Re: Apt-pinning confusion

On Fri, 23 Mar 2012 18:55:13 +0000, Camaleón wrote:

On Fri, 23 Mar 2012 12:15:10 +0000, Ramon Hofer wrote:

I'm trying to put the MythTV PVR XBMC version on my Shuttle box. I need
a newer alsa version than the one from Squeeze because the stable
version doesn't see the soundcard. So I wanted to install alsa from
And because I use a SSD I thought it would be a good idea to use the
squeeze-backports kernel.

What do you think it would be better to completely go with testing.

Testing is currently quite stable but there are significant differences
between wheezy and squeeze, like the gnome environment.

I think this won't make any difference for me because I will only use the
base system with xorg and xbmc without any window manager.

There are two reasons why I didn't want to do this:

First I need to compile the jme module manually to be able to use the
network interface. So I thought the less changes to the kernel makes me
less often compile that module again.

My wild guess is that wheezy kernel is not going to change much since
now (3.2.12 is the current one) and IIRC, wheezy will be relased with
this (3.2.x) branch but well... this can change at any time so yes, you
will have to recompile the kernel module for every kernel change.

Ok, so at least I don't have to expect kernel changes every day :-)

Second the XBMC version I want to install needs libboost version 1.47
or older.

Any specific reason for you to stick with a specific version of XBMC?

Yes, I want to use xbmc as a frontend for mythtv. And there's a branch of
xbmc pvr that can connect to mythbackend:

Unfortunately this doesn't work. I suppose my problem are wrong apt-
preferences numbers or something like this.

Here's my sources.list: And apt

(tip: when sending a pastebin link, I prefer to use the "raw" mode, it
reads better, i.e.:

Didn't know that. Thanks for the tip I will post it like this from now
on :-)

I'm not going to make any comments about pinning because I've never used
but just a question: have you considered in using pinning only for the
packages you want to be kept for a specific flavour? That is, being more
"selective" to avoid additional problems or messing up too many

This sounds good.
I thought I can do that by installing via "apt-get -t wheezy alsa-utils".

And here's the error I get when I try to install linux-headers-686-pae
from squeeze-backports:

The following packages have unmet dependencies:
linux-headers-686-pae : Depends: linux-headers-3.2.0-0.bpo.2-686-pae
it is not going to be installed
E: Broken packages

Mmm... "linux-headers-686-pae" is a metapackage that has to pull "linux-
headers-3.2.0-0.bpo.2-686-pae" automatically, I would open Sypatic to
see what's going on with this although manually installing "linux-
headers-3.2.0-0.bpo.2-686-pae" in addition to the metacpake should work.

I installed the two metapackages linux-headers-686-pae and linux-
image-686-pae so that I always have the newest backport kernel with the
matching headers.

Unfortunately I don't have synaptic. I only have the terminal since I
don't want to use any window manager for xbmc.

I can't as well install build-essential. There are many dependencies
which usually are solved automatically.
I think this is something that shouldn't be. When I want to install build-
essential it asks for libc6-dev which depends on libc but a newer version
is to be installed:

It seems as if I made a mess because there already is a libc6 package
from testing installed.

Yesterday I had the problem with alsa but today witchcraft made the
problem with alsa disappear but the one with the kernel header and as
well build-essential appear.

Is this really a problem of the apt pinning numbers? Or what can you
suggest me to do?
Maybe stick with the stable kernel and compile alsa from source?

Your first plan seems good, it may just need to be polished a bit :-)

Ok, thanks.
I will try to again maybe with a clean install again. Like that the mess
with the package dependencies should be gone.

Best regards

To UNSUBSCRIBE, email to debian-user-REQUEST@xxxxxxxxxxxxxxxx
with a subject of "unsubscribe". Trouble? Contact listmaster@xxxxxxxxxxxxxxxx