On Tue, Jan 11, 2000 at 09:37:23PM +0000, Steve Haslam wrote: > On Tue, Jan 11, 2000 at 10:26:03PM +0100, Gergely Madarasz wrote: > > I see... so I think there should be some policy about logrotate... > > I agree. > > My position, FWIW, is that packages should depend on logrotate unless > they provide a method of rotating the logs when logrotate is not > present, in which case they should either recommend or suggest it. > > The point being, that if a package maintainer is providing a logrotate > config file, then s/he intended the logs to be rotated, and the logs > not being rotated is a situation where the package is malfunctioning.
I am not sure I like the idea of packages rotation their logs on their own, if logrotate isn't installed; I think of it as, if the logs are going to be rotated, that is a job for a general purpose log-rotator, so why should the wheel be reinvented with each package? Also, using logrotate allows for a centralization of configuring the rotating of logs -- one of my complaints with other linux distros is that sometimes the setup info is spread in many different places. (It gets worse with commercial unices, the ones I have used in any event... /usr/mmdf/mmdftailor, &c) So to my way of thinking, no logs should be rotated if there isn't a logrotate package installed. If there is one or not is up to the administrator, though I strongly suggest one... (Where this breaks down is of course programs that are smart enough to trim their logs at X bytes, to avoid eating a whole partition with logs... how should they be handled? I don't know. :) ObDisclaimer: I am not a debian developer (yet). -- Seth Arnold | http://www.willamette.edu/~sarnold/ Hate spam? See http://maps.vix.com/rbl/ for help Hi! I'm a .signature virus! Copy me into your ~/.signature to help me spread!
