Issue #25262. Added support for BINBYTES8 opcode in Python implementation of

unpickler.  Highest 32 bits of 64-bit size for BINUNICODE8 and BINBYTES8
opcodes no longer silently ignored on 32-bit platforms in C implementation.

1 files changed, 161 insertions, 108 deletions

diff --git a/Modules/_io/stringio.c b/Modules/_io/stringio.c
index 9d73884f73..8315ab8e70 100644
--- a/Modules/_io/stringio.c
+++ b/Modules/_io/stringio.c
@@ -11,6 +11,12 @@
 #define STATE_REALIZED 1
 #define STATE_ACCUMULATING 2
 
+/*[clinic input]
+module _io
+class _io.StringIO "stringio *" "&PyStringIO_Type"
+[clinic start generated code]*/
+/*[clinic end generated code: output=da39a3ee5e6b4b0d input=c17bc0f42165cd7d]*/
+
 typedef struct {
     PyObject_HEAD
     Py_UCS4 *buf;
@@ -39,6 +45,8 @@ typedef struct {
     PyObject *weakreflist;
 } stringio;
 
+static int _io_StringIO___init__(PyObject *self, PyObject *args, PyObject *kwargs);
+
 #define CHECK_INITIALIZED(self) \
     if (self->ok <= 0) { \
         PyErr_SetString(PyExc_ValueError, \
@@ -58,12 +66,6 @@ typedef struct {
         return NULL; \
     }
 
-PyDoc_STRVAR(stringio_doc,
-    "Text I/O implementation using an in-memory buffer.\n"
-    "\n"
-    "The initial_value argument sets the value of object.  The newline\n"
-    "argument is like the one of TextIOWrapper's constructor.");
-
 
 /* Internal routine for changing the size, in terms of characters, of the
    buffer of StringIO objects.  The caller should ensure that the 'size'
@@ -264,11 +266,15 @@ fail:
     return -1;
 }
 
-PyDoc_STRVAR(stringio_getvalue_doc,
-    "Retrieve the entire contents of the object.");
+/*[clinic input]
+_io.StringIO.getvalue
+
+Retrieve the entire contents of the object.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_getvalue(stringio *self)
+_io_StringIO_getvalue_impl(stringio *self)
+/*[clinic end generated code: output=27b6a7bfeaebce01 input=d23cb81d6791cf88]*/
 {
     CHECK_INITIALIZED(self);
     CHECK_CLOSED(self);
@@ -278,33 +284,40 @@ stringio_getvalue(stringio *self)
                                      self->string_size);
 }
 
-PyDoc_STRVAR(stringio_tell_doc,
-    "Tell the current file position.");
+/*[clinic input]
+_io.StringIO.tell
+
+Tell the current file position.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_tell(stringio *self)
+_io_StringIO_tell_impl(stringio *self)
+/*[clinic end generated code: output=2e87ac67b116c77b input=ec866ebaff02f405]*/
 {
     CHECK_INITIALIZED(self);
     CHECK_CLOSED(self);
     return PyLong_FromSsize_t(self->pos);
 }
 
-PyDoc_STRVAR(stringio_read_doc,
-    "Read at most n characters, returned as a string.\n"
-    "\n"
-    "If the argument is negative or omitted, read until EOF\n"
-    "is reached. Return an empty string at EOF.\n");
+/*[clinic input]
+_io.StringIO.read
+    size as arg: object = None
+    /
+
+Read at most size characters, returned as a string.
+
+If the argument is negative or omitted, read until EOF
+is reached. Return an empty string at EOF.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_read(stringio *self, PyObject *args)
+_io_StringIO_read_impl(stringio *self, PyObject *arg)
+/*[clinic end generated code: output=3676864773746f68 input=9a319015f6f3965c]*/
 {
     Py_ssize_t size, n;
     Py_UCS4 *output;
-    PyObject *arg = Py_None;
 
     CHECK_INITIALIZED(self);
-    if (!PyArg_ParseTuple(args, "|O:read", &arg))
-        return NULL;
     CHECK_CLOSED(self);
 
     if (PyNumber_Check(arg)) {
@@ -373,20 +386,23 @@ _stringio_readline(stringio *self, Py_ssize_t limit)
     return PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, start, len);
 }
 
-PyDoc_STRVAR(stringio_readline_doc,
-    "Read until newline or EOF.\n"
-    "\n"
-    "Returns an empty string if EOF is hit immediately.\n");
+/*[clinic input]
+_io.StringIO.readline
+    size as arg: object = None
+    /
+
+Read until newline or EOF.
+
+Returns an empty string if EOF is hit immediately.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_readline(stringio *self, PyObject *args)
+_io_StringIO_readline_impl(stringio *self, PyObject *arg)
+/*[clinic end generated code: output=99fdcac03a3dee81 input=e0e0ed4042040176]*/
 {
-    PyObject *arg = Py_None;
     Py_ssize_t limit = -1;
 
     CHECK_INITIALIZED(self);
-    if (!PyArg_ParseTuple(args, "|O:readline", &arg))
-        return NULL;
     CHECK_CLOSED(self);
     ENSURE_REALIZED(self);
 
@@ -441,22 +457,25 @@ stringio_iternext(stringio *self)
     return line;
 }
 
-PyDoc_STRVAR(stringio_truncate_doc,
-    "Truncate size to pos.\n"
-    "\n"
-    "The pos argument defaults to the current file position, as\n"
-    "returned by tell().  The current file position is unchanged.\n"
-    "Returns the new absolute position.\n");
+/*[clinic input]
+_io.StringIO.truncate
+    pos as arg: object = None
+    /
+
+Truncate size to pos.
+
+The pos argument defaults to the current file position, as
+returned by tell().  The current file position is unchanged.
+Returns the new absolute position.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_truncate(stringio *self, PyObject *args)
+_io_StringIO_truncate_impl(stringio *self, PyObject *arg)
+/*[clinic end generated code: output=6072439c2b01d306 input=748619a494ba53ad]*/
 {
     Py_ssize_t size;
-    PyObject *arg = Py_None;
 
     CHECK_INITIALIZED(self);
-    if (!PyArg_ParseTuple(args, "|O:truncate", &arg))
-        return NULL;
     CHECK_CLOSED(self);
 
     if (PyNumber_Check(arg)) {
@@ -490,49 +509,51 @@ stringio_truncate(stringio *self, PyObject *args)
     return PyLong_FromSsize_t(size);
 }
 
-PyDoc_STRVAR(stringio_seek_doc,
-    "Change stream position.\n"
-    "\n"
-    "Seek to character offset pos relative to position indicated by whence:\n"
-    "    0  Start of stream (the default).  pos should be >= 0;\n"
-    "    1  Current position - pos must be 0;\n"
-    "    2  End of stream - pos must be 0.\n"
-    "Returns the new absolute position.\n");
+/*[clinic input]
+_io.StringIO.seek
+    pos: Py_ssize_t
+    whence: int = 0
+    /
+
+Change stream position.
+
+Seek to character offset pos relative to position indicated by whence:
+    0  Start of stream (the default).  pos should be >= 0;
+    1  Current position - pos must be 0;
+    2  End of stream - pos must be 0.
+Returns the new absolute position.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_seek(stringio *self, PyObject *args)
+_io_StringIO_seek_impl(stringio *self, Py_ssize_t pos, int whence)
+/*[clinic end generated code: output=e9e0ac9a8ae71c25 input=e3855b24e7cae06a]*/
 {
-    Py_ssize_t pos;
-    int mode = 0;
-
     CHECK_INITIALIZED(self);
-    if (!PyArg_ParseTuple(args, "n|i:seek", &pos, &mode))
-        return NULL;
     CHECK_CLOSED(self);
 
-    if (mode != 0 && mode != 1 && mode != 2) {
+    if (whence != 0 && whence != 1 && whence != 2) {
         PyErr_Format(PyExc_ValueError,
-                     "Invalid whence (%i, should be 0, 1 or 2)", mode);
+                     "Invalid whence (%i, should be 0, 1 or 2)", whence);
         return NULL;
     }
-    else if (pos < 0 && mode == 0) {
+    else if (pos < 0 && whence == 0) {
         PyErr_Format(PyExc_ValueError,
                      "Negative seek position %zd", pos);
         return NULL;
     }
-    else if (mode != 0 && pos != 0) {
+    else if (whence != 0 && pos != 0) {
         PyErr_SetString(PyExc_IOError,
                         "Can't do nonzero cur-relative seeks");
         return NULL;
     }
 
-    /* mode 0: offset relative to beginning of the string.
-       mode 1: no change to current position.
-       mode 2: change position to end of file. */
-    if (mode == 1) {
+    /* whence = 0: offset relative to beginning of the string.
+       whence = 1: no change to current position.
+       whence = 2: change position to end of file. */
+    if (whence == 1) {
         pos = self->pos;
     }
-    else if (mode == 2) {
+    else if (whence == 2) {
         pos = self->string_size;
     }
 
@@ -541,14 +562,20 @@ stringio_seek(stringio *self, PyObject *args)
     return PyLong_FromSsize_t(self->pos);
 }
 
-PyDoc_STRVAR(stringio_write_doc,
-    "Write string to file.\n"
-    "\n"
-    "Returns the number of characters written, which is always equal to\n"
-    "the length of the string.\n");
+/*[clinic input]
+_io.StringIO.write
+    s as obj: object
+    /
+
+Write string to file.
+
+Returns the number of characters written, which is always equal to
+the length of the string.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_write(stringio *self, PyObject *obj)
+_io_StringIO_write(stringio *self, PyObject *obj)
+/*[clinic end generated code: output=0deaba91a15b94da input=cf96f3b16586e669]*/
 {
     Py_ssize_t size;
 
@@ -569,14 +596,20 @@ stringio_write(stringio *self, PyObject *obj)
     return PyLong_FromSsize_t(size);
 }
 
-PyDoc_STRVAR(stringio_close_doc,
-    "Close the IO object. Attempting any further operation after the\n"
-    "object is closed will raise a ValueError.\n"
-    "\n"
-    "This method has no effect if the file is already closed.\n");
+/*[clinic input]
+_io.StringIO.close
+
+Close the IO object.
+
+Attempting any further operation after the object is closed
+will raise a ValueError.
+
+This method has no effect if the file is already closed.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_close(stringio *self)
+_io_StringIO_close_impl(stringio *self)
+/*[clinic end generated code: output=04399355cbe518f1 input=cbc10b45f35d6d46]*/
 {
     self->closed = 1;
     /* Free up some memory */
@@ -644,19 +677,25 @@ stringio_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
     return (PyObject *)self;
 }
 
+/*[clinic input]
+_io.StringIO.__init__
+    initial_value as value: object(c_default="NULL") = ''
+    newline as newline_obj: object(c_default="NULL") = '\n'
+
+Text I/O implementation using an in-memory buffer.
+
+The initial_value argument sets the value of object.  The newline
+argument is like the one of TextIOWrapper's constructor.
+[clinic start generated code]*/
+
 static int
-stringio_init(stringio *self, PyObject *args, PyObject *kwds)
+_io_StringIO___init___impl(stringio *self, PyObject *value,
+                           PyObject *newline_obj)
+/*[clinic end generated code: output=a421ea023b22ef4e input=cee2d9181b2577a3]*/
 {
-    char *kwlist[] = {"initial_value", "newline", NULL};
-    PyObject *value = NULL;
-    PyObject *newline_obj = NULL;
     char *newline = "\n";
     Py_ssize_t value_len;
 
-    if (!PyArg_ParseTupleAndKeywords(args, kwds, "|OO:__init__", kwlist,
-                                     &value, &newline_obj))
-        return -1;
-
     /* Parse the newline argument. This used to be done with the 'z'
        specifier, however this allowed any object with the buffer interface to
        be converted. Thus we have to parse it manually since we only want to
@@ -761,33 +800,45 @@ stringio_init(stringio *self, PyObject *args, PyObject *kwds)
 
 /* Properties and pseudo-properties */
 
-PyDoc_STRVAR(stringio_readable_doc,
-"readable() -> bool. Returns True if the IO object can be read.");
+/*[clinic input]
+_io.StringIO.readable
 
-PyDoc_STRVAR(stringio_writable_doc,
-"writable() -> bool. Returns True if the IO object can be written.");
-
-PyDoc_STRVAR(stringio_seekable_doc,
-"seekable() -> bool. Returns True if the IO object can be seeked.");
+Returns True if the IO object can be read.
+[clinic start generated code]*/
 
 static PyObject *
-stringio_seekable(stringio *self, PyObject *args)
+_io_StringIO_readable_impl(stringio *self)
+/*[clinic end generated code: output=b19d44dd8b1ceb99 input=39ce068b224c21ad]*/
 {
     CHECK_INITIALIZED(self);
     CHECK_CLOSED(self);
     Py_RETURN_TRUE;
 }
 
+/*[clinic input]
+_io.StringIO.writable
+
+Returns True if the IO object can be written.
+[clinic start generated code]*/
+
 static PyObject *
-stringio_readable(stringio *self, PyObject *args)
+_io_StringIO_writable_impl(stringio *self)
+/*[clinic end generated code: output=13e4dd77187074ca input=7a691353aac38835]*/
 {
     CHECK_INITIALIZED(self);
     CHECK_CLOSED(self);
     Py_RETURN_TRUE;
 }
 
+/*[clinic input]
+_io.StringIO.seekable
+
+Returns True if the IO object can be seeked.
+[clinic start generated code]*/
+
 static PyObject *
-stringio_writable(stringio *self, PyObject *args)
+_io_StringIO_seekable_impl(stringio *self)
+/*[clinic end generated code: output=4d20b4641c756879 input=4c606d05b32952e6]*/
 {
     CHECK_INITIALIZED(self);
     CHECK_CLOSED(self);
@@ -809,7 +860,7 @@ stringio_writable(stringio *self, PyObject *args)
 static PyObject *
 stringio_getstate(stringio *self)
 {
-    PyObject *initvalue = stringio_getvalue(self);
+    PyObject *initvalue = _io_StringIO_getvalue_impl(self);
     PyObject *dict;
     PyObject *state;
 
@@ -857,7 +908,7 @@ stringio_setstate(stringio *self, PyObject *state)
     initarg = PyTuple_GetSlice(state, 0, 2);
     if (initarg == NULL)
         return NULL;
-    if (stringio_init(self, initarg, NULL) < 0) {
+    if (_io_StringIO___init__((PyObject *)self, initarg, NULL) < 0) {
         Py_DECREF(initarg);
         return NULL;
     }
@@ -959,19 +1010,21 @@ stringio_newlines(stringio *self, void *context)
     return PyObject_GetAttr(self->decoder, _PyIO_str_newlines);
 }
 
+#include "clinic/stringio.c.h"
+
 static struct PyMethodDef stringio_methods[] = {
-    {"close",    (PyCFunction)stringio_close,    METH_NOARGS,  stringio_close_doc},
-    {"getvalue", (PyCFunction)stringio_getvalue, METH_NOARGS,  stringio_getvalue_doc},
-    {"read",     (PyCFunction)stringio_read,     METH_VARARGS, stringio_read_doc},
-    {"readline", (PyCFunction)stringio_readline, METH_VARARGS, stringio_readline_doc},
-    {"tell",     (PyCFunction)stringio_tell,     METH_NOARGS,  stringio_tell_doc},
-    {"truncate", (PyCFunction)stringio_truncate, METH_VARARGS, stringio_truncate_doc},
-    {"seek",     (PyCFunction)stringio_seek,     METH_VARARGS, stringio_seek_doc},
-    {"write",    (PyCFunction)stringio_write,    METH_O,       stringio_write_doc},
-
-    {"seekable", (PyCFunction)stringio_seekable, METH_NOARGS, stringio_seekable_doc},
-    {"readable", (PyCFunction)stringio_readable, METH_NOARGS, stringio_readable_doc},
-    {"writable", (PyCFunction)stringio_writable, METH_NOARGS, stringio_writable_doc},
+    _IO_STRINGIO_CLOSE_METHODDEF
+    _IO_STRINGIO_GETVALUE_METHODDEF
+    _IO_STRINGIO_READ_METHODDEF
+    _IO_STRINGIO_READLINE_METHODDEF
+    _IO_STRINGIO_TELL_METHODDEF
+    _IO_STRINGIO_TRUNCATE_METHODDEF
+    _IO_STRINGIO_SEEK_METHODDEF
+    _IO_STRINGIO_WRITE_METHODDEF
+
+    _IO_STRINGIO_SEEKABLE_METHODDEF
+    _IO_STRINGIO_READABLE_METHODDEF
+    _IO_STRINGIO_WRITABLE_METHODDEF
 
     {"__getstate__", (PyCFunction)stringio_getstate, METH_NOARGS},
     {"__setstate__", (PyCFunction)stringio_setstate, METH_O},
@@ -1013,7 +1066,7 @@ PyTypeObject PyStringIO_Type = {
     0,                                         /*tp_as_buffer*/
     Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE
                        | Py_TPFLAGS_HAVE_GC,   /*tp_flags*/
-    stringio_doc,                              /*tp_doc*/
+    _io_StringIO___init____doc__,              /*tp_doc*/
     (traverseproc)stringio_traverse,           /*tp_traverse*/
     (inquiry)stringio_clear,                   /*tp_clear*/
     0,                                         /*tp_richcompare*/
@@ -1028,7 +1081,7 @@ PyTypeObject PyStringIO_Type = {
     0,                                         /*tp_descr_get*/
     0,                                         /*tp_descr_set*/
     offsetof(stringio, dict),                  /*tp_dictoffset*/
-    (initproc)stringio_init,                   /*tp_init*/
+    _io_StringIO___init__,                     /*tp_init*/
     0,                                         /*tp_alloc*/
     stringio_new,                              /*tp_new*/
 };