csg_stat segfault on h5md files
junghans opened this issue · comments
From:
docker run -it votca/buildenv:fedora_rawhide /bin/bash
git clone --recursive -b stable_fixes https://github.com/votca/votca
cd votca/
git -C csg checkout for/stable/cmake_typo
mkdir build
cd build/
cmake -DENABLE_REGRESSION_TESTING=ON ..
make
ctest -V -R espresso
...
segfault
Running the isolated csg_stat command:
../../../../csg/src/tools/csg_stat --nt 2 --options /home/votca/votca/build/csg-tutorials/spce/ibi_espresso/settings.xml --top topol.xml --trj traj.h5 --begin 0 --first-frame 0
begin to calculate distribution functions
# of bonded interactions: 0
# of non-bonded interactions: 1
WARNING: topology created from .gro file, masses and charges are wrong!
I have 2180 beads in 2180 molecules
WARNING: topology created from .gro file, masses and charges are wrong!
H5MD: static box
H5MD: Found box 40.31 x 40.31 x 40.31
H5MD: has /force
H5MD: has /velocity
H5MD: has /id group
Reading frame: 0
Segmentation fault
This already includes the fixed from #481 and votca/votca#238.
What was the problem?
Two more Index back to int conversions were needed. see #481