这是2013年写的一篇旧文,放在gegahost.net上面 http://raison.gegahost.net/?p=39

February 20, 2013

casting in C++

Filed under: c++ — Tags: C++ internal, dynamic_cast, reinterpret_cast, static_cast — Raison @ 6:17 am

(original work by Peixu Zhu)

Unlike in C language, C++ language offers richer casting features, however, at the same time, it also brings complexity.

Keep in mind that in C++, casting may return a different address rather than the original one !

1.  static_cast

  • mainly focus on syntax checking at compile time, it can be used to cast a base class into a derived class.
  • if it fails, the compiler will give error information.
  • in case of casting from a class with multiple parent class, it may return an address offset to the sourcing address. However, casting to void* will not change the address.

2.  dynamic_cast

  • when it fails, it will return NULL value at run time.
  • in case of casting from a class with multiple parent class, it may return an address offset to the sourcing address. However, casting to void* will not change the address.

3.  reinterpret_cast

  • will not change returned pointers.
  • the target must have enough size to accept the source value.

4.  casting classes with multiple parent classes

  • for force casting/static_cast/dynamic_cast, if the casted class has more than one parent classes, the result may be not the original address, but an address having an offset to the original address. However, though the resulted address is different to the original address, comparative operator will restore the original address temporary, in case of compare to the pointer of original class. That is, the compiler will make both pointer to be the same class type in memory layout.
  • However, if the pointer is casted to void*, it will not change the sourcing address.

5.  internal of static_cast/dynamic a derived class which has multiple parent classes into a base class:

  • get the address of the derived instance.
  • if the address is zero, return it.
  • if the address is non-zero, compiler will return an address with an offset to the instance address.

6.  `==’ operator on two pointers of which one is subobject of another one.

  • the compiler check whether the subobject one is non-zero, if it is non-zero, promote it like another one, by changing it’s address temporarily with an offset.
  • when the two pointers are compared, compiler will temporary `upgrade’ the subobject to same grade class, then perform the comparison, and if the pointer is NULL, the upgrade is not necessary and then omitted.

casting in C++的更多相关文章

  1. 读书笔记 effective c++ Item 27 尽量少使用转型(casting)

    C++设计的规则是用来保证使类型相关的错误不再可能出现.理论上来说,如果你的程序能够很干净的通过编译,它就不会尝试在任何对象上执行任何不安全或无意义的操作.这个保证很有价值,不要轻易放弃它. 不幸的是 ...

  2. Laravel attribute casting 导致的 Indirect modification of overloaded property

    在 Laravel model 中,设置了某个属性做 array casting. protected $casts = [ 'rounds' => 'array', ]; 但是在 contro ...

  3. C++进阶--显式类型转换(casting)

    //############################################################################ /* * 显式类型转换 * * 类型转换 ...

  4. Reference Type Casting

    5.5.1. Reference Type Casting Given a compile-time reference type S (source) and a compile-time refe ...

  5. Welcome-to-Swift-18类型转换(Type Casting)

    类型转换是一种检查类实例的方式,并且哦或者也是让实例作为它的父类或者子类的一种方式. Type casting is a way to check the type of an instance, a ...

  6. java中Number Type Casting(数字类型强转)的用法

    4.5 Number Type Casting(数字类型强转)隐式 casting(from small to big) byte a = 111; int b = a;显式 casting(from ...

  7. 关于“No loop matching the specified signature and casting was found for ufunc lstsq_n”问题的解决

    下面这段代码是使用MatPlotLib绘制数据随时间变化的趋势. import datetime as dt import numpy as np import pandas as pd import ...

  8. system verilog中的类型转换(type casting)、位宽转换(size casting)和符号转换(sign casting)

    类型转换 verilog中,任何类型的任何数值都用来给任何类型赋值.verilog使用赋值语句自动将一种类型的数值转换为另一种类型. 例如,当一个wire类型赋值给一个reg类型的变量时,wire类型 ...

  9. 对象转型 casting

    一个基类的引用类型变量可以"指向"其子类对象. 一个基类的引用不可以访问其子类对象新增加的成员(属性和方法). 基类强制转型成子类,则能访问子类独有的成员. 可以使用 引用变量in ...

随机推荐

  1. docker 初试---常用命令

    http://blog.csdn.net/wsscy2004/article/details/25878363 常用命令 查看容器的root用户密码 docker logs <容器名orID&g ...

  2. linux防火墙过滤规则

    一.linux防火墙基础 防火墙分为硬件防火墙和软件防火墙. 1.概述 linux 防火墙体系主要工作在网络层,针对TCP/IP数据包实施过滤和限制,属于典型的包过滤防火墙. 包过滤机制:netfil ...

  3. Objective-C Runtime(一)预备知识

    很早就知道了Objective-C Runtime这个概念,「Objective-C奇技淫巧」「iOS黑魔法」各种看起来很屌的主题中总会有它的身影:但一直没有深入去学习,一来觉得目前在实际项目中还没有 ...

  4. bzoj1222

    奇怪的dp 思路清奇 dp[i][j]表示当前做完了i个任务,1机器花了j秒,2机器花费的最少时间,然后转移就行了. #include<bits/stdc++.h> using names ...

  5. Gym 100962G Green Day (找规律)

    题意:你用k 个生成树构成一个完全图. 析:n 个点的完全图有n(n-1)/2个边,一个生成树有n-1个边,你有k 个生成树 即边数等于 K(n-1) ,即  n(n-1)/2 == k(n-1)   ...

  6. MongoDb Samus c# Find函数的使用说明

    长活短说, 网上有一些是不对的 比如 Op.GreaterThan(...).LessThan(..) 不能这么用来表示 ( , ) 而应该这么用: var doc = new Document( a ...

  7. poj 3710 Christmas Game【博弈论+SG】

    也就是转换到树形删边游戏,详见 https://wenku.baidu.com/view/25540742a8956bec0975e3a8.html #include<iostream> ...

  8. 洛谷P3825 [NOI2017]游戏(2-SAT)

    传送门 果然图论的题永远建图最麻烦……看着题解代码的建图过程真的很珂怕…… 先不考虑地图$x$,那么每一个地图都只能用两种赛车,于是我们可以用2-SAT来搞,用$i$表示这个地图能用的第一辆车,$i' ...

  9. C# BitmapData使用说明

    C# BitmapData使用说明msdn关于BitmapData原文解释地址:http://msdn.microsoft.com/zh-cn/library/5ey6h79d(v=vs.110).a ...

  10. zabbix数据库存储和性能

    1 基本认识 1.1 history zabbix默认每分钟采集一次item的数据,然后会存入history表中.主机越多,需要采集的item值就越多,history表的增长速度就越快. 在histo ...