在開始类的编写之前我们依旧须要回想整理一下前面所说的内容,(前面尽管是一个自己定义数据类型的实现过程,可是内容有点繁杂).

先看一段代码:

/** @file calssStruct.cpp */

/** Member Functions for Class point */

#include <cmath> // for sqrt and atan

using namespace  std;

struct point

{

  point()

  : x_(0.0), y_(0.0)

  {}

  point(double x, double y)

  : x_(x), y_(y)

  {}

  point(point const& pt)

  : x_(pt.x_), y_(pt.y_)

  {}

  /// Distance to the origin.

  double distance()

  {

    return std::sqrt(x*x + y*y);

  }

  /// Angle relative to x-axis.

  double angle()

  {

    return std::atan2(y, x);

  }

  /// Add an offset to x and y.

  void offset(double off)

  {

    offset(off, off);

  }

  /// Add an offset to x and an offset to y

  void offset(double  xoff, double yoff)

  {

    x = x + xoff;

    y = y + yoff;

  }

  /// Scale x and y.

  void scale(double mult)

  {

    this->scale(mult, mult);

  }

  /// Scale x and y.

  void scale(double xmult, double ymult)

  {

    this->x = this->x * xmult;

    this->y = this->y * ymult;

  }

  double x_;

  double y_;

};

上面代码中有构造函数,有成员函数,有数据成员.我们如今须要分析一个在上面和之前的成员函数中出现的一个this引用.

在C++中,每一个成员函数都有一个隐含的形參this,当成员函数被调用时,其对象本身就被编译器作为实參隐式传入.在成员函数中能够使用*this表达式訪问该对象.在C++语法中点操作符比星操作符优先级高,因此须要使用圆括号(如(*this).x).还有一种等价的调用时使用”箭头”,即是this->x.

可是事实上编译器能够自己主动识别那些是成员名字,所以this->是可选的,即是能够省略的.可是有时候为了代码清晰总是加上它.而一些程序猿则为了表明是数据成员,有时候往往在数据成员名称前面加上前缀,或者在后面加上后缀.我倾向于后缀,如上面的代码.

关于构造函数,初始化时类和内置类型的一个重要差别,假设定义一个内置类型而不提供初始化序列,那么它的值就是无意义的;而定义类的对象时,该对象一定会被构造器初始化,因此总有机会初始化数据成员.

虽然构造函数的冒号后面的初始化列表是可选的,可是建议总是加上.

依据数据成员的类型是类或者内置类型,编译器会做出不同的处理.每一个成员的初始化表有以下的三种情况:

以下再来一个构造函数的验证代码:

/** @file constructFun.cpp */

/** Visual Constructors */

#include <iostream>

#include <ostream>

using namespace std;

struct demo

{

  demo()      : x_(0) { std::cout << "default constructor\n"; }

  demo(int x) : x_(x) { std::cout << "constructor(" << x << ")\n"; }

  demo(demo const& that)

  : x_(that.x_)

  {

    std::cout << "copy constructor(" << x_ << ")\n";

  }

  int x_;

};

demo addone(demo d)

{

  ++d.x_;

  return d;

}

int main()

{

  demo d1;

  demo d2(d1);

  demo d3(42);

  demo d4(addone(d3));

}

结果例如以下:

可是是否是学习到这里我们就把构造函数什么的搞明确了呢,以下我们来測试一下,先看代码:

/** @file Test_Construct.cpp */

/** Mystery Program */

#include <iostream>

#include <ostream>

using namespace std;

struct point

{

  point()

  : x_(0.0), y_(0.0)

  {

    cout << "default constructor\n";

  }

  point(double x, double y)

  : x_(x), y_(y)

  {

    cout << "constructor(" << x << ", " << y << ")\n";

  }

  double x_;

  double y_;

};

int main()

{

  point pt();

}

你觉得这个编译后会输出什么呢?是default constructor么?

可是你错了……这就是输出:

即是根本未定义变量,编译器觉得你编写了一个无參数的返回值为point的名字是pt的函数声明!

假设你不理解那么把point改成int,如今是intpt();这样是不是就更像一个函数声明了呢.可是假设要定义一个变量应该怎么样的呢?

是 point pt;即是不带括号这个时候调用了默认的无參数构造函数初始化.

所以在定义变量的时候注意哪些括号是必须的,不然非常可能就会误导编译器将你的所谓的”变量声明”当成函数声明.

练习项目一

身体质量指数BMI小程序

学习了那么多能够做个小项目练习一下了.要计算BMI须要知道一个人的身高体重,BMI的计算公式是体重/(身高^2),结果是一个无单位的值.如今的任务是编写一个程序,使其能够读取记录,打印记录并计算一些统计值.该程序以请求一个BMI上极限開始,仅打印BMI值大雨或等于此极限值的记录,每条记录包括姓名(能够有空格),体重,身高(单位cm),以及性别(M或F,不限大写和小写).读取完每一个人的记录后要马上打印该记录的BMI值,手机全部的记录后,基于数据打印两个表------男性一个,女性一个.在BMI值后面用*标记超过界限的BMI记录.并打印BMI的均值和中值.

当中的一个演示样例代码例如以下.

/** @file BMI_Again.cpp */

/** New BMI Program */

#include <algorithm>

#include <cstdlib>

#include <iomanip>

#include <ios>

#include <iostream>

#include <istream>

#include <limits>

#include <locale>

#include <ostream>

#include <string>

#include <vector>

using namespace std;

/// Compute body-mass index from height in centimeters and weight in kilograms.

int compute_bmi(int height, int weight)

{

   return static_cast<int>(weight * 10000 / (height * height) + 0.5);

}

/// Skip the rest of the input line.

void skip_line(istream& in)

{

  in.ignore(numeric_limits<int>::max(), '\n');

}

/// Represent one person’s record, storing the person’s name, height, weight,

/// sex, and body-mass index (BMI), which is computed from the height and weight.

struct record

{

  record() : height_(0), weight_(0), bmi_(0), sex_('?'), name_()

  {}

  /// Get this record, overwriting the data members.

  /// Error-checking omitted for brevity.

  /// @return true for success or false for eof or input failure

  bool read(istream& in, int num)

  {

    cout << "Name " << num << ": ";

    string name;

    if (not getline(in, name))

      return false;

    cout << "Height (cm): ";

    int height;

    if (not (in >> height))

      return false;

    skip_line(in);

    cout << "Weight (kg): ";

    int weight;

    if (not (in >> weight))

      return false;

    skip_line(in);

    cout << "Sex (M or F): ";

    char sex;

    if (not (in >> sex))

      return false;

    skip_line(in);

    sex = toupper(sex, locale());

    // Store information into data members only after reading

    // everything successfully.

    name_ = name;

    height_ = height;

    weight_ = weight;

    sex_ = sex;

    bmi_ = compute_bmi(height_, weight_);

    return true;

  }

  /// Print this record to @p out.

  void print(ostream& out, int threshold)

  {

    out << setw(6) << height_

        << setw(7) << weight_

        << setw(3) << sex_

        << setw(6) << bmi_;

    if (bmi_ >= threshold)

      out << '*';

    else

      out << ' ';

    out << ' ' << name_ << '\n';

  }

  int height_;       ///< height in centimeters

  int weight_;       ///< weight in kilograms

  int bmi_;          ///< Body-mass index

  char sex_;         ///< 'M' for male or 'F' for female

  string name_; ///< Person’s name

};

/** Print a table.

 * Print a table of height, weight, sex, BMI, and name.

 * Print only records for which sex matches @p sex.

 * At the end of each table, print the mean and median BMI.

 */

void print_table(char sex, vector<record>& records, int threshold)

{

  cout << "Ht(cm) Wt(kg) Sex  BMI  Name\n";

  float bmi_sum(0);

  long int bmi_count(0);

  vector<int> tmpbmis; // store only the BMIs that are printed

                            // in order to compute the median

  for (vector<record>::iterator iter(records.begin());

       iter != records.end();

       ++iter)

  {

    if (iter->sex_ == sex)

    {

      bmi_sum = bmi_sum + iter->bmi_;

      ++bmi_count;

      tmpbmis.push_back(iter->bmi_);

      iter->print(cout, threshold);

    }

  }

  // If the vectors are not empty, print basic statistics.

  if (bmi_count != 0)

  {

    cout << "Mean BMI = "

              << setprecision(1) << fixed << bmi_sum / bmi_count

              << '\n';

    // Median BMI is trickier. The easy way is to sort the

    // vector and pick out the middle item or items.

    sort(tmpbmis.begin(), tmpbmis.end());

    cout << "Median BMI = ";

    // Index of median item.

    int i(tmpbmis.size() / 2);

    if (tmpbmis.size() % 2 == 0)

      cout << (tmpbmis.at(i) + tmpbmis.at(i-1)) / 2.0 << '\n';

    else

      cout << tmpbmis.at(i) << '\n';

  }

}

/** Main program to compute BMI. */

int main()

{

  locale::global(locale(""));

  cout.imbue(locale());

  cin.imbue(locale());

  vector<record> records;

  int threshold;

  cout << "Enter threshold BMI: ";

  if (not (cin >> threshold))

    return EXIT_FAILURE;

  skip_line(cin);

  cout << "Enter name, height (in cm),"

               " and weight (in kg) for each person:\n";

  record rec;

  while (rec.read(cin, records.size()+1))

  {

    records.push_back(rec);

    cout << "BMI = " << rec.bmi_ << '\n';

  }

  // Print the data.

  cout << "\n\nMale data\n";

  print_table('M', records, threshold);

  cout << "\nFemale data\n";

  print_table('F', records, threshold);

}

执行结果例如以下:

watermark/2/text/aHR0cDovL2Jsb2cuY3Nkbi5uZXQvc3Vvb2w=/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70/gravity/SouthEast" alt="">

可是看到上面的这个函数:

你是否认为函数參数传递应该是const类型更合适呢,我们改动了一下,測试,发现,,,不行,出现了错误.可是显然应该是const类型的引用传递.那怎么实现呢.

应该记得在每一个成员函数的内部都有一个this隐藏參数,上面的代码段中,print_table调用print成员函数,可是经过改动,this引用了一个const对象.虽然你知道print函数不会改动不论什么的数据成员,可是编译器不知道,因此你须要让编译器知道,怎么做呢,仅仅要在函数头和函数体之间加上一个const修饰符就可以.例如以下:

而一个通用的规则是,为全部的不改动成员变量的函数加入const修饰符,它确保程序在有const对象的时候就能够调用成员函数.