Libguestfs - Dec 2013 - Re: [PATCH] sysprep: handle distro specific sysv scripts

On Thursday 12 December 2013 15:01:07 Olaf Hering wrote:
> On Thu, Dec 12, Pino Toscano wrote:
> > On Thursday 12 December 2013 14:49:36 Olaf Hering wrote:
> > > On Thu, Dec 12, Pino Toscano wrote:
> > > > What do you think?
> > > 
> > > I think that your Should-Start handling is broken. Required means
> > > the
> > > given file can not properly work without the listed servers,
> > > insserv
> > > will error out. Should means it can very well work without them
if
> > > they are not present or enabled. Otherwise the given file has to
> > > be
> > > scheduled after the listed services.
> > 
> > Sure, and I said there's no guarantee that the firstboot command
can
> > work without any of the other system services.
> > What if my command requires the system time to be set, but it is
> > scheduled by insserv before $time is available?
> 
> $all should imply $time.
Sure, but as a Should-Start means it is a weak dependency, and can be 
skipped, which is what I don't want.
> But looking at the patch again, there is
> appearently no insserv involved. Its just a hardcoded ln command.
> So why is the script not exectued at the very end in your setup?
As said in my first email:
> (Firstboot scripts are not working in Debian anyway, see #1019388.)
aka, an invocation of update-rc.d (which on Debian invokes insserv) or 
insserv directly is needed to get them working.

-- 
Pino Toscano

On Thu, Dec 12, Pino Toscano wrote:
> Sure, but as a Should-Start means it is a weak dependency, and can be 
> skipped, which is what I don't want.
Are you saying in Debian the Should-Start is not handled properly?
Even if Should-Start is weak it still has to be taken into account.

Olaf

On Thursday 12 December 2013 16:27:30 Olaf Hering wrote:
> On Thu, Dec 12, Pino Toscano wrote:
> > Sure, but as a Should-Start means it is a weak dependency, and can
> > be
> > skipped, which is what I don't want.
> 
> Are you saying in Debian the Should-Start is not handled properly?
> Even if Should-Start is weak it still has to be taken into account.
No, I'm saying a weak dependency is no guarantee the provided firstboot 
commands can run successfully, and running them after $all should 
provide a better safety.

-- 
Pino Toscano