[CMake] Contents of link.txt differ between Solaris 8 and Linux
Tom Kacvinsky
tom.kacvinsky at vectorcast.com
Tue Dec 10 08:47:48 EST 2013
Hi,
I am using cmake 2.8.11.2 on Solaris 8 and Ubuntu 12.04.2 LTS, built from
source using the GCC tool chain.
I noted that the link.txt for one of our binaries on Solaris 8 are:
/home/tjk/gnatpro-7.1.2/bin/gnatmake
-aI/home/Users/tjk/vector/build/cmake/SunOS/.common-enums
-aI/home/Users/tjk/vector/build/cmake/SunOS/.chop_Linux
-aI/home/Users/tjk/vector/build/cmake/SunOS/.chop
-aI/home/Users/tjk/vector/build/cmake/SunOS/.chop_ada
-aI/home/Users/tjk/vector/source/cmake/vcast/
-aI/home/Users/tjk/vector/source/cmake/vcast/TGT_IO
-aO/home/Users/tjk/vector/build/cmake/SunOS/vcast/CMakeFiles/commonada.dir
-aO/home/Users/tjk/vector/build/cmake/SunOS/vcast/CMakeFiles/commonada.dir/__/.chop
-aO/home/Users/tjk/vector/build/cmake/SunOS/vcast/CMakeFiles/commonada.dir/__/.common-enums
monitor -o ../../vc/IO/monitor -cargs -largs
-L/home/Users/tjk/vector/build/cmake/SunOS
-L/home/Users/tjk/vector/build/cmake/SunOS/vc/lib
-L/home/Users/tjk/kdchart-2.3.0-qt-4.8.5/lib
-L/home/FLEXlm/FLEXlm_10/sun4_u5
-L/home/FLEXlm/FLEXlm_10/sun4_u5/activation/lib -L/home/tjk/python/libs
-L/usr/ccs/lib
-L/export/home/tjk/gnatpro-7.1.2/lib/gcc/sparc-sun-solaris2.8/4.7.3
-L/export/home/tjk/gnatpro-7.1.2/lib/gcc
-L/export/home/tjk/gnatpro-7.1.2/lib ../../vc/lib/libcinterfacetoada.a
../../vc/lib/libcpputilsforada.a ../../libcommonada.a
/home/TOOLS/libxml2/libxml2-2.6.27/lib/libxml2.a -llmgr_nomt -lcrvs -lsb
/home/tjk/gnatpro-7.1.2/lib/gcc/sparc-sun-solaris2.8/4.7.3/adalib/libgnat.a
-lsocket -lnsl -lintl -ldl -lpthread -lrt ../../vc/lib/libvxml_c.a
../../vc/lib/libcutilsforada.a /home/FLEXlm/FLEXlm_10/sun4_u5/lm_new.o -ldl
-lm /home/TOOLS/libxml2/libxml2-2.6.27/lib/libxml2.a
../../vc/lib/libcinterfacetoadastubs.a
../../vc/lib/libcinterfacetoada_ada.a -lpthread -lrt
/home/FLEXlm/FLEXlm_10/sun4_u5/lm_new.o -lm -lc -lc -lstdc++ -lm -lc -lc
Note the duplication of lm_new.o and several system libraries (-lm and
-lc). This causes problems with GNU binutils as there are multiply defined
symbols when the lm_new.o object file is duplicated. What I noted is that
this duplication does not happen on Linux:
/home/GCC/gcc-4.1.2/bin/gnatmake
-aI/home/Users/tjk/vector/build/cmakeLinux/Linux/.common-enums
-aI/home/Users/tjk/vector/build/cmakeLinux/Linux/.chop_Linux
-aI/home/Users/tjk/vector/build/cmakeLinux/Linux/.chop
-aI/home/Users/tjk/vector/build/cmakeLinux/Linux/.chop_ada
-aI/home/Users/tjk/vector/source/cmakeLinux/vcast/
-aI/home/Users/tjk/vector/source/cmakeLinux/vcast/TGT_IO
-aO/home/Users/tjk/vector/build/cmakeLinux/Linux/vcast/CMakeFiles/commonada.dir
-aO/home/Users/tjk/vector/build/cmakeLinux/Linux/vcast/CMakeFiles/commonada.dir/__/.chop
-aO/home/Users/tjk/vector/build/cmakeLinux/Linux/vcast/CMakeFiles/commonada.dir/__/.common-enums
monitor -o ../../vc/IO/monitor -cargs -largs
-L/home/Users/tjk/vector/build/cmakeLinux/Linux
-L/home/Users/tjk/vector/build/cmakeLinux/Linux/vc/lib
-L/home/TOOLS/kdchart-2.3.0-qt-4.6.2-release-gcc-4.1.2-release/lib
-L/home/FLEXlm/FLEXlm_11_9/i86_lsb
-L/home/FLEXlm/FLEXlm_11_9/i86_lsb/activation/lib
-L/home/TOOLS/python-2.7.5/libs
-L/Unix_Volume/GCC/Linux/gcc-4.1.2/lib/gcc/i686-pc-linux-gnu/4.1.2
-L/Unix_Volume/GCC/Linux/gcc-4.1.2/lib/gcc
-L/home/GCC/gcc-4.1.2/lib/gcc/i686-pc-linux-gnu/4.1.2
-L/Unix_Volume/GCC/Linux/gcc-4.1.2/lib -L/home/GCC/gcc-4.1.2/lib -rdynamic
../../vc/lib/libcinterfacetoada.a ../../vc/lib/libcpputilsforada.a
../../libcommonada.a /home/TOOLS/libxml2/libxml2-2.6.27/lib/libxml2.a
-llmgr_nomt -lcrvs -lsb -lnoact -llmgr_dongle_stub
/home/GCC/gcc-4.1.2/lib/gcc/i686-pc-linux-gnu/4.1.2/adalib/libgnat.a -lutil
../../vc/lib/libvxml_c.a ../../vc/lib/libcutilsforada.a
/home/FLEXlm/FLEXlm_11_9/i86_lsb/lm_new.o -ldl -lm
/home/TOOLS/libxml2/libxml2-2.6.27/lib/libxml2.a
../../vc/lib/libcompat_glib_workaround.a
../../vc/lib/libcinterfacetoadastubs.a
../../vc/lib/libcinterfacetoada_ada.a -lc -lstdc++ -lm -lc
What do you need from me in terms of helping diagnose this issue?
Thanks,
Tom
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.cmake.org/pipermail/cmake/attachments/20131210/89a70492/attachment-0001.htm>
More information about the CMake
mailing list