gh-99612: Fix PyUnicode_DecodeUTF8Stateful() for ASCII-only data (GH-99613)

Previously *consumed was not set in this case.

1 files changed, 36 insertions, 0 deletions

diff --git a/Modules/_testcapi/unicode.c b/Modules/_testcapi/unicode.c
index 4c5049d..2d23993 100644
--- a/Modules/_testcapi/unicode.c
+++ b/Modules/_testcapi/unicode.c
@@ -239,6 +239,40 @@ unicode_asutf8andsize(PyObject *self, PyObject *args)
     return Py_BuildValue("(Nn)", result, utf8_len);
 }
 
+/* Test PyUnicode_DecodeUTF8() */
+static PyObject *
+unicode_decodeutf8(PyObject *self, PyObject *args)
+{
+    const char *data;
+    Py_ssize_t size;
+    const char *errors = NULL;
+
+    if (!PyArg_ParseTuple(args, "y#|z", &data, &size, &errors))
+        return NULL;
+
+    return PyUnicode_DecodeUTF8(data, size, errors);
+}
+
+/* Test PyUnicode_DecodeUTF8Stateful() */
+static PyObject *
+unicode_decodeutf8stateful(PyObject *self, PyObject *args)
+{
+    const char *data;
+    Py_ssize_t size;
+    const char *errors = NULL;
+    Py_ssize_t consumed = 123456789;
+    PyObject *result;
+
+    if (!PyArg_ParseTuple(args, "y#|z", &data, &size, &errors))
+        return NULL;
+
+    result = PyUnicode_DecodeUTF8Stateful(data, size, errors, &consumed);
+    if (!result) {
+        return NULL;
+    }
+    return Py_BuildValue("(Nn)", result, consumed);
+}
+
 /* Test PyUnicode_Concat() */
 static PyObject *
 unicode_concat(PyObject *self, PyObject *args)
@@ -1025,6 +1059,8 @@ static PyMethodDef TestMethods[] = {
     {"unicode_asucs4",           unicode_asucs4,                 METH_VARARGS},
     {"unicode_asutf8",           unicode_asutf8,                 METH_VARARGS},
     {"unicode_asutf8andsize",    unicode_asutf8andsize,          METH_VARARGS},
+    {"unicode_decodeutf8",       unicode_decodeutf8,             METH_VARARGS},
+    {"unicode_decodeutf8stateful",unicode_decodeutf8stateful,    METH_VARARGS},
     {"unicode_concat",           unicode_concat,                 METH_VARARGS},
     {"unicode_splitlines",       unicode_splitlines,             METH_VARARGS},
     {"unicode_split",            unicode_split,                  METH_VARARGS},