numericCast.h File Reference

#include "pxr/pxr.h"
#include "pxr/base/gf/traits.h"
#include <cmath>
#include <limits>
#include <optional>
#include <type_traits>

Include dependency graph for numericCast.h:

Go to the source code of this file.

Enumerations
enum	GfNumericCastFailureType { GfNumericCastPosOverflow, GfNumericCastNegOverflow, GfNumericCastNaN }

Functions
template<class T , class U >
PXR_NAMESPACE_OPEN_SCOPE constexpr bool	GfIntegerCompareLess (T t, U u) noexcept
template<class To , class From >
std::optional< To >	GfNumericCast (From from, GfNumericCastFailureType *failType=nullptr)

Enumeration Type Documentation

enum GfNumericCastFailureType

Enumerator
GfNumericCastPosOverflow	Value too high to convert.
GfNumericCastNegOverflow	Value too low to convert.
GfNumericCastNaN	Value is a floating-point NaN.

Definition at line 45 of file numericCast.h.

Function Documentation

template<class T , class U >

PXR_NAMESPACE_OPEN_SCOPE constexpr bool GfIntegerCompareLess ( T  t, U  u  )

noexcept

Return true if integer t compares logically less-than integer u in a mathematical sense. The comparison is safe against non-value-preserving integral conversion.

This mimics the C++20 std::cmp_less function for comparing integers of different types where negative signed integers always compare less than (and not equal to) unsigned integers.

Definition at line 30 of file numericCast.h.

template<class To , class From >

std::optional<To> GfNumericCast	(	From	from,
		GfNumericCastFailureType *	failType = nullptr
	)

Attempt to convert from to a value of type To "safely". From and To must be arithmetic types according to GfIsArithmetic – either integral or floating-point types (including GfHalf). Return a std::optional holding the converted value if conversion succeeds, otherwise the empty optional. The optional out-parameter failType can be used to determine why conversion failed if desired.

What "safely" means depends on the types From and To. If From and To are both integral types, then from can safely convert to To if from is in To's range. For example if from is an int32_t and To is uint16_t, then from can successfully convert if it is in the range [0, 65535].

If To is an integral type and From is a floating-point type (including GfHalf), then from can safely convert to To if it is neither a NaN nor an infinity, and after truncation to integer its value is in To's range, as above.

Following boost::numeric_cast's behavior, no range checking is performed converting from integral to floating-point or from floating-point to other floating-point types. Note that converting an integral value that is out of GfHalf's finite range will produce a +/- inf GfHalf.

Definition at line 75 of file numericCast.h.