Issue #28126: Replace Py_MEMCPY with memcpy(). Visual Studio can properly optimize memcpy().
diff --git a/Objects/abstract.c b/Objects/abstract.c
index 17da5c9..36f2242 100644
--- a/Objects/abstract.c
+++ b/Objects/abstract.c
@@ -2353,7 +2353,7 @@
 
     /* use borrowed references */
     stack[0] = obj;
-    Py_MEMCPY(&stack[1],
+    memcpy(&stack[1],
               &PyTuple_GET_ITEM(args, 0),
               argcount * sizeof(PyObject *));
 
@@ -2428,7 +2428,7 @@
     }
 
     /* Copy position arguments (borrowed references) */
-    Py_MEMCPY(stack, args, nargs * sizeof(stack[0]));
+    memcpy(stack, args, nargs * sizeof(stack[0]));
 
     kwstack = stack + nargs;
     pos = i = 0;
diff --git a/Objects/bytesobject.c b/Objects/bytesobject.c
index 4d14451..1550083 100644
--- a/Objects/bytesobject.c
+++ b/Objects/bytesobject.c
@@ -120,7 +120,7 @@
     if (str == NULL)
         return (PyObject *) op;
 
-    Py_MEMCPY(op->ob_sval, str, size);
+    memcpy(op->ob_sval, str, size);
     /* share short strings */
     if (size == 1) {
         characters[*str & UCHAR_MAX] = op;
@@ -163,7 +163,7 @@
         return PyErr_NoMemory();
     (void)PyObject_INIT_VAR(op, &PyBytes_Type, size);
     op->ob_shash = -1;
-    Py_MEMCPY(op->ob_sval, str, size+1);
+    memcpy(op->ob_sval, str, size+1);
     /* share short strings */
     if (size == 0) {
         nullstring = op;
@@ -437,7 +437,7 @@
         str = _PyBytesWriter_Prepare(writer, str, len);
         if (str == NULL)
             return NULL;
-        Py_MEMCPY(str, p, len);
+        memcpy(str, p, len);
         PyMem_Free(p);
         str += len;
         return str;
@@ -626,7 +626,7 @@
                 len = format_len - (fmt - format);
             assert(len != 0);
 
-            Py_MEMCPY(res, fmt, len);
+            memcpy(res, fmt, len);
             res += len;
             fmt += len;
             fmtcnt -= (len - 1);
@@ -1009,7 +1009,7 @@
             }
 
             /* Copy bytes */
-            Py_MEMCPY(res, pbuf, len);
+            memcpy(res, pbuf, len);
             res += len;
 
             /* Pad right with the fill character if needed */
@@ -1473,12 +1473,12 @@
     }
     i = 0;
     if (i < size) {
-        Py_MEMCPY(op->ob_sval, a->ob_sval, Py_SIZE(a));
+        memcpy(op->ob_sval, a->ob_sval, Py_SIZE(a));
         i = Py_SIZE(a);
     }
     while (i < size) {
         j = (i <= size-i)  ?  i  :  size-i;
-        Py_MEMCPY(op->ob_sval+i, op->ob_sval, j);
+        memcpy(op->ob_sval+i, op->ob_sval, j);
         i += j;
     }
     return (PyObject *) op;
@@ -2765,7 +2765,7 @@
     n = PyBytes_GET_SIZE(tmp);
     pnew = type->tp_alloc(type, n);
     if (pnew != NULL) {
-        Py_MEMCPY(PyBytes_AS_STRING(pnew),
+        memcpy(PyBytes_AS_STRING(pnew),
                   PyBytes_AS_STRING(tmp), n+1);
         ((PyBytesObject *)pnew)->ob_shash =
             ((PyBytesObject *)tmp)->ob_shash;
@@ -3237,7 +3237,7 @@
                 dest = PyByteArray_AS_STRING(writer->buffer);
             else
                 dest = PyBytes_AS_STRING(writer->buffer);
-            Py_MEMCPY(dest,
+            memcpy(dest,
                       writer->small_buffer,
                       pos);
         }
@@ -3372,7 +3372,7 @@
     if (str == NULL)
         return NULL;
 
-    Py_MEMCPY(str, bytes, size);
+    memcpy(str, bytes, size);
     str += size;
 
     return str;
diff --git a/Objects/stringlib/join.h b/Objects/stringlib/join.h
index 90f966d..6f314e1 100644
--- a/Objects/stringlib/join.h
+++ b/Objects/stringlib/join.h
@@ -107,7 +107,7 @@
         for (i = 0; i < nbufs; i++) {
             Py_ssize_t n = buffers[i].len;
             char *q = buffers[i].buf;
-            Py_MEMCPY(p, q, n);
+            memcpy(p, q, n);
             p += n;
         }
         goto done;
@@ -116,12 +116,12 @@
         Py_ssize_t n;
         char *q;
         if (i) {
-            Py_MEMCPY(p, sepstr, seplen);
+            memcpy(p, sepstr, seplen);
             p += seplen;
         }
         n = buffers[i].len;
         q = buffers[i].buf;
-        Py_MEMCPY(p, q, n);
+        memcpy(p, q, n);
         p += n;
     }
     goto done;
diff --git a/Objects/stringlib/transmogrify.h b/Objects/stringlib/transmogrify.h
index 9903912..a314572 100644
--- a/Objects/stringlib/transmogrify.h
+++ b/Objects/stringlib/transmogrify.h
@@ -108,7 +108,7 @@
     if (u) {
         if (left)
             memset(STRINGLIB_STR(u), fill, left);
-        Py_MEMCPY(STRINGLIB_STR(u) + left,
+        memcpy(STRINGLIB_STR(u) + left,
                STRINGLIB_STR(self),
                STRINGLIB_LEN(self));
         if (right)
@@ -275,13 +275,13 @@
 
     if (to_len > 1) {
         /* Lay the first one down (guaranteed this will occur) */
-        Py_MEMCPY(result_s, to_s, to_len);
+        memcpy(result_s, to_s, to_len);
         result_s += to_len;
         count -= 1;
 
         for (i = 0; i < count; i++) {
             *result_s++ = *self_s++;
-            Py_MEMCPY(result_s, to_s, to_len);
+            memcpy(result_s, to_s, to_len);
             result_s += to_len;
         }
     }
@@ -297,7 +297,7 @@
     }
 
     /* Copy the rest of the original string */
-    Py_MEMCPY(result_s, self_s, self_len - i);
+    memcpy(result_s, self_s, self_len - i);
 
     return result;
 }
@@ -337,11 +337,11 @@
         next = findchar(start, end - start, from_c);
         if (next == NULL)
             break;
-        Py_MEMCPY(result_s, start, next - start);
+        memcpy(result_s, start, next - start);
         result_s += (next - start);
         start = next + 1;
     }
-    Py_MEMCPY(result_s, start, end - start);
+    memcpy(result_s, start, end - start);
 
     return result;
 }
@@ -390,12 +390,12 @@
             break;
         next = start + offset;
 
-        Py_MEMCPY(result_s, start, next - start);
+        memcpy(result_s, start, next - start);
 
         result_s += (next - start);
         start = next + from_len;
     }
-    Py_MEMCPY(result_s, start, end - start);
+    memcpy(result_s, start, end - start);
     return result;
 }
 
@@ -427,7 +427,7 @@
         return NULL;
     }
     result_s = STRINGLIB_STR(result);
-    Py_MEMCPY(result_s, self_s, self_len);
+    memcpy(result_s, self_s, self_len);
 
     /* change everything in-place, starting with this one */
     start =  result_s + (next - self_s);
@@ -477,11 +477,11 @@
         return NULL;
     }
     result_s = STRINGLIB_STR(result);
-    Py_MEMCPY(result_s, self_s, self_len);
+    memcpy(result_s, self_s, self_len);
 
     /* change everything in-place, starting with this one */
     start =  result_s + offset;
-    Py_MEMCPY(start, to_s, from_len);
+    memcpy(start, to_s, from_len);
     start += from_len;
     end = result_s + self_len;
 
@@ -491,7 +491,7 @@
                                 0);
         if (offset == -1)
             break;
-        Py_MEMCPY(start + offset, to_s, from_len);
+        memcpy(start + offset, to_s, from_len);
         start += offset + from_len;
     }
 
@@ -544,20 +544,20 @@
 
         if (next == start) {
             /* replace with the 'to' */
-            Py_MEMCPY(result_s, to_s, to_len);
+            memcpy(result_s, to_s, to_len);
             result_s += to_len;
             start += 1;
         } else {
             /* copy the unchanged old then the 'to' */
-            Py_MEMCPY(result_s, start, next - start);
+            memcpy(result_s, start, next - start);
             result_s += (next - start);
-            Py_MEMCPY(result_s, to_s, to_len);
+            memcpy(result_s, to_s, to_len);
             result_s += to_len;
             start = next + 1;
         }
     }
     /* Copy the remainder of the remaining bytes */
-    Py_MEMCPY(result_s, start, end - start);
+    memcpy(result_s, start, end - start);
 
     return result;
 }
@@ -613,20 +613,20 @@
         next = start + offset;
         if (next == start) {
             /* replace with the 'to' */
-            Py_MEMCPY(result_s, to_s, to_len);
+            memcpy(result_s, to_s, to_len);
             result_s += to_len;
             start += from_len;
         } else {
             /* copy the unchanged old then the 'to' */
-            Py_MEMCPY(result_s, start, next - start);
+            memcpy(result_s, start, next - start);
             result_s += (next - start);
-            Py_MEMCPY(result_s, to_s, to_len);
+            memcpy(result_s, to_s, to_len);
             result_s += to_len;
             start = next + from_len;
         }
     }
     /* Copy the remainder of the remaining bytes */
-    Py_MEMCPY(result_s, start, end - start);
+    memcpy(result_s, start, end - start);
 
     return result;
 }
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index aaebfd0..85cdbb7 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -1048,7 +1048,7 @@
             return NULL;
         copy_length = _PyUnicode_WSTR_LENGTH(unicode);
         copy_length = Py_MIN(copy_length, length);
-        Py_MEMCPY(_PyUnicode_WSTR(w), _PyUnicode_WSTR(unicode),
+        memcpy(_PyUnicode_WSTR(w), _PyUnicode_WSTR(unicode),
                   copy_length * sizeof(wchar_t));
         return w;
     }
@@ -1435,7 +1435,7 @@
             if (max_char >= 128)
                 return -1;
         }
-        Py_MEMCPY((char*)to_data + to_kind * to_start,
+        memcpy((char*)to_data + to_kind * to_start,
                   (char*)from_data + from_kind * from_start,
                   to_kind * how_many);
     }
@@ -2024,7 +2024,7 @@
         break;
     case PyUnicode_2BYTE_KIND:
 #if Py_UNICODE_SIZE == 2
-        Py_MEMCPY(PyUnicode_2BYTE_DATA(unicode), u, size * 2);
+        memcpy(PyUnicode_2BYTE_DATA(unicode), u, size * 2);
 #else
         _PyUnicode_CONVERT_BYTES(Py_UNICODE, Py_UCS2,
                                 u, u + size, PyUnicode_2BYTE_DATA(unicode));
@@ -2037,7 +2037,7 @@
         unicode_convert_wchar_to_ucs4(u, u + size, unicode);
 #else
         assert(num_surrogates == 0);
-        Py_MEMCPY(PyUnicode_4BYTE_DATA(unicode), u, size * 4);
+        memcpy(PyUnicode_4BYTE_DATA(unicode), u, size * 4);
 #endif
         break;
     default:
@@ -2348,7 +2348,7 @@
         return NULL;
     assert(PyUnicode_KIND(copy) == PyUnicode_KIND(unicode));
 
-    Py_MEMCPY(PyUnicode_DATA(copy), PyUnicode_DATA(unicode),
+    memcpy(PyUnicode_DATA(copy), PyUnicode_DATA(unicode),
               length * PyUnicode_KIND(unicode));
     assert(_PyUnicode_CheckConsistency(copy, 1));
     return copy;
@@ -2454,7 +2454,7 @@
     }
     else {
         assert(kind == PyUnicode_4BYTE_KIND);
-        Py_MEMCPY(target, data, len * sizeof(Py_UCS4));
+        memcpy(target, data, len * sizeof(Py_UCS4));
     }
     if (copy_null)
         target[len] = 0;
@@ -2963,7 +2963,7 @@
             size = res + 1;
         else
             res = size;
-        Py_MEMCPY(w, wstr, size * sizeof(wchar_t));
+        memcpy(w, wstr, size * sizeof(wchar_t));
         return res;
     }
     else
@@ -3986,7 +3986,7 @@
             return NULL;
         }
         _PyUnicode_UTF8_LENGTH(unicode) = PyBytes_GET_SIZE(bytes);
-        Py_MEMCPY(_PyUnicode_UTF8(unicode),
+        memcpy(_PyUnicode_UTF8(unicode),
                   PyBytes_AS_STRING(bytes),
                   _PyUnicode_UTF8_LENGTH(unicode) + 1);
         Py_DECREF(bytes);
@@ -5473,7 +5473,7 @@
         }
 
         if (PyBytes_Check(rep)) {
-            Py_MEMCPY(out, PyBytes_AS_STRING(rep), repsize);
+            memcpy(out, PyBytes_AS_STRING(rep), repsize);
             out += moreunits;
         } else /* rep is unicode */ {
             assert(PyUnicode_KIND(rep) == PyUnicode_1BYTE_KIND);
@@ -5825,7 +5825,7 @@
         }
 
         if (PyBytes_Check(rep)) {
-            Py_MEMCPY(out, PyBytes_AS_STRING(rep), repsize);
+            memcpy(out, PyBytes_AS_STRING(rep), repsize);
             out += moreunits;
         } else /* rep is unicode */ {
             assert(PyUnicode_KIND(rep) == PyUnicode_1BYTE_KIND);
@@ -10012,7 +10012,7 @@
 
             /* Copy item, and maybe the separator. */
             if (i && seplen != 0) {
-                Py_MEMCPY(res_data,
+                memcpy(res_data,
                           sep_data,
                           kind * seplen);
                 res_data += kind * seplen;
@@ -10020,7 +10020,7 @@
 
             itemlen = PyUnicode_GET_LENGTH(item);
             if (itemlen != 0) {
-                Py_MEMCPY(res_data,
+                memcpy(res_data,
                           PyUnicode_DATA(item),
                           kind * itemlen);
                 res_data += kind * itemlen;
@@ -12396,11 +12396,11 @@
         Py_ssize_t done = PyUnicode_GET_LENGTH(str);
         const Py_ssize_t char_size = PyUnicode_KIND(str);
         char *to = (char *) PyUnicode_DATA(u);
-        Py_MEMCPY(to, PyUnicode_DATA(str),
+        memcpy(to, PyUnicode_DATA(str),
                   PyUnicode_GET_LENGTH(str) * char_size);
         while (done < nchars) {
             n = (done <= nchars-done) ? done : nchars-done;
-            Py_MEMCPY(to + (done * char_size), to, n * char_size);
+            memcpy(to + (done * char_size), to, n * char_size);
             done += n;
         }
     }
@@ -13531,7 +13531,7 @@
         const Py_UCS1 *str = (const Py_UCS1 *)ascii;
         Py_UCS1 *data = writer->data;
 
-        Py_MEMCPY(data + writer->pos, str, len);
+        memcpy(data + writer->pos, str, len);
         break;
     }
     case PyUnicode_2BYTE_KIND:
@@ -14928,7 +14928,7 @@
         _PyUnicode_WSTR(self) = (wchar_t *)data;
     }
 
-    Py_MEMCPY(data, PyUnicode_DATA(unicode),
+    memcpy(data, PyUnicode_DATA(unicode),
               kind * (length + 1));
     assert(_PyUnicode_CheckConsistency(self, 1));
 #ifdef Py_DEBUG