bpo-39943: Add the const qualifier to pointers on non-mutable PyUnicode data. (GH-19345)

Python/_warnings.c

@@ -435,7 +435,7 @@ normalize_module(PyObject *filename)
 {
     PyObject *module;
     int kind;
-    void *data;
+    const void *data;
     Py_ssize_t len;
 
     len = PyUnicode_GetLength(filename);
@@ -519,7 +519,7 @@ show_warning(PyObject *filename, int lineno, PyObject *text,
     /* Print "  source_line\n" */
     if (sourceline) {
         int kind;
-        void *data;
+        const void *data;
         Py_ssize_t i, len;
         Py_UCS4 ch;
         PyObject *truncated;

Python/ast.c

@@ -4588,7 +4588,7 @@ decode_unicode_with_escapes(struct compiling *c, const node *n, const char *s,
         if (*s & 0x80) { /* XXX inefficient */
             PyObject *w;
             int kind;
-            void *data;
+            const void *data;
             Py_ssize_t len, i;
             w = decode_utf8(c, &s, end);
             if (w == NULL) {

Python/codecs.c

@@ -701,8 +701,7 @@ PyObject *PyCodec_ReplaceErrors(PyObject *exc)
 
     if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeEncodeError)) {
         PyObject *res;
-        int kind;
-        void *data;
+        Py_UCS1 *outp;
         if (PyUnicodeEncodeError_GetStart(exc, &start))
             return NULL;
         if (PyUnicodeEncodeError_GetEnd(exc, &end))
@@ -711,10 +710,10 @@ PyObject *PyCodec_ReplaceErrors(PyObject *exc)
         res = PyUnicode_New(len, '?');
         if (res == NULL)
             return NULL;
-        kind = PyUnicode_KIND(res);
-        data = PyUnicode_DATA(res);
+        assert(PyUnicode_KIND(res) == PyUnicode_1BYTE_KIND);
+        outp = PyUnicode_1BYTE_DATA(res);
         for (i = 0; i < len; ++i)
-            PyUnicode_WRITE(kind, data, i, '?');
+            outp[i] = '?';
         assert(_PyUnicode_CheckConsistency(res, 1));
         return Py_BuildValue("(Nn)", res, end);
     }
@@ -727,8 +726,7 @@ PyObject *PyCodec_ReplaceErrors(PyObject *exc)
     }
     else if (PyObject_TypeCheck(exc, (PyTypeObject *)PyExc_UnicodeTranslateError)) {
         PyObject *res;
-        int kind;
-        void *data;
+        Py_UCS2 *outp;
         if (PyUnicodeTranslateError_GetStart(exc, &start))
             return NULL;
         if (PyUnicodeTranslateError_GetEnd(exc, &end))
@@ -737,10 +735,10 @@ PyObject *PyCodec_ReplaceErrors(PyObject *exc)
         res = PyUnicode_New(len, Py_UNICODE_REPLACEMENT_CHARACTER);
         if (res == NULL)
             return NULL;
-        kind = PyUnicode_KIND(res);
-        data = PyUnicode_DATA(res);
-        for (i=0; i < len; i++)
-            PyUnicode_WRITE(kind, data, i, Py_UNICODE_REPLACEMENT_CHARACTER);
+        assert(PyUnicode_KIND(res) == PyUnicode_2BYTE_KIND);
+        outp = PyUnicode_2BYTE_DATA(res);
+        for (i = 0; i < len; i++)
+            outp[i] = Py_UNICODE_REPLACEMENT_CHARACTER;
         assert(_PyUnicode_CheckConsistency(res, 1));
         return Py_BuildValue("(Nn)", res, end);
     }
@@ -759,7 +757,7 @@ PyObject *PyCodec_XMLCharRefReplaceErrors(PyObject *exc)
         Py_ssize_t start;
         Py_ssize_t end;
         PyObject *res;
-        unsigned char *outp;
+        Py_UCS1 *outp;
         Py_ssize_t ressize;
         Py_UCS4 ch;
         if (PyUnicodeEncodeError_GetStart(exc, &start))
@@ -855,7 +853,7 @@ PyObject *PyCodec_BackslashReplaceErrors(PyObject *exc)
     Py_ssize_t start;
     Py_ssize_t end;
     PyObject *res;
-    unsigned char *outp;
+    Py_UCS1 *outp;
     int ressize;
     Py_UCS4 c;
 
@@ -966,7 +964,7 @@ PyObject *PyCodec_NameReplaceErrors(PyObject *exc)
         Py_ssize_t start;
         Py_ssize_t end;
         PyObject *res;
-        unsigned char *outp;
+        Py_UCS1 *outp;
         Py_ssize_t ressize;
         int replsize;
         Py_UCS4 c;

Python/formatter_unicode.c

@@ -62,7 +62,7 @@ get_integer(PyObject *str, Py_ssize_t *ppos, Py_ssize_t end,
     Py_ssize_t accumulator, digitval, pos = *ppos;
     int numdigits;
     int kind = PyUnicode_KIND(str);
-    void *data = PyUnicode_DATA(str);
+    const void *data = PyUnicode_DATA(str);
 
     accumulator = numdigits = 0;
     for (; pos < end; pos++, numdigits++) {
@@ -170,7 +170,7 @@ parse_internal_render_format_spec(PyObject *format_spec,
 {
     Py_ssize_t pos = start;
     int kind = PyUnicode_KIND(format_spec);
-    void *data = PyUnicode_DATA(format_spec);
+    const void *data = PyUnicode_DATA(format_spec);
     /* end-pos is used throughout this code to specify the length of
        the input string */
 #define READ_spec(index) PyUnicode_READ(kind, data, index)
@@ -443,7 +443,7 @@ parse_number(PyObject *s, Py_ssize_t pos, Py_ssize_t end,
 {
     Py_ssize_t remainder;
     int kind = PyUnicode_KIND(s);
-    void *data = PyUnicode_DATA(s);
+    const void *data = PyUnicode_DATA(s);
 
     while (pos<end && Py_ISDIGIT(PyUnicode_READ(kind, data, pos)))
         ++pos;

Python/getargs.c

@@ -923,7 +923,7 @@ convertsimple(PyObject *arg, const char **p_format, va_list *p_va, int flags,
     case 'C': {/* unicode char */
         int *p = va_arg(*p_va, int *);
         int kind;
-        void *data;
+        const void *data;
 
         if (!PyUnicode_Check(arg))
             return converterr("a unicode character", arg, msgbuf, bufsize);

Python/traceback.c

@@ -376,7 +376,7 @@ _Py_DisplaySourceLine(PyObject *f, PyObject *filename, int lineno, int indent)
     int fd;
     int i;
     char *found_encoding;
-    char *encoding;
+    const char *encoding;
     PyObject *io;
     PyObject *binary;
     PyObject *fob = NULL;
@@ -384,7 +384,7 @@ _Py_DisplaySourceLine(PyObject *f, PyObject *filename, int lineno, int indent)
     PyObject *res;
     char buf[MAXPATHLEN+1];
     int kind;
-    void *data;
+    const void *data;
 
     /* open the file */
     if (filename == NULL)