[Debian-med-packaging] Debian Med package for MrBayes

Tue Jun 26 15:11:26 BST 2018

Thanks for the log! Much appreciated!

The error is due to a declaration in one of our header files.  This
declaration is commented out in our code on the 'develop' branch (since
it's of a Beagle function and thus already declared in the Beagle
header).  This leads me to believe that you are in fact compiling the
'master' branch.  Could you please retry with the 'develop' branch?

Andreas

On Tue, Jun 26, 2018 at 03:16:32PM +0200, Andreas Tille wrote:
> Dear Johan,
> 
> On Tue, Jun 26, 2018 at 11:22:35AM +0000, Johan Nylander wrote:
> > Dear Andreas!
> > 
> > The code for the MrBayes package is coming closer to a new release. 
> > Again, the previous release (currently on Debian) was hosted on 
> > sourceforge, but we now moved everything to github. During the move, we 
> > also did a major overhaul of the build system.
> 
> Thanks a lot for the early warning about changes.  That's really
> appreciated.
>  
> > Would it be OK for you to take a quick look at what we currently have, 
> > and maybe give us some pointers to what we can improve in order for the 
> > incorporation into debian repos to be as smooth as possible?
> > 
> > Here is the repo for the the next upcoming release ("develop" branch):
> > 
> > https://github.com/NBISweden/MrBayes
> 
> I've checked this out, did some adaptations to the Debian build system
> but finally I get a build error where I have no clue how to fix it:
> 
> ...
> gcc -DHAVE_CONFIG_H -I.   -Wdate-time -D_FORTIFY_SOURCE=2 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -mabm -I/usr/include/libhmsbeagle-1 -g -O2 -std=c99 -pedantic -g -O2 -fdebug-prefix-map=/build/mrbayes-3.2.7~pre=. -fstack-protector-strong -Wformat -Werror=format-security -Wall -c -o mb-bayes.o `test -f 'bayes.c' || echo './'`bayes.c
> gcc -DHAVE_CONFIG_H -I.   -Wdate-time -D_FORTIFY_SOURCE=2 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -mabm -I/usr/include/libhmsbeagle-1 -g -O2 -std=c99 -pedantic -g -O2 -fdebug-prefix-map=/build/mrbayes-3.2.7~pre=. -fstack-protector-strong -Wformat -Werror=format-security -Wall -c -o mb-best.o `test -f 'best.c' || echo './'`best.c
> gcc -DHAVE_CONFIG_H -I.   -Wdate-time -D_FORTIFY_SOURCE=2 -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -mabm -I/usr/include/libhmsbeagle-1 -g -O2 -std=c99 -pedantic -g -O2 -fdebug-prefix-map=/build/mrbayes-3.2.7~pre=. -fstack-protector-strong -Wformat -Werror=format-security -Wall -c -o mb-command.o `test -f 'command.c' || echo './'`command.c
> In file included from command.c:40:0:
> mbbeagle.h:24:14: error: conflicting types for 'beagleGetVersion'
>  extern char *beagleGetVersion (void);
>               ^~~~~~~~~~~~~~~~
> In file included from bayes.h:58:0,
>                  from command.c:38:
> /usr/include/libhmsbeagle-1/libhmsbeagle/beagle.h:240:30: note: previous declaration of 'beagleGetVersion' was here
>  BEAGLE_DLLEXPORT const char* beagleGetVersion(void);
>                               ^~~~~~~~~~~~~~~~
> command.c: In function 'FreeCharacters':
> command.c:9050:9: warning: this 'if' clause does not guard... [-Wmisleading-indentation]
>          if (numChar > numTaxa)
>          ^~
> command.c:9052:13: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the 'if'
>              tempSetNeg = (int *) SafeRealloc ((void *)tempSetNeg, (size_t)numTaxa*sizeof(int));
>              ^~~~~~~~~~
> make[4]: *** [Makefile:463: mb-command.o] Error 1
> 
> 
> I've attached the full debian build log - may be you scroll
> down until "dh_auto_configure" or so.
> 
> 
> > Thank you very much
> 
> Thanks to you
> 
>       Andreas.
> 
>  
> > Yours
> > 
> > Johan Nylander & Andreas Kusalananda Kähäri
> > 
> > 
> > 
> > On 2018-02-16 09:59, Andreas Tille wrote:
> > > Dear Johan,
> > >
> > > On Thu, Feb 15, 2018 at 01:32:24PM +0000, Johan Nylander wrote:
> > >> I saw on the Debian pages that you are involved in packaging the
> > >> software MrBayes for Debian Med.
> > >>
> > >> I just wanted to inform you that we are currently in the process of
> > >> moving all code base from SourceForge to Github.
> > > Thanks a lot for this very helpful information!  That's really welcome
> > > since our automatic system to seek for new upstream versions would not
> > > notice this.
> > >   
> > >> The latest version of the code is in pre-release state at
> > >> https://github.com/NBISweden/MrBayes, and we are hoping to complete the
> > >> next release soon.
> > > A short ping would be helpful to let us issue an updated package right
> > > after this.
> > >   
> > >> Furthermore, is there anything  we can do in order for the building
> > >> process to be as smooth as possible on your side?
> > > There is one thing I should have probably reported earlier:  In Debian
> > > the "source" of all "binaries" is required which also affects the source
> > > of PDF documentation.  While this is part of your repository it was not
> > > contained inside the release tarball.  We fixed this by the following
> > > extra code:
> > >
> > >
> > > # remove PDF without source
> > > rm -f $documentation/commref*.pdf
> > > # fetch sources for remaining doc from SVN into temporary dir
> > > svn export https://svn.code.sf.net/p/mrbayes/code/documentation tmp
> > > # remove unneeded stuff from documentation SVN
> > > rm -f tmp/Manual*.pdf
> > > rm -f tmp/*.rtf
> > > rm -f tmp/*.rb
> > > rm -f tmp/Release_instructions.txt
> > > # move remaining doc to source dir
> > > mv tmp/* documentation
> > > rmdir tmp
> > >
> > >
> > > It would be really helpful if you would include all the sources that are
> > > needed to build the PDF documents would be included into the download
> > > tarball which would safe us the extra code.  I realised that the
> > > prerelease MrBayes-pre-3.2.7.tar.gz contains a restructured doc directory
> > > including Makefile.{am,in} - may be you realised my request even before
> > > I asked for it.
> > >
> > > Kind regards and thanks again
> > >
> > >         Andreas.
> > >
> > 
> > -- 
> > 
> > Johan A. A. Nylander, PhD.
> > Department of Bioinformatics and Genetics
> > Swedish Museum of Natural History
> > BOX 50007, 10405 Stockholm, Sweden
> > Visiting addr.: Frescativägen 40, 11418 Stockholm
> > Ph.: +46(0)851954056, e-mail:Johan.Nylander at nrm.se
> 
> -- 
> http://fam-tille.de

-- 
Andreas Kusalananda Kähäri,
National Bioinformatics Infrastructure Sweden (NBIS),
Uppsala University, Sweden.