如何实现std::array的绑定检查? 我已经扩展了C++ 11的STD::数组,它是工作文件,但是当我试图重载运算符[]时,我得到这个错误: error: lvalue required as left operand of assignment array[0] = 911; ^~~
是否可以实现运算符[]为std::数组类型添加绑定检查 代码如下:如何实现std::array的绑定检查? 我已经扩展了C++ 11的STD::数组,它是工作文件,但是当我试图重载运算符[]时,我得到这个错误: error: lvalue required as left operand of assignment array[0] = 911; ^~~,c++,arrays,c++11,operator-overloading,C++,Arrays,C++11,Operator Overloading,是否可以实现运算符[]为std::数组类型添加绑定检查 代码如下: #include <array> #include <cassert> #include <iostream> template <unsigned int array_size, typename array_datatype=long int> struct Array : public std::array<array_datatype, array_size>
#include <array>
#include <cassert>
#include <iostream>
template <unsigned int array_size, typename array_datatype=long int>
struct Array : public std::array<array_datatype, array_size>
{
Array()
{
}
// std::array constructor inheritance
// https://stackoverflow.com/questions/24280521/stdarray-constructor-inheritance
Array(std::initializer_list< array_datatype > new_values)
{
unsigned int data_size = new_values.size();
unsigned int column_index = 0;
// std::cout << data_size << std::endl;
if( data_size == 1 )
{
this->clear(*(new_values.begin()));
}
else
{
assert(data_size == array_size);
for( auto column : new_values )
{
(*this)[column_index] = column;
column_index++;
}
}
}
array_datatype operator[](unsigned int line)
{
assert(line < array_size);
assert(line > -1);
return (*this)[line];
}
/**
* Prints a more beauty version of the array when called on `std::cout<< array << std::end;`
*/
friend std::ostream& operator<<( std::ostream &output, const Array &array )
{
unsigned int column;
output << "{";
for( column=0; column < array_size; column++ )
{
output << array[column];
if( column != array_size-1 )
{
output << ", ";
}
}
output << "}";
return output;
}
}
相关的:
您可以使用:
array_datatype& operator[](unsigned int line)&
array_datatype const& operator[](unsigned int line)const&
array_datatype operator[](unsigned int line)&&
您可以使用:
array_datatype& operator[](unsigned int line)&
array_datatype const& operator[](unsigned int line)const&
array_datatype operator[](unsigned int line)&&
如果要在赋值的左侧使用运算符[]的返回值,则必须通过引用而不是通过值返回数组元素 您还有一个递归循环,因为您正在从自身内部调用自己的运算符[]。您希望调用基类的运算符[],因此需要对其进行限定 试试这个:
array_datatype& operator[](unsigned int line)
{
assert(line < array_size);
assert(line > -1);
return std::array<array_datatype, array_size>::operator[](line);
}
如果要在赋值的左侧使用运算符[]的返回值,则必须通过引用而不是通过值返回数组元素 您还有一个递归循环,因为您正在从自身内部调用自己的运算符[]。您希望调用基类的运算符[],因此需要对其进行限定 试试这个:
array_datatype& operator[](unsigned int line)
{
assert(line < array_size);
assert(line > -1);
return std::array<array_datatype, array_size>::operator[](line);
}
如果希望对数组元素进行绑定检查访问,只需使用std::vector或std::array的at方法,而不是[]运算符。它就是为了这个目的而存在的,不要重新发明轮子:
有关数组边界检查的文档,请参阅。如果希望对数组元素进行绑定检查访问,只需使用std::vector或std::array的at方法,而不是[]运算符。它就是为了这个目的而存在的,不要重新发明轮子:
有关数组边界检查的文档,请参阅。这是我针对一维数组的解决方案:
#include <array>
#include <cassert>
#include <iostream>
template <unsigned int array_width, typename array_datatype=long int>
struct Array
{
/**
* Is it okay to inherit implementation from STL containers, rather than delegate?
* https://stackoverflow.com/questions/2034916/is-it-okay-to-inherit-implementation-from-stl-containers-rather-than-delegate
*/
std::array<array_datatype, array_width> _data;
/**
* std::array constructor inheritance
* https://stackoverflow.com/questions/24280521/stdarray-constructor-inheritance
*/
Array()
{
}
Array(std::initializer_list< array_datatype > new_values)
{
unsigned int data_size = new_values.size();
unsigned int column_index = 0;
// std::cout << data_size << std::endl;
if( data_size == 1 )
{
this->clear(*(new_values.begin()));
}
else
{
assert(data_size == array_width);
for( auto column : new_values )
{
this->_data[column_index] = column;
column_index++;
}
}
}
/**
* Overloads the `[]` array access operator, allowing you to access this class objects as the
* where usual `C` arrays.
*
* How to implement bound checking for std::array?
* https://stackoverflow.com/questions/49419089/how-to-implement-bound-checking-for-stdarray
*
* @param line the current line you want to access
* @return a pointer to the current line
*/
array_datatype operator[](unsigned int line)&&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
array_datatype const& operator[](unsigned int line)const&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
array_datatype& operator[](unsigned int line)&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
void clear(array_datatype initial = 0)
{
unsigned int column_index = 0;
for( ; column_index < array_width; column_index++ )
{
this->_data[column_index] = initial;
}
}
/**
* The Array<> type includes the Matrix<> type, because you can multiply a `Array` by an `Matrix`,
* but not a vice-versa.
*/
void multiply(Array< array_width, Array< array_width, array_datatype > > &matrix)
{
unsigned int column;
unsigned int step;
array_datatype old_array[array_width];
for(column = 0; column < array_width; column++)
{
old_array [column] = this->_data[column];
this->_data[column] = 0;
}
for(column = 0; column < array_width; column++)
{
for(step = 0; step < array_width; step++)
{
this->_data[column] += old_array[step] * matrix._data[step][column];
}
}
// If you would like to preserve the original value, it can be returned here
// return old_array;
}
/**
* Prints a more beauty version of the array when called on `std::cout<< array << std::end;`
*/
friend std::ostream& operator<<( std::ostream &output, const Array &array )
{
unsigned int column;
output << "{";
for( column=0; column < array_width; column++ )
{
output << array._data[column];
if( column != array_width-1 )
{
output << ", ";
}
}
output << "}";
return output;
}
};
这是我针对一维阵列的解决方案:
#include <array>
#include <cassert>
#include <iostream>
template <unsigned int array_width, typename array_datatype=long int>
struct Array
{
/**
* Is it okay to inherit implementation from STL containers, rather than delegate?
* https://stackoverflow.com/questions/2034916/is-it-okay-to-inherit-implementation-from-stl-containers-rather-than-delegate
*/
std::array<array_datatype, array_width> _data;
/**
* std::array constructor inheritance
* https://stackoverflow.com/questions/24280521/stdarray-constructor-inheritance
*/
Array()
{
}
Array(std::initializer_list< array_datatype > new_values)
{
unsigned int data_size = new_values.size();
unsigned int column_index = 0;
// std::cout << data_size << std::endl;
if( data_size == 1 )
{
this->clear(*(new_values.begin()));
}
else
{
assert(data_size == array_width);
for( auto column : new_values )
{
this->_data[column_index] = column;
column_index++;
}
}
}
/**
* Overloads the `[]` array access operator, allowing you to access this class objects as the
* where usual `C` arrays.
*
* How to implement bound checking for std::array?
* https://stackoverflow.com/questions/49419089/how-to-implement-bound-checking-for-stdarray
*
* @param line the current line you want to access
* @return a pointer to the current line
*/
array_datatype operator[](unsigned int line)&&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
array_datatype const& operator[](unsigned int line)const&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
array_datatype& operator[](unsigned int line)&
{
assert(line < array_width);
assert(line >= 0);
return this->_data[line];
}
void clear(array_datatype initial = 0)
{
unsigned int column_index = 0;
for( ; column_index < array_width; column_index++ )
{
this->_data[column_index] = initial;
}
}
/**
* The Array<> type includes the Matrix<> type, because you can multiply a `Array` by an `Matrix`,
* but not a vice-versa.
*/
void multiply(Array< array_width, Array< array_width, array_datatype > > &matrix)
{
unsigned int column;
unsigned int step;
array_datatype old_array[array_width];
for(column = 0; column < array_width; column++)
{
old_array [column] = this->_data[column];
this->_data[column] = 0;
}
for(column = 0; column < array_width; column++)
{
for(step = 0; step < array_width; step++)
{
this->_data[column] += old_array[step] * matrix._data[step][column];
}
}
// If you would like to preserve the original value, it can be returned here
// return old_array;
}
/**
* Prints a more beauty version of the array when called on `std::cout<< array << std::end;`
*/
friend std::ostream& operator<<( std::ostream &output, const Array &array )
{
unsigned int column;
output << "{";
for( column=0; column < array_width; column++ )
{
output << array._data[column];
if( column != array_width-1 )
{
output << ", ";
}
}
output << "}";
return output;
}
};
这确实需要一些解释。第三个看起来非常可选,除非它将绑定到从函数调用返回的auto&@lili literal数组,否则使用[]获取元素,按值返回是正确的。例如,处理引用生存期扩展。这确实需要一些解释。第三个看起来非常可选,除非它将绑定到从函数调用返回的auto&@lili literal数组,否则使用[]获取元素,按值返回是正确的。例如,处理引用生存期延长。请注意的风险。您还应注意使用assertline>-1;对Francois来说,从POD类或聚合类派生出general通常是一个坏主意,也是一个错误的设计。数组是一个POD类。没有构造函数,没有赋值方法,没有虚函数,没有虚析构函数等等。它可能是功能性的,但这是一种危险的设计,不推荐使用。@FERcsI,你能参考一些解释它是如何危险的吗?如果你在原POD类的基础上创建一个新的POD类,应该没有问题。但由于不建议使用POD作为基类,理论上它可能导致未定义的行为。但是,如果您打破此限制并创建构造函数、析构函数、赋值运算符,编译器的行为可能会与预期不同。请注意的风险。您还应注意使用assertline>-1;对Francois来说,从POD类或聚合类派生出general通常是一个坏主意,也是一个错误的设计。数组是一个POD类。没有构造函数,没有赋值方法,没有虚函数,没有虚析构函数等等。它可能是功能性的,但这是一种危险的设计,不推荐使用。@FERcsI,你能参考一些解释它是如何危险的吗?如果你在原POD类的基础上创建一个新的POD类,应该没有问题。但由于不建议使用POD作为基类,理论上它可能导致未定义的行为。然而,若你们打破这一点,创建构造函数、析构函数、赋值运算符,编译器的行为可能会和预期的不同。