Merge lp://staging/~vorlon/ubuntu/natty/eglibc/multiarch-support into lp://staging/ubuntu/natty/eglibc

Proposed by Steve Langasek
Status: Merged
Merged at revision: 134
Proposed branch: lp://staging/~vorlon/ubuntu/natty/eglibc/multiarch-support
Merge into: lp://staging/ubuntu/natty/eglibc
Diff against target: 129 lines (+44/-6)
5 files modified
debian/changelog (+10/-0)
debian/control (+18/-4)
debian/control.in/libc (+1/-1)
debian/control.in/main (+14/-0)
debian/rules (+1/-1)
To merge this branch: bzr merge lp://staging/~vorlon/ubuntu/natty/eglibc/multiarch-support
Reviewer Review Type Date Requested Status
Colin Watson (community) Approve
Martin Pitt Approve
Ubuntu Release Team Pending
Review via email: mp+54135@code.staging.launchpad.net

Description of the change

Currently, trying to install any foreign arch libraries with apt will fail with this error:

  E: Couldn't configure pre-depend multiarch-support for libgcc1, probably a dependency cycle.

This is correct behavior on the part of apt. libc6 Provides: multiarch-support, libgcc1 Pre-Depends: multiarch-support, libc6 Depends: libgcc1.

The proposed branch will break this dependency loop for foreign arch packages by making 'multiarch-support' a real, Multi-Arch: foreign package depending on the native libc to satisfy these dependencies. I believe this should have no effect on the native system, which avoids this dependency loop by having the necessary libraries installed before apt is ever called; and I believe it's safe for foreign architectures because a) there's no way you're getting a foreign libc unpacked, *period*, unless it's already Multi-Arch: same, and there should be no libc package that's Multi-Arch: same that doesn't also support the multiarch paths; and b) it's the foreign arch, so if I'm wrong about this there's no danger to the package manager's integrity (you can't cross-grade the base system yet).

But it's late in the cycle and this is a new (if contentless) binary package being added to the base system, so I'm requesting the release team's review before landing this change.

To post a comment you must log in.
Revision history for this message
Steve Langasek (vorlon) wrote :

Hmm, just noticed that requesting review from 'ubuntu-release' didn't seem to actually generate any useful mails. Requesting review from Colin and/or Martin explicitly.

Revision history for this message
Steve Langasek (vorlon) wrote :

BTW, have upgrade-tested this in ppa and maverick->natty and natty->natty both go fine.

Revision history for this message
Martin Pitt (pitti) wrote :

The reasoning makes sense to me. However, what I don't understand why the new "multiarch-support" real package is described as being transitional? As far as I understood, the reason for existing won't go away after an upgrade?

Revision history for this message
Steve Langasek (vorlon) wrote :

It's transitional because after a full release cycle or so, there will be no libc6 that doesn't provide this support, so we can drop the pre-depends.

Revision history for this message
Martin Pitt (pitti) wrote :

Ah, thanks for the clarification. Seems fine to me then.

Revision history for this message
Martin Pitt (pitti) :
review: Approve
Revision history for this message
Colin Watson (cjwatson) wrote :

Makes sense to me too, for the record.

review: Approve

Preview Diff

[H/L] Next/Prev Comment, [J/K] Next/Prev File, [N/P] Next/Prev Hunk
The diff is not available at this time. You can reload the page or download it.

Subscribers

People subscribed via source and target branches