Bug#695229: chromium-inspector: should be 'Multi-Arch: foreign' so that chromium:{arch} is installable
Hi,
Peter Samuelson wrote:
In order to install chromium:i386 on my amd64 system, its dependency,
chromium-inspector, would need to be Multi-Arch: foreign. This is
because it is Architecture: all, which in a multi-arch context is
calculated as Architecture: {primary installed arch}.
Are you sure? Could you send output from trying to install it?
(I had always thought that in the multi-arch world Arch: all meant
with the same architecture as its dependencies.)
Thanks and hope that helps,
Jonathan
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#695229: chromium-inspector: should be 'Multi-Arch: foreign' so that chromium:{arch} is installable
[Jonathan Nieder]
Are you sure? Could you send output from trying to install it?
I used interactive aptitude and it can't install chromium:i386 because
it can't find a candidate for the chromium-inspector dependency.
(I had always thought that in the multi-arch world Arch: all meant
with the same architecture as its dependencies.)
That's what they want you to think! No, for dependency resolution
purposes, arch:all is equivalent to arch:{dpkg's primary arch}. There
are arcane reasons for this, which I only half-understand so I can't
easily explain. The short version:
https://wiki.ubuntu.com/MultiarchSpec#Dependencies_involving_Architecture:_all_packages
The workaround is, for an arch:all package that really should be
Multi-Arch: foreign, to explicitly state that it is Multi-Arch:
foreign.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#695229: chromium-inspector: should be 'Multi-Arch: foreign' so that chromium:{arch} is installable
Peter Samuelson wrote:
[Jonathan Nieder]
(I had always thought that in the multi-arch world Arch: all meant
with the same architecture as its dependencies.)
That's what they want you to think! No, for dependency resolution
purposes, arch:all is equivalent to arch:{dpkg's primary arch}. There
are arcane reasons for this, which I only half-understand so I can't
easily explain. The short version:
https://wiki.ubuntu.com/MultiarchSpec#Dependencies_involving_Architecture:_all_packages
Thanks for the pointer. Yuck.
Multiarch folks: would it make sense to tweak the spec to require
behavior matching the rationale
architecture-dependent packages may depend on Architecture:
all packages and assume that the transitive dependencies will
be resolved using packages of the same architecture or other
packages that are Architecture: all
so that dpkg, apt, and aptitude can agree on what to do here?
Confused,
Jonathan
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#695229: chromium-inspector: should be 'Multi-Arch: foreign' so that chromium:{arch} is installable
Daniel Hartwig wrote: On 6 December 2012 15:26, Jonathan Nieder jrnie...@gmail.com wrote: so that dpkg, apt, and aptitude can agree on what to do here? Do they currently disagree? No, I was just confused. Contrary to my expectation, dpkg follows the spec. In my mental model, there is no such thing as the native architecture except when building source packages. Looks like I need a new mental model. ;-) Sorry for the nonsense. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
