[英]How to emulate higher order function with C++ templates?
Here's a the template lessThan
which works as a function. 这是一个模板lessThan
,可以用作功能。
template<int n>
struct Box
{
};
template<typename T>
struct Unbox_;
template<int n>
struct Unbox_<Box<n>>
{
static constexpr int value = n;
};
template<typename T>
constexpr int unbox = Unbox_<T>::value;
template<typename T, typename U>
struct LessThan_
{
static constexpr bool value = unbox<T> < unbox<U>;
};
template<typename T, typename U>
constexpr bool lessThan = LessThan_<T, U>::value;
#include <iostream>
int main()
{
std::cout << lessThan<Box<1>, Box<2>> << '\n';
std::cout << lessThan<Box<3>, Box<2>> << '\n';
}
I now want to do something like this 我现在想做这样的事情
lessThan<Box<1>><Box<2>> == true
which is of course not valid C++. 这当然不是有效的C ++。 Why do I need this? 我为什么需要这个? Consider below. 考虑下面。
template<typename T>
struct LessThanOne_
{
static constexpr bool value = unbox<T> < 1;
};
template<typename T>
constexpr bool lessThanOne = LessThanOne_<T>::value;
In some places where I need to pass a template with one parameter, instead of passing lessThanOne
, I want to pass something like lessThan<Box<1>>
, so that I don't need to hardcode all cases. 在某些地方,我需要传递一个参数的模板,而不是传递lessThanOne
,而不是传递lessThan<Box<1>>
,这样就不需要对所有情况进行硬编码。 Is there any workaround? 有什么解决方法吗?
You are looking for the concept calked currying. 您正在寻找“ ked缩”的概念。 Look it up. 查一下 Here's a quickly thrown together sample implementation: 这是一个快速汇总的示例实现:
template<template<class,class> class fn>
struct curry
{
template <class A>
struct apply1
{
template <class B>
using apply = fn<A,B>;
};
template<class A>
using apply = apply1<A>;
};
// That's it. Below is a test rig.
template <class>
struct test1 {};
template <template<class>class>
struct test2{};
// a meta function to test
template <class, class>
struct myfn {};
// same function, curried
using myfngood = curry<myfn>;
// fully applied myfngood is a type
test1 <myfngood::apply<int>::apply<char*>> t1;
// partially applied myfngood is a template
test2 <myfngood::apply<int>::apply> t2;
As RedX's comment: 正如RedX的评论:
template<typename RHS>
struct LessThanN {
template <typename LHS>
struct inner {
static constexpr bool value = unbox<LHS> < unbox<RHS>;
};
};
and use it like: 并像这样使用它:
template <typename LHS>
using LessThanOne = LessThanN<Box<1>>::inner<LHS>;
int main() {
auto x = LessThanN<Box<2>>::inner<Box<1>> {};
cout << "1<2 " << boolalpha << x.value << '\n';
cout << "3<2 " << boolalpha << LessThanN<Box<2>>::inner<Box<3>>::value << '\n';
cout << "0<1 " << boolalpha << LessThanOne<Box<0>>::value << '\n';
}
Having said that, your original motivational example seems to be perfectly workable if you correct the syntax error, although it doesn't deal with partial application: 话虽如此,但是如果您纠正语法错误,那么您的原始动机示例似乎是完全可行的,尽管它不能处理部分应用程序:
template<typename T, typename U>
struct LessThan_ {
static constexpr bool value = unbox<T> < unbox<U>;
};
template<typename T, typename U>
constexpr bool lessThan = LessThan_<T, U>::value;
static_assert(lessThan<Box<1>, Box<2>> == true, "check");
// NOT lessThan<Box<1>><Box<2>> == true
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.