Bug#915573: libhdf5-mpich-103:amd64 should conflict with libhdf5-mpich-101:amd64

[Witold Baryluk]

Oh. I see. The -101 was from experimental. I think installed it long time
ago, because I needed a specific version is from octave-dev that had
important fix for me, and it probably also pulled hdf from experimental at
the same time.

On Wed, 5 Dec 2018, 11:11 Witold Baryluk <witold.baryluk at gmail.com wrote:

> I didn't realise it was coming from experimental.
>
> I was simply doing apt dist-upgrade.
>
> I uninstalled old (and not installed new) package, that triggered removal
> of octave-dev, then reinstalled octave-dev, and it installed fine, but also
> pulled this new version of hdf5-mpich-103. I am not at the computer right
> now, but I will check apt and dpkg logs.
>
>
>
> On Wed, 5 Dec 2018, 11:05 Gilles Filippini <pini at debian.org wrote:
>
>> On 2018-12-05 10:53, Adrian Bunk wrote:
>> > On Wed, Dec 05, 2018 at 10:35:17AM +0100, Gilles Filippini wrote:
>> >> Control: severity -1 wishlist
>> >> Control: tags -1 + wontfix
>> >>
>> >> Hi,
>> >>
>> >> On 2018-12-05 00:34, Witold Baryluk wrote:
>> >> > Package: libhdf5-mpich-103
>> >> > Severity: important
>> >> >
>> >> >
>> >> > The following additional packages will be installed:
>> >> >   libhdf5-mpich-103
>> >> > The following NEW packages will be installed:
>> >> >   libhdf5-mpich-103
>> >> > 0 upgraded, 1 newly installed, 0 to remove and 141 not upgraded.
>> >> > 2 not fully installed or removed.
>> >> > Need to get 0 B/1365 kB of archives.
>> >> > After this operation, 4688 kB of additional disk space will be used.
>> >> > Do you want to continue? [Y/n]
>> >> > (Reading database ... 551515 files and directories currently
>> installed.)
>> >> > Preparing to unpack .../libhdf5-mpich-103_1.10.4+repack-1_amd64.deb
>> ...
>> >> > Unpacking libhdf5-mpich-103:amd64 (1.10.4+repack-1) ...
>> >> > dpkg: error processing archive
>> >> > /var/cache/apt/archives/libhdf5-mpich-103_1.10.4+repack-1_amd64.deb
>> >> > (--unpack):
>> >> >  trying to overwrite
>> >> > '/usr/lib/x86_64-linux-gnu/libhdf5_mpich_fortran.so.100', which is
>> >> > also in package libhdf5-mpich-101:amd64 1.10.2+repack-1~exp1
>> >> > Errors were encountered while processing:
>> >> >  /var/cache/apt/archives/libhdf5-mpich-103_1.10.4+repack-1_amd64.deb
>> >> > E: Sub-process /usr/bin/dpkg returned an error code (1)
>> >>
>> >> I do not agree: HDF5 1.10.2 was uploaded to experimental only. While
>> >> this
>> >> conflict do exist, there is no unhandled conflict with previous
>> >> releases
>> >> from testing or unstable.
>> >
>> > What is the problem with adding Replaces+Conflicts
>> > also for libhdf5-mpich-101?
>> >
>> > I agree that the severity looks less clear if this was
>> > experimental-only,
>> > but it is also pretty easy to do.
>>
>> Sure, this is an unfortunate oversight from me. But (1) hdf5 1.10.4 is
>> currently
>> transitioning and I don't want to delay that anymore, and (2) I expect
>> that
>> anybody installing packages from experimental is able to deal with this
>> kind
>> of conflict.
>>
>> > In theory (likely not applicable here) it is even possible that some
>> > downstream distribution like Ubuntu shipped the experimental version
>> > in a stable release.
>>
>> No, ubuntu didn't ship any HDF5 release from experimental. And packages
>> from
>> experimental must not be used in any stable release. Quoting the FAQ:
>> > Experimental is used for packages which are still being developed, and
>> > with
>> > a high risk of breaking your system. It's used by developers who'd like
>> > to
>> > study and test bleeding edge software. Users shouldn't be using
>> > packages
>> > from there, because they can be dangerous and harmful even for the most
>> > experienced people.
>>
>> Thanks,
>>
>> _g.
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-grass-devel/attachments/20181205/3267f256/attachment.html>