Summary: | [science overlay] sci-mathematics/scilab-5.4.0 does not link with CC - fixed using CXX as linker | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | cmuelle8 <cmue81> |
Component: | Current packages | Assignee: | Gentoo Science Related Packages <sci> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | guillaume.horel, kripton |
Priority: | Normal | Keywords: | PATCH |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 466112 | ||
Bug Blocks: | |||
Attachments: |
build log 1366161542-install-sci-mathematics_scilab-5.4.0:0::science.out.gz
scilab-5.4.0.ebuild.patch |
Description
cmuelle8
2013-04-17 17:23:54 UTC
Created attachment 345810 [details]
build log 1366161542-install-sci-mathematics_scilab-5.4.0:0::science.out.gz
Created attachment 345812 [details, diff]
scilab-5.4.0.ebuild.patch
Add
+ # use CXX as a linker for scilab-bin
+ sed -e '/^CCLD/ s/CC)/CXX)/' \
+ -i Makefile.in || die "failed to sed Makefile.in"
prettify some other sed lines..
Can you show which error you get with lapack and blas during configure? The ebuild exports the variables BLAS_LIBS ans LAPACK_LIBS which contain the output pkg-config --libs blas and pkg-config --libs lapack respectively. If these variables are in the environment the configure script use them overy anything else, so there shouldn't be an issue (and this is unrelated to bug #466112). The package links fine with CCLD on my machine. It also works with your patch, so I could apply it I guess, but I don't see why it should be needed. Can you report it upstream? Maybe they can shed light on this. (In reply to comment #3) > The package links fine with CCLD on my machine. It also works with your > patch, so I could apply it I guess, but I don't see why it should be needed. > Can you report it upstream? Maybe they can shed light on this. I'd point upstream here. Which compiler version do you use? I tested with 4.7.2-r1 - the errors in the build log clearly show that libstdc++ is not linked, which is due to gcc being used as a linker that does not default to use -lstdc++ Exporting BLAS_LIBS to the environment of the ebuild might be a portage/paludis issue, I'll have a look and report back.. (In reply to comment #3) > Can you show which error you get with lapack and blas during configure? The > ebuild exports the variables BLAS_LIBS ans LAPACK_LIBS which contain the > output pkg-config --libs blas and pkg-config --libs lapack respectively. This issue seems to be gone, configure script shows that variables are taken just as you've stated, see snippet below. I do not get linking errors anymore. But the CCLD issue remains. -->8-- checking if BLAS, ATLAS or MKL is available... checking how to get verbose linking output from x86_64-pc-linux-gnu-gfortran... -v checking for Fortran 77 libraries of x86_64-pc-linux-gnu-gfortran... -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.2 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.2/../../../../lib64 -L/lib/../lib64 -L/usr/lib/../lib64 -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.2/../../../../x86_64-pc-linux-gnu/lib -L/usr/lib/gcc/x86_64-pc-linux-gnu/4.7.2/../../.. -lpthread -ldl -lcurses -lgfortran -lm -lquadmath checking for dummy main to link with Fortran 77 libraries... none checking for Fortran 77 name-mangling scheme... lower case, underscore, no extra underscore checking for sgemm_ in -L//opt/acml5.3.0/gfortran64_mp/lib -lacml_mp ... yes Using BLAS_LIBS environment variable found checking if LAPACK is available... checking for cheev_ in -L//opt/acml5.3.0/gfortran64_mp/lib -lacml_mp ... yes LAPACK_LIBS env variable found checking if ARPACK-NG is available... checking for znaupd_ in -larpack... yes ARPACK-NG library found --8<-- |