Summary: | sci-libs/netcdf-4.1.3 fails with -l-L/usr/lib/gcc/x86_64-pc... | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Juergen Rose <rose> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | RESOLVED FIXED | ||
Severity: | normal | ||
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
/var/tmp/portage/sci-libs/netcdf-4.1.3/temp/build.log
/usr/local/portage/sci-libs/netcdf/netcdf-4.1.3.ebuild /usr/local/portage/sci-libs/netcdf/files/netcdf-postdeps.patch /usr/local/portage/sci-libs/netcdf/netcdf-4.1.3.ebuild |
Description
Juergen Rose
2012-09-01 11:59:17 UTC
Yes I got that too Juergen. I had to switched to gcc-4.5.x to compile netcdf. I slacked in not filling the bug myself. It is a rather strange flag problem and I don't have the expertise for it. For info the exact same thing happens with netcdf-4.1.1-r4. Just to set the record straight on this, paraview ship with netcdf-4.1.2 (not in portage). Paraview won't compile against netcdf-4.1.1-r4 or 4.2+. Just 4.1.3. Do we really need that version of netcdf for paraview? Can't we patch paraviewto support newer netcdf? It seems there is some automagic to use mpi in this version of netcdf (In reply to comment #2) > Do we really need that version of netcdf for paraview? Can't we patch > paraviewto support newer netcdf? > > It seems there is some automagic to use mpi in this version of netcdf Well it is using the c++ interface, and more and more c++ makes me want to vomit. OK, the first thing about it now that I thought a little bit about it is that the c++ headers are split in netcdf-4.2 - vtk/paraview only look at netcdfcpp.h. Getting the headers in would be the first step. If we are lucky that may be enough. What about sci-libs/netcdf-cxx? OK - I gave a go at 4.2 on Friday and I used both netcdf and netcdf-cxx, that's why I can talk about netcdfcpp.h being split. You would expect it to be in netcdf-cxx but it isn't, there is a number of headers instead if I remember well. Had another look at using 4.2 patching for the appropriate netcdf-cxx headers but I am not going anywhere with that: [ 23%] Building CXX object VTK/IO/CMakeFiles/vtkIO.dir/vtkMPASReader.cxx.o cd /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/IO && /usr/bin/x86_64-pc-linux-gnu-g++ -DvtkIO_EXPORTS -DVTK_PYTHON_BUILD -DVTK_IN_VTK -DNDEBUG -O1 -march=native -pipe -ggdb -Wno-deprecated -Wno-deprecated -fPIC -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Common -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/VolumeRendering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Rendering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Charts -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Chemistry -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/vtkalglib -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Wrapping/Python -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Wrapping/Python -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Infovis -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/TextAnalysis -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Geovis -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Views -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Parallel -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/VolumeRendering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Hybrid -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Widgets -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Rendering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Charts -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Chemistry -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Rendering/Testing/Cxx -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Imaging -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Graphics -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/GenericFiltering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Filtering -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Common -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Common/Testing/Cxx -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/vtklibproj4 -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/vtklibproj4 -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/DICOMParser -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/DICOMParser -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/LSDyna -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/LSDyna -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/MaterialLibrary -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/MaterialLibrary -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/vtkmetaio -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/vtkmetaio -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/verdict -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/verdict -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/Cosmo -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/Cosmo -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/VPIC -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/VPIC -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/utf8/source -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/GUISupport/Qt -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/GUISupport/Qt -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/GUISupport/Qt/Chart -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/GUISupport/Qt/Chart -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/vtkalglib -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/Utilities/ftgl/src -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/Utilities/ftgl -I/usr/include/libxml2 -I/usr/include/freetype2 -I/scratch/portage/sci-visualization/paraview-3.14.1-r2/work/paraview-3.14.1_build/VTK/IO -I/usr/include/mysql -I/usr/include/python2.7 -fvisibility=hidden -fvisibility-inlines-hidden -o CMakeFiles/vtkIO.dir/vtkMPASReader.cxx.o -c /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:129:3: error: 'NcFile' does not name a type /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:130:3: error: 'NcVar' does not name a type /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:131:3: error: 'NcVar' does not name a type /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx: In constructor 'vtkMPASReader::Internal::Internal()': /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:113:5: error: class 'vtkMPASReader::Internal' does not have any field named 'ncFile' /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:117:13: error: 'class vtkMPASReader::Internal' has no member named 'cellVars' /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:118:13: error: 'class vtkMPASReader::Internal' has no member named 'pointVars' /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx: In destructor 'vtkMPASReader::Internal::~Internal()': /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:123:14: error: 'class vtkMPASReader::Internal' has no member named 'ncFile' /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:125:14: error: 'ncFile' was not declared in this scope /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx: At global scope: /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:175:14: error: 'NcFile' was not declared in this scope /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:175:14: note: suggested alternative: /usr/include/ncFile.h:18:9: note: 'netCDF::NcFile' /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:175:22: error: 'ncFile' was not declared in this scope /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:175:30: error: 'NcToken' was not declared in this scope /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:175:42: error: expression list treated as compound expression in initializer [-fpermissive] /scratch/portage/sci-visualization/paraview-3.14.1-r2/work/ParaView-3.14.1-Source/VTK/IO/vtkMPASReader.cxx:176:1: error: expected ',' or ';' before '{' token make[2]: *** [VTK/IO/CMakeFiles/vtkIO.dir/vtkMPASReader.cxx.o] Error 1 Just a quick note. The compilation problem only occur in the fortran interface. If USE is set to -fortran netcdf-4.1.3 will happily compile and install. I wonder if it is somehow related to bug #348922? I have a very dirty hack. I have talked with people that are actually doing weather research (NZ metservice and New Zealand institute for weather and atmospheric aka NIWA) and all kind of stuff using netCDF. The word is don't use 4.2. 4.1.3 while using hdf5 still use some bits of netcdf3 underneath so porting is non-trivial. OK so this only happens with the fortran interface and hdf5[mpi]. In my case openmpi, I haven't tried other. The problem ultimately is in the generated libtool file where we get the following: predep_objects="" postdep_objects="/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../lib64/crti.o /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/crtbeginS.o /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/crtendS.o /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../lib64/crtn.o" predeps="" postdeps="-l -l -L/usr/lib64 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/lib -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../.. -lmpi_f90 -lmpi_f77 -lmpi -lopen-rte -lopen-pal -ldl -lnsl -lutil -lgfortran -lm -ldl -lgfortran -lm -lgcc_s -lquadmath -lm -lgcc_s -lpthread -lc -lgcc_s" # The library search path used internally by the compiler when linking # a shared library. compiler_lib_search_path="-L/usr/lib64" # ### END LIBTOOL TAG CONFIG: FC Or a variation thereof. I initially had something in "-l -lm -ldl -l -l" in there. So the libtool generation is messed up. Eliminating the "-l " from libtool solve the problem. Fedora has a patch for configure to deal with this (I presume): http://pkgs.fedoraproject.org/cgit/netcdf.git/diff/netcdf-postdeps.patch?h=f17&id=b9d7608ce6081b097a9631efeef527238eb9a01e Is this some kind of bug in autoconf/libtool or more subtle? If I try to use the patch from http://pkgs.fedoraproject.org/cgit/netcdf.git/diff/netcdf-postdeps.patch?h=f17&id=b9d7608ce6081b097a9631efeef527238eb9a01e , I get now a slot conflict: root@impala:/usr/local/portage/sci-libs(9)# emerge -vuDN world These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild U ] sci-libs/proj-4.8.0 [4.7.0] USE="java%* -static-libs" 2,968 kB [ebuild U ] net-fs/openafs-kernel-1.6.1-r1 [1.6.1] 0 kB [ebuild U ] net-print/gtklp-1.2.9 [1.2.8a] USE="nls ssl" 0 kB [ebuild U ] net-fs/openafs-1.6.1-r1 [1.6.1] USE="doc kerberos pam" 0 kB [ebuild U ] dev-perl/HTML-Mason-1.500.0 [1.480.0] USE="doc -modperl {-test}" 0 kB [ebuild UD ] sci-libs/netcdf-4.1.3::x-portage [4.2.1.1::gentoo] USE="cxx%* doc fortran%* hdf5 -dap* -static-libs (-examples%*) (-hdf%*) (-mpi%*) (-szip%*) (-test%) (-tools%*)" 0 kB [ebuild U ] media-gfx/sane-backends-1.0.23 [1.0.22-r2] USE="doc gphoto2 ipv6 usb -avahi -v4l -xinetd%" SANE_BACKENDS="abaton agfafocus apple artec artec_eplus48u as6e avision bh canon canon630u canon_dr canon_pp cardscan coolscan coolscan2 coolscan3 dc210 dc240 dc25 dell1600n_net dmc epjitsu epson epson2 fujitsu genesys gt68xx hp hp3500 hp3900 hp4200 hp5400 hp5590 hpljm1005 hpsj5s hs2p ibm kodak kodakaio%* kvs1025 kvs20xx kvs40xx%* leo lexmark ma1509 magicolor matsushita microtek microtek2 mustek mustek_pp mustek_usb nec net niash p5 pie pixma plustek plustek_pp qcam ricoh rts8891 s9036 sceptre sharp sm3600 sm3840 snapscan sp15c st400 stv680 tamarack teco1 teco2 teco3 {test} u12 umax umax1220u umax_pp xerox_mfp -pnm" 0 kB [ebuild U #] sci-visualization/paraview-3.14.1-r4::science [3.14.1::science] USE="boost cg doc examples ffmpeg gui mpi mysql plugins python theora webkit -coprocessing -gl2ps% -nvcontrol" 0 kB Total: 8 packages (7 upgrades, 1 downgrade), Size of downloads: 2,968 kB !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: sci-libs/proj:0 (sci-libs/proj-4.8.0::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (sci-libs/proj-4.7.0::gentoo, installed) pulled in by <sci-libs/proj-4.8 required by (dev-perl/PDL-2.4.11::gentoo, installed) (and 1 more with the same problem) sci-libs/netcdf:0 (sci-libs/netcdf-4.1.3::x-portage, ebuild scheduled for merge) pulled in by ~sci-libs/netcdf-4.1.3[cxx,hdf5] required by (sci-visualization/paraview-3.14.1-r4::science, ebuild scheduled for merge) (sci-libs/netcdf-4.2.1.1::gentoo, installed) pulled in by >=sci-libs/netcdf-4.2 required by (sci-libs/netcdf-cxx-4.2::gentoo, installed) If I delete netcdf-cxx, the slot conflicts remains, because depends dev-lang/gdl-0.9.2-r2 on netcdf-cxx: root@impala:/usr/local/portage/sci-libs(11)# emerge -vuDNt world These are the packages that would be merged, in reverse order: Calculating dependencies... done! [nomerge ] dev-lang/gdl-0.9.2-r2 USE="fftw hdf hdf5 imagemagick netcdf openmp postscript proj python wxwidgets -grib -gshhs -static-libs {-test} -udunits" [ebuild N ] sci-libs/netcdf-cxx-4.2 USE="examples -static-libs" 386 kB [nomerge ] media-gfx/xsane-0.998-r1 USE="gimp jpeg lcms nls ocr png tiff" ... [ebuild UD ] sci-libs/netcdf-4.1.3::x-portage [4.2.1.1::gentoo] USE="cxx%* doc fortran%* hdf5 -dap* -static-libs (-examples%*) (-hdf%*) (-mpi%* ... !!! Multiple package instances within a single package slot have been pulled !!! into the dependency graph, resulting in a slot conflict: sci-libs/proj:0 (sci-libs/proj-4.8.0::gentoo, ebuild scheduled for merge) pulled in by (no parents that aren't satisfied by other packages in this slot) (sci-libs/proj-4.7.0::gentoo, installed) pulled in by <sci-libs/proj-4.8 required by (dev-perl/PDL-2.4.11::gentoo, installed) (and 1 more with the same problem) sci-libs/netcdf:0 (sci-libs/netcdf-4.1.3::x-portage, ebuild scheduled for merge) pulled in by ~sci-libs/netcdf-4.1.3[cxx,hdf5] required by (sci-visualization/paraview-3.14.1-r4::science, ebuild scheduled for merge) (sci-libs/netcdf-4.2.1.1::gentoo, installed) pulled in by >=sci-libs/netcdf-4.2 required by (sci-libs/netcdf-cxx-4.2::gentoo, ebuild scheduled for merge) If I then remove all netcdf-4.2*.ebuild from my /usr/local/portage/sci-libs/netcdf overlay, the slot conflict disappears, but 'emerge netcdf' fails with the old error: >>> Emerging (1 of 6) sci-libs/netcdf-4.1.3 from x-portage * netcdf-4.1.3.tar.gz SHA256 SHA512 WHIRLPOOL size ;-) ... [ ok ] ... /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/include/netcdf.h:1448:1: note: expected 'long long int *' but argument is of type 'int *' /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/fortran/fort-nc4.c: In function 'nf_get_var_int64_': /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/fortran/fort-nc4.c:223:1: warning: passing argument 3 of 'nc_get_var_longlong' from incompatible pointer type [enabled by default] /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/include/netcdf.h:1539:1: note: expected 'long long int *' but argument is of type 'int *' /bin/sh ../libtool --tag=FC --mode=link mpif90 -O2 -pipe -version-number 5:1:0 -Wl,-O1 -Wl,--as-needed -o libnetcdff.la -rpath /usr/lib64 fort-attio.lo fort-control.lo fort-dim.lo fort-genatt.lo fort-geninq.lo fort-genvar.lo fort-lib.lo fort-misc.lo fort-v2compat.lo fort-vario.lo fort-var1io.lo fort-varaio.lo fort-varmio.lo fort-varsio.lo fort-nc4.lo ../liblib/libnetcdf.la ../f90/libnetcdff90.la -lsz -lhdf5_hl -lhdf5 -lm -lz -lcurl libtool: link: mpif90 -shared -fPIC .libs/fort-attio.o .libs/fort-control.o .libs/fort-dim.o .libs/fort-genatt.o .libs/fort-geninq.o .libs/fort-genvar.o .libs/fort-lib.o .libs/fort-misc.o .libs/fort-v2compat.o .libs/fort-vario.o .libs/fort-var1io.o .libs/fort-varaio.o .libs/fort-varmio.o .libs/fort-varsio.o .libs/fort-nc4.o -Wl,--whole-archive ../f90/.libs/libnetcdff90.a -Wl,--no-whole-archive -Wl,-rpath -Wl,/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3_build/liblib/.libs -Wl,--as-needed ../liblib/.libs/libnetcdf.so -lsz -lhdf5_hl -lhdf5 -lz -lcurl -L/usr/lib64 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3 -l -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/lib -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../.. /usr/lib64/libmpi_f90.so /usr/lib64/libmpi_f77.so /usr/lib64/libmpi.so -lrt -lnsl -lutil -lhwloc -ldl /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/libgfortran.so /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/libquadmath.so -lm -lpthread -lc -lgcc_s -O2 -Wl,-O1 -pthread -Wl,-soname -Wl,libnetcdff.so.5 -o .libs/libnetcdff.so.5.1.0 /usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -l-L/usr/lib/gcc/x86_64-pc-linux-gnu/4.6.3/../../../../lib64 collect2: ld returned 1 exit status make[3]: *** [libnetcdff.la] Error 1 Created attachment 325980 [details]
/usr/local/portage/sci-libs/netcdf/netcdf-4.1.3.ebuild
The ebuild I used.
Created attachment 325982 [details]
/usr/local/portage/sci-libs/netcdf/files/netcdf-postdeps.patch
The patch again.
The patch from fedora cannot work as is. We run autoreconfig in src_prepare so the configure script is regenerated however you patched it. You could possibly try to apply it - or an updated version of it after autotools-utils_src_prepare but I don't think we can guaranty that a patch to configure will apply generally after autoreconfig. Possibly you could try sed -i 's:test $p = "-R":test $p = "-R" || test $p = "-l":' configure after autotools-utils_src_prepare. I am working on something else right now so cannot test. Thanks Francois, the patch corresonding to Comment 14 works for me, at least on three systems. Created attachment 325988 [details] /usr/local/portage/sci-libs/netcdf/netcdf-4.1.3.ebuild netcdf-4.1.3.ebuild corresponding to Comment 14 It's extremely dirty but that's the best workaround I found so far. + 21 Aug 2013; Christoph Junghans <ottxor@gentoo.org> netcdf-4.1.3.ebuild: + fixed linking issue (bug #433647) + Now 'emerge netcdf' fails with the netcdf with: ... /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90.texi:2085: warning: @item missing argument /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90.texi:6982: raising the section level of @unnumberedsubsec which is too low mv: cannot move ‘.am23616/netcdf-f90.info’ to ‘.//var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/’: No such file or directory Makefile:484: recipe for target '/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90.info' failed make: *** [/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90.info] Error 1 make: Leaving directory '/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3_build/man4' * ERROR: sci-libs/netcdf-4.1.3::x-portage failed (compile phase): * emake failed (In reply to Juergen Rose from comment #19) > Now 'emerge netcdf' fails with the netcdf with: > ... > /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90. > texi:2085: warning: @item missing argument > /var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90. > texi:6982: raising the section level of @unnumberedsubsec which is too low > mv: cannot move ‘.am23616/netcdf-f90.info’ to > ‘.//var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/’: No such > file or directory > Makefile:484: recipe for target > '/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90. > info' failed > make: *** > [/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3/man4/netcdf-f90. > info] Error 1 > make: Leaving directory > '/var/tmp/portage/sci-libs/netcdf-4.1.3/work/netcdf-4.1.3_build/man4' > * ERROR: sci-libs/netcdf-4.1.3::x-portage failed (compile phase): > * emake failed please open a new bug with full build.log. Looks like a different issue. |