don't overload meaning of info in Sparse Cholesky factorization functions * chol.cc (Fchol): New variable, force. Always pass natural and force to SparseCHOL and SparseComplexCHOL constructors. * SparsedbleCHOL.h, SparsedbleCHOL.cc (SparseCHOL::SparseCHOL): New arg, force. Pass it to sparse_base_chol constructor. * SparseCmplxCHOL.h, SparseCmplxCHOL.cc (SparseComplexCHOL::SparseComplexCHOL): Likewise. * sparse-base-chol.h (sparse_base_chol<>::sparse_base_chol_rep::init): * sparse-base-chol.h, sparse-base-chol.cc (sparse_base_chol<>::sparse_base_chol_rep::init): Replace nargout argument with force. Check force, not nargout > 1. * sparse-base-chol.h (sparse_base_chol::sparse_base_chol_rep::sparse_base_chol_rep): New arg, force. Pass it to init. (sparse_base_chol::sparse_base_chol): New arg force. Pass it to rep constructor.

--- a/liboctave/sparse-base-chol.cc	Thu Aug 30 20:52:40 2012 +0200
+++ b/liboctave/sparse-base-chol.cc	Thu Aug 30 16:57:24 2012 -0400
@@ -80,9 +80,10 @@
 template <class chol_type, class chol_elt, class p_type>
 octave_idx_type
 sparse_base_chol<chol_type, chol_elt, p_type>::sparse_base_chol_rep::init
-  (const chol_type& a, bool natural, octave_idx_type nargout)
+  (const chol_type& a, bool natural, bool force)
 {
   volatile octave_idx_type info = 0;
+
 #ifdef HAVE_CHOLMOD
   octave_idx_type a_nr = a.rows ();
   octave_idx_type a_nc = a.cols ();
@@ -170,7 +171,7 @@
   is_pd = cm->status == CHOLMOD_OK;
   info = (is_pd ? 0 : cm->status);
 
-  if (is_pd || nargout > 1)
+  if (is_pd || force)
     {
       BEGIN_INTERRUPT_IMMEDIATELY_IN_FOREIGN_CODE;
       cond = CHOLMOD_NAME(rcond) (Lfactor, cm);