[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Roundup Robot added the comment: New changeset 379aad232000 by Victor Stinner in branch '3.4': Issue #11453, #18174: Fix leak of file descriptor in test_asyncore http://hg.python.org/cpython/rev/379aad232000 New changeset 0ced2d2325fb by Victor Stinner in branch 'default': (Merge 3.4) Issue #11453, #18174: Fix leak of file descriptor in test_asyncore http://hg.python.org/cpython/rev/0ced2d2325fb -- ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Roundup Robot added the comment: New changeset ae12a926e680 by Victor Stinner in branch '3.4': Issue #11453: asyncore: emit a ResourceWarning when an unclosed file_wrapper http://hg.python.org/cpython/rev/ae12a926e680 -- nosy: +python-dev ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
STINNER Victor added the comment: I fixed the issue in Python 3.4 and 3.5, thanks for the report. In Python 3.4+, it's safe to add a destructor (__del__ method): even if the object is part of a reference cycle, it will be destroyed. It's not the case in Python 2.7. I prefer to leave Python 2.7 unchanged to limit the risk of regression. -- nosy: +haypo resolution: - fixed status: open - closed versions: +Python 3.4, Python 3.5 -Python 2.7, Python 3.2, Python 3.3 ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Roundup Robot added the comment: New changeset 7c9335d97628 by Victor Stinner in branch 'default': (Merge 3.4) Issue #11453: asyncore: emit a ResourceWarning when an unclosed http://hg.python.org/cpython/rev/7c9335d97628 -- ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Changes by Terry J. Reedy tjre...@udel.edu: -- versions: -Python 2.5, Python 2.6, Python 3.1 ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Aldona Majorek amajo...@google.com added the comment: Adding __exit__ will not make asyncore.file_wrapper close file descriptor when garbage collected. Here is clone of socket.py solution for the same problem. def close(self): if self.fd: os.close(self.fd) self.fd = None # or maybe self.fd = 0 will be better def __del__(self): try: self.close() except: # close() may fail if __init__ didn't complete pass -- ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Ben Hayden hayden...@gmail.com added the comment: Adding a patch that adds an __exit__ function much like the one that socket.socket implements. Passes the test_asyncore also doesn't raise a resource warning when I explicitly comment out some close() calls on file wrapper objects in the test. -- keywords: +patch nosy: +beardedp Added file: http://bugs.python.org/file21236/issue11453.patch ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
Changes by Giampaolo Rodola' g.rod...@gmail.com: -- nosy: +giampaolo.rodola ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue11453] asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does.
New submission from Aldona Majorek pyt...@ada.majorek.org: asyncore.file_wrapper duplicates file descriptor of given file and closes it in it's close method. But unlike socket.socket class it does not automatically call close when object is garbage collected. Users of regular sockets and asyncore.dispatcher do not experience resource leaks when they forget to call self.close() in handle_close(). But people using file_dispatcher do loose file descriptor every time file_wrapper object is garbage collected without calling self.close() first. -- components: Library (Lib) messages: 130458 nosy: amajorek priority: normal severity: normal status: open title: asyncore.file_wrapper should implement __del__ and call close there to prevent resource leaks and behave like socket.socket does. type: resource usage versions: Python 2.5, Python 2.6, Python 2.7, Python 3.1, Python 3.2, Python 3.3 ___ Python tracker rep...@bugs.python.org http://bugs.python.org/issue11453 ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
