Re: Testing a new udeb: best method?
Quoting Frans Pop (aragorn@tiscali.nl):
> On Tuesday 15 November 2005 18:37, Christian Perrier wrote:
> > I began working on user-setup, a udeb package which supposedly will
> > replace the very temporary initial-passwd-udeb package we released
> > along with shadow.
>
> I talked with Joey about this on IRC and he is not convinced that this
> actually needs a udeb.
> The alternative would probably be a regular package that is run in
> the /target chroot (using the passthrough interface). However, maybe
> doing it in an udeb is still be the best solution if there is no usecase
> for ever running the package on an installed system.
> Could the package be useful, e.g. after setting up a system with
> debootstrap?
Well, as far as I know (I'm deeply incompetent when it comes at
debootstrap) nothing was used up to now....and I just concluded that
the only use we currently have of passwd.config is when we use it
during a D-I 2nd stage run...
Hence a udeb....
More precisely, what I think is better is making this package
maintained by the D-I team in our main tree....Whether it produces a
deb or a udeb actually doesn't really matter....
>
> > Am I using the best method for these testings and in fact, why do my
> > tests fail?
>
> AFAIK, if the udeb is correct, including it in the mini iso should work.
> Did you look in /var/lib/cdebconf/templates to see if the templates are
> there?
Well, when I name my templates file properly, yes....When I use a
wrong name, they aren't there. Funny isn't it ?
/me beats self with a cluebat....
--
Reply to: