Debian Bug report logs - #842135
nmu: slepc4py_3.7.0-2+b3

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Drew Parsons <dparsons@debian.org>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: release.debian.org: asdasd

Date: Wed, 26 Oct 2016 16:23:21 +0800

Package: release.debian.org
Severity: normal

Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
From: Drew Parsons <dparsons@debian.org>
To: Debian Bug Tracking System <submit@bugs.debian.org>
Subject: nmu: slepc4py_3.7.0-2+b3
Message-ID: <147746397798.2727.2988347581649171064.reportbug@grendel.emerall.com>
X-Mailer: reportbug 6.6.6
Date: Wed, 26 Oct 2016 14:39:37 +0800

Package: release.debian.org
Severity: normal
User: release.debian.org@packages.debian.org
Usertags: binnmu

nmu slepc4py_3.7.0-2+b3 . amd64 arm64 armel armhf i386 mips mipsel powerpc ppc64el s390x hppa hurd-i386 kfreebsd-amd64 kfreebsd-i386 x32 . unstable . -m "Rebuild against slepc 3.7.3+dfsg1-2 (Provides virtual libslepc3.7, libslepc-complex-3.7)"

thanks

Provision of virtual packages libslepc3.7, libslepc-complex-3.7 makes
shlib dependencies more sane, reducing the urgency of binNMUs against
SLEPc in the future.

alpha, mips64el, ppc64, sparc64 are already built against 
slepc 3.7.3+dfsg1-2 

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.7.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_AU.UTF-8, LC_CTYPE=en_AU.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

-- System Information:
Debian Release: stretch/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 4.7.0-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_AU.UTF-8, LC_CTYPE=en_AU.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Message #12 received at 842135-done@bugs.debian.org (full text, mbox, reply):

From: Emilio Pozuelo Monfort <pochu@debian.org>

To: Drew Parsons <dparsons@debian.org>, 842135-done@bugs.debian.org

Subject: Re: Bug#842135: release.debian.org: asdasd

Date: Wed, 26 Oct 2016 10:37:40 +0200

On 26/10/16 10:23, Drew Parsons wrote:
> Package: release.debian.org
> Severity: normal
> 
> Content-Type: text/plain; charset="us-ascii"
> MIME-Version: 1.0
> Content-Transfer-Encoding: 7bit
> From: Drew Parsons <dparsons@debian.org>
> To: Debian Bug Tracking System <submit@bugs.debian.org>
> Subject: nmu: slepc4py_3.7.0-2+b3
> Message-ID: <147746397798.2727.2988347581649171064.reportbug@grendel.emerall.com>
> X-Mailer: reportbug 6.6.6
> Date: Wed, 26 Oct 2016 14:39:37 +0800
> 
> Package: release.debian.org
> Severity: normal
> User: release.debian.org@packages.debian.org
> Usertags: binnmu
> 
> nmu slepc4py_3.7.0-2+b3 . amd64 arm64 armel armhf i386 mips mipsel powerpc ppc64el s390x hppa hurd-i386 kfreebsd-amd64 kfreebsd-i386 x32 . unstable . -m "Rebuild against slepc 3.7.3+dfsg1-2 (Provides virtual libslepc3.7, libslepc-complex-3.7)"
> 
> thanks
> 
> Provision of virtual packages libslepc3.7, libslepc-complex-3.7 makes
> shlib dependencies more sane, reducing the urgency of binNMUs against
> SLEPc in the future.
> 
> alpha, mips64el, ppc64, sparc64 are already built against 
> slepc 3.7.3+dfsg1-2 

Done, including deal.ii getdp slepc4py and dolfin.

I don't quite understand why slepc and petsc have versioned packages and use
alternatives to provide the shared library, but I haven't looked closely at it.
Not sure I want to though :P

Cheers,
Emilio

Message #17 received at 842135@bugs.debian.org (full text, mbox, reply):

From: Drew Parsons <dparsons@debian.org>

To: 842135@bugs.debian.org

Subject: Re: Bug#842135: release.debian.org: nmu: slepc4py_3.7.0-2+b3

Date: Wed, 26 Oct 2016 16:58:19 +0800

On Wed, 26 Oct 2016 10:37:40 +0200 Emilio Pozuelo Monfort <pochu@debian
.org> wrote:
> 
> Done, including deal.ii getdp slepc4py and dolfin.

Thanks.
 
> I don't quite understand why slepc and petsc have versioned packages
and use
> alternatives to provide the shared library, but I haven't looked
closely at it.
> Not sure I want to though :P


Since you asked ;) ...

PETSc upstream has arranged the library so that the preferred version
can be easily identified with env variable PETSC_DIR.

Each installation (referenced by PETSC_DIR) is patch specific, i.e. the
library is located in /usr/lib/petscdir/3.7.3 or
/usr/lib/petscdir/3.7.4 etc. This allows the "flexibility" of parallel
installation of different patch versions.

But the soname only contains the minor version, 3.7 not 3.7.4.

slepc just follows the structure of petsc.

I guess what you're asking is why upstream bothers providing patch-
specific installations when the soname is the same between them. 
Upstream will have to speak for themselves, but I could imagine that
doing it that way makes it easier to manage installations on high
performance supercomputers.  Installations on the supercomputers tend
to be conservative. They worry about differences in the results of
computations arising from different patch versions, even though the
client software will happily run with any version with the same soname.

Whatever upstream's reasons might be,  I've set up the Debian packaging
to reflect this patch-version "flexibility" provided by the upstream
directory structure.

Drew

Send a report that this bug log contains spam.