TiO2eels test -> illegal k point set
In the TiO2eels and TiO2eelsXML tests the k point set consists of 2 points. But in evaldos first triang is called and then make_tetra. If triang is called with k point sets of less than 3 points the subroutine does not do anything but only returns. In make_tetra some the variable ntria is used that should be initialized in triang but in the given test case it stays uninitialized.
What can we do about this? Is it reasonable to just set ntria to 0 in such cases? Or should we throw an error that the k point set is too small? Or is there another solution?
This is probably an important issue for larger unit cells that typically come with rather small k point sets.