Universal newline support for the read method of ZipFileExt objects created by calling the ZipFile.open method is not working. The attached archive contains a test script for demonstrating the behavior; unzip the files into the same directory and run the script. It tests opening a text file with all three types of line endings in five different ways: with the file.read method in modes 'rU', 'r', and 'rb', and with the zipfile.ZipFileExt.read and readlines methods in mode 'rU'.

The first two and last methods both read the file with the expected universal newline support; the '\r' and '\r\n' endings are translated to '\n'. But zipfile.ZipFileExt.read, even with universal newline support, leaves these endings untouched, just like the 'rb' mode.

This issue seems to have been raised in the past (http://bugs.python.org/issue6759) for Python 3.3, with the decision apparently made to deprecate universal newline support for this method. Is this also the case for Python 2.7, and if so, why does the zipfile documentation for both versions not mention this missing functionality? As a user, I can say that this deprecation is certainly disrupting my workflow; it is not at all the behavior I expected.

I am running Python 2.7.11 on a 64-bit copy of Windows 7 Professional SP1.

The decision was to deprecate 'U' *everywhere* in python3.

Yes, the 2.7 method you are concerned about is broken with respect to its documentation, which is what #6759 was about originally.  That issue was deemed too disruptive to the codebase to address effectively, but in fixing things properly for python3 we apparently lost track of the 2.7 bug.

You can fix the problem (in a forward compatible way) by wrapping the ZipExtFile in an io.TextIOWrapper.

The 2.7 documentation should be changed to address this issue and the solution.

The support of the "U" mode will be removed in 3.6 (issue27029).

Python 2.7 is no longer supported.