bpo-31718: Fix io.IncrementalNewlineDecoder SystemErrors and segfaults

Lib/test/test_io.py

@@ -3871,7 +3871,15 @@ def test_translate(self):
         self.assertEqual(decoder.decode(b"\r\r\n"), "\r\r\n")
 
 class CIncrementalNewlineDecoderTest(IncrementalNewlineDecoderTest):
-    pass
+    @support.cpython_only
+    def test_uninitialized(self):
+        uninitialized = self.IncrementalNewlineDecoder.__new__(
+            self.IncrementalNewlineDecoder)
+        self.assertRaises(ValueError, uninitialized.decode, b'bar')
+        self.assertRaises(ValueError, uninitialized.getstate)
+        self.assertRaises(ValueError, uninitialized.setstate, (b'foo', 0))
+        self.assertRaises(ValueError, uninitialized.reset)
+
 
 class PyIncrementalNewlineDecoderTest(IncrementalNewlineDecoderTest):
     pass

Misc/NEWS.d/next/Core and Builtins/2020-02-23-23-48-15.bpo-31718.sXko5e.rst

@@ -0,0 +1,3 @@
+Raise :exc:`ValueError` instead of :exc:`SystemError` when methods of
+uninitialized :class:`io.IncrementalNewlineDecoder` objects are called.
+Patch by Oren Milman.

Modules/_io/textio.c

@@ -248,19 +248,19 @@ _io_IncrementalNewlineDecoder___init___impl(nldecoder_object *self,
                                             PyObject *errors)
 /*[clinic end generated code: output=fbd04d443e764ec2 input=89db6b19c6b126bf]*/
 {
-    self->decoder = decoder;
-    Py_INCREF(decoder);
-
     if (errors == NULL) {
-        self->errors = _PyUnicode_FromId(&PyId_strict);
+        Py_XSETREF(self->errors, _PyUnicode_FromId(&PyId_strict));
         if (self->errors == NULL)
             return -1;
     }
     else {
-        self->errors = errors;
+        Py_XSETREF(self->errors, errors);
     }
     Py_INCREF(self->errors);
 
+    Py_INCREF(decoder);
+    Py_XSETREF(self->decoder, decoder);
+
     self->translate = translate ? 1 : 0;
     self->seennl = 0;
     self->pendingcr = 0;
@@ -295,6 +295,13 @@ check_decoded(PyObject *decoded)
     return 0;
 }
 
+#define CHECK_INITIALIZED_DECODER(self) \
+    if (self->errors == NULL) { \
+        PyErr_SetString(PyExc_ValueError, \
+                        "IncrementalNewlineDecoder.__init__() not called"); \
+        return NULL; \
+    }
+
 #define SEEN_CR   1
 #define SEEN_LF   2
 #define SEEN_CRLF 4
@@ -308,11 +315,7 @@ _PyIncrementalNewlineDecoder_decode(PyObject *myself,
     Py_ssize_t output_len;
     nldecoder_object *self = (nldecoder_object *) myself;
 
-    if (self->decoder == NULL) {
-        PyErr_SetString(PyExc_ValueError,
-                        "IncrementalNewlineDecoder.__init__ not called");
-        return NULL;
-    }
+    CHECK_INITIALIZED_DECODER(self);
 
     /* decode input (with the eventual \r from a previous pass) */
     if (self->decoder != Py_None) {
@@ -526,6 +529,8 @@ _io_IncrementalNewlineDecoder_getstate_impl(nldecoder_object *self)
     PyObject *buffer;
     unsigned long long flag;
 
+    CHECK_INITIALIZED_DECODER(self);
+
     if (self->decoder != Py_None) {
         PyObject *state = PyObject_CallMethodNoArgs(self->decoder,
            _PyIO_str_getstate);
@@ -570,6 +575,8 @@ _io_IncrementalNewlineDecoder_setstate(nldecoder_object *self,
     PyObject *buffer;
     unsigned long long flag;
 
+    CHECK_INITIALIZED_DECODER(self);
+
     if (!PyTuple_Check(state)) {
         PyErr_SetString(PyExc_TypeError, "state argument must be a tuple");
         return NULL;
@@ -598,6 +605,8 @@ static PyObject *
 _io_IncrementalNewlineDecoder_reset_impl(nldecoder_object *self)
 /*[clinic end generated code: output=32fa40c7462aa8ff input=728678ddaea776df]*/
 {
+    CHECK_INITIALIZED_DECODER(self);
+
     self->seennl = 0;
     self->pendingcr = 0;
     if (self->decoder != Py_None)
@@ -609,6 +618,8 @@ _io_IncrementalNewlineDecoder_reset_impl(nldecoder_object *self)
 static PyObject *
 incrementalnewlinedecoder_newlines_get(nldecoder_object *self, void *context)
 {
+    CHECK_INITIALIZED_DECODER(self);
+
     switch (self->seennl) {
     case SEEN_CR:
         return PyUnicode_FromString("\r");