Problem with libigl when built from source about tetwild HOT 5 OPEN

Additionally, I encountered a problem with CGAL while making libigl (see log).

Maybe you could have a look at it, too?

log_cmake.txt

from tetwild.

Hi,

I checked your log file, the error is caused by missing GMP. You need to set up the path of GMP first (set up GMP_LIBRARIES GMP_INCLUDE_DIR).

The error is:
Could NOT find GMP (missing: GMP_LIBRARIES GMP_INCLUDE_DIR)

from tetwild.

Yes, you are right!

Now I tried for a second attempt. At first I have changed in cmake/TetWildDownloadExternal.cmake the following:

## libigl
function(tetwild_download_libigl)
    tetwild_download_project(libigl
        GIT_REPOSITORY https://github.com/libigl/libigl.git
        GIT_TAG        v2.2.0
    )
endfunction()

That was necessary because the old line

GIT_TAG        e699d257fe860d4fd82709835e72438bd0cd1533

has become obsolete. With this version the Eigen library could not be found because it moved from Bitbucket to GitLab recently.

I installed GMP and MFRP libraries and passed the CMake arguments

-DGMP_LIBRARIES="path/to/gmp-6.2.0" \
-DGMP_INCLUDE_DIR="path/to/gmp-6.2.0" \
-DMPFR_LIBRARIES="path/to/mpfr-4.1.0" \
-DMPFR_INCLUDE_DIR="path/to/mpfr-4.1.0" \

but then it got stuck when looking for the Boost library (see logfile). Similar CMake arguments could not help.

Do you have an idea where the problem lies?

log_boost_200904.log

from tetwild.

Finally, I made it work by compiling each of the necessary libraries (GMP, MPFR, CGAL). Maybe you could give a hint in the instructions.
So you can close the issue.
...and maybe change the Git-Source of LibIGL in cmake/TetWildDownloadExternal.cmake as proposed. ;)

Best regards

from tetwild.

Hi,

About the Boost question, you may try: "If you do not have Boost installed (which is needed for CGAL), you can enable the cmake option -DTETWILD_WITH_HUNTER=ON" as we mentioned in the documentation. But since you've made it work, you may not need it anymore.

I've update the version of libigl. Thanks for the advice!

For easier usage, you may try fTetWild (https://github.com/wildmeshing/fTetWild) which also has a python wrapper and can be installed via anaconda.

Hope this helps!

from tetwild.