[pull] main from python:main

Doc/c-api/bytes.rst

@@ -231,6 +231,7 @@ called with a non-bytes parameter.
    Resize a bytes object. *newsize* will be the new length of the bytes object.
    You can think of it as creating a new bytes object and destroying the old
    one, only more efficiently.
+
    Pass the address of an
    existing bytes object as an lvalue (it may be written into), and the new size
    desired.  On success, *\*bytes* holds the resized bytes object and ``0`` is
@@ -239,6 +240,11 @@ called with a non-bytes parameter.
    *\*bytes* is set to ``NULL``, :exc:`MemoryError` is set, and ``-1`` is
    returned.
 
+   While bytes objects are usually immutable in Python, this special C API
+   allows mutating a bytes object in-place. The returned bytes object can still
+   be mutated using :c:func:`PyBytesWriter_GetData`; except if *newsize* is
+   zero in which case it returns the immutable empty bytes string.
+
    .. soft-deprecated:: 3.15
       Use the :c:type:`PyBytesWriter` API instead.
 
@@ -290,10 +296,10 @@ object.
 
 .. c:type:: PyBytesWriter
 
-   A bytes writer instance.
+   A bytes writer object.
 
-   The API is **not thread safe**: a writer should only be used by a single
-   thread at the same time.
+   The API is **not thread safe**. A :c:type:`PyBytesWriter` object must only
+   be used by a single thread, it must not be shared between threads.
 
    The instance must be destroyed by :c:func:`PyBytesWriter_Finish` on
    success, or :c:func:`PyBytesWriter_Discard` on error.
@@ -429,7 +435,7 @@ Low-level API
    On success, return ``0``.
    On error, set an exception and return ``-1``.
 
-   *size* can be negative to shrink the writer.
+   *grow* can be negative to shrink the writer.
 
 .. c:function:: void* PyBytesWriter_GrowAndUpdatePointer(PyBytesWriter *writer, Py_ssize_t size, void *buf)
 

Include/internal/pycore_bytesobject.h

@@ -77,6 +77,10 @@ PyAPI_FUNC(PyObject *) _PyBytes_Repeat(PyObject *self, Py_ssize_t n);
 
 extern int _PyBytes_ResizeKeepOnError(PyObject **pv, Py_ssize_t newsize);
 
+#ifndef NDEBUG
+extern int _PyBytes_IsMutable(PyObject *obj);
+#endif
+
 /* --- PyBytesWriter ------------------------------------------------------ */
 
 struct PyBytesWriter {

Lib/test/test_bytes.py

@@ -1130,13 +1130,14 @@ def test_translate(self):
         self.assertRaises(ValueError, b.translate, bytes(range(255)))
 
         c = b.translate(rosetta, b'hello')
-        self.assertEqual(b, b'hello')
-        self.assertIsInstance(c, self.type2test)
+        self.assertEqual(c, b'')
+        self.assertEqual(type(c), self.type2test)
 
         c = b.translate(rosetta)
         d = b.translate(rosetta, b'')
-        self.assertEqual(c, d)
         self.assertEqual(c, b'helle')
+        self.assertEqual(type(c), self.type2test)
+        self.assertEqual(d, b'helle')
 
         c = b.translate(rosetta, b'l')
         self.assertEqual(c, b'hee')

Lib/test/test_capi/test_bytes.py

@@ -461,6 +461,24 @@ def test_grow(self):
         writer.grow(0)  # noop
         self.assertEqual(writer.finish(), b'number=123')
 
+        for size in (self.SMALL_BUFFER, self.LARGE_BUFFER):
+            with self.subTest(size=size):
+                # Truncate the last byte
+                data = b'x' * size
+                writer = self.create_writer(size)
+                writer.write(0, data)
+                self.assertEqual(writer.get_data(), data)
+                writer.grow(-1)
+                self.assertEqual(writer.get_data(), data[:-1])
+                self.assertEqual(writer.finish(),  data[:-1])
+
+                # Make the buffer empty
+                writer = self.create_writer(size)
+                writer.write(0, data)
+                writer.grow(-size)
+                self.assertEqual(writer.get_data(), b'')
+                self.assertEqual(writer.finish(),  b'')
+
         # Switch from small buffer to large buffer
         writer = self.create_writer()
         small, large = self.SMALL_BUFFER, self.LARGE_BUFFER
@@ -476,8 +494,8 @@ def test_grow(self):
             with self.subTest(size=size):
                 writer = self.create_writer()
                 writer.write_bytes(b'x' * size, -1)
-                with self.assertRaisesRegex(ValueError, 'size must be >= 0'):
-                    writer.grow(-1)
+                with self.assertRaisesRegex(ValueError, 'invalid size'):
+                    writer.grow(-size - 1)
                 with self.assertRaises(MemoryError):
                     writer.grow(_testcapi.PY_SSIZE_T_MAX)
                 self.assertEqual(writer.finish(), b'x' * size)

Misc/NEWS.d/next/Library/2026-09-12-15-50-23.gh-issue-157366.djwD2M.rst

@@ -0,0 +1,7 @@
+Remove the accidental acceptance of :class:`bytes` in the C implementation of
+:mod:`xml.etree.ElementTree`, a leftover of the Python 2 to Python 3 migration,
+for paths of :meth:`~xml.etree.ElementTree.Element.find`
+and similar methods, for the tag of :meth:`~xml.etree.ElementTree.Element.iter`,
+and for the names of events of :class:`~xml.etree.ElementTree.XMLPullParser`
+and :func:`~xml.etree.ElementTree.iterparse`.  It now raises the same
+exceptions as the Python implementation.

Modules/_elementtree.c

@@ -1230,24 +1230,6 @@ checkpath(PyObject* tag)
         }
         return 0;
     }
-    if (PyBytes_Check(tag)) {
-        const char *p = PyBytes_AS_STRING(tag);
-        const Py_ssize_t len = PyBytes_GET_SIZE(tag);
-        if (len >= 3 && p[0] == '{' && (
-                p[1] == '}' || (p[1] == '*' && p[2] == '}'))) {
-            /* wildcard: '{}tag' or '{*}tag' */
-            return 1;
-        }
-        for (i = 0; i < len; i++) {
-            if (p[i] == '{')
-                check = 0;
-            else if (p[i] == '}')
-                check = 1;
-            else if (check && PATHCHAR(p[i]))
-                return 1;
-        }
-        return 0;
-    }
 
     return 1; /* unknown type; might be path expression */
 }
@@ -1552,10 +1534,6 @@ _elementtree_Element_iter_impl(ElementObject *self, PyTypeObject *cls,
         if (PyUnicode_GET_LENGTH(tag) == 1 && PyUnicode_READ_CHAR(tag, 0) == '*')
             tag = Py_None;
     }
-    else if (PyBytes_Check(tag)) {
-        if (PyBytes_GET_SIZE(tag) == 1 && *PyBytes_AS_STRING(tag) == '*')
-            tag = Py_None;
-    }
 
     elementtreestate *st = get_elementtree_state_by_cls(cls);
     return create_elementiter(st, self, tag, 0);
@@ -2935,17 +2913,7 @@ treebuilder_handle_data(TreeBuilderObject* self, PyObject* data)
         self->data = Py_NewRef(data);
     } else {
         /* more than one item; use a list to collect items */
-        if (PyBytes_CheckExact(self->data)
-            && _PyObject_IsUniquelyReferenced(self->data)
-            && PyBytes_CheckExact(data) && PyBytes_GET_SIZE(data) == 1) {
-            /* XXX this code path unused in Python 3? */
-            /* expat often generates single character data sections; handle
-               the most common case by resizing the existing string... */
-            Py_ssize_t size = PyBytes_GET_SIZE(self->data);
-            if (_PyBytes_Resize(&self->data, size + 1) < 0)
-                return NULL;
-            PyBytes_AS_STRING(self->data)[size] = PyBytes_AS_STRING(data)[0];
-        } else if (PyList_CheckExact(self->data)) {
+        if (PyList_CheckExact(self->data)) {
             if (PyList_Append(self->data, data) < 0)
                 return NULL;
         } else {
@@ -4363,18 +4331,14 @@ _elementtree_XMLParser__setevents_impl(XMLParserObject *self,
 
     for (i = 0; i < PySequence_Fast_GET_SIZE(events_seq); ++i) {
         PyObject *event_name_obj = PySequence_Fast_GET_ITEM(events_seq, i);
-        const char *event_name = NULL;
-        if (PyUnicode_Check(event_name_obj)) {
-            event_name = PyUnicode_AsUTF8(event_name_obj);
-        } else if (PyBytes_Check(event_name_obj)) {
-            event_name = PyBytes_AS_STRING(event_name_obj);
+        if (!PyUnicode_Check(event_name_obj)) {
+            goto unknown_event;
         }
+        const char *event_name = PyUnicode_AsUTF8(event_name_obj);
         if (event_name == NULL) {
             Py_DECREF(events_seq);
-            PyErr_Format(PyExc_ValueError, "invalid events sequence");
             return NULL;
         }
-
         if (strcmp(event_name, "start") == 0) {
             Py_XSETREF(target->start_event_obj, Py_NewRef(event_name_obj));
         } else if (strcmp(event_name, "end") == 0) {
@@ -4406,7 +4370,8 @@ _elementtree_XMLParser__setevents_impl(XMLParserObject *self,
                 (XML_ProcessingInstructionHandler) expat_pi_handler
                 );
         } else {
-            PyErr_Format(PyExc_ValueError, "unknown event '%s'", event_name);
+unknown_event:
+            PyErr_Format(PyExc_ValueError, "unknown event %R", event_name_obj);
             Py_DECREF(events_seq);
             return NULL;
         }

Objects/bytearrayobject.c

@@ -256,6 +256,7 @@ bytearray_resize_storage(PyByteArrayObject *self,
         bytearray_write_trailing_null_byte(self);
         return -1;
     }
+    assert(_PyBytes_IsMutable(self->ob_bytes_object));
     return 0;
 }