Factor object messed up in CHOLMOD

segfault when deallocating a cholmod_factor CHOLMOD struct in CholmodSolverBase_@index@_@type@. Seemed to work before, not anymore. Why?

Is this related to messing around with a cholmod_sparse object directly? Because I only use CHOLMOD's own methods to create, manipulate and delete the cholmod_factor object.

What is also strange is that CHOLMOD recognized the cholmod_factor object to be in good state:

• after analyze():

True
CHOLMOD factor:  Internal CHOLMOD factor struct:  5-by-5  simplicial, LDL'.  nz 0  OK

• after factorize():

True
CHOLMOD factor:  Internal CHOLMOD factor struct:  5-by-5  simplicial, LDL'. nzmax 15.  nz 15  OK

If I only do a analyze(), then no problem to delete the cholmod_factor object. It is only after a factorize() operation that the segfault happens.

solve() works like a charm... so I guess that the cholmod_sparse is not that messed up after all...