bpo-39674: Revert "bpo-37330: open() no longer accept 'U' in file mode (GH-16959)" (GH-18767)

vstinner · web-flow · commit 942f7a2dea2e · 2020-03-04T18:50:22.000+01:00
This reverts commit e471e72. The mode will be removed from Python 3.10.
diff --git a/Doc/library/codecs.rst b/Doc/library/codecs.rst
@@ -197,9 +197,6 @@ wider range of codecs when working with binary files:
    *buffering* has the same meaning as for the built-in :func:`open` function.
    It defaults to -1 which means that the default buffer size will be used.
 
-   .. versionchanged:: 3.9
-      The ``'U'`` mode has been removed.
-
 
 .. function:: EncodedFile(file, data_encoding, file_encoding=None, errors='strict')
 
diff --git a/Doc/library/fileinput.rst b/Doc/library/fileinput.rst
@@ -148,8 +148,8 @@ available for subclassing as well:
    The sequence must be accessed in strictly sequential order; random access
    and :meth:`~io.TextIOBase.readline` cannot be mixed.
 
-   With *mode* you can specify which file mode will be passed to :func:`open`.
-   It must be ``'r'`` or ``'rb'``.
+   With *mode* you can specify which file mode will be passed to :func:`open`. It
+   must be one of ``'r'``, ``'rU'``, ``'U'`` and ``'rb'``.
 
    The *openhook*, when given, must be a function that takes two arguments,
    *filename* and *mode*, and returns an accordingly opened file-like object. You
@@ -166,14 +166,15 @@ available for subclassing as well:
    .. versionchanged:: 3.2
       Can be used as a context manager.
 
+   .. deprecated:: 3.4
+      The ``'rU'`` and ``'U'`` modes.
+
    .. deprecated:: 3.8
       Support for :meth:`__getitem__` method is deprecated.
 
    .. versionchanged:: 3.8
       The keyword parameter *mode* and *openhook* are now keyword-only.
 
-   .. versionchanged:: 3.9
-      The ``'rU'`` and ``'U'`` modes have been removed.
 
 
 **Optional in-place filtering:** if the keyword argument ``inplace=True`` is
diff --git a/Doc/library/functions.rst b/Doc/library/functions.rst
@@ -1090,6 +1090,12 @@ are always available.  They are listed here in alphabetical order.
    first decoded using a platform-dependent encoding or using the specified
    *encoding* if given.
 
+   There is an additional mode character permitted, ``'U'``, which no longer
+   has any effect, and is considered deprecated. It previously enabled
+   :term:`universal newlines` in text mode, which became the default behaviour
+   in Python 3.0. Refer to the documentation of the
+   :ref:`newline <open-newline-parameter>` parameter for further details.
+
    .. note::
 
       Python doesn't depend on the underlying operating system's notion of text
@@ -1246,6 +1252,10 @@ are always available.  They are listed here in alphabetical order.
 
          * The file is now non-inheritable.
 
+   .. deprecated-removed:: 3.4 3.10
+
+      The ``'U'`` mode.
+
    .. versionchanged::
       3.5
 
@@ -1261,10 +1271,6 @@ are always available.  They are listed here in alphabetical order.
          * On Windows, opening a console buffer may return a subclass of
            :class:`io.RawIOBase` other than :class:`io.FileIO`.
 
-   .. versionchanged:: 3.9
-      The ``'U'`` mode has been removed.
-
-
 .. function:: ord(c)
 
    Given a string representing one Unicode character, return an integer
diff --git a/Doc/whatsnew/3.9.rst b/Doc/whatsnew/3.9.rst
@@ -625,14 +625,6 @@ that may require changes to your code.
 Changes in the Python API
 -------------------------
 
-* :func:`open`, :func:`io.open`, :func:`codecs.open` and
-  :class:`fileinput.FileInput` no longer accept ``'U'`` ("universal newline")
-  in the file mode. This flag was deprecated since Python 3.3. In Python 3, the
-  "universal newline" is used by default when a file is open in text mode.  The
-  :ref:`newline parameter <open-newline-parameter>` of :func:`open` controls
-  how universal newlines works.
-  (Contributed by Victor Stinner in :issue:`37330`.)
-
 * :func:`__import__` and :func:`importlib.util.resolve_name` now raise
   :exc:`ImportError` where it previously raised :exc:`ValueError`. Callers
   catching the specific exception type and supporting both Python 3.9 and
diff --git a/Lib/_pyio.py b/Lib/_pyio.py
@@ -71,6 +71,7 @@ def open(file, mode="r", buffering=-1, encoding=None, errors=None,
     'b'       binary mode
     't'       text mode (default)
     '+'       open a disk file for updating (reading and writing)
+    'U'       universal newline mode (deprecated)
     ========= ===============================================================
 
     The default mode is 'rt' (open for reading text). For binary random
@@ -86,6 +87,10 @@ def open(file, mode="r", buffering=-1, encoding=None, errors=None,
     returned as strings, the bytes having been first decoded using a
     platform-dependent encoding or using the specified encoding if given.
 
+    'U' mode is deprecated and will raise an exception in future versions
+    of Python.  It has no effect in Python 3.  Use newline to control
+    universal newlines mode.
+
     buffering is an optional integer used to set the buffering policy.
     Pass 0 to switch buffering off (only allowed in binary mode), 1 to select
     line buffering (only usable in text mode), and an integer > 1 to indicate
@@ -171,7 +176,7 @@ def open(file, mode="r", buffering=-1, encoding=None, errors=None,
     if errors is not None and not isinstance(errors, str):
         raise TypeError("invalid errors: %r" % errors)
     modes = set(mode)
-    if modes - set("axrwb+t") or len(mode) > len(modes):
+    if modes - set("axrwb+tU") or len(mode) > len(modes):
         raise ValueError("invalid mode: %r" % mode)
     creating = "x" in modes
     reading = "r" in modes
@@ -180,6 +185,13 @@ def open(file, mode="r", buffering=-1, encoding=None, errors=None,
     updating = "+" in modes
     text = "t" in modes
     binary = "b" in modes
+    if "U" in modes:
+        if creating or writing or appending or updating:
+            raise ValueError("mode U cannot be combined with 'x', 'w', 'a', or '+'")
+        import warnings
+        warnings.warn("'U' mode is deprecated",
+                      DeprecationWarning, 2)
+        reading = True
     if text and binary:
         raise ValueError("can't have text and binary mode at once")
     if creating + reading + writing + appending > 1:
diff --git a/Lib/fileinput.py b/Lib/fileinput.py
@@ -209,10 +209,15 @@ def __init__(self, files=None, inplace=False, backup="", *,
         self._isstdin = False
         self._backupfilename = None
         # restrict mode argument to reading modes
-        if mode not in ('r', 'rb'):
-            raise ValueError("FileInput opening mode must be 'r' or 'rb'")
+        if mode not in ('r', 'rU', 'U', 'rb'):
+            raise ValueError("FileInput opening mode must be one of "
+                             "'r', 'rU', 'U' and 'rb'")
+        if 'U' in mode:
+            import warnings
+            warnings.warn("'U' mode is deprecated",
+                          DeprecationWarning, 2)
         self._mode = mode
-        self._write_mode = mode.replace('r', 'w')
+        self._write_mode = mode.replace('r', 'w') if 'U' not in mode else 'w'
         if openhook:
             if inplace:
                 raise ValueError("FileInput cannot use an opening hook in inplace mode")
diff --git a/Lib/imp.py b/Lib/imp.py
@@ -225,7 +225,7 @@ def load_module(name, file, filename, details):
 
     """
     suffix, mode, type_ = details
-    if mode and (not mode.startswith('r') or '+' in mode):
+    if mode and (not mode.startswith(('r', 'U')) or '+' in mode):
         raise ValueError('invalid file open mode {!r}'.format(mode))
     elif file is None and type_ in {PY_SOURCE, PY_COMPILED}:
         msg = 'file object required for import (type code {})'.format(type_)
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
@@ -712,23 +712,11 @@ def test_bug691291(self):
         self.addCleanup(support.unlink, support.TESTFN)
         with open(support.TESTFN, 'wb') as fp:
             fp.write(s)
-        with codecs.open(support.TESTFN, 'r',
-                         encoding=self.encoding) as reader:
+        with support.check_warnings(('', DeprecationWarning)):
+            reader = codecs.open(support.TESTFN, 'U', encoding=self.encoding)
+        with reader:
             self.assertEqual(reader.read(), s1)
 
-    def test_invalid_modes(self):
-        for mode in ('U', 'rU', 'r+U'):
-            with self.assertRaises(ValueError) as cm:
-                codecs.open(support.TESTFN, mode, encoding=self.encoding)
-            self.assertIn('invalid mode', str(cm.exception))
-
-        for mode in ('rt', 'wt', 'at', 'r+t'):
-            with self.assertRaises(ValueError) as cm:
-                codecs.open(support.TESTFN, mode, encoding=self.encoding)
-            self.assertIn("can't have text and binary mode at once",
-                          str(cm.exception))
-
-
 class UTF16LETest(ReadTest, unittest.TestCase):
     encoding = "utf-16-le"
     ill_formed_sequence = b"\x80\xdc"
diff --git a/Lib/test/test_fileinput.py b/Lib/test/test_fileinput.py
@@ -226,11 +226,19 @@ def test_fileno(self):
         self.assertEqual(fi.fileno(), -1)
 
     def test_opening_mode(self):
-        # invalid modes
-        for mode in ('w', 'rU', 'U'):
-            with self.subTest(mode=mode):
-                with self.assertRaises(ValueError):
-                    FileInput(mode=mode)
+        try:
+            # invalid mode, should raise ValueError
+            fi = FileInput(mode="w")
+            self.fail("FileInput should reject invalid mode argument")
+        except ValueError:
+            pass
+        # try opening in universal newline mode
+        t1 = self.writeTmp(b"A\nB\r\nC\rD", mode="wb")
+        with check_warnings(('', DeprecationWarning)):
+            fi = FileInput(files=t1, mode="U")
+        with check_warnings(('', DeprecationWarning)):
+            lines = list(fi)
+        self.assertEqual(lines, ["A\n", "B\n", "C\n", "D"])
 
     def test_stdin_binary_mode(self):
         with mock.patch('sys.stdin') as m_stdin:
@@ -977,6 +985,10 @@ def check(mode, expected_lines):
             self.assertEqual(lines, expected_lines)
 
         check('r', ['A\n', 'B\n', 'C\n', 'D\u20ac'])
+        with self.assertWarns(DeprecationWarning):
+            check('rU', ['A\n', 'B\n', 'C\n', 'D\u20ac'])
+        with self.assertWarns(DeprecationWarning):
+            check('U', ['A\n', 'B\n', 'C\n', 'D\u20ac'])
         with self.assertRaises(ValueError):
             check('rb', ['A\n', 'B\r\n', 'C\r', 'D\u20ac'])
 
diff --git a/Lib/test/test_io.py b/Lib/test/test_io.py
@@ -3900,6 +3900,16 @@ def test_attributes(self):
         self.assertEqual(f.mode, "wb")
         f.close()
 
+        with support.check_warnings(('', DeprecationWarning)):
+            f = self.open(support.TESTFN, "U")
+        self.assertEqual(f.name,            support.TESTFN)
+        self.assertEqual(f.buffer.name,     support.TESTFN)
+        self.assertEqual(f.buffer.raw.name, support.TESTFN)
+        self.assertEqual(f.mode,            "U")
+        self.assertEqual(f.buffer.mode,     "rb")
+        self.assertEqual(f.buffer.raw.mode, "rb")
+        f.close()
+
         f = self.open(support.TESTFN, "w+")
         self.assertEqual(f.mode,            "w+")
         self.assertEqual(f.buffer.mode,     "rb+") # Does it really matter?
@@ -3913,13 +3923,6 @@ def test_attributes(self):
         f.close()
         g.close()
 
-    def test_removed_u_mode(self):
-        # "U" mode has been removed in Python 3.9
-        for mode in ("U", "rU", "r+U"):
-            with self.assertRaises(ValueError) as cm:
-                self.open(support.TESTFN, mode)
-            self.assertIn('invalid mode', str(cm.exception))
-
     def test_open_pipe_with_append(self):
         # bpo-27805: Ignore ESPIPE from lseek() in open().
         r, w = os.pipe()
diff --git a/Misc/NEWS.d/next/Library/2020-03-03-16-21-41.bpo-39674.HJVkD5.rst b/Misc/NEWS.d/next/Library/2020-03-03-16-21-41.bpo-39674.HJVkD5.rst
@@ -0,0 +1,3 @@
+Revert "bpo-37330: open() no longer accept 'U' in file mode". The "U" mode of
+open() is kept in Python 3.9 to ease transition from Python 2.7, but will be
+removed in Python 3.10.
diff --git a/Modules/_io/_iomodule.c b/Modules/_io/_iomodule.c
@@ -138,6 +138,7 @@ Character Meaning
 'b'       binary mode
 't'       text mode (default)
 '+'       open a disk file for updating (reading and writing)
+'U'       universal newline mode (deprecated)
 ========= ===============================================================
 
 The default mode is 'rt' (open for reading text). For binary random
@@ -153,6 +154,10 @@ bytes objects without any decoding. In text mode (the default, or when
 returned as strings, the bytes having been first decoded using a
 platform-dependent encoding or using the specified encoding if given.
 
+'U' mode is deprecated and will raise an exception in future versions
+of Python.  It has no effect in Python 3.  Use newline to control
+universal newlines mode.
+
 buffering is an optional integer used to set the buffering policy.
 Pass 0 to switch buffering off (only allowed in binary mode), 1 to select
 line buffering (only usable in text mode), and an integer > 1 to indicate
@@ -228,12 +233,12 @@ static PyObject *
 _io_open_impl(PyObject *module, PyObject *file, const char *mode,
               int buffering, const char *encoding, const char *errors,
               const char *newline, int closefd, PyObject *opener)
-/*[clinic end generated code: output=aefafc4ce2b46dc0 input=1543f4511d2356a5]*/
+/*[clinic end generated code: output=aefafc4ce2b46dc0 input=7295902222e6b311]*/
 {
     unsigned i;
 
     int creating = 0, reading = 0, writing = 0, appending = 0, updating = 0;
-    int text = 0, binary = 0;
+    int text = 0, binary = 0, universal = 0;
 
     char rawmode[6], *m;
     int line_buffering, is_number;
@@ -291,6 +296,10 @@ _io_open_impl(PyObject *module, PyObject *file, const char *mode,
         case 'b':
             binary = 1;
             break;
+        case 'U':
+            universal = 1;
+            reading = 1;
+            break;
         default:
             goto invalid_mode;
         }
@@ -313,6 +322,18 @@ _io_open_impl(PyObject *module, PyObject *file, const char *mode,
     *m = '\0';
 
     /* Parameters validation */
+    if (universal) {
+        if (creating || writing || appending || updating) {
+            PyErr_SetString(PyExc_ValueError,
+                            "mode U cannot be combined with 'x', 'w', 'a', or '+'");
+            goto error;
+        }
+        if (PyErr_WarnEx(PyExc_DeprecationWarning,
+                         "'U' mode is deprecated", 1) < 0)
+            goto error;
+        reading = 1;
+    }
+
     if (text && binary) {
         PyErr_SetString(PyExc_ValueError,
                         "can't have text and binary mode at once");
diff --git a/Modules/_io/clinic/_iomodule.c.h b/Modules/_io/clinic/_iomodule.c.h
@@ -36,6 +36,7 @@ PyDoc_STRVAR(_io_open__doc__,
 "\'b\'       binary mode\n"
 "\'t\'       text mode (default)\n"
 "\'+\'       open a disk file for updating (reading and writing)\n"
+"\'U\'       universal newline mode (deprecated)\n"
 "========= ===============================================================\n"
 "\n"
 "The default mode is \'rt\' (open for reading text). For binary random\n"
@@ -51,6 +52,10 @@ PyDoc_STRVAR(_io_open__doc__,
 "returned as strings, the bytes having been first decoded using a\n"
 "platform-dependent encoding or using the specified encoding if given.\n"
 "\n"
+"\'U\' mode is deprecated and will raise an exception in future versions\n"
+"of Python.  It has no effect in Python 3.  Use newline to control\n"
+"universal newlines mode.\n"
+"\n"
 "buffering is an optional integer used to set the buffering policy.\n"
 "Pass 0 to switch buffering off (only allowed in binary mode), 1 to select\n"
 "line buffering (only usable in text mode), and an integer > 1 to indicate\n"
@@ -318,4 +323,4 @@ _io_open_code(PyObject *module, PyObject *const *args, Py_ssize_t nargs, PyObjec
 exit:
     return return_value;
 }
-/*[clinic end generated code: output=680e4b488c7da8a1 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=3df6bc6d91697545 input=a9049054013a1b77]*/
