gh-111089: Revert PyUnicode_AsUTF8() changes (#111833)

* Revert "gh-111089: Use PyUnicode_AsUTF8() in Argument Clinic (#111585)"

This reverts commit d9b606b3d04fc56fb0bcc479d7d6c14562edb5e2.

* Revert "gh-111089: Use PyUnicode_AsUTF8() in getargs.c (#111620)"

This reverts commit cde1071b2a72e8261ca66053ef61431b7f3a81fd.

* Revert "gh-111089: PyUnicode_AsUTF8() now raises on embedded NUL (#111091)"

This reverts commit d731579bfb9a497cfb0076cb6b221058a20088fe.

* Revert "gh-111089: Add PyUnicode_AsUTF8() to the limited C API (#111121)"

This reverts commit d8f32be5b6a736dc2fc9dca3f1bf176c82fc9b44.

* Revert "gh-111089: Use PyUnicode_AsUTF8() in sqlite3 (#111122)"

This reverts commit 37e4e20eaa8f27ada926d49e5971fecf0477ad26.

7 files changed, 92 insertions, 29 deletions

diff --git a/Objects/clinic/bytearrayobject.c.h b/Objects/clinic/bytearrayobject.c.h
index 355ae49..d952450 100644
--- a/Objects/clinic/bytearrayobject.c.h
+++ b/Objects/clinic/bytearrayobject.c.h
@@ -68,10 +68,15 @@ bytearray___init__(PyObject *self, PyObject *args, PyObject *kwargs)
             _PyArg_BadArgument("bytearray", "argument 'encoding'", "str", fastargs[1]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(fastargs[1]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(fastargs[1], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -80,10 +85,15 @@ bytearray___init__(PyObject *self, PyObject *args, PyObject *kwargs)
         _PyArg_BadArgument("bytearray", "argument 'errors'", "str", fastargs[2]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(fastargs[2]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(fastargs[2], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = bytearray___init___impl((PyByteArrayObject *)self, arg, encoding, errors);
 
@@ -950,10 +960,15 @@ bytearray_decode(PyByteArrayObject *self, PyObject *const *args, Py_ssize_t narg
             _PyArg_BadArgument("decode", "argument 'encoding'", "str", args[0]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(args[0]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(args[0], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -962,10 +977,15 @@ bytearray_decode(PyByteArrayObject *self, PyObject *const *args, Py_ssize_t narg
         _PyArg_BadArgument("decode", "argument 'errors'", "str", args[1]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(args[1]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(args[1], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = bytearray_decode_impl(self, encoding, errors);
 
@@ -1241,4 +1261,4 @@ bytearray_sizeof(PyByteArrayObject *self, PyObject *Py_UNUSED(ignored))
 {
     return bytearray_sizeof_impl(self);
 }
-/*[clinic end generated code: output=5a7de6295a7ce6cc input=a9049054013a1b77]*/
+/*[clinic end generated code: output=0797a5e03cda2a16 input=a9049054013a1b77]*/
diff --git a/Objects/clinic/bytesobject.c.h b/Objects/clinic/bytesobject.c.h
index 042d0bf..1e45be3 100644
--- a/Objects/clinic/bytesobject.c.h
+++ b/Objects/clinic/bytesobject.c.h
@@ -720,10 +720,15 @@ bytes_decode(PyBytesObject *self, PyObject *const *args, Py_ssize_t nargs, PyObj
             _PyArg_BadArgument("decode", "argument 'encoding'", "str", args[0]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(args[0]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(args[0], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -732,10 +737,15 @@ bytes_decode(PyBytesObject *self, PyObject *const *args, Py_ssize_t nargs, PyObj
         _PyArg_BadArgument("decode", "argument 'errors'", "str", args[1]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(args[1]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(args[1], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = bytes_decode_impl(self, encoding, errors);
 
@@ -987,10 +997,15 @@ bytes_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
             _PyArg_BadArgument("bytes", "argument 'encoding'", "str", fastargs[1]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(fastargs[1]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(fastargs[1], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -999,14 +1014,19 @@ bytes_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
         _PyArg_BadArgument("bytes", "argument 'errors'", "str", fastargs[2]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(fastargs[2]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(fastargs[2], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = bytes_new_impl(type, x, encoding, errors);
 
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=97aab3f6ae398664 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=8a49dbbd78914a6f input=a9049054013a1b77]*/
diff --git a/Objects/clinic/floatobject.c.h b/Objects/clinic/floatobject.c.h
index 8c33119..10f6149 100644
--- a/Objects/clinic/floatobject.c.h
+++ b/Objects/clinic/floatobject.c.h
@@ -275,10 +275,15 @@ float___getformat__(PyTypeObject *type, PyObject *arg)
         _PyArg_BadArgument("__getformat__", "argument", "str", arg);
         goto exit;
     }
-    typestr = PyUnicode_AsUTF8(arg);
+    Py_ssize_t typestr_length;
+    typestr = PyUnicode_AsUTF8AndSize(arg, &typestr_length);
     if (typestr == NULL) {
         goto exit;
     }
+    if (strlen(typestr) != (size_t)typestr_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
     return_value = float___getformat___impl(type, typestr);
 
 exit:
@@ -313,4 +318,4 @@ float___format__(PyObject *self, PyObject *arg)
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=01f6fbd082eefead input=a9049054013a1b77]*/
+/*[clinic end generated code: output=c79743c8551c30d9 input=a9049054013a1b77]*/
diff --git a/Objects/clinic/memoryobject.c.h b/Objects/clinic/memoryobject.c.h
index ebc1e06..f199434 100644
--- a/Objects/clinic/memoryobject.c.h
+++ b/Objects/clinic/memoryobject.c.h
@@ -305,10 +305,15 @@ memoryview_tobytes(PyMemoryViewObject *self, PyObject *const *args, Py_ssize_t n
         order = NULL;
     }
     else if (PyUnicode_Check(args[0])) {
-        order = PyUnicode_AsUTF8(args[0]);
+        Py_ssize_t order_length;
+        order = PyUnicode_AsUTF8AndSize(args[0], &order_length);
         if (order == NULL) {
             goto exit;
         }
+        if (strlen(order) != (size_t)order_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
     }
     else {
         _PyArg_BadArgument("tobytes", "argument 'order'", "str or None", args[0]);
@@ -408,4 +413,4 @@ skip_optional_pos:
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=abd8c0ce804d8992 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=7e76a09106921ba2 input=a9049054013a1b77]*/
diff --git a/Objects/clinic/unicodeobject.c.h b/Objects/clinic/unicodeobject.c.h
index 83e3bf2..7711434 100644
--- a/Objects/clinic/unicodeobject.c.h
+++ b/Objects/clinic/unicodeobject.c.h
@@ -203,10 +203,15 @@ unicode_encode(PyObject *self, PyObject *const *args, Py_ssize_t nargs, PyObject
             _PyArg_BadArgument("encode", "argument 'encoding'", "str", args[0]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(args[0]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(args[0], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -215,10 +220,15 @@ unicode_encode(PyObject *self, PyObject *const *args, Py_ssize_t nargs, PyObject
         _PyArg_BadArgument("encode", "argument 'errors'", "str", args[1]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(args[1]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(args[1], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = unicode_encode_impl(self, encoding, errors);
 
@@ -1463,10 +1473,15 @@ unicode_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
             _PyArg_BadArgument("str", "argument 'encoding'", "str", fastargs[1]);
             goto exit;
         }
-        encoding = PyUnicode_AsUTF8(fastargs[1]);
+        Py_ssize_t encoding_length;
+        encoding = PyUnicode_AsUTF8AndSize(fastargs[1], &encoding_length);
         if (encoding == NULL) {
             goto exit;
         }
+        if (strlen(encoding) != (size_t)encoding_length) {
+            PyErr_SetString(PyExc_ValueError, "embedded null character");
+            goto exit;
+        }
         if (!--noptargs) {
             goto skip_optional_pos;
         }
@@ -1475,14 +1490,19 @@ unicode_new(PyTypeObject *type, PyObject *args, PyObject *kwargs)
         _PyArg_BadArgument("str", "argument 'errors'", "str", fastargs[2]);
         goto exit;
     }
-    errors = PyUnicode_AsUTF8(fastargs[2]);
+    Py_ssize_t errors_length;
+    errors = PyUnicode_AsUTF8AndSize(fastargs[2], &errors_length);
     if (errors == NULL) {
         goto exit;
     }
+    if (strlen(errors) != (size_t)errors_length) {
+        PyErr_SetString(PyExc_ValueError, "embedded null character");
+        goto exit;
+    }
 skip_optional_pos:
     return_value = unicode_new_impl(type, x, encoding, errors);
 
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=20313d6339272ddc input=a9049054013a1b77]*/
+/*[clinic end generated code: output=873d8b3d09af3095 input=a9049054013a1b77]*/
diff --git a/Objects/typeobject.c b/Objects/typeobject.c
index f44e30c..557464c 100644
--- a/Objects/typeobject.c
+++ b/Objects/typeobject.c
@@ -3501,14 +3501,13 @@ type_new_set_doc(PyTypeObject *type)
         return 0;
     }
 
-    Py_ssize_t doc_size;
-    const char *doc_str = PyUnicode_AsUTF8AndSize(doc, &doc_size);
+    const char *doc_str = PyUnicode_AsUTF8(doc);
     if (doc_str == NULL) {
         return -1;
     }
 
     // Silently truncate the docstring if it contains a null byte
-    Py_ssize_t size = doc_size + 1;
+    Py_ssize_t size = strlen(doc_str) + 1;
     char *tp_doc = (char *)PyObject_Malloc(size);
     if (tp_doc == NULL) {
         PyErr_NoMemory();
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index 87636ef..53e1e56 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -3844,13 +3844,7 @@ PyUnicode_AsUTF8AndSize(PyObject *unicode, Py_ssize_t *psize)
 const char *
 PyUnicode_AsUTF8(PyObject *unicode)
 {
-    Py_ssize_t size;
-    const char *utf8 = PyUnicode_AsUTF8AndSize(unicode, &size);
-    if (utf8 != NULL && strlen(utf8) != (size_t)size) {
-        PyErr_SetString(PyExc_ValueError, "embedded null character");
-        return NULL;
-    }
-    return utf8;
+    return PyUnicode_AsUTF8AndSize(unicode, NULL);
 }
 
 /*