On Sun, Mar 14, 2010 at 06:46:14PM +0100, Frans Pop wrote:
>
> > Changing d-i because of c-s seems wrong to me.
>
> It's not actually a "change" in D-I. There's simply more than one way to
> Rome. The difference is this:
> --- kbd-chooser/post-base-installer.d/20kbd-chooser (revision 62565)
> +++ kbd-chooser/post-base-installer.d/20kbd-chooser (working copy)
> @@ -27,6 +27,6 @@
> d-i debian-installer/keymap string $KEYMAP
> EOF
>
> -apt-install console-setup || true
> +apt-install --with-recommends console-setup || true
>
> exit 0
>
> Trivial, no?
Yes, trivial, but somehow I don't like this. I don't see why c-s has to
be the package that ensures that kbd is installed. What if for some
reason some future version of c-s reduces the recommendation to only a
suggestion? I'd do this:
if bds_or_hurd; then
apt-install keyboard-configuration || true
else
apt-install kbd || true
apt-install console-setup || true
fi
BTW, since currently neither console-setup-udeb nor kbd-chooser work on
hurd or bsd, how is the installer going to configure the keyboard there?
Kbd-chooser doesn't exist on these architectures.
> > So there are two options:
> > 1. Changing c-s to depend on kbd | console-tools and installing
> > keyboard-configuration instead of console-setup on some
> > architectures.
>
> That's certainly possible. It's a fairly simple change in kbd-chooser.
> Seems like the best option as it means that the --with-recommends is not
> needed and will also avoid accidental incomplete manual installs of c-s on
> Linux.
>
> Does that mean console-setup itself is essentially useless on kfreebsd and
> hurd?
Almost useless. I think the fonts can be used and one can also use the
package to build linux keymaps on a hurd or bsd host. But this is not
that important to make the strong dependency a bad thing. Tell me if
you want this.
Anton Zinoviev
--
To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100315133126.ga30...@logic.fmi.uni-sofia.bg
