Class 方法专业化

Question

Currently using g++-11.3.0, C++20.

I'm trying to make this class that accepts two parameters at construction: a void pointer that points to a pre-allocated block of memory, and the size of the memory block. The class then iterates through the data block with calls to the next() method, which returns a reference of the next available slot in the data block, enclosed in the specified type T .

class DataArrayIterator
{
private:
    
    void* data;
    size_t maxSize;
    size_t size;

public:

    template<typename T>
    T& next<T>()
    {
        assert(sizeof(T) + size <= maxSize);

        size_t offset = size;

        size += sizeof(T);

        return *((T*)(static_cast<char*>(data) + offset));
    }

    DataArrayIterator(void* data, size_t maxSize)
      :  data{data}, maxSize{maxSize}, size{0}
    {

    }
};

Compiling the code gives me the error non-class, non-variable partial specialization 'next<T>' is not allowed . Is there anything I could do to mimic the same functionality as closely as possible? Or are there more suitable alternatives?

Answer 1

You don't need any partial specialisation in this case, just introduce a function template:

template<typename T>
T& next() {
    ... // implementation goes here
}

And instantiate the function when needed:

auto i = 8;
// val is an int reference with value of 8
auto& val = DataArrayIterator{ &i, sizeof(i) }.next<int>();