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

Bug#893754: autopkgtest 5.1 "autopkgtest-default-release" breaks tests that exercise apt



Hi,

On 22-03-18 10:43, Iain Lane wrote:
> On Wed, Mar 21, 2018 at 05:18:53PM -0700, Steve Langasek wrote:
>> When autopkgtest 5.2 was recently rolled out to the Ubuntu infrastructure,
>> various package-manager-related test suites began to fail as a result of the
>> use of APT::Default-Release introduced in 5.1:
>> [?]
> 
> I see autopkgtest already does:
> 
>   for default in default_releases:
>       script += 'printf "\nPackage: *\\nPin: release a=%(default)s\\nPin-Priority: 990\\n" >> /etc/apt/preferences.d/autopkgtest-%(release)s; ' % \
>           {'release': release, 'default': default}
> 
> if there are default releases passed to this function, which there are
> only for -updates. But AIUI (jak?) this is effectively the same as
> APT::Default-Release anyway, so I'm proposing that we always do this for
> the default release instead of the apt.conf.d snippet.

autopkgtest works great with the new functionality, and it is really
needed for Debian (the Ubuntu way of archives isn't robust, a= and s=
are not the same except in Ubuntu). I tried, but decided against it, to
have logic to convert e.g. sid to unstable (or the other way around), as
it isn't robust for Debian downstreams.

In my opinion, python(3|)-apt in Ubuntu is broken. It needs to accept
valid APT::Default-Release options in it's own distribution.
(Interestingly, python(3|)-apt doesn't depend on python-debian, but it
does declare the Test-Depends in its autopkgtest).

Paul

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/autopkgtest-devel/attachments/20180322/e7e47e93/attachment.sig>


Reply to: