对，我查Stack Overflow上说，这个std::abs和abs是两套东西。c的abs，只支持int
cppreference的std::abs页面上Defect reports说，

The following behavior-changing defect reports were applied retroactively to previously published C++ standards.
DR            LWG 2192    
Applied to     C++98    
Behavior as published      overloads of std::abs were inconsistently declared in two headers     -- 这是说定义不致 <cmath> <cstdlib>
Correct behavior         declared these overloads in both headers

另一个页面，std::fabs，说
DR            LWG 2735    
Applied to     C++11      
Behavior as published      overloads of std::abs for integer types returning double was erroneously required  -- 这是说 double fabs(int)，进int出double，易错。
Correct behavior         removed the requirement  


https://en.cppreference.com/w/cpp/numeric/math/abs
https://en.cppreference.com/w/cpp/numeric/math/fabs

琢磨来琢磨去，我觉得 abs , std::abs留给整型， fabs()留给float/double最佳实践。
至于abs支持float/double，应该是第一个Defect Report引入的bug。

【 在 origin008 的大作中提到: 】
: 有的编译器，abs就自动int了
: 很久以前遇到过这个问题
: 不过那个时候还没用std::这么高级的货
: ...................
--
FROM 61.185.187.*