＆＃x27; fpClassify＆＃x27;：含糊不清的呼叫对超载功能

发布于 2025-01-21 05:32:19 字数 1479 浏览 1 评论 0 原文

我成功地在MacBook上以及带有Intel编译器的服务器上编译了代码库。当我通过公司的连续集成测试（具有其自身的构建系统）运行它时，我会在编译中遇到以下错误。

..\corecrt_math.h(415): error C2668: 'fpclassify': ambiguous call to overloaded function
..\corecrt_math.h(300): note: could be 'int fpclassify(long double) noexcept'
..\corecrt_math.h(295): note: or       'int fpclassify(double) noexcept'
..\corecrt_math.h(290): note: or       'int fpclassify(float) noexcept'
..\corecrt_math.h(415): note: while trying to match the argument list '(_Ty)'
        with
        [
            _Ty=kc20::wrapped_int8::underlying
        ]
/path/to/file/kernel.cpp(69): note: see reference to function template instantiation 'bool isnan<integer_type>(_Ty) noexcept' being compiled
        with
        [
            integer_type=kc20::wrapped_int8::underlying,
            _Ty=kc20::wrapped_int8::underlying
        ]

请注意， kc20 是一个名称空间，包含 wrapped_int8 struct。 wrapped_int8 定义为

struct alignas(1) wrapped_int8 {
  using underlying = int8_t;
  int8_t value;
  // some other function definitions
};

kc20 :: wrapped_int8 :: ustlying 只是 int8_t 。在错误消息中，它似乎仅列出 fpClassify in =“ nofollow noreferrer”> https://en.cppreference.com/w/cpp/numeric/math/math/fpclassify 。我的代码应该使用此链接的情况（4），因为它是一个积分类型。这里的歧义是什么？

原文

I successfully compiled a codebase on my macbook and also on a server with intel compilers. When I run it through my company's continuous integrating testing, which has its own build system, I am getting the following error in the compilation.

..\corecrt_math.h(415): error C2668: 'fpclassify': ambiguous call to overloaded function
..\corecrt_math.h(300): note: could be 'int fpclassify(long double) noexcept'
..\corecrt_math.h(295): note: or       'int fpclassify(double) noexcept'
..\corecrt_math.h(290): note: or       'int fpclassify(float) noexcept'
..\corecrt_math.h(415): note: while trying to match the argument list '(_Ty)'
        with
        [
            _Ty=kc20::wrapped_int8::underlying
        ]
/path/to/file/kernel.cpp(69): note: see reference to function template instantiation 'bool isnan<integer_type>(_Ty) noexcept' being compiled
        with
        [
            integer_type=kc20::wrapped_int8::underlying,
            _Ty=kc20::wrapped_int8::underlying
        ]

Note that kc20 is a namespace that contains the wrapped_int8 struct. wrapped_int8 is defined as

struct alignas(1) wrapped_int8 {
  using underlying = int8_t;
  int8_t value;
  // some other function definitions
};

so kc20::wrapped_int8::underlying is just int8_t. In the error message, it seems to only list the first 3 overloads of fpclassify in https://en.cppreference.com/w/cpp/numeric/math/fpclassify . My code should be using case (4) of this link since it's an integral type. What is the ambiguity here?