bpo-22908: Add seek and tell functionality to ZipExtFile (GH-4966)

This allows for nested zip files, tar files within zip files, zip files within tar files, etc.

Contributed by: John Jolly

diff --git a/Doc/library/zipfile.rst b/Doc/library/zipfile.rst
index d58efe0b4175168cfb3756ce81e83a3ee8dc5926..7c9a8c80225491a8a7d8e9fc83ff324464c662a5 100644 (file)
--- a/Doc/library/zipfile.rst
+++ b/Doc/library/zipfile.rst
@@ -246,9 +246,9 @@ ZipFile Objects
    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__`,
-   :meth:`~iterator.__next__`.  These objects can operate independently of
-   the ZipFile.
+   :meth:`~io.IOBase.readlines`, :meth:`~io.IOBase.seek`,
+   :meth:`~io.IOBase.tell`, :meth:`__iter__`, :meth:`~iterator.__next__`.
+   These objects can operate independently of the ZipFile.
 
    With ``mode='w'``, a writable file handle is returned, which supports the
    :meth:`~io.BufferedIOBase.write` method.  While a writable file handle is open,

diff --git a/Lib/test/test_zipfile.py b/Lib/test/test_zipfile.py
index 94db858a1517c4d753c3205ba3e29f5977425492..61c3e349a69ef4f827c8653e091b597c054e6b07 100644 (file)
--- a/Lib/test/test_zipfile.py
+++ b/Lib/test/test_zipfile.py
@@ -1628,6 +1628,40 @@ class OtherTests(unittest.TestCase):
             self.assertEqual(zipf.read('baz'), msg3)
             self.assertEqual(zipf.namelist(), ['foo', 'bar', 'baz'])
 
+    def test_seek_tell(self):
+        # Test seek functionality
+        txt = b"Where's Bruce?"
+        bloc = txt.find(b"Bruce")
+        # Check seek on a file
+        with zipfile.ZipFile(TESTFN, "w") as zipf:
+            zipf.writestr("foo.txt", txt)
+        with zipfile.ZipFile(TESTFN, "r") as zipf:
+            with zipf.open("foo.txt", "r") as fp:
+                fp.seek(bloc, os.SEEK_SET)
+                self.assertEqual(fp.tell(), bloc)
+                fp.seek(-bloc, os.SEEK_CUR)
+                self.assertEqual(fp.tell(), 0)
+                fp.seek(bloc, os.SEEK_CUR)
+                self.assertEqual(fp.tell(), bloc)
+                self.assertEqual(fp.read(5), txt[bloc:bloc+5])
+                fp.seek(0, os.SEEK_END)
+                self.assertEqual(fp.tell(), len(txt))
+        # Check seek on memory file
+        data = io.BytesIO()
+        with zipfile.ZipFile(data, mode="w") as zipf:
+            zipf.writestr("foo.txt", txt)
+        with zipfile.ZipFile(data, mode="r") as zipf:
+            with zipf.open("foo.txt", "r") as fp:
+                fp.seek(bloc, os.SEEK_SET)
+                self.assertEqual(fp.tell(), bloc)
+                fp.seek(-bloc, os.SEEK_CUR)
+                self.assertEqual(fp.tell(), 0)
+                fp.seek(bloc, os.SEEK_CUR)
+                self.assertEqual(fp.tell(), bloc)
+                self.assertEqual(fp.read(5), txt[bloc:bloc+5])
+                fp.seek(0, os.SEEK_END)
+                self.assertEqual(fp.tell(), len(txt))
+
     def tearDown(self):
         unlink(TESTFN)
         unlink(TESTFN2)

diff --git a/Lib/zipfile.py b/Lib/zipfile.py
index f9db45f58a2bde79d82aa722353a6f01a88fd985..5df7b1bf75b9d9aefb989efb4fe5a39421d66bd5 100644 (file)
--- a/Lib/zipfile.py
+++ b/Lib/zipfile.py
@@ -696,6 +696,18 @@ class _SharedFile:
         self._close = close
         self._lock = lock
         self._writing = writing
+        self.seekable = file.seekable
+        self.tell = file.tell
+
+    def seek(self, offset, whence=0):
+        with self._lock:
+            if self.writing():
+                raise ValueError("Can't reposition in the ZIP file while "
+                        "there is an open writing handle on it. "
+                        "Close the writing handle before trying to read.")
+            self._file.seek(self._pos)
+            self._pos = self._file.tell()
+            return self._pos
 
     def read(self, n=-1):
         with self._lock:
@@ -746,6 +758,9 @@ class ZipExtFile(io.BufferedIOBase):
     # Read from compressed files in 4k blocks.
     MIN_READ_SIZE = 4096
 
+    # Chunk size to read during seek
+    MAX_SEEK_READ = 1 << 24
+
     def __init__(self, fileobj, mode, zipinfo, decrypter=None,
                  close_fileobj=False):
         self._fileobj = fileobj
@@ -778,6 +793,17 @@ class ZipExtFile(io.BufferedIOBase):
         else:
             self._expected_crc = None
 
+        self._seekable = False
+        try:
+            if fileobj.seekable():
+                self._orig_compress_start = fileobj.tell()
+                self._orig_compress_size = zipinfo.compress_size
+                self._orig_file_size = zipinfo.file_size
+                self._orig_start_crc = self._running_crc
+                self._seekable = True
+        except AttributeError:
+            pass
+
     def __repr__(self):
         result = ['<%s.%s' % (self.__class__.__module__,
                               self.__class__.__qualname__)]
@@ -963,6 +989,62 @@ class ZipExtFile(io.BufferedIOBase):
         finally:
             super().close()
 
+    def seekable(self):
+        return self._seekable
+
+    def seek(self, offset, whence=0):
+        if not self._seekable:
+            raise io.UnsupportedOperation("underlying stream is not seekable")
+        curr_pos = self.tell()
+        if whence == 0: # Seek from start of file
+            new_pos = offset
+        elif whence == 1: # Seek from current position
+            new_pos = curr_pos + offset
+        elif whence == 2: # Seek from EOF
+            new_pos = self._orig_file_size + offset
+        else:
+            raise ValueError("whence must be os.SEEK_SET (0), "
+                             "os.SEEK_CUR (1), or os.SEEK_END (2)")
+
+        if new_pos > self._orig_file_size:
+            new_pos = self._orig_file_size
+
+        if new_pos < 0:
+            new_pos = 0
+
+        read_offset = new_pos - curr_pos
+        buff_offset = read_offset + self._offset
+
+        if buff_offset >= 0 and buff_offset < len(self._readbuffer):
+            # Just move the _offset index if the new position is in the _readbuffer
+            self._offset = buff_offset
+            read_offset = 0
+        elif read_offset < 0:
+            # Position is before the current position. Reset the ZipExtFile
+
+            self._fileobj.seek(self._orig_compress_start)
+            self._running_crc = self._orig_start_crc
+            self._compress_left = self._orig_compress_size
+            self._left = self._orig_file_size
+            self._readbuffer = b''
+            self._offset = 0
+            self._decompressor = zipfile._get_decompressor(self._compress_type)
+            self._eof = False
+            read_offset = new_pos
+
+        while read_offset > 0:
+            read_len = min(self.MAX_SEEK_READ, read_offset)
+            self.read(read_len)
+            read_offset -= read_len
+
+        return self.tell()
+
+    def tell(self):
+        if not self._seekable:
+            raise io.UnsupportedOperation("underlying stream is not seekable")
+        filepos = self._orig_file_size - self._left - len(self._readbuffer) + self._offset
+        return filepos
+
 
 class _ZipWriteFile(io.BufferedIOBase):
     def __init__(self, zf, zinfo, zip64):

diff --git a/Misc/NEWS.d/next/Library/2017-12-21-22-00-11.bpo-22908.cVm89I.rst b/Misc/NEWS.d/next/Library/2017-12-21-22-00-11.bpo-22908.cVm89I.rst
new file mode 100644 (file)
index 0000000..4f3cc01
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2017-12-21-22-00-11.bpo-22908.cVm89I.rst
@@ -0,0 +1,2 @@
+Added seek and tell to the ZipExtFile class. This only works if the file
+object used to open the zipfile is seekable.