logging.shutdown() doesn't clean up the dict of handlers at the module level and the list of handlers maintained and at the level of the manager of loggers. So after calling logging.getLogger() etc. for the second time, you still have the closed filehandle in the list of loghandlers. The new FileHandler with the open filehandle is appended to this list of handlers. Then when you call ...info("...") this call is dispatched to all handlers in the handlerlist, i.e. also to the one with the closed filehandle.Neil Benn wrote:
Hello,
I'm running a test and having issues with logging, if I call logging.shutdown() and then want to start the logging going again then I get a problem as if I call shutdown, I can't get the root logger again, such as :
.<snip>
Previous code missed out a line :
.>>> import logging
.>>> objTestLogger = logging.getLogger()
.>>> objTestLogger.setLevel(logging.INFO)
.>>> objTestLogger.addHandler(logging.FileHandler('c:\\test.log'))
.>>> objTestLogger.info("THIS IS A TEST")
.>>> logging.shutdown()
.>>> objTestLogger = logging.getLogger()
.>>> objTestLogger.setLevel(logging.INFO)
.>>> objTestLogger.addHandler(logging.FileHandler('c:\\test.log'))
.>>> objTestLogger.info("THIS IS A TEST")
.Traceback (most recent call last):
. File "c:\program files\python23\lib\logging\__init__.py", line 679, in emit
. self.stream.write("%s\n" % msg)
.ValueError: I/O operation on closed file
.>>>
I don't know if this has to be considered a bug in the logging module or if your use case is simply not covered.
If you add something like
.>>> logging._handlers.clear()
.>>> logging.root.handlers = []
.>>> for l in logging.Logger.manager.loggerDict.values():
.>>> l.handlers = []
after logging.shutdown() and before getting the new logger, your script will probably run without errors.
hth,
Samuel
-- http://mail.python.org/mailman/listinfo/python-list
