On Sat, Jul 03, 2010 at 08:03:25PM -0500, Raphael Geissert wrote:
> > @@ -7897,25 +7899,28 @@ ln -fs ../sbin/sendmail debian/tmp/usr/bin/runq
> > section="8">):
> > <example compact="compact">
> > /var/log/foo/*.log {
> > -rotate 12
> > -weekly
> > -compress
> > -postrotate
> > -/etc/init.d/foo force-reload
> > -endscript
> > + rotate 12
> > + weekly
> > + compress
> > + missingok
> > + postrotate
> > + /etc/init.d/foo force-reload
> > + endscript
> > }
> While we are at it please let's use invoke-rc.d, which is what should be used.
> The current example could actually be considered as a violation of section
> 9.3.3.2 (if a logrotate file is considered as a "package maintainer script,")
> so please take this email as an objection.
> If invoke-rc.d is not used and the init script follows Policy's requirements
> for 'force-reload' and 'restart,' the service could be started in a runlevel
> where it has been explicitly disabled by the administrator.
> This does not address the case where a service has been stopped but is not
> disabled for the current runlevel, but that's an issue with invoke-rc.d.
This force-reload also caught my eye when I first read the patch, but I
think this should be treated as a separate bug. logrotate scripts are *not*
maintainer scripts, and recommending use of invoke-rc.d for non-maintainer
scripts is a separate substantive change that ought to be discussed in its
own bug.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
slangasek@ubuntu.com vorlon@debian.org
Attachment:
signature.asc
Description: Digital signature