Goswin von Brederlow writes ("Re: Please test gzip -9n - related to dpkg with
multiarch support"):
> Steve Langasek <vor...@debian.org> writes:
> > the [pam] module packages should be installed
> > for all archs, not just a subset[1].
>
> Ok, that is acceptable. We just lack any technical means to ensure this
> so far. Same problem as for input method plugins for example.
So we need a new mechanism for this.
Where should this fact be declared ? Is it a property of a package
that it makes sense to install it only on all configured architectures
or none ? Or is it a property of the dependency from the depending
package ?
I'm inclined to the former view. After all if you think about
installing some plugin, there may not even /be/ a dependency in
question. And I'm finding it difficult to imagine a package which has
this all-arches-needed multiarch property for some of its purposes but
not others.
This situation can only arise for a m-a:same package, since only those
are coinstallable. I would suggest that the right answer is a new
value for the Multi-Arch field, let's call it "all". It would work
like Multi-Arch: same except that:
* Dependencies on multiarch:all package are not satisfied unless
the package is suitably installed and configured on all configured
architectures. Ie with
Package: plugin
Multi-Arch: all
Architecture: i386
Package: plugin
Multi-Arch: all
Architecture: amd64
this dependency
Depends: plugin
is read as
Depends: plugin:i386, plugin:amd64
* The higher-level package manager, when it is asked to install a
multiarch:all package, will install it for all configured
architectures.
Ian.
--
To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20281.609.605532.103...@chiark.greenend.org.uk
