Re: Bug#407132: hal: Circumvents invoke-rc.d (and thus policy-rc.d) in postinst
On Wed, Jan 17, 2007 at 10:57:29AM +0100, Jonas Smedegaard wrote:
> Sjoerd Simons wrote:
> > On Tue, Jan 16, 2007 at 01:29:12PM +0100, Jonas Smedegaard wrote:
> >> I can only interpret the dbus subscript as indirectly being part of the
> >> `/etc/init.d/*' initscripts.
> > Maybe, but is is _not_ managed by sysvinit but by the dbus init scripts, thus
> > it can't use invoke-rc.d. While i agree it's not as nice it coulde be, i don't
> > see a real solution untill we get an init system that properly supports events
> The dbus init scripts _are_ subscripts of sysvinit scripts.
"subscripts of sysvinit scripts" are not what is governed by the invoke-rc.d
> Policy describes how system daemons should be started only through the
> use of invoke-rc.d, and the dbus subscripts are in fact started that way.
It describes how system daemons *that have sysv init scripts* must be
started using invoke-rc.d. In fact, hal's postinst also does a force-reload
on dbus, and in that case it *does* use invoke-rc.d correctly.
> hal restarts one of those subscripts directly, which causes problems for
> systems relying on policy-rc.d.
Yes, clearly it is a problem, but that doesn't make it an RC policy
violation. You're extrapolating policy to apply to cases it doesn't cover
-- that's fine as a starting point for discussing how this should be fixed,
but it doesn't make it an RC 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.