[New-bugs-announce] [issue11444] logging FileHandler.close should acquire its lock before closing stream
report at bugs.python.org
Tue Mar 8 20:54:34 CET 2011
New submission from Adam Ernst <adamjernst at gmail.com>:
This is my first report, apologies if I missed any tracker etiquette.
The logging module registers shutdown() to run via atexit. shutdown() calls flush() and close() on each handler.
However if a FileHandler is writing to a file while shutdown() is executing, an IOError will be raised as the file is being closed while written to. (This can happen if you use daemon threads, which can still be running while exiting.)
Traceback (most recent call last):
File "/usr/local/lib/python2.7/atexit.py", line 24, in _run_exitfuncs
File "/usr/local/lib/python2.7/logging/__init__.py", line 1616, in shutdown
File "/usr/local/lib/python2.7/logging/__init__.py", line 898, in close
IOError: close() called during concurrent operation on the same file object.
I'm not familiar with the internals of logging, but it seems that FileHandler should call self.acquire()/self.release() around its close and flush operations. Otherwise a daemon thread might be emitting a record while closing the handler.
components: Library (Lib)
title: logging FileHandler.close should acquire its lock before closing stream
versions: Python 2.7
Python tracker <report at bugs.python.org>
More information about the New-bugs-announce