Round, rather than truncate, in wxSize and wxPoint operations

It seems to make more sense to round the result to int rather than to
truncate it when scaling wxSize or wxPoint coordinates by a floating
point number.

include/wx/gdicmn.h

@@ -292,8 +292,8 @@ public:
     wxSize& operator*=(long i) { x *= i; y *= i; return *this; }
     wxSize& operator/=(unsigned long i) { x /= i; y /= i; return *this; }
     wxSize& operator*=(unsigned long i) { x *= i; y *= i; return *this; }
-    wxSize& operator/=(double i) { x = int(x/i); y = int(y/i); return *this; }
-    wxSize& operator*=(double i) { x = int(x*i); y = int(y*i); return *this; }
+    wxSize& operator/=(double i) { x = wxRound(x/i); y = wxRound(y/i); return *this; }
+    wxSize& operator*=(double i) { x = wxRound(x*i); y = wxRound(y*i); return *this; }
 
     void IncTo(const wxSize& sz)
         { if ( sz.x > x ) x = sz.x; if ( sz.y > y ) y = sz.y; }
@@ -319,7 +319,7 @@ public:
 
 
     wxSize& Scale(double xscale, double yscale)
-        { x = (int)(x*xscale); y = (int)(y*yscale); return *this; }
+        { x = wxRound(x*xscale); y = wxRound(y*yscale); return *this; }
 
     // accessors
     void Set(int xx, int yy) { x = xx; y = yy; }
@@ -428,12 +428,12 @@ inline wxSize operator*(unsigned long i, const wxSize& s)
 
 inline wxSize operator*(const wxSize& s, double i)
 {
-    return wxSize(int(s.x * i), int(s.y * i));
+    return wxSize(wxRound(s.x * i), wxRound(s.y * i));
 }
 
 inline wxSize operator*(double i, const wxSize& s)
 {
-    return wxSize(int(s.x * i), int(s.y * i));
+    return wxSize(wxRound(s.x * i), wxRound(s.y * i));
 }
 
 
@@ -567,7 +567,7 @@ public:
 
     wxPoint() : x(0), y(0) { }
     wxPoint(int xx, int yy) : x(xx), y(yy) { }
-    wxPoint(const wxRealPoint& pt) : x(int(pt.x)), y(int(pt.y)) { }
+    wxPoint(const wxRealPoint& pt) : x(wxRound(pt.x)), y(wxRound(pt.y)) { }
 
     // no copy ctor or assignment operator - the defaults are ok