QTest::toString(const T&), class template, explicit template specialization

若有定義好==運算子，QCOMPARE巨集可比較自定義類別。

舉例來說:
class Employee {
public:
  Employee() { id_ = count_++; }
  unsigned id() const { return id_; }
private:
  static unsigned count_;
  unsigned id_;
}

unsigned Employee::count_ = 0;

定義==運算子
bool operator ==(const Employee& lhs, const Employee& rhs)
{
  return lhs.id() == rhs.id();
}


於Test Case
Employee employee1;
Employee employee2;
QCOMPARE(empolyee1, employee2);


如此，執行Unit Test會跑出Compare values are not the same。但不會像比較double等內定型別一樣，列出Actual與Expected。


若要希望QCOMPARE自定類別也可列出Actual與Expected，需explicit template specialization QTest::toString(const T&):

namespace QTest {
template <>
char* toString(const Employee& employee)
{
  QByteArray ba = "Employee, id = ";
  ba += QByteArray::number(employee.id());
  return qstrdup(ba.data());
}
}

如此，會跑出
Actual   (employee1): Employee, id = 0
Expected (employee2): Employee, id = 1



假使希望對class template做相同的事:

template <unsigned _rowCount>
class Vector {
public:
  double& operator ()(unsigned row) { return data_[row-1]; }
  double operator ()(unsigned row) const {return data_[row-1]; }
private:
  double data_[_rowCount];
}

==運算子:
template <unsigned _rowCount>
bool operator ==(const Vector<_rowcount>& lhs, const Vector<_rowcount>& rhs>
{
  for (unsigned row = 1; row <= _rowCount; ++row) {
    if (lhs(row) != rhs(row)) return false;
  }
  return true;
}

class template之toString之寫法較特殊:
namespace QTest {
template <typename T, unsigned _rowCount>
char* toString(const Vector<_rowcount>& v)
{
  QByteArray ba = "Vector(";
  for (unsigned row = 1; row <= _rowCount; ++row) {
    ba += QByteArray::number(v(row));
    ba += (row != _rowCount) ? ", " : ")";
  }
}
}
儘管T未使用，仍不可只寫<>，否則無法正確explicit template specialization toString(const T&)，Actual與Expected也不會正確顯示。

先前Employee之toString也可改寫為<typename T>，脫褲子放屁就是了。

於Test Case
Vector<3> vector1;
vector1(1) = vector1(2) = vector1(3) = 1;
Vector<3> vector2;
vector2(1) = vector2(2) = vector2(3) = 2;
QCOMPARE(vector1, vector2);


如此，會跑出
Actual   (vector1): Vector(1, 1, 1)
Expected (vector2): Vector(2, 2, 2)