[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: MBF: Multi-Arch: same violations

On 2017-08-28 09:14, Helmut Grohne wrote:
> Hi,
> I analyzed file conflicts in binary packages marked Multi-Arch: same. To
> do that, I created a branch of https://dedup.debian.net/ specifically
> for analyzing multiarchy features of packages. The resulting issues are
> fed into tracker.d.o already, but little happens about them. File
> conflicts in Multi-Arch: same packages are particularly annoying,
> because they result in unpack errors and leave the system in a broken
> state. We typically file bugs of severity serious for unpack errors, but
> for multiarchy issues we tend to reduce that to important.
> I attempted to reproduce all file conflicts using "apt-get install
> pkg:arch1 pkg:arch2" in a clean sid chroot. I am only reporting those
> that actually result in unpack errors.
> Attached to this mail you can find a file "fail" with four columns:
>  * a binary package name
>  * a release architecture
>  * a different release architecture
>  * optionally an existing bug number
> For each combination "apt-get install $1:$2 $1:$3" fails with an unpack
> error. The corresponding dd-list is attached as well. Some bugs are
> already filed (mostly by Jakub Wilk, thanks!), but still included in
> both lists. Where the fourth column is non-empty, I won't file a new
> bug.
> Here is the bug template:
> v v v v v
> Package: $1
> Severity: important
> User: multiarch-devel@lists.alioth.debian.org
> Usertags: multiarch
> $1 is marked Multi-Arch: same, but fails to coinstall with itself on $2
> and $3. You can find a failing installation log attached. Please
> consider removing the Multi-Arch: same header or declaring the relevant
> architecture-qualified Conflicts.
> ^ ^ ^ ^ ^

architecture-qualified self-conflicts is not something supported. See
bug #747261. This makes somehow difficult to fix this kind of bugs.


Aurelien Jarno                          GPG: 4096R/1DDD8C9B
aurelien@aurel32.net                 http://www.aurel32.net

Reply to: