remove/delete method for zipfile/tarfile objects

[loewis]

http://bugs.python.org/review/6818/diff/2034/3872
File Doc/library/zipfile.rst (right):

http://bugs.python.org/review/6818/diff/2034/3872#newcode255
Doc/library/zipfile.rst:255: :exc:`RuntimeError`.
This needs a versionadded macro.

http://bugs.python.org/review/6818/diff/2034/3873
File Lib/test/test_zipfile.py (right):

http://bugs.python.org/review/6818/diff/2034/3873#newcode1494
Lib/test/test_zipfile.py:1494: self.assertLess(new_size, size)
This should be stronger: the lengths should have decreased by the compressed
size+header size+data descriptor size.

http://bugs.python.org/review/6818/diff/2034/3873#newcode1510
Lib/test/test_zipfile.py:1510: 
There should also be tests that remove a file from the middle, or as the first
and last file.

http://bugs.python.org/review/6818/diff/2034/3874
File Lib/zipfile.py (right):

http://bugs.python.org/review/6818/diff/2034/3874#newcode1213
Lib/zipfile.py:1213: zlen = len(zinfo.FileHeader()) + zinfo.compress_size
This needs to take the data descriptor into account, see
http://www.pkware.com/documents/casestudies/APPNOTE.TXT

http://bugs.python.org/review/6818/diff/2034/3874#newcode1231
Lib/zipfile.py:1231: del self.NameToInfo[fname]
IIUC, the file pointer is now at the wrong position - namely after the end of
the original data (including central directory), which means that .close() will
write a second central directory. I believe that fp should be positioned after
the last file.

[ubershmekel]

These were all taken care of in zipfile.remove.2.patch

http://bugs.python.org/review/6818/diff/2034/3872
File Doc/library/zipfile.rst (right):

http://bugs.python.org/review/6818/diff/2034/3872#newcode255
Doc/library/zipfile.rst:255: :exc:`RuntimeError`.
On 2011/03/14 01:13:59, loewis wrote:
> This needs a versionadded macro.

Done.

http://bugs.python.org/review/6818/diff/2034/3873
File Lib/test/test_zipfile.py (right):

http://bugs.python.org/review/6818/diff/2034/3873#newcode1494
Lib/test/test_zipfile.py:1494: self.assertLess(new_size, size)
On 2011/03/14 01:13:59, loewis wrote:
> This should be stronger: the lengths should have decreased by the compressed
> size+header size+data descriptor size.

Done.

http://bugs.python.org/review/6818/diff/2034/3873#newcode1510
Lib/test/test_zipfile.py:1510: 
On 2011/03/14 01:13:59, loewis wrote:
> There should also be tests that remove a file from the middle, or as the first
> and last file.

Done.

http://bugs.python.org/review/6818/diff/2034/3874
File Lib/zipfile.py (right):

http://bugs.python.org/review/6818/diff/2034/3874#newcode1213
Lib/zipfile.py:1213: zlen = len(zinfo.FileHeader()) + zinfo.compress_size
On 2011/03/14 01:13:59, loewis wrote:
> This needs to take the data descriptor into account, see
> http://www.pkware.com/documents/casestudies/APPNOTE.TXT

Done.

http://bugs.python.org/review/6818/diff/2034/3874#newcode1231
Lib/zipfile.py:1231: del self.NameToInfo[fname]
On 2011/03/14 01:13:59, loewis wrote:
> IIUC, the file pointer is now at the wrong position - namely after the end of
> the original data (including central directory), which means that .close()
will
> write a second central directory. I believe that fp should be positioned after
> the last file.

Done.