k-order MEX: fix crash on Octave when higher-order derivatives have numerical zero

In Octave, when some values given to the sparse() function are numerically
zero, then the nzmax of the generated sparse matrix is shrinked accordingly;
while under MATLAB, the nzmax is the length of the vector of values, zeros
included.

The check at the top of
DynamicModelMFile::unpackSparseMatrixAndCopyIntoTwoDMatData() would then fail
under Octave if some higher-derivatives had an element which is symbolically
non-zero but numerically zero.

We therefore relax the check, and accordingly adapt the code that handles
numerical zeros.

This bug was uncovered by tests/pruning/AnSchorfheide_pruned_state_space.mod,
which was failing under Octave.

(cherry picked from commit 144bdf34)

mex/sources/k_order_perturbation/dynamic_m.cc

 /*
- * Copyright © 2010-2019 Dynare Team
+ * Copyright © 2010-2020 Dynare Team
  *
  * This file is part of Dynare.
  *
@@ -35,11 +35,13 @@ DynamicModelMFile::unpackSparseMatrixAndCopyIntoTwoDMatData(mxArray *sparseMat,
 {
   int totalCols = mxGetN(sparseMat);
   mwIndex *rowIdxVector = mxGetIr(sparseMat);
-  mwSize sizeRowIdxVector = mxGetNzmax(sparseMat);
   mwIndex *colIdxVector = mxGetJc(sparseMat);
 
   assert(tdm.ncols() == 3);
-  assert(tdm.nrows() == sizeRowIdxVector);
+  /* Under MATLAB, the following check always holds at equality; under Octave,
+     there may be an inequality, because Octave diminishes nzmax if one gives
+     zeros in the values vector when calling sparse(). */
+  assert(tdm.nrows() >= mxGetNzmax(sparseMat));
 
   double *ptr = mxGetPr(sparseMat);
 
@@ -55,11 +57,11 @@ DynamicModelMFile::unpackSparseMatrixAndCopyIntoTwoDMatData(mxArray *sparseMat,
         output_row++;
       }
 
-  /* If there are less elements than Nzmax (that might happen if some
+  /* If there are less elements than expected (that might happen if some
      derivative is symbolically not zero but numerically zero at the evaluation
      point), then fill in the matrix with empty entries, that will be
      recognized as such by KordpDynare::populateDerivativesContainer() */
-  while (output_row < static_cast<int>(sizeRowIdxVector))
+  while (output_row < tdm.nrows())
     {
       tdm.get(output_row, 0) = 0;
       tdm.get(output_row, 1) = 0;