Haven't been able to make world in about a year

[Vaaf]

At 14:17 07.03.2006, Stijn Hoop wrote:
>On Mon, Mar 06, 2006 at 05:29:36PM +0100, Kristian Vaaf wrote:
> > At 22:08 04.03.2006, Stijn Hoop wrote:
> > >On Sat, Mar 04, 2006 at 05:45:29PM +0100, Kristian Vaaf wrote:
> > >> At 12:05 03.03.2006, Stijn Hoop wrote:
> > >> > On Fri, Mar 03, 2006 at 11:58:37AM +0100, Kristian Vaaf wrote:
> > >> > > I run the script to save time.
> > >> > > Basically I'd run the exact same chain of commands otherwise.
> > >> >
> > >> > You're missing the point: you'd run the exact same chain of commands
> > >> >
> > >> > --> _if everything goes according to plan_ <--
> > >> >
> > >> > What this list has been telling you is that it sometimes doesn't work
> > >> > like anyone expects to, and you need to make an informed 
> decision about
> > >> > the next command to enter instead of having the script proceed.
> > >>
> > >> Whether I have my commands in my script or in my head doesn't
> > >> make any difference. Yes I do read UPDATING and if I notice any
> > >> changes they will be applied respectively.
> > >
> > >The moment one step does NOT work in the command sequence, you need to
> > >alter your next move. No script can be prepared for all the things
> > >that can happen. Which is why everyone is recommending you NOT to run
> > >things in a script.
> >
> > I understand what you mean.
>
>No you apparently do not.
>
> > What I'm saying is, I do not expect a script to be prepared.
>
>Good.
>
> > I am the one reading UPDATING and modifying the script if there 
> is a change.
> > Manually. Whether I write the sequence in the command line or into
> > a script that I execute doesn't make no difference!
>
>It doesn't, but you're missing the point again.
>
>Let's say for the sake of argument that your computer's internal clock
>broke down. FreeBSD will keep the time just fine so long as your
>machine has power, so you don't notice anything.
>
>Now you run your script.
>
>It will go through the stages of building world, building kernel,
>installing the kernel, and then rebooting.
>
>At this point your computer loses power for a few seconds and *presto*
>your clock is set to 1970.
>
>Now, your script proceeds after the reboot with the new kernel.
>Because your clock is WAY off, 'make installworld' will complain (well
>it could work, but let's assume it doesn't).
>
>Your script however, doesn't see this.
>
>It therefore proceeds with mergemaster, and rebooting again.
>
>Now you're running a new kernel with older binaries, that may or may
>not work.
>
>If this had happened to me [1], I would have stopped at the point where
>'make installworld' throws an error, and wondered what went wrong. I could
>backout my kernel install, reboot into my working configuration, and figure
>things out before it all got out of hand (edited configuration files etc).
>
>And this is just ONE example of a admittedly minor thing that can go
>wrong.
>
>This is why you should not automate the installation yourself.
>
> > What do you mean, mailing list in the loop?
>
>I mean that you should keep CC'ing questions at FreeBSD.org so that other
>who wonder why automating this is not a good idea can search the
>archives to find out my answer to you.
>
> > Hartelijk dank, Stijn!
>
>Geen probleem.
>
>--Stijn
>
>--
> > Thus again, we have succesfully proven that I cannot read minds.
>It doesn't help. Almost all you ever get is "This mind intentionally left
>blank."
>                 -- Steve VanDevender, alt.sysadmin.recovery
>

Interesting statement.

I'd say that, however, is kinda like the chance of getting shot by African
guerilla when you live on the North Pole. Ain't gonna happen :)

So, to everyone who has tried helping me:

I guess there is no solution to this?

FreeBSD will NOT successfully make my world no matter what I do.

I've tried all the sequences.
Please stop flaming me for using the wrong sequence.
I've used ALL sequences. So I know that's not the problem.

But then, what can it be?

Is it time for the list to admit "there's nothing we can do"?

All the best,
Vaaf