalue(' \t,:'), None)
+        self.assertEqual(buildvalue('O,', 1), 1)
         self.assertEqual(buildvalue('   O   ', 1), 1)
         self.assertEqual(buildvalue('\tO\t', 1), 1)
         self.assertEqual(buildvalue('O,O', 1, 2), (1, 2))
@@ -327,17 +328,18 @@ class CAPITest(unittest.TestCase):
         self.assertEqual(buildvalue('O O', 1, 2), (1, 2))
         self.assertEqual(buildvalue('O\tO', 1, 2), (1, 2))
         self.assertEqual(buildvalue('(O,O)', 1, 2), (1, 2))
-        self.assertEqual(buildvalue('(O, O)', 1, 2), (1, 2))
+        self.assertEqual(buildvalue('(O, O,)', 1, 2), (1, 2))
-        self.assertEqual(buildvalue(' ( O O) ', 1, 2), (1, 2))
+        self.assertEqual(buildvalue(' ( O O ) ', 1, 2), (1, 2))
-        self.assertEqual(buildvalue('\t(\tO\tO)\t', 1, 2), (1, 2))
+        self.assertEqual(buildvalue('\t(\tO\tO\t)\t', 1, 2), (1, 2))
         self.assertEqual(buildvalue('[O,O]', 1, 2), [1, 2])
-        self.assertEqual(buildvalue('[O, O]', 1, 2), [1, 2])
+        self.assertEqual(buildvalue('[O, O,]', 1, 2), [1, 2])
-        self.assertEqual(buildvalue(' [ O O] ', 1, 2), [1, 2])
+        self.assertEqual(buildvalue(' [ O O ] ', 1, 2), [1, 2])
+        self.assertEqual(buildvalue(' [\tO\tO\t] ', 1, 2), [1, 2])
         self.assertEqual(buildvalue('{O:O}', 1, 2), {1: 2})
         self.assertEqual(buildvalue('{O:O,O:O}', 1, 2, 3, 4), {1: 2, 3: 4})
-        self.assertEqual(buildvalue('{O: O, O: O}', 1, 2, 3, 4), {1: 2, 3: 4})
+        self.assertEqual(buildvalue('{O: O, O: O,}', 1, 2, 3, 4), {1: 2, 3: 4})
-        self.assertEqual(buildvalue(' { O O O O} ', 1, 2, 3, 4), {1: 2, 3: 4})
+        self.assertEqual(buildvalue(' { O O O O } ', 1, 2, 3, 4), {1: 2, 3: 4})
-        self.assertEqual(buildvalue('\t{\tO\tO\tO\tO}\t', 1, 2, 3, 4), {1: 2, 3: 4})
+        self.assertEqual(buildvalue('\t{\tO\tO\tO\tO\t}\t', 1, 2, 3, 4), {1: 2, 3: 4})
 
         self.assertRaises(SystemError, buildvalue, 'O', NULL)
         self.assertRaises(SystemError, buildvalue, '(O)', NULL)
@@ -378,6 +380,12 @@ class CAPITest(unittest.TestCase):
         self.assertEqual(buildvalue('C', sys.maxunicode), chr(sys.maxunicode))
         self.assertRaises(ValueError, buildvalue, 'C', -1)
         self.assertRaises(ValueError, buildvalue, 'C', sys.maxunicode+1)
+
+        # gh-84489
+        self.assertRaises(ValueError, buildvalue, '(C )i', -1, 2)
+        self.assertRaises(ValueError, buildvalue, '[C ]i', -1, 2)
+        self.assertRaises(ValueError, buildvalue, '{Ci }i', -1, 2, 3)
+
     def test_buildvalue_N(self):
         _testcapi.test_buildvalue_N()
 

Misc/NEWS.d/next/C API/2020-06-25-09-44-59.bpo-40309.CuoGoQ.rst

@@ -0,0 +1 @@
+Properly handle trailing spaces before closing parenthesis in :c:func:`Py_BuildValue` format strings.

Python/modsupport.c

@@ -88,6 +88,24 @@ static PyObject *do_mklist(const char**, va_list *, char, Py_ssize_t);
 static PyObject *do_mkdict(const char**, va_list *, char, Py_ssize_t);
 static PyObject *do_mkvalue(const char**, va_list *);
 
+static int
+check_end(const char **p_format, char endchar)
+{
+    const char *f = *p_format;
+    while (*f != endchar) {
+        if (*f != ' ' && *f != '\t' && *f != ',' && *f != ':') {
+            PyErr_SetString(PyExc_SystemError,
+                            "Unmatched paren in format");
+            return 0;
+        }
+        f++;
+    }
+    if (endchar) {
+        f++;
+    }
+    *p_format = f;
+    return 1;
+}
 
 static void
 do_ignore(const char **p_format, va_list *p_va, char endchar, Py_ssize_t n)
@@ -108,14 +126,9 @@ do_ignore(const char **p_format, va_list *p_va, char endchar, Py_ssize_t n)
         }
     }
     Py_XDECREF(v);
-    if (**p_format != endchar) {
+    if (!check_end(p_format, endchar)) {
-        PyErr_SetString(PyExc_SystemError,
-                        "Unmatched paren in format");
         return;
     }
-    if (endchar) {
-        ++*p_format;
-    }
 }
 
 static PyObject *
@@ -157,14 +170,10 @@ do_mkdict(const char **p_format, va_list *p_va, char endchar, Py_ssize_t n)
         Py_DECREF(k);
         Py_DECREF(v);
     }
-    if (**p_format != endchar) {
+    if (!check_end(p_format, endchar)) {
         Py_DECREF(d);
-        PyErr_SetString(PyExc_SystemError,
-                        "Unmatched paren in format");
         return NULL;
     }
-    if (endchar)
-        ++*p_format;
     return d;
 }
 
@@ -191,14 +200,10 @@ do_mklist(const char **p_format, va_list *p_va, char endchar, Py_ssize_t n)
         }
         PyList_SET_ITEM(v, i, w);
     }
-    if (**p_format != endchar) {
+    if (!check_end(p_format, endchar)) {
         Py_DECREF(v);
-        PyErr_SetString(PyExc_SystemError,
-                        "Unmatched paren in format");
         return NULL;
     }
-    if (endchar)
-        ++*p_format;
     return v;
 }
 
@@ -221,14 +226,9 @@ do_mkstack(PyObject **stack, const char **p_format, va_list *p_va,
         }
         stack[i] = w;
     }
-    if (**p_format != endchar) {
+    if (!check_end(p_format, endchar)) {
-        PyErr_SetString(PyExc_SystemError,
-                        "Unmatched paren in format");
         goto error;
     }
-    if (endchar) {
-        ++*p_format;
-    }
     return 0;
 
 error:
@@ -261,14 +261,10 @@ do_mktuple(const char **p_format, va_list *p_va, char endchar, Py_ssize_t n)
         }
         PyTuple_SET_ITEM(v, i, w);
     }
-    if (**p_format != endchar) {
+    if (!check_end(p_format, endchar)) {
         Py_DECREF(v);
-        PyErr_SetString(PyExc_SystemError,
-                        "Unmatched paren in format");
         return NULL;
     }
-    if (endchar)
-        ++*p_format;
     return v;
 }