Issue #27029: Removed deprecated support of universal newlines mode from ZipFile.open().

serhiy-storchaka · serhiy-storchaka · commit e670be2273f9 · 2016-06-11T19:32:44.000+03:00
diff --git a/Doc/library/zipfile.rst b/Doc/library/zipfile.rst
@@ -204,18 +204,13 @@ ZipFile Objects
    Return a list of archive members by name.
 
 
-.. index::
-   single: universal newlines; zipfile.ZipFile.open method
-
 .. method:: ZipFile.open(name, mode='r', pwd=None, *, force_zip64=False)
 
-   Access a member of the archive as a file-like object.  *name*
-   is the name of the file in the archive, or a :class:`ZipInfo` object.  The
-   *mode* parameter, if included, must be one of the following: ``'r'`` (the
-   default), ``'U'``, ``'rU'`` or ``'w'``.  Choosing ``'U'`` or  ``'rU'`` will
-   enable :term:`universal newlines` support in the read-only object.  *pwd* is
-   the password used to decrypt encrypted ZIP files.  Calling  :meth:`.open` on
-   a closed ZipFile will raise a  :exc:`RuntimeError`.
+   Access a member of the archive as a binary file-like object.  *name*
+   can be either the name of a file within the archive or a :class:`ZipInfo`
+   object.  The *mode* parameter, if included, must be ``'r'`` (the default)
+   or ``'w'``.  *pwd* is the password used to decrypt encrypted ZIP files.
+   Calling :meth:`.open` on a closed ZipFile will raise a :exc:`RuntimeError`.
 
    :meth:`~ZipFile.open` is also a context manager and therefore supports the
    :keyword:`with` statement::
@@ -224,7 +219,7 @@ ZipFile Objects
           with myzip.open('eggs.txt') as myfile:
               print(myfile.read())
 
-   With *mode* ``'r'``, ``'U'`` or ``'rU'``, the file-like object
+   With *mode* ``'r'`` the file-like object
    (``ZipExtFile``) is read-only and provides the following methods:
    :meth:`~io.BufferedIOBase.read`, :meth:`~io.IOBase.readline`,
    :meth:`~io.IOBase.readlines`, :meth:`__iter__`,
@@ -248,8 +243,8 @@ ZipFile Objects
       or a :class:`ZipInfo` object.  You will appreciate this when trying to read a
       ZIP file that contains members with duplicate names.
 
-   .. deprecated-removed:: 3.4 3.6
-      The ``'U'`` or  ``'rU'`` mode.  Use :class:`io.TextIOWrapper` for reading
+   .. versionchanged:: 3.6
+      Removed support of ``mode='U'``.  Use :class:`io.TextIOWrapper` for reading
       compressed text files in :term:`universal newlines` mode.
 
    .. versionchanged:: 3.6
diff --git a/Doc/whatsnew/3.6.rst b/Doc/whatsnew/3.6.rst
@@ -580,6 +580,11 @@ API and Feature Removals
   :mod:`tkinter` widget classes were removed (corresponding Tk commands
   were obsolete since Tk 4.0).
 
+* The :meth:`~zipfile.ZipFile.open` method of the :class:`zipfile.ZipFile`
+  class no longer supports the ``'U'`` mode (was deprecated since Python 3.4).
+  Use :class:`io.TextIOWrapper` for reading compressed text files in
+  :term:`universal newlines` mode.
+
 
 Porting to Python 3.6
 =====================
diff --git a/Lib/test/test_zipfile.py b/Lib/test/test_zipfile.py
@@ -38,10 +38,6 @@ def get_files(test):
         yield f
         test.assertFalse(f.closed)
 
-def openU(zipfp, fn):
-    with check_warnings(('', DeprecationWarning)):
-        return zipfp.open(fn, 'rU')
-
 class AbstractTestsWithSourceFile:
     @classmethod
     def setUpClass(cls):
@@ -1035,32 +1031,6 @@ def test_open_via_zip_info(self):
                 data += zipfp.read(info)
             self.assertIn(data, {b"foobar", b"barfoo"})
 
-    def test_universal_deprecation(self):
-        f = io.BytesIO()
-        with zipfile.ZipFile(f, "w") as zipfp:
-            zipfp.writestr('spam.txt', b'ababagalamaga')
-
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for mode in 'U', 'rU':
-                with self.assertWarns(DeprecationWarning):
-                    zipopen = zipfp.open('spam.txt', mode)
-                zipopen.close()
-
-    def test_universal_readaheads(self):
-        f = io.BytesIO()
-
-        data = b'a\r\n' * 16 * 1024
-        with zipfile.ZipFile(f, 'w', zipfile.ZIP_STORED) as zipfp:
-            zipfp.writestr(TESTFN, data)
-
-        data2 = b''
-        with zipfile.ZipFile(f, 'r') as zipfp, \
-             openU(zipfp, TESTFN) as zipopen:
-            for line in zipopen:
-                data2 += line
-
-        self.assertEqual(data, data2.replace(b'\n', b'\r\n'))
-
     def test_writestr_extended_local_header_issue1202(self):
         with zipfile.ZipFile(TESTFN2, 'w') as orig_zip:
             for data in 'abcdefghijklmnop':
@@ -1268,9 +1238,12 @@ def test_bad_open_mode(self):
             zipf.writestr("foo.txt", "O, for a Muse of Fire!")
 
         with zipfile.ZipFile(TESTFN, mode="r") as zipf:
-        # read the data to make sure the file is there
+            # read the data to make sure the file is there
             zipf.read("foo.txt")
             self.assertRaises(RuntimeError, zipf.open, "foo.txt", "q")
+            # universal newlines support is removed
+            self.assertRaises(RuntimeError, zipf.open, "foo.txt", "U")
+            self.assertRaises(RuntimeError, zipf.open, "foo.txt", "rU")
 
     def test_read0(self):
         """Check that calling read(0) on a ZipExtFile object returns an empty
@@ -2011,138 +1984,6 @@ def tearDown(self):
             unlink(TESTFN)
 
 
-class AbstractUniversalNewlineTests:
-    @classmethod
-    def setUpClass(cls):
-        cls.line_gen = [bytes("Test of zipfile line %d." % i, "ascii")
-                        for i in range(FIXEDTEST_SIZE)]
-        cls.seps = (b'\r', b'\r\n', b'\n')
-        cls.arcdata = {}
-        for n, s in enumerate(cls.seps):
-            cls.arcdata[s] = s.join(cls.line_gen) + s
-
-    def setUp(self):
-        self.arcfiles = {}
-        for n, s in enumerate(self.seps):
-            self.arcfiles[s] = '%s-%d' % (TESTFN, n)
-            with open(self.arcfiles[s], "wb") as f:
-                f.write(self.arcdata[s])
-
-    def make_test_archive(self, f, compression):
-        # Create the ZIP archive
-        with zipfile.ZipFile(f, "w", compression) as zipfp:
-            for fn in self.arcfiles.values():
-                zipfp.write(fn, fn)
-
-    def read_test(self, f, compression):
-        self.make_test_archive(f, compression)
-
-        # Read the ZIP archive
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for sep, fn in self.arcfiles.items():
-                with openU(zipfp, fn) as fp:
-                    zipdata = fp.read()
-                self.assertEqual(self.arcdata[sep], zipdata)
-
-    def test_read(self):
-        for f in get_files(self):
-            self.read_test(f, self.compression)
-
-    def readline_read_test(self, f, compression):
-        self.make_test_archive(f, compression)
-
-        # Read the ZIP archive
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for sep, fn in self.arcfiles.items():
-                with openU(zipfp, fn) as zipopen:
-                    data = b''
-                    while True:
-                        read = zipopen.readline()
-                        if not read:
-                            break
-                        data += read
-
-                        read = zipopen.read(5)
-                        if not read:
-                            break
-                        data += read
-
-            self.assertEqual(data, self.arcdata[b'\n'])
-
-    def test_readline_read(self):
-        for f in get_files(self):
-            self.readline_read_test(f, self.compression)
-
-    def readline_test(self, f, compression):
-        self.make_test_archive(f, compression)
-
-        # Read the ZIP archive
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for sep, fn in self.arcfiles.items():
-                with openU(zipfp, fn) as zipopen:
-                    for line in self.line_gen:
-                        linedata = zipopen.readline()
-                        self.assertEqual(linedata, line + b'\n')
-
-    def test_readline(self):
-        for f in get_files(self):
-            self.readline_test(f, self.compression)
-
-    def readlines_test(self, f, compression):
-        self.make_test_archive(f, compression)
-
-        # Read the ZIP archive
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for sep, fn in self.arcfiles.items():
-                with openU(zipfp, fn) as fp:
-                    ziplines = fp.readlines()
-                for line, zipline in zip(self.line_gen, ziplines):
-                    self.assertEqual(zipline, line + b'\n')
-
-    def test_readlines(self):
-        for f in get_files(self):
-            self.readlines_test(f, self.compression)
-
-    def iterlines_test(self, f, compression):
-        self.make_test_archive(f, compression)
-
-        # Read the ZIP archive
-        with zipfile.ZipFile(f, "r") as zipfp:
-            for sep, fn in self.arcfiles.items():
-                with openU(zipfp, fn) as fp:
-                    for line, zipline in zip(self.line_gen, fp):
-                        self.assertEqual(zipline, line + b'\n')
-
-    def test_iterlines(self):
-        for f in get_files(self):
-            self.iterlines_test(f, self.compression)
-
-    def tearDown(self):
-        for sep, fn in self.arcfiles.items():
-            unlink(fn)
-        unlink(TESTFN)
-        unlink(TESTFN2)
-
-
-class StoredUniversalNewlineTests(AbstractUniversalNewlineTests,
-                                  unittest.TestCase):
-    compression = zipfile.ZIP_STORED
-
-@requires_zlib
-class DeflateUniversalNewlineTests(AbstractUniversalNewlineTests,
-                                   unittest.TestCase):
-    compression = zipfile.ZIP_DEFLATED
-
-@requires_bz2
-class Bzip2UniversalNewlineTests(AbstractUniversalNewlineTests,
-                                 unittest.TestCase):
-    compression = zipfile.ZIP_BZIP2
-
-@requires_lzma
-class LzmaUniversalNewlineTests(AbstractUniversalNewlineTests,
-                                unittest.TestCase):
-    compression = zipfile.ZIP_LZMA
-
 class ZipInfoTests(unittest.TestCase):
     def test_from_file(self):
         zi = zipfile.ZipInfo.from_file(__file__)
diff --git a/Lib/zipfile.py b/Lib/zipfile.py
@@ -743,9 +743,6 @@ class ZipExtFile(io.BufferedIOBase):
     # Read from compressed files in 4k blocks.
     MIN_READ_SIZE = 4096
 
-    # Search for universal newlines or line chunks.
-    PATTERN = re.compile(br'^(?P<chunk>[^\r\n]+)|(?P<newline>\n|\r\n?)')
-
     def __init__(self, fileobj, mode, zipinfo, decrypter=None,
                  close_fileobj=False):
         self._fileobj = fileobj
@@ -762,7 +759,6 @@ def __init__(self, fileobj, mode, zipinfo, decrypter=None,
         self._readbuffer = b''
         self._offset = 0
 
-        self._universal = 'U' in mode
         self.newlines = None
 
         # Adjust read size for encrypted files since the first 12 bytes
@@ -799,49 +795,15 @@ def readline(self, limit=-1):
         If limit is specified, at most limit bytes will be read.
         """
 
-        if not self._universal and limit < 0:
+        if limit < 0:
             # Shortcut common case - newline found in buffer.
             i = self._readbuffer.find(b'\n', self._offset) + 1
             if i > 0:
                 line = self._readbuffer[self._offset: i]
                 self._offset = i
                 return line
 
-        if not self._universal:
-            return io.BufferedIOBase.readline(self, limit)
-
-        line = b''
-        while limit < 0 or len(line) < limit:
-            readahead = self.peek(2)
-            if readahead == b'':
-                return line
-
-            #
-            # Search for universal newlines or line chunks.
-            #
-            # The pattern returns either a line chunk or a newline, but not
-            # both. Combined with peek(2), we are assured that the sequence
-            # '\r\n' is always retrieved completely and never split into
-            # separate newlines - '\r', '\n' due to coincidental readaheads.
-            #
-            match = self.PATTERN.search(readahead)
-            newline = match.group('newline')
-            if newline is not None:
-                if self.newlines is None:
-                    self.newlines = []
-                if newline not in self.newlines:
-                    self.newlines.append(newline)
-                self._offset += len(newline)
-                return line + b'\n'
-
-            chunk = match.group('chunk')
-            if limit >= 0:
-                chunk = chunk[: limit - len(line)]
-
-            self._offset += len(chunk)
-            line += chunk
-
-        return line
+        return io.BufferedIOBase.readline(self, limit)
 
     def peek(self, n=1):
         """Returns buffered bytes without advancing the position."""
@@ -1360,12 +1322,8 @@ def open(self, name, mode="r", pwd=None, *, force_zip64=False):
         files.  If the size is known in advance, it is best to pass a ZipInfo
         instance for name, with zinfo.file_size set.
         """
-        if mode not in {"r", "w", "U", "rU"}:
-            raise RuntimeError('open() requires mode "r", "w", "U", or "rU"')
-        if 'U' in mode:
-            import warnings
-            warnings.warn("'U' mode is deprecated",
-                          DeprecationWarning, 2)
+        if mode not in {"r", "w"}:
+            raise RuntimeError('open() requires mode "r" or "w"')
         if pwd and not isinstance(pwd, bytes):
             raise TypeError("pwd: expected bytes, got %s" % type(pwd))
         if pwd and (mode == "w"):
diff --git a/Misc/NEWS b/Misc/NEWS
@@ -38,6 +38,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #27029: Removed deprecated support of universal newlines mode from
+  ZipFile.open().
+
 - Issue #27030: Unknown escapes consisting of ``'\'`` and ASCII letter in
   regular expressions now are errors.  The re.LOCALE flag now can be used
   only with bytes patterns.
