A RetroSearch Logo

Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Search Query:

Showing content from https://en.cppreference.com/w/cpp/language/../keywords/../types/numeric_limits/max.html below:

std::numeric_limits<T>::max - cppreference.com

static T max() throw();

(until C++11)

static constexpr T max() noexcept;

(since C++11)

Returns the maximum finite value representable by the numeric type T. Meaningful for all bounded types.

[edit] Return value [edit] Example

Demonstrates the use of max() with some fundamental types and some standard library typedefs (the output is system-specific):

#include <boost/type_index.hpp>
#include <cstddef>
#include <iomanip>
#include <iostream>
#include <limits>
#include <type_traits>
 
template<typename T>
void print_max_value_of()
{
    constexpr T max{std::numeric_limits<T>::max()};
 
    std::cout << std::setw(16) << boost::typeindex::type_id<T>() << ": ";
    if constexpr (std::is_floating_point_v<T>)
        std::cout << std::defaultfloat << max << " = " << std::hexfloat << max << '\n';
    else
    {
        constexpr auto m{static_cast<unsigned long long>(max)};
        std::cout << std::dec << m << " = " << std::hex << m << '\n';
    }
}
 
int main()
{
    std::cout << std::showbase;
 
    print_max_value_of<bool>();
    print_max_value_of<short>();
    print_max_value_of<int>();
    print_max_value_of<std::streamsize>();
    print_max_value_of<std::size_t>();
    print_max_value_of<char>();
    print_max_value_of<char16_t>();
    print_max_value_of<wchar_t>();
    print_max_value_of<float>();
    print_max_value_of<double>();
    print_max_value_of<long double>();
}

Possible output:

            bool: 1 = 0x1
           short: 32767 = 0x7fff
             int: 2147483647 = 0x7fffffff
            long: 9223372036854775807 = 0x7fffffffffffffff
   unsigned long: 18446744073709551615 = 0xffffffffffffffff
            char: 127 = 0x7f
        char16_t: 65535 = 0xffff
         wchar_t: 2147483647 = 0x7fffffff
           float: 3.40282e+38 = 0x1.fffffep+127
          double: 1.79769e+308 = 0x1.fffffffffffffp+1023
     long double: 1.18973e+4932 = 0xf.fffffffffffffffp+16380
[edit] See also returns the lowest finite value of the given type, i.e. the most negative value for signed types, ​0​ for unsigned types
(public static member function) [edit] returns the smallest finite value of the given non-floating-point type, or the smallest positive normal value of the given floating-point type
(public static member function) [edit]

RetroSearch is an open source project built by @garambo | Open a GitHub Issue

Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo

HTML: 3.2 | Encoding: UTF-8 | Version: 0.7.4