Proposed Migration - Reference Edition
Steve Langasek
steve.langasek at ubuntu.com
Thu Dec 12 17:10:05 UTC 2019
On Thu, Dec 12, 2019 at 11:56:48AM -0300, Andreas Hasenack wrote:
> On Wed, Dec 11, 2019 at 9:22 PM Bryce Harrington
> <bryce.harrington at canonical.com> wrote:
> > #######################
> > ### bind9 (Andreas) ###
> > #######################
> >
> > bind9 blocking netbase (5.6 to 5.8) for 5 days
> > Regressions
> > bind9/1:9.11.5.P4+dfsg-5.1ubuntu4: i386 (log, history)
> >
> > * bind9 (i386)
> > - Looks like maybe something to do with netbase/5.8?
> > - Andreas writes, "Blocking migration: I spotted bind9/i386 blocking
> > the migration of netbase, so I'll take a look at that tomorrow. It
> > was also in Steve's last email about the i386 status in focal."
> >
> > * See https://lists.ubuntu.com/archives/ubuntu-devel/2019-December/040859.html
>
> Since then, the test was hinted and is a known failure now, not
> blocking the migration anymore. I took a look in any case, and
> launching a focal/i386 container and running that test works just
> fine, for now (since lxc still has focal/i386 images). But I spotted
> this in the autopkgtest log:
> autopkgtest [19:05:04]: testbed running kernel: Linux 5.3.0-24-generic
> #26-Ubuntu SMP Thu Nov 14 01:33:18 UTC 2019
> autopkgtest [19:05:04]: testbed dpkg architecture: amd64
> autopkgtest [19:05:04]: testbed target architecture: i386
> I'll re-read Steve's series of emails about how these tests are run.
> The actual failure was a dependency problem involving amd64 packages,
> and this is supposed to be an i386 test run, so maybe something is
> wrong in how the environment is setup to run these i386 tests without
> a base i386 image.
Many tests need changes to annotate their dependencies as to whether they
need a version of the package that matches the target arch or not. In the
case of bind9, it looks like the problematic dependency chain is bind9 ->
bind9utils -> python3.
python3 is Multi-Arch: allowed, which means that packages which only use it
as an interpreter can /in theory/ depend on python3:any instead of on
python3 to declare that they're fine with the interpreter being from a
different arch. But while dh-python correctly gives bind9utils a dependency
on python3:any, there is also a hard-coded dependency in debian/control on
'python3', which forces the selection of the i386 python3, which is not
cross-installable.
So please fix bind9utils to drop this hard-coded dependency on python3.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer https://www.debian.org/
slangasek at ubuntu.com vorlon at debian.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.ubuntu.com/archives/ubuntu-server/attachments/20191212/b4cad3a9/attachment.sig>
More information about the ubuntu-server
mailing list