Fix warnins reported by -Wunused-variable

This trick is a bit nasty but I do not want to mess up to much with the
XMPSample application at this moment. Once we have more unit tets for
that part of the code, the xmpsample could disappear.

samples/xmpsample.cpp

@@ -11,6 +11,8 @@
 #include <cassert>
 #include <cmath>
 
+#define UNUSED(expr) do { (void)(expr); } while (0)
+
 bool isEqual(float a, float b)
 {
     double d = std::fabs(a - b);
@@ -54,21 +56,25 @@ try {
     assert(xmpData["Xmp.dc.one"].value().ok());
 
     const Exiv2::Value &getv1 = xmpData["Xmp.dc.one"].value();
+    UNUSED(getv1);
     assert(isEqual(getv1.toFloat(), -1));
     assert(getv1.ok());
     assert(getv1.toRational() == Exiv2::Rational(-1, 1));
     assert(getv1.ok());
 
     const Exiv2::Value &getv2 = xmpData["Xmp.dc.two"].value();
+    UNUSED(getv2);
     assert(isEqual(getv2.toFloat(), 3.1415f));
     assert(getv2.ok());
     assert(getv2.toLong() == 3);
     assert(getv2.ok());
     Exiv2::Rational R = getv2.toRational();
+    UNUSED(R);
     assert(getv2.ok());
     assert(isEqual(static_cast<float>(R.first) / R.second, 3.1415f ));
 
     const Exiv2::Value &getv3 = xmpData["Xmp.dc.three"].value();
+    UNUSED(getv3);
     assert(isEqual(getv3.toFloat(), 5.0f/7.0f));
     assert(getv3.ok());
     assert(getv3.toLong() == 0);  // long(5.0 / 7.0)
@@ -77,6 +83,7 @@ try {
     assert(getv3.ok());
 
     const Exiv2::Value &getv6 = xmpData["Xmp.dc.six"].value();
+    UNUSED(getv6);
     assert(getv6.toLong() == 0);
     assert(getv6.ok());
     assert(getv6.toFloat() == 0.0f);
@@ -89,6 +96,7 @@ try {
     assert(!getv7.ok());
 
     const Exiv2::Value &getv8 = xmpData["Xmp.dc.eight"].value();
+    UNUSED(getv8);
     assert(getv8.toLong() == 1);
     assert(getv8.ok());
     assert(getv8.toFloat() == 1.0f);