Segmentation fault for wrong stat

[frederic-michaud]

I just realized that trying to compute a non-existing statistic leads to a segfault, if only one statistic is computed. Though it took me a few seconds to track down the error in my input file, I guess that for new users, it might be difficult to know where the problem comes since there is no indication whatsoever of what is wrong with the input file.
The expected behaviors is probably to just skip the stat, and send a warning.

generations	 	10000	 	
patch_capacity	 	1000	 	
ntrl_loci	 	1	 		
stat	 	{n.adlt.nbAl}	

[frederic-michaud]

If only one, misspelled, statistic is given, some pointers are not well defined (e.g. _pStat_db) and lead to segmentation fault at various places. Testing that the pointer exists before calling it (at various places) solves the segmentation but then no Warning is emitted about the misspelling. Still looking for a better way to solve the problem.