In functional programming, functions are objects and can be processed like objects. With Boost.Phoenix, it is possible for a function to return another function as a result. It is also possible to pass a function as a parameter to another function. Because functions are objects, it's possible to distinguish between instantiation and execution. Accessing a function isn't equal to executing it.

Boost.Phoenix supports functional programming with function objects: Functions are objects based on classes which overload the operator(). That way function objects behave like other objects in c++. For example, they can be copied and stored in a container. However, they also behave like functions because they can be called.

1. phoenix

#include <boost/phoenix/phoenix.hpp>
#include <vector>
#include <algorithm>
#include <iostream> bool is_odd(int i) { return i % == ; } int main()
{
std::vector<int> v{, , , , }; std::cout << std::count_if(v.begin(), v.end(), is_odd) << std::endl; auto lambda = [](int i){ return i % == ; };
std::cout << std::count_if(v.begin(), v.end(), lambda) << std::endl; using namespace boost::phoenix::placeholders;
auto phoenix = arg1 % == ;
std::cout << std::count_if(v.begin(), v.end(), phoenix) << std::endl;
  std::vector<long> v2;
v2.insert(v2.begin(), v.begin(), v.end());
auto phoenix = arg1 % 2 == 1;
std::cout << std::count_if(v.begin(), v.end(), phoenix) << '\n';
std::cout << std::count_if(v2.begin(), v2.end(), phoenix) << '\n';
return ; }

The Phoenix function differs from free-standing and lambda functions because it has no frame. While the other two functions have a function header with a signature, the Phoenix function seems to consist of a function body only. The crucial component of the Phoenix function is boost::phoenix::placeholders::arg1. arg1 is a global instance of a function object.

arg1 is used to define an unary function. The expression arg1 % 2 == 1 creates a new function that expects one parameter. The function isn’t executed immediately but stored in phoenix. phoenix is passed to std::count_if() which calls the predicate for every number in v.

Phoenix function parameters have no types. The lambda function lambda expects a parameter of type int. The Phoenix function phoenix will accept any type that the modulo operator can handle. Think of Phoenix functions as function templates. Like function templates, Phoenix functions can accept any type.

You can think of Phoenix functions as C++ code that isn’t executed immediately. The Phoenix function in example above looks like a condition that uses multiple logical and arithmetic operators. However, the condition isn’t executed immediately. It is only executed when it is accessed from within std::count_if(). The access in std::count_if() is a normal function call.

2. boost::phoenix::val()

#include <boost/phoenix/phoenix.hpp>
#include <vector>
#include <algorithm>
#include <iostream> int main()
{
std::vector<int> v{, , , , }; using namespace boost::phoenix;
using namespace boost::phoenix::placeholders;
auto phoenix = arg1 > val() && arg1 % val() == val();
std::cout << std::count_if(v.begin(), v.end(), phoenix) << std::endl;

std::cout << arg1(1, 2, 3, 4, 5) << std::endl;
auto v = boost::phoenix::val(2);
std::cout << v() << std::endl;
return ; }

boost::phoenix::val() returns a function object initialized with the values passed to boost::phoenix::val(). The acutal type of the function object dosen't matter.

arg1 is an instance of a function object. It can be used directly and called like a function. You can pass as many parameters as you like - arg1 returns the first one.

val() is a function to create an instance of a function object. The function object is initialized with the value passed as a parameter. Is the instance is accessed like a function, the value is returned.

3. create your own phoenix functions

#include <boost/phoenix/phoenix.hpp>
#include <vector>
#include <algorithm>
#include <iostream> struct is_odd_impl
{
typedef bool result_type; template <typename T>
bool operator()(T t) const { return t % == ; }
}; boost::phoenix::function<is_odd_impl> is_odd; int main()
{
std::vector<int> v{, , , , }; using namespace boost::phoenix::placeholders;
std::cout << std::count_if(v.begin(), v.end(), is_odd(arg1)) << std::endl;
return ;
}

Class should overload the operator operator(): when an odd number is passed in, the operator returns true. Please note that you must define the type result_type. Boost.Phoenix uses it to detect the type of the return value of the operator operator().

4. transform

#include <boost/phoenix/phoenix.hpp>
#include <vector>
#include <algorithm>
#include <iostream> bool is_odd_function(int i) { return i % == ; } BOOST_PHOENIX_ADAPT_FUNCTION(bool, is_odd, is_odd_function, )

bool is_odd(int i) {
return i % 2 == 1;
}
int main()
{
std::vector<int> v{, , , , };
using namespace boost::phoenix;
using namespace boost::phoenix::placeholders;
std::cout << std::count_if(v.begin(), v.end(), is_odd(arg1)) << std::endl;
 std::cout << std::count_if(v.begin(), v.end(), bind(is_odd, arg1)) << std::endl;
return ; }

Use the macro BOOST_PHOENIX_ADAPT_FUNCTION to turn a free-standing function into a Phoenix function. Pass the type of the return value, the name of the Phoenix function to define, the name of the free-standing function, and the number of parameters to the macro.

boost phoenix的更多相关文章

  1. boost 的函数式编程库 Phoenix入门学习

    这篇文章是我学习boost phoenix的总结. 序言 Phoenix是一个C++的函数式编程(function programming)库.Phoenix的函数式编程是构建在函数对象上的.因此,了 ...

  2. boost::spirit unicode 简用记录

    本文简单记录使用boost::spirit解析有中文关键字的字符串并执行响应动作,类似于语法分析+执行. 关键字:字符串解析 boost::spirit::qi::parse qi::unicode: ...

  3. 一个c++剧情脚本指令系统

    项目希望能够实现一些剧情动画,类似角色移动,镜头变化,台词展现等.剧情动画这东西随时需要修改调整,不能写死在代码里.考虑之后认为需要做一个简单的DSL来定制剧情脚本,策划在脚本里按顺序写入命令,然后我 ...

  4. 译:Boost Property Maps

    传送门:Boost Graph Library 快速入门 原文:Boost Property Map 图的抽象数学性质与它们被用来解决具体问题之间的主要联系就是被附加在图的顶点和边上的属性(prope ...

  5. Boost简介

    原文链接:  吴豆豆http://www.cnblogs.com/gdutbean/archive/2012/03/30/2425201.html Boost库 Boost库是为C++语言标准库提供扩 ...

  6. Boost 1.61.0 Library Documentation

    http://www.boost.org/doc/libs/1_61_0/ Boost 1.61.0 Library Documentation Accumulators Framework for ...

  7. boost库的安装,使用,介绍,库分类

    1)首先去官网下载boost源码安装包:http://www.boost.org/ 选择下载对应的boost源码包.本次下载使用的是 boost_1_60_0.tar.gz (2)解压文件:tar - ...

  8. C++ Boost库分类总结

    c# 程序员写c++,各种不适应.尤其是被内存操作和几十种字符串类型的转换,简直疯了,大小写转换竟然要手动写代码实现. Boost看介绍不错,也不知道能不能跨平台.过几天要上linux写c++, 也不 ...

  9. 小试 boost spirit

    解释文本文件是日常编程中太平常的一件事情了,一般来说,土鳖点的做法可以直接手写 parser 用循环暴力地去 map 文本上的关键字从而提取相关信息,想省力一点则可以使用 tokenizer 或正则表 ...

随机推荐

  1. python保存selenium的cookies写入和读出

    def write_cookie(self, cookie): try: with open("cookies%s" % self.uid, "wb+") as ...

  2. linux的shell脚本运行python程序

    可以说和windows里的bat是一样的. python3 /opt/pyweibo/get_user_info.py 104501 104502 104503

  3. 【ABAP系列】SAP ABAP基础-程序优化及响应速度之LOOP

    公众号:SAP Technical 本文作者:matinal 原文出处:http://www.cnblogs.com/SAPmatinal/ 原文链接:[ABAP系列]SAP ABAP基础-程序优化及 ...

  4. 回调-> 观察者模式->反应堆模式

    关于回调: 回调是观察者模式以及反应堆模式的基础 一句话,回调就是一种双向调用模式,什么意思呢,就是说,被调用方在被调用时也会调用对方,这就叫回调.“If you call me, i will ca ...

  5. tbox协程使用之切换与等待

    tbox的协程实现,是stackfull模式的,需要指定独立堆栈和协程函数,目前暂时还不能像golang那样实现堆栈的动态增长,之后会对其进行支持. 目前提供下面一些功能特性: 1. 提供yield切 ...

  6. 基于Netty重构RPC框架

    下面的这张图,大概很多小伙伴都见到过,这是Dubbo 官网中的一张图描述了项目架构的演进过程.随着互联网的发展,网站应用的规模不断扩大,常规的垂直应用架构已无法应对,分布式服务架构以及流动计算架构势在 ...

  7. java序列化的相关介绍

    1.什么是序列化?为什么要用序列化? 序列化就是将对象状态转换为可保持或传输的格式的过程.与序列化相对的就是反序列化,他将流转换成对象.这两个过程结合起来,可以轻松地存储和传输数据. 注意:对象序列化 ...

  8. mybatis 批量update报语法错误解决方法

    1.为什么会报语法错误 原因:在 *.xml文件内使用了循环,在mybatis中默认是不允许使用批量修改. <update id="setMaxMin" parameterT ...

  9. CentnOS7安装Nginx“No package available”

    Nginx相对Apache有轻量级,简洁的优点,算得上Apache的优秀替代品了,但是由于Nginx不在yum的官方源中,因此安装时总会出现失败的现象,只需: yum install epel-rel ...

  10. JS的两种函数声明方式的区别

    ---恢复内容开始--- js中常见的两种函数声明方式如下: // 函数表达式 var f = function() { console.log(1); } // 直接声明 function f () ...