better use of templates for Cholesky factorization * liboctave/numeric/chol.h, liboctave/numeric/chol.cc: New files generated from CmplxCHOL.cc, fCmplxCHOL.cc, floatCHOL.cc, CmplxCHOL.h, dbleCHOL.cc, dbleCHOL.h, fCmplxCHOL.h, and floatCHOL.h and converted to templates. * liboctave/numeric/module.mk: Update. * __qp__.cc, chol.cc, CMatrix.cc, CMatrix.h, dMatrix.cc, dMatrix.h, fCMatrix.cc, fCMatrix.h, fMatrix.cc, fMatrix.h, eigs-base.cc, mx-defs.h, mx-ext.h: Use new classes.

--- a/liboctave/array/fMatrix.h	Tue Feb 16 00:32:29 2016 -0500
+++ b/liboctave/array/fMatrix.h	Tue Feb 16 02:47:29 2016 -0500
@@ -49,6 +49,9 @@
   typedef FloatMatrix real_matrix_type;
   typedef FloatComplexMatrix complex_matrix_type;
 
+  typedef float real_elt_type;
+  typedef FloatComplex complex_elt_type;
+
   typedef void (*solve_singularity_handler) (float rcon);
 
   FloatMatrix (void) : FloatNDArray () { }