Raise ValueError instead of UnicodeDecodeError.

Author: serhiy-storchaka

Lib/test/test_struct.py

@@ -930,13 +930,13 @@ def __init__(self, newargs, initargs):
                 MyStruct(('>h',), ('$',))
         with self.assertRaises(ValueError):
             MyStruct(('\udc00',), ('>h',))
-        with self.assertRaises(UnicodeDecodeError):
+        with self.assertRaises(ValueError):
             MyStruct((b'\xa4',), ('>h',))
         with self.assertWarns(FutureWarning):
             with self.assertRaises(ValueError):
                 MyStruct(('>h',), ('\udc00',))
         with self.assertWarns(FutureWarning):
-            with self.assertRaises(UnicodeDecodeError):
+            with self.assertRaises(ValueError):
                 MyStruct(('>h',), (b'\xa4',))
         with self.assertWarns(FutureWarning):
             my_struct = MyStruct(('>h',), ('<h',))
@@ -961,7 +961,7 @@ class MyStruct(struct.Struct):
             MyStruct('$')
         with self.assertRaises(ValueError):
             MyStruct('\udc00')
-        with self.assertRaises(UnicodeDecodeError):
+        with self.assertRaises(ValueError):
             MyStruct(b'\xa4')
         with self.assertRaises(TypeError):
             MyStruct('>h', 42)

Misc/NEWS.d/next/Library/2026-03-12-12-17-39.gh-issue-145850.uW3stt.rst

@@ -1,6 +1,6 @@
 Changed some implementation details in :class:`struct.Struct`: calling it
 with non-ASCII string format will now raise a :exc:`ValueError` instead of
 :exc:`UnicodeEncodeError`, calling it with non-ASCII bytes format will now
-raise a :exc:`UnicodeDecodeError` instead of :exc:`struct.error`, getting
+raise a :exc:`ValueError` instead of :exc:`struct.error`, getting
 the :attr:`!format` attribute of uninitialized object will now raise an
 :exc:`AttributeError` instead of :exc:`RuntimeError`.

Modules/_struct.c

@@ -1635,7 +1635,10 @@ prepare_s(PyStructObject *self, PyObject *format)
 
     _structmodulestate *state = get_struct_state_structinst(self);
 
-    assert(PyUnicode_IS_ASCII(format));
+    if (!PyUnicode_IS_ASCII(format)) {
+        PyErr_SetString(PyExc_ValueError, "non-ASCII character in struct format");
+        return -1;
+    }
     fmt = (const char *)PyUnicode_1BYTE_DATA(format);
     if (strlen(fmt) != (size_t)PyUnicode_GET_LENGTH(format)) {
         PyErr_SetString(state->StructError,
@@ -1781,15 +1784,11 @@ static int
 set_format(PyStructObject *self, PyObject *format)
 {
     if (PyUnicode_Check(format)) {
-        if (!PyUnicode_IS_ASCII(format)) {
-            PyErr_SetString(PyExc_ValueError, "non-ASCII character in struct format");
-            return -1;
-        }
         format = PyUnicode_FromObject(format);
     }
     else if (PyBytes_Check(format)) {
         format = PyUnicode_DecodeASCII(PyBytes_AS_STRING(format),
-                                       PyBytes_GET_SIZE(format), NULL);
+                                       PyBytes_GET_SIZE(format), "surrogateescape");
     }
     else  {
         PyErr_Format(PyExc_TypeError,