The log method no longer errors out on non-positive values

For a zero, it returns -Inf.
For a negative number, it returns NaN.

This is not exactly the same behavior as MATLAB's log function, because we want
to enforce type stability (i.e. never return a complex).

src/@dseries/log.m

@@ -25,10 +25,6 @@ function o = log(o) % --*-- Unitary tests --*--
 % You should have received a copy of the GNU General Public License
 % along with Dynare.  If not, see <http://www.gnu.org/licenses/>.
 
-if any(o.data<eps)
-    error('dseries:WrongInputArguments', 'Variables in %s have be strictly positive!', inputname(1))
-end
-
 o = copy(o);
 o.log_;
 

src/@dseries/log_.m

@@ -25,9 +25,8 @@ function o = log_(o) % --*-- Unitary tests --*--
 % You should have received a copy of the GNU General Public License
 % along with Dynare.  If not, see <http://www.gnu.org/licenses/>.
 
-if any(o.data<eps)
-    error('dseries:WrongInputArguments', 'Variables in %s have be strictly positive!', inputname(1))
-end
+% Ensure that we never return a complex number (rather return a NaN)
+o.data(find(o.data < 0)) = NaN;
 
 o.data = log(o.data);