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

Re: Open issue with libbpp-core



Dear all,

These .all files are groups of header files, provided for convenience of the user (was an old request of a reviewer who complained that their was too many files to include!). Bio++ classes are sorted in directories, such as Bpp/Phyl/Model/Protein for instance, which contains all classes of phylogenetic models for protein sequences. It was kind of a trick to emulate "subpackages", as g++ did not allow to include a directory: #include <Bpp/Phyl/Model.all> would include all phylogenetic models, while #include <Bpp/Phyl/Model/Protein.all> would only include all protein models. As more libs were introduced, however, these .all files could not be pre-generated and had to be generated while installing, which led to several post-install and post-uninstall scripts. We had agreed however to remove them some time ago, and they will not be present in the next version of the libraries. In theory they can be removed from the current version I think, only some unit tests might not compile anymore...

I apologize that we're so delayed with the next release. For sure it will be much easier to maintain, and I'm including all your suggestions for easier packaging. As for gcc6, it should normally work: at least I could compile the latest code in a virtual machine and did not face any issue.

Best,

J.


On Wed, Jun 29, 2016 at 9:58 PM, Andreas Tille <andreas@an3as.eu> wrote:
Hi Mattia,

On Wed, Jun 29, 2016 at 03:54:14PM +0000, Mattia Rizzolo wrote:
> > I understand this, but all the other things happaning in the script
> > seem to be related to automatically create some header files, I'm not
> > sure whether these scripts should be run by both, the lib*-dev and the
> > lib${soname] package. Since it seems that Andreas started this, I'd ask
> > hit to clarify this. 

Hmmmm, I was actually not starting this but I took over an unmaintained
package which was once created by Julien Yann Dutheil (in CC) who is
also upstream.  I tried to make some sense out of the packaging and
moved it into team maintenance.  I'm happy that somebody else is having
an additional look since I might have overlooked something easily.

> yep, I don't have a clue what the rest of the script does (expecially,
> on postrm during remove?!?).  But still, the ldconfig call still looks
> superfluos to me.
> But let's wait on Andreas :)

I have removed the ldconfig call in any case.  From my point of view:
The postinst is really only needed inside the *-dev package since only
there are any header files which are assembled in the *.all files.  I
have no idea for what purpose those *.all files are good for.  Once I
considered to create those files inside the package Build process which
would simplify the postinst issue but then I decided to change as less
as possible of things I might not fully understand.

Julien, could you please comment on those *.all files?

Kind regards

      Andreas.

PS: Julien, the repository now contains a patch for a gcc-6 build issue
which you might be interested in.

--
http://fam-tille.de



--
Julien Y. Dutheil, Ph-D
0 (+49) 6421 178 986

§ Max Planck Institute for Evolutionary Biology
Molecular Systems Evolution
Department of Evolutionary Genetics
Plön -- GERMANY

§ Institute of Evolutionary Sciences - Montpellier
University of Montpellier 2 -- FRANCE

Reply to: