c++ - 一类嵌套类不能作为模板函数typename

Question

对于嵌套类，我想使用一类嵌套类作为模板类型名称，请看下面的代码:

#include <stdio.h>
#include <stdint.h>

struct H264 {
    struct NAL_UNIT {
        uint8_t nal_ref_idc;
        uint8_t nal_unit_type;
    };
};

struct H265 {
    struct NAL_UNIT {
        uint8_t nal_unit_type;
        uint8_t nuh_layer_id;
        uint8_t nuh_temporal_id_plus1;
    };
};

template <class T> void PrintNALUnitType(T::NAL_UNIT& nal_unit) {
    printf("nal_unit_type: %d.\n", nal_unit.nal_unit_type);
}

int main() {
    H264::NAL_UNIT h264_nu;
    h264_nu.nal_ref_idc = 2;
    h264_nu.nal_unit_type = 5;

    H265::NAL_UNIT h265_nu;
    h265_nu.nal_unit_type = 35;
    h265_nu.nuh_layer_id = 0;
    h265_nu.nuh_temporal_id_plus1 = 1;

    PrintNALUnitType(h264_nu);
    return 0;
}

但是在gcc中编译失败，

namespace.cpp:22:26: error: variable or field ‘PrintNALUnitType’ declared void
 void PrintNALUnitType(T::NAL_UNIT& nal_unit)
                          ^~~~~~~~
namespace.cpp:22:36: error: ‘nal_unit’ was not declared in this scope
 void PrintNALUnitType(T::NAL_UNIT& nal_unit)
                                    ^~~~~~~~
namespace.cpp:22:36: note: suggested alternative: ‘__unix’
 void PrintNALUnitType(T::NAL_UNIT& nal_unit)
                                    ^~~~~~~~
                                    __unix
namespace.cpp: In function ‘int main()’:
namespace.cpp:38:5: error: ‘PrintNALUnitType’ was not declared in this scope
     PrintNALUnitType(h264_nu);
     ^~~~~~~~~~~~~~~~

我知道它可以通过更改来修复

template <class T>
void PrintNALUnitType(T::NAL_UNIT& nal_unit)

到

template <class T>
void PrintNALUnitType(T& nal_unit)

但我只是想知道为什么它会破坏 c++ 规范，有人可以给出一些提示吗？

Answer 1

确实，对于从函数调用中推导模板参数，不能从嵌套类名中推导出封闭类名。在您的模板函数中:

template <class T>
void PrintNALUnitType(typename T::NAL_UNIT& nal_unit) {
   printf("nal_unit_type: %d.\n", nal_unit.nal_unit_type);
}

当像 PrintNALUnitType(h264_nu); 那样调用函数时，无法从参数类型推导出模板参数 T。

[temp.deduct.type]/8列出模板函数参数类型必须可扣除的可能形式:

A template type argument T, a template template argument TT or a template non-type argument i can be deduced if P and A have one of the following forms:

T

cv-list T

T*

T&

T&&

T[integer-constant]

template-name (where template-name refers to a class template)

type(T)

T()

T(T)

T type::*

type T::*

T T::*

T (type::*)()

type (T::*)()

type (type::*)(T)

type (T::*)(T)

T (type::*)(T)

T (T::*)()

T (T::*)(T)

type[i]

template-name (where template-name refers to a class template)

TT

TT

TT<>

where (T) represents a parameter-type-list ([dcl.fct]) where at least one parameter type contains a T, and () represents a parameter-type-list where no parameter type contains a T. Similarly, represents template argument lists where at least one argument contains a T, represents template argument lists where at least one argument contains an i and <> represents template argument lists where no argument contains a T or an i.