On Sat, Sep 09, 2017 at 08:22:08PM +0200, Ralf Treinen wrote:
> Package: libatlas3-base,libcblas3
> Version: libatlas3-base/3.10.3-4
> Version: libcblas3/3.2.1+dfsg-1
> Severity: serious
> User: trei...@debian.org
> Usertags: edos-file-overwrite
> 
> Date: 2017-09-09
> Architecture: amd64
> Distribution: sid

> automatic installation tests of packages that share a file and at the
> same time do not conflict by their package dependency relationships has
> detected the following problem:

> dpkg: error processing archive 
> /var/cache/apt/archives/libcblas3_3.2.1+dfsg-1_amd64.deb (--unpack):
>  trying to overwrite '/usr/lib/x86_64-linux-gnu/libcblas.so.3', which is also 
> in package libatlas3-base:amd64 3.10.3-4
> Processing triggers for libc-bin (2.24-17) ...
> Errors were encountered while processing:
>  /var/cache/apt/archives/libcblas3_3.2.1+dfsg-1_amd64.deb
> E: Sub-process /usr/bin/dpkg returned an error code (1)

First, note that the problem has appeared because I have multiarchified atlas,
and therefore the libcblas.so.3 that it provides has moved from /usr/lib to
/usr/lib/<multiarch>. But in a sense the situation was already problematic
before that change , because two libraries with the same SONAME were in the
dynamic linker search path.

Same reasoning for #874803.

Now, I think there are two options for fixing this issue:

- completely dropping libcblas3 (and libcblas-dev). These two packages are
  actually useless, because they provide the same C interface to BLAS than the
  other BLAS implementations (libblas3, libatlas3-base, libopenblas-base), but
  in a less optimized way;

- or making the two packages conflict.

Andreas (and others), what do you think?

Best,

-- 
⢀⣴⠾⠻⢶⣦⠀  Sébastien Villemot
⣾⠁⢠⠒⠀⣿⡁  Debian Developer
⢿⡄⠘⠷⠚⠋⠀  http://sebastien.villemot.name
⠈⠳⣄⠀⠀⠀⠀  http://www.debian.org

Attachment: signature.asc
Description: PGP signature

-- 
debian-science-maintainers mailing list
debian-science-maintainers@lists.alioth.debian.org
http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/debian-science-maintainers

Reply via email to