Extends c++ type traits and implements a few more to boot.
Provides some cross platform type traits above and beyond what is provided by c++.
Include the following at the top of any translation unit which requires this library:
#include <ecl/type_traits.hpp> // The error interfaces using ecl::numeric_limits; using ecl::is_integral; using ecl::is_float; using ecl::is_signed; using ecl::is_unsigned; using ecl::is_byte;
You will also need to link to -lecl_type_traits.
This extends c++'s numeric_limits class for fundamental types, providing in addition, static constants that can be utilised as template arguments (as opposed to c macros or c++ functions in numeric_limits). For integral types, an example of the extras available:
ecl::uint16 bits = NumericLimits<int>::bits;
ecl::uint16 bytes = NumericLimits<int>::bytes;
int j = NumericLimits<int>::one;
int i = NumericLimits<int>::minimum;
int k = NumericLimits<int>::maximum;
An example of some extras for a float type:
ecl::uint16 bits = NumericLimits<float>::bits;
ecl::uint16 bytes = NumericLimits<float>::bytes;
// precision setting is a rough, but useful aid to bounding the accuracy for some math calculations
NumericLimits<float>::Precision p = NumericLimits<float>::dummy_precision();
// note, numeric_limits<float>::min() is the smallest
// positive value, this is the smallest negative value.
float minimum = NumericLimits<float>::minimum;
float maximum = NumericLimits<float>::maximum;
Note that it inherits the numeric_limits class, so all the functions therein can also be used.
These are in the upcoming C++0x standard, so are only a temporary feature in the ecl. They are a subset and work just like boosts type traits. More will be added as needed. The code snippet below is for runtime code, but they're actually more useful with enable_if in determining which compile time template specialisations to instantiate.
if ( is_float<double>::value ) { // take some action }
The current list of type traits include: