I have a problem with an xml file locking after read and/or write.

I have a short sample code with unit test available here:
http://bugs.jython.org/file741/lock.py

Below are the
outputs from Jython and CPython (v 2.5.4).
 
I am using Jython 2.5.1 on Windows XP and java version "1.6.0_16".

>jython lock.py
E
======================================================================
ERROR: runTest (__main__.TestDelete)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "W:\testTeam\SIT\JySit\sit\scripts\lock.py", line 73, in tearDown
    os.remove(self.fileName)
  File "C:\jython2.5.1\Lib\os.py", line 342, in remove
    raise OSError(0, "couldn't delete file", path)
OSError: [Errno 0] couldn't delete file: 'C:\\testdir\\test.xml'
 
----------------------------------------------------------------------
Ran 1 test in 0.454s
 
FAILED (errors=1)
 
 
>python lock.py
.
----------------------------------------------------------------------
Ran 1 test in 0.015s
 
OK


The problem can be cured by applying the changes as per file:
http://bugs.jython.org/file742/ElementTree.py

This is on the Jython list:
http://bugs.jython.org/issue1479

The patch proposed on msg95349:
 * need tests
 * not reviewed (yet)

Patch with tests (which can only really test it on Windows) here

http://bitbucket.org/pjenvey/et-2009-provolone/changeset/8292a06090a3/

Florent, any chance of signing off on this for 3.2? I was waiting for the patch to go through your authorized elementtree fork

Changeset cff78ffb932a fixed the issue for 3.2.
I will consider backporting.

New changeset eceaa31252b3 by Florent Xicluna in branch '2.7':
Closes #7334: close source files on ElementTree.parse and iterparse (partial backport of issue #10093 from 3.2).
http://hg.python.org/cpython/rev/eceaa31252b3

I think the cElementTree portion of the applied changes is incorrect. `close_source` is never actually set to `True`.

http://hg.python.org/cpython/file/50722d2f08c7/Modules/_elementtree.c#l2926

Indeed, the 2.7 backport was not correctly applied for _elementtree.c, leaving files open because the close_source flag is set to False *again* when opening a filename.

Should a new issue be opened or should this ticket be re-opened?

> Should a new issue be opened or should this ticket be re-opened?

It's a 3 years old backport, I'd say open a new issue.