Open thovide opened 2 years ago
You're right, although I believe this is valid C++ and not a bug; see, for example, this Stack Overflow post.
Since there is an associated array called mep
that is also allocated from i=0, it made sense to start the mne
array from i=0 also. But it could likely be adjusted to i=1, since the zero entry should never get accessed.
Thanks for your answer. Actually, it is detected as a programming error by our automated quality checking tools. It's valid but bad 🙂.
Hello,
in cell.cpp or cell3d.cpp (depending on the version), the instruction in bold leads to a call to new int[0] for the first iteration...
void voronoicell_neighbor_3d::memory_setup() { int i; mne=new int[current_vertex_order]; ne=new int[current_vertices]; *for(i=0;i<3;i++) mne[i]=new int[init_n_verticesi];* mne[3]=new int[init_3_vertices3]; for(i=4;i<current_vertex_order;i++) mne[i]=new int[init_n_vertices*i]; }