AUR (en) - gmsh

Git Clone URL:	https://aur.archlinux.org/gmsh.git (read-only, click to copy)
Submitter:	S1G1
Maintainer:	gborzi (carlosal1015, gpettinello)
Last Packager:	carlosal1015
Votes:	62
Popularity:	0.41
First Submitted:	2006-04-04 23:31 (UTC)
Last Updated:	2024-05-25 16:15 (UTC)

Latest Comments

« First ‹ Previous 1 .. 10 11 12 13 14 15 16 17 18 19 20 21 Next › Last »

gborzi commented on 2016-03-29 21:31 (UTC)

@mickele Tried opencascade 6.9.0 and 6.8.0. None worked.

gborzi commented on 2016-03-29 21:17 (UTC)

@mickele Maybe it worked with the older opencascade version (6.9.0)? I don't have this version in my cache, so I'll have to download it to check.

mickele commented on 2016-03-29 20:43 (UTC)

Saving mesh file in med format I have the following error _MEDmeshDatagroupOpen.c [45] : Erreur à l'ouverture du groupe _MEDmeshDatagroupOpen.c [45] : du maillage _MEDmeshDatagroupOpen.c [46] : meshname = "1D_6" MEDmeshNodeCoordinateWr.c [73] : Erreur à l'ouverture du groupe MEDmeshNodeCoordinateWr.c [73] : du maillage MEDmeshNodeCoordinateWr.c [74] : _meshpath = "" MEDmeshNodeWr.c [72] : Erreur d'appel de l'API MEDmeshNodeWr.c [72] : MEDmeshNodeCoordinateWr _MEDmeshDatagroupOpen.c [45] : Erreur à l'ouverture du groupe _MEDmeshDatagroupOpen.c [45] : du maillage _MEDmeshDatagroupOpen.c [46] : meshname = "1D_6" _MEDmeshAdvancedWr.c [125] : Erreur à l'ouverture du groupe _MEDmeshAdvancedWr.c [125] : du maillage _MEDmeshAdvancedWr.c [126] : _meshpath = "" MEDmeshElementWr.c [79] : Erreur d'appel de l'API MEDmeshElementWr.c [79] : MEDmeshElementConnectivityWr _MEDmeshDatagroupOpen.c [45] : Erreur à l'ouverture du groupe _MEDmeshDatagroupOpen.c [45] : du maillage _MEDmeshDatagroupOpen.c [46] : meshname = "1D_6" _MEDmeshAdvancedWr.c [125] : Erreur à l'ouverture du groupe _MEDmeshAdvancedWr.c [125] : du maillage _MEDmeshAdvancedWr.c [126] : _meshpath = "" MEDmeshElementWr.c [79] : Erreur d'appel de l'API MEDmeshElementWr.c [79] : MEDmeshElementConnectivityWr Some time ago the same model gave no problem in saving MED file. Any idea about possible causes?

ian11213 commented on 2015-11-04 08:13 (UTC)

Many thanks - issue solved. In fact, recompile/reinstall of med seems sufficient - don't need to recompile and install gmsh again.

gborzi commented on 2015-11-04 00:27 (UTC)

@ian11213 You have to recompile and reinstall med, before recompiling gmsh.

ian11213 commented on 2015-11-04 00:00 (UTC)

As of today, attempts to run gmsh on my system (from console) gives the error: gmsh: error while loading shared libraries: libhdf5.so.9: cannot open shared object file: No such file or directory Error persists on updating the package to the latest version. I find only /lib64/libhdf5.settings /lib64/libhdf5.so /lib64/libhdf5.so.10 /lib64/libhdf5.so.10.0.1 /lib64/libhdf5_hl.so /lib64/libhdf5_hl.so.10 /lib64/libhdf5_hl.so.10.0.1 on my system. Grateful for any help to get up and running again.

gborzi commented on 2015-10-01 16:47 (UTC)

@nivata Thanks for your efforts. Maybe there is some compilation option that can solve this problem when trilinos is installed, but I've not seen anything in CMakeList.

nivata commented on 2015-09-30 11:36 (UTC)

So, just a quick wrap-up on this story. Turns out I also had the Petsc packaged installed with Trilinos support. After removing Trilinos I reinstalled Petsc without it, and now Gmsh compiles fine with Petsc support (but not Trilinos). At the moment Trilinos from AUR doesn't compile on my computer, so the package I had was outdated (libml.so is a library form Trilinos, and it was compiled against an older version of openmp I guess). I hope this helps other people who might encounter this issue!

nivata commented on 2015-09-29 19:13 (UTC)

My libml.so was owned by the package trilinos (from aur and was compiled a while ago). I removed it but somehow gmsh still wants to build using the "-lml" option. Otherwise I got the same result as you when I do "locate libmpi.so". This is strange, I'll look into it and keep you informed about what I find.

gborzi commented on 2015-09-29 11:54 (UTC)

@nivata I've just recompiled the package but it doesn't show any problem, i.e. I'm unable to reproduce the bug. I think there is something strange about your openmpi files: I don't have a libmpi.so.1, "locate libmpi.so" returns /usr/lib/openmpi/libmpi.so /usr/lib/openmpi/libmpi.so.12 /usr/lib/openmpi/libmpi.so.12.0.0 that don't point to any libml.so library (which isn't in my system).