EclipseWriter: don't convert the unit inside the rate() function

because if that is done, the integration for the production total is
wrong because the time step length is days and not seconds.

Note that the conversion now uses Opm::unit::convert::to(<>,
1/conversionFactor) instead of simply multiplying with the conversion
factor. I still think this obfuscates things more than it helps, but
[at] bska will hopefully want it this way...

opm/core/io/eclipse/EclipseWriter.cpp

@@ -547,14 +547,13 @@ protected:
         flatIdx_ = timeStepWellIdx_*phaseUses_.num_phases + phaseUses_.phase_pos[phaseIdx_];
     }
 
+    // return m^3/s of injected or produced fluid
     double rate(const WellState& wellState)
     {
-        // convert m^3/s of injected fluid to m^3/d of produced fluid
-        const double convFactor = Opm::unit::day;
         double value = 0;
         if (wellState.wellRates().size() > 0) {
             assert(int(wellState.wellRates().size()) > flatIdx_);
-            value = sign_ * wellState.wellRates()[flatIdx_] * convFactor;
+            value = sign_ * wellState.wellRates()[flatIdx_];
         }
         return value;
     }
@@ -668,7 +667,9 @@ public:
         }
 
         // TODO: Why only positive rates?
-        return std::max(0., rate(wellState));
+        using namespace Opm::unit;
+        return convert::to(std::max(0., rate(wellState)),
+                           cubic(meter)/day);
     }
 };