gh-101703: use PyOS_snprintf instead of sprintf

Modules/_ctypes/_ctypes.c

@@ -392,9 +392,9 @@ _ctypes_alloc_format_string_with_shape(int ndim, const Py_ssize_t *shape,
         strcat(new_prefix, "(");
         for (k = 0; k < ndim; ++k) {
             if (k < ndim-1) {
-                sprintf(buf, "%zd,", shape[k]);
+                PyOS_snprintf(buf, sizeof(buf), "%zd,", shape[k]);
             } else {
-                sprintf(buf, "%zd)", shape[k]);
+                PyOS_snprintf(buf, sizeof(buf), "%zd)", shape[k]);
             }
             strcat(new_prefix, buf);
         }
@@ -2639,7 +2639,8 @@ unique_key(CDataObject *target, Py_ssize_t index)
     size_t bytes_left;
 
     Py_BUILD_ASSERT(sizeof(string) - 1 > sizeof(Py_ssize_t) * 2);
-    cp += sprintf(cp, "%x", Py_SAFE_DOWNCAST(index, Py_ssize_t, int));
+    cp += PyOS_snprintf(cp, sizeof(string), "%x",
+                        Py_SAFE_DOWNCAST(index, Py_ssize_t, int));
     while (target->b_base) {
         bytes_left = sizeof(string) - (cp - string) - 1;
         /* Hex format needs 2 characters per byte */
@@ -2648,7 +2649,8 @@ unique_key(CDataObject *target, Py_ssize_t index)
                             "ctypes object structure too deep");
             return NULL;
         }
-        cp += sprintf(cp, ":%x", Py_SAFE_DOWNCAST(target->b_index, Py_ssize_t, int));
+        cp += PyOS_snprintf(cp, bytes_left, ":%x",
+                            Py_SAFE_DOWNCAST(target->b_index, Py_ssize_t, int));
         target = target->b_base;
     }
     return PyUnicode_FromStringAndSize(string, cp-string);
@@ -3360,11 +3362,12 @@ static PPROC FindAddress(void *handle, const char *name, PyObject *type)
        funcname -> _funcname@<n>
        where n is 0, 4, 8, 12, ..., 128
      */
-    mangled_name = alloca(strlen(name) + 1 + 1 + 1 + 3); /* \0 _ @ %d */
+    int toalloc = strlen(name) + 1 + 1 + 1 + 3; /* \0 _ @ %d */
+    mangled_name = alloca(toalloc);
     if (!mangled_name)
         return NULL;
     for (i = 0; i < 32; ++i) {
-        sprintf(mangled_name, "_%s@%d", name, i*4);
+        PyOS_snprintf(mangled_name, toalloc, "_%s@%d", name, i*4);
         Py_BEGIN_ALLOW_THREADS
         address = (PPROC)GetProcAddress(handle, mangled_name);
         Py_END_ALLOW_THREADS
@@ -4851,10 +4854,10 @@ PyCArrayType_from_ctype(PyObject *itemtype, Py_ssize_t length)
         return NULL;
     }
 #ifdef MS_WIN64
-    sprintf(name, "%.200s_Array_%Id",
+    PyOS_snprintf(name, sizeof(name), "%.200s_Array_%Id",
         ((PyTypeObject *)itemtype)->tp_name, length);
 #else
-    sprintf(name, "%.200s_Array_%ld",
+    PyOS_snprintf(name, sizeof(name), "%.200s_Array_%ld",
         ((PyTypeObject *)itemtype)->tp_name, (long)length);
 #endif
 

Modules/_ctypes/stgdict.c

@@ -612,7 +612,7 @@ PyCStructUnionType_update_stgdict(PyObject *type, PyObject *fields, int isStruct
                 PyErr_NoMemory();
                 return -1;
             }
-            sprintf(buf, "%s:%s:", fieldfmt, fieldname);
+            PyOS_snprintf(buf, len + 2 + 1, "%s:%s:", fieldfmt, fieldname);
 
             ptr = stgdict->format;
             if (dict->shape != NULL) {

Modules/_datetimemodule.c

@@ -1559,11 +1559,13 @@ make_freplacement(PyObject *object)
 {
     char freplacement[64];
     if (PyTime_Check(object))
-        sprintf(freplacement, "%06d", TIME_GET_MICROSECOND(object));
+        PyOS_snprintf(freplacement, sizeof(freplacement),
+                      "%06d", TIME_GET_MICROSECOND(object));
     else if (PyDateTime_Check(object))
-        sprintf(freplacement, "%06d", DATE_GET_MICROSECOND(object));
+        PyOS_snprintf(freplacement, sizeof(freplacement),
+                      "%06d", DATE_GET_MICROSECOND(object));
     else
-        sprintf(freplacement, "%06d", 0);
+        PyOS_snprintf(freplacement, sizeof(freplacement), "%06d", 0);
 
     return PyBytes_FromStringAndSize(freplacement, strlen(freplacement));
 }

Modules/_pickle.c

@@ -2098,7 +2098,7 @@ save_long(PicklerObject *self, PyObject *obj)
             }
         }
         else {
-            sprintf(pdata, "%c%ld\n", INT,  val);
+            PyOS_snprintf(pdata, sizeof(pdata), "%c%ld\n", INT,  val);
             len = strlen(pdata);
         }
         if (_Pickler_Write(self, pdata, len) < 0)

Modules/_ssl.c

@@ -1324,8 +1324,9 @@ _get_peer_alt_names (_sslmodulestate *state, X509 *certificate) {
                 } else if (name->d.ip->length == 16) {
                     /* PyUnicode_FromFormat() does not support %X */
                     unsigned char *p = name->d.ip->data;
-                    len = sprintf(
+                    len = PyOS_snprintf(
                         buf,
+                        sizeof(buf),
                         "%X:%X:%X:%X:%X:%X:%X:%X",
                         p[0] << 8 | p[1],
                         p[2] << 8 | p[3],

Modules/_testcapimodule.c

@@ -1097,7 +1097,8 @@ test_capsule(PyObject *self, PyObject *Py_UNUSED(ignored))
 #undef FAIL
 #define FAIL(x) \
         { \
-        sprintf(buffer, "%s module: \"%s\" attribute: \"%s\"", \
+        PyOS_snprintf(buffer, sizeof(buffer), \
+            "%s module: \"%s\" attribute: \"%s\"", \
             x, known->module, known->attribute); \
         error = buffer; \
         goto exit; \

Modules/getnameinfo.c

@@ -139,7 +139,7 @@ getnameinfo(sa, salen, host, hostlen, serv, servlen, flags)
     if (serv == NULL || servlen == 0) {
         /* what we should do? */
     } else if (flags & NI_NUMERICSERV) {
-        sprintf(numserv, "%d", ntohs(port));
+        PyOS_snprintf(numserv, sizeof(numserv), "%d", ntohs(port));
         if (strlen(numserv) > servlen)
             return ENI_MEMORY;
         strcpy(serv, numserv);

Modules/socketmodule.c

@@ -1302,11 +1302,11 @@ makebdaddr(bdaddr_t *bdaddr)
         octets[i] = ((*bdaddr) >> (8 * i)) & 0xFF;
     }
 
-    sprintf(buf, "%02X:%02X:%02X:%02X:%02X:%02X",
+    PyOS_snprintf(buf, sizeof(buf), "%02X:%02X:%02X:%02X:%02X:%02X",
         octets[5], octets[4], octets[3],
         octets[2], octets[1], octets[0]);
 #else
-    sprintf(buf, "%02X:%02X:%02X:%02X:%02X:%02X",
+    PyOS_snprintf(buf, sizeof(buf), "%02X:%02X:%02X:%02X:%02X:%02X",
         bdaddr->b[5], bdaddr->b[4], bdaddr->b[3],
         bdaddr->b[2], bdaddr->b[1], bdaddr->b[0]);
 #endif

Modules/unicodedata.c

@@ -1108,7 +1108,7 @@ _getucname(PyObject *self,
         if (buflen < 28)
             /* Worst case: CJK UNIFIED IDEOGRAPH-20000 */
             return 0;
-        sprintf(buffer, "CJK UNIFIED IDEOGRAPH-%X", code);
+        PyOS_snprintf(buffer, buflen, "CJK UNIFIED IDEOGRAPH-%X", code);
         return 1;
     }
 

Objects/bytesobject.c

@@ -270,40 +270,45 @@ PyBytes_FromFormatV(const char *format, va_list vargs)
 
         case 'd':
             if (longflag) {
-                sprintf(buffer, "%ld", va_arg(vargs, long));
+                PyOS_snprintf(buffer, sizeof(buffer),
+                              "%ld", va_arg(vargs, long));
             }
             else if (size_tflag) {
-                sprintf(buffer, "%zd", va_arg(vargs, Py_ssize_t));
+                PyOS_snprintf(buffer, sizeof(buffer),
+                              "%zd", va_arg(vargs, Py_ssize_t));
             }
             else {
-                sprintf(buffer, "%d", va_arg(vargs, int));
+                PyOS_snprintf(buffer, sizeof(buffer), "%d", va_arg(vargs, int));
             }
             assert(strlen(buffer) < sizeof(buffer));
             WRITE_BYTES(buffer);
             break;
 
         case 'u':
             if (longflag) {
-                sprintf(buffer, "%lu", va_arg(vargs, unsigned long));
+                PyOS_snprintf(buffer, sizeof(buffer),
+                              "%lu", va_arg(vargs, unsigned long));
             }
             else if (size_tflag) {
-                sprintf(buffer, "%zu", va_arg(vargs, size_t));
+                PyOS_snprintf(buffer, sizeof(buffer),
+                              "%zu", va_arg(vargs, size_t));
             }
             else {
-                sprintf(buffer, "%u", va_arg(vargs, unsigned int));
+                PyOS_snprintf(buffer, sizeof(buffer),
+                              "%u", va_arg(vargs, unsigned int));
             }
             assert(strlen(buffer) < sizeof(buffer));
             WRITE_BYTES(buffer);
             break;
 
         case 'i':
-            sprintf(buffer, "%i", va_arg(vargs, int));
+            PyOS_snprintf(buffer, sizeof(buffer), "%i", va_arg(vargs, int));
             assert(strlen(buffer) < sizeof(buffer));
             WRITE_BYTES(buffer);
             break;
 
         case 'x':
-            sprintf(buffer, "%x", va_arg(vargs, int));
+            PyOS_snprintf(buffer, sizeof(buffer), "%x", va_arg(vargs, int));
             assert(strlen(buffer) < sizeof(buffer));
             WRITE_BYTES(buffer);
             break;
@@ -329,7 +334,7 @@ PyBytes_FromFormatV(const char *format, va_list vargs)
         }
 
         case 'p':
-            sprintf(buffer, "%p", va_arg(vargs, void*));
+            PyOS_snprintf(buffer, sizeof(buffer), "%p", va_arg(vargs, void*));
             assert(strlen(buffer) < sizeof(buffer));
             /* %p is ill-defined:  ensure leading 0x. */
             if (buffer[1] == 'X')

Objects/typeobject.c

@@ -4730,8 +4730,9 @@ type_traverse(PyTypeObject *type, visitproc visit, void *arg)
        for heaptypes. */
     if (!(type->tp_flags & Py_TPFLAGS_HEAPTYPE)) {
         char msg[200];
-        sprintf(msg, "type_traverse() called on non-heap type '%.100s'",
-                type->tp_name);
+        PyOS_snprintf(msg, sizeof(msg),
+                      "type_traverse() called on non-heap type '%.100s'",
+                      type->tp_name);
         _PyObject_ASSERT_FAILED_MSG((PyObject *)type, msg);
     }
 

Objects/unicodeobject.c

@@ -683,13 +683,38 @@ backslashreplace(_PyBytesWriter *writer, char *str,
     return str;
 }
 
+static Py_ssize_t
+xmlcharrefreplace_get_incr(Py_UCS4 ch)
+{
+    Py_ssize_t incr;
+
+    if (ch < 10)
+        incr = 2+1+1;
+    else if (ch < 100)
+        incr = 2+2+1;
+    else if (ch < 1000)
+        incr = 2+3+1;
+    else if (ch < 10000)
+        incr = 2+4+1;
+    else if (ch < 100000)
+        incr = 2+5+1;
+    else if (ch < 1000000)
+        incr = 2+6+1;
+    else {
+        assert(ch <= MAX_UNICODE);
+        incr = 2+7+1;
+    }
+
+    return incr;
+}
+
 /* Implementation of the "xmlcharrefreplace" error handler for 8-bit encodings:
    ASCII, Latin1, UTF-8, etc. */
 static char*
 xmlcharrefreplace(_PyBytesWriter *writer, char *str,
                   PyObject *unicode, Py_ssize_t collstart, Py_ssize_t collend)
 {
-    Py_ssize_t size, i;
+    Py_ssize_t size, incr, i;
     Py_UCS4 ch;
     int kind;
     const void *data;
@@ -700,25 +725,9 @@ xmlcharrefreplace(_PyBytesWriter *writer, char *str,
     size = 0;
     /* determine replacement size */
     for (i = collstart; i < collend; ++i) {
-        Py_ssize_t incr;
-
         ch = PyUnicode_READ(kind, data, i);
-        if (ch < 10)
-            incr = 2+1+1;
-        else if (ch < 100)
-            incr = 2+2+1;
-        else if (ch < 1000)
-            incr = 2+3+1;
-        else if (ch < 10000)
-            incr = 2+4+1;
-        else if (ch < 100000)
-            incr = 2+5+1;
-        else if (ch < 1000000)
-            incr = 2+6+1;
-        else {
-            assert(ch <= MAX_UNICODE);
-            incr = 2+7+1;
-        }
+        incr = xmlcharrefreplace_get_incr(ch);
+
         if (size > PY_SSIZE_T_MAX - incr) {
             PyErr_SetString(PyExc_OverflowError,
                             "encoded result is too long for a Python string");
@@ -733,7 +742,9 @@ xmlcharrefreplace(_PyBytesWriter *writer, char *str,
 
     /* generate replacement */
     for (i = collstart; i < collend; ++i) {
-        size = sprintf(str, "&#%d;", PyUnicode_READ(kind, data, i));
+        ch = PyUnicode_READ(kind, data, i);
+        incr = xmlcharrefreplace_get_incr(ch);
+        size = PyOS_snprintf(str, incr + 1,  "&#%d;", ch);
         if (size < 0) {
             return NULL;
         }
@@ -2455,33 +2466,42 @@ unicode_fromformat_arg(_PyUnicodeWriter *writer,
 
         if (*f == 'u') {
             if (longflag) {
-                len = sprintf(buffer, "%lu", va_arg(*vargs, unsigned long));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%lu", va_arg(*vargs, unsigned long));
             }
             else if (longlongflag) {
-                len = sprintf(buffer, "%llu", va_arg(*vargs, unsigned long long));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%llu", va_arg(*vargs, unsigned long long));
             }
             else if (size_tflag) {
-                len = sprintf(buffer, "%zu", va_arg(*vargs, size_t));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%zu", va_arg(*vargs, size_t));
             }
             else {
-                len = sprintf(buffer, "%u", va_arg(*vargs, unsigned int));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%u", va_arg(*vargs, unsigned int));
             }
         }
         else if (*f == 'x') {
-            len = sprintf(buffer, "%x", va_arg(*vargs, int));
+            len = PyOS_snprintf(buffer, sizeof(buffer),
+                                "%x", va_arg(*vargs, int));
         }
         else {
             if (longflag) {
-                len = sprintf(buffer, "%li", va_arg(*vargs, long));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%li", va_arg(*vargs, long));
             }
             else if (longlongflag) {
-                len = sprintf(buffer, "%lli", va_arg(*vargs, long long));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%lli", va_arg(*vargs, long long));
             }
             else if (size_tflag) {
-                len = sprintf(buffer, "%zi", va_arg(*vargs, Py_ssize_t));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%zi", va_arg(*vargs, Py_ssize_t));
             }
             else {
-                len = sprintf(buffer, "%i", va_arg(*vargs, int));
+                len = PyOS_snprintf(buffer, sizeof(buffer),
+                                    "%i", va_arg(*vargs, int));
             }
         }
         assert(len >= 0);
@@ -2530,7 +2550,8 @@ unicode_fromformat_arg(_PyUnicodeWriter *writer,
     {
         char number[MAX_LONG_LONG_CHARS];
 
-        len = sprintf(number, "%p", va_arg(*vargs, void*));
+        len = PyOS_snprintf(number, sizeof(number),
+                            "%p", va_arg(*vargs, void*));
         assert(len >= 0);
 
         /* %p is ill-defined:  ensure leading 0x. */
@@ -8104,7 +8125,8 @@ charmap_encoding_error(
         for (collpos = collstartpos; collpos < collendpos; ++collpos) {
             char buffer[2+29+1+1];
             char *cp;
-            sprintf(buffer, "&#%d;", (int)PyUnicode_READ_CHAR(unicode, collpos));
+            PyOS_snprintf(buffer, sizeof(buffer),
+                          "&#%d;", (int)PyUnicode_READ_CHAR(unicode, collpos));
             for (cp = buffer; *cp; ++cp) {
                 x = charmapencode_output(*cp, mapping, res, respos);
                 if (x==enc_EXCEPTION)

Parser/string_parser.c

@@ -118,7 +118,7 @@ decode_unicode_with_escapes(Parser *parser, const char *s, size_t len, Token *t)
             w_len = PyUnicode_GET_LENGTH(w);
             for (i = 0; i < w_len; i++) {
                 Py_UCS4 chr = PyUnicode_READ(kind, data, i);
-                sprintf(p, "\\U%08x", chr);
+                PyOS_snprintf(p, 10 + 1, "\\U%08x", chr);
                 p += 10;
             }
             /* Should be impossible to overflow */

Programs/_freeze_module.c

@@ -124,7 +124,7 @@ static PyObject *
 compile_and_marshal(const char *name, const char *text)
 {
     char *filename = (char *) malloc(strlen(name) + 10);
-    sprintf(filename, "<frozen %s>", name);
+    PyOS_snprintf(filename, strlen(name) + 10, "<frozen %s>", name);
     PyObject *code = Py_CompileStringExFlags(text, filename,
                                              Py_file_input, NULL, 0);
     free(filename);