i have a very basic setup of logging in a long running daemon app..
I use TimedRotatingFileHandler with rotating at midnight.
The bug:
The last open logging file was for 2009-05-25..
app was running without logging anything for about a week or more.
After a week past something happened and daemon started to log messages,
but.. for each new message logging system creates new file.
In other words.. suppose we was idle since 2009-05-25, now it is 
2009-06-10 and we want to log 10 messages.. 
current opened file is for 2009-05-25.. so when a first messages
arrives into logging system current must be closed. and a new file must 
be opened where all 10 messages should be..
but instead logging system will log each subsequent message in files 
2009-05-26, 2009-05-27, 2009-05-28..(in fact it will log in file 
without a date and then when nect message arives rename current to 
2009-05-26, 2009-05-27, 2009-05-28.. but is still the same) and so on, 
one message per file.. i think until it reaches current date.
It is a logic error.


my logging setup looks like this:
---------
import logging
from logging import debug,warning,info,error,critical
from logging.handlers import TimedRotatingFileHandler

log_path = '/var/log/cherry/smsd_reg'
log_level = logging.DEBUG

basic_log_handler = TimedRotatingFileHandler(log_path + "/
app_log_smsd_reg_server",'midnight',1)
basic_log_handler.setLevel(log_level)
basic_log_formatter = logging.Formatter('%(asctime)s pid: %(process)d, 
%(levelname)s %(message)s')
basic_log_handler.setFormatter(basic_log_formatter)
logging.getLogger('').addHandler(basic_log_handler)
logging.getLogger('').setLevel(log_level)

----------------

system info:
Python 2.5.1
FreeBSD 6.2-RELEASE-p7 FreeBSD 6.2-RELEASE-p7