Bug#778106: rheolef: ftbfs with GCC-5

Cyril Brulebois kibi at debian.org
Mon Aug 3 18:43:55 UTC 2015


Hi,

I haven't touched this:

Matthias Klose <doko at debian.org> (2015-07-25):
> 
> after fixing
> 
> --- a/skit/plib2/load_chunk.h
> +++ b/skit/plib2/load_chunk.h
> @@ -28,7 +28,7 @@ load_chunk (std::istream& s, RandomItera
>  {
>      while (iter != last)
>          if (!(s >> *iter++)) return false;
> -    return s;
> +    return bool(s);
>  }
>  template <class RandomIterator, class GetFunction>
>  inline
> @@ -37,7 +37,7 @@ load_chunk (std::istream& s, RandomItera
>  {
>      while (iter != last)
>          if (! get_element (s, *iter++)) return false;
> -    return s;
> +    return bool(s);
>  }
>  } // namespace rheolef
>  #endif // _RHEOLEF_LOAD_CHUNK_H

but I had to patch configure.ac because hardening-cc can get in the way
(binaries exist but can't compile anything if the relevant package isn't
installed):
| --- rheolef-6.5/configure.ac~
| +++ rheolef-6.5/configure.ac
| @@ -459,9 +459,9 @@ dnl-------------------------------------
|  USER_CFLAGS=${CFLAGS-"-O2"}
|  USER_CXXFLAGS="$CXXFLAGS"
| 
| -AC_PROG_CC(gcc-4.5 gcc cc icc cl)
| +AC_PROG_CC(gcc cc icc cl)
|  AC_PROG_CPP
| -AC_PROG_CXX(g++-4.5 g++ c++ cxx icpc KCC CC CC cc++ xlC aCC)
| +AC_PROG_CXX(g++ c++ cxx icpc KCC CC CC cc++ xlC aCC)
|  AC_PROG_CXXCPP
| 
|  # when GNU C/C++ is recognized, force "-O2 -g" => skip it

> I get: […]

I'm getting something similar in unstable right now, w/o wrapping:
| make  all-am
| /bin/bash ../../libtool  --tag=CXX   --mode=compile g++ -DHAVE_CONFIG_H -I. -I../../config  -fno-strict-aliasing -I../../include -I/usr/include/suitesparse        -D_FORTIFY_SOURCE=2  -O3 -D_FORTIFY_SOURCE=2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -std=c++11 -Wall -Wno-unused -Werror -Wno-strict-aliasing -O2  -MT index_set.lo -MD -MP -MF .deps/index_set.Tpo -c -o index_set.lo index_set.cc
| libtool: compile:  g++ -DHAVE_CONFIG_H -I. -I../../config -fno-strict-aliasing -I../../include -I/usr/include/suitesparse -D_FORTIFY_SOURCE=2 -O3 -D_FORTIFY_SOURCE=2 -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -std=c++11 -Wall -Wno-unused -Werror -Wno-strict-aliasing -O2 -MT index_set.lo -MD -MP -MF .deps/index_set.Tpo -c index_set.cc  -fPIC -DPIC -o .libs/index_set.o
| In file included from /usr/include/boost/serialization/set.hpp:26:0,
|                  from ../../include/rheolef/index_set.h:38,
|                  from index_set.cc:21:
| /usr/include/boost/serialization/detail/stack_constructor.hpp: In constructor 'boost::serialization::detail::stack_construct<Archive, T>::stack_construct(Archive&, unsigned int)':
| /usr/include/boost/serialization/detail/stack_constructor.hpp:54:9: error: 'load_construct_data_adl' is not a member of 'boost::serialization'
|          boost::serialization::load_construct_data_adl(
|          ^
| Makefile:727: recipe for target 'index_set.lo' failed

One might need to take a closer look at this package anyway, given one
can see this in debian/rules:
|         # Prepare the debian control files necessary for the creation of the packages
|         cat debian/shlibs.librheolef1.X | \
|           sed -e "s/\@LIBRHEOLEF\@/$(LIBRHEOLEF)/g" \
|               -e "s/\@MAJOR_VERSION\@/$(MAJOR_VERSION)/g" \
|               -e "s/\@MINOR_VERSION\@/$(MINOR_VERSION)/g" \
|               -e "s/\@VERSION\@/$(VERSION)/g" \
|               -e "s/\@SONAME\@/$(SONAME)/g" > debian/shlibs.$(LIBRHEOLEF)
|         echo 'shlibs:Depends=libc6 (>= 2.7-1), libgcc1 (>= 1:4.3.1-9), libstdc++6 (>= 4.3.1-9)' >  debian/$(LIBRHEOLEF).substvars

and I'm not sure this is exactly correct…

This package looks like a good candidate for a removal if it gets in the
way.

Mraw,
KiBi.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/debian-science-maintainers/attachments/20150803/7ee63abc/attachment.sig>


More information about the debian-science-maintainers mailing list