summaryrefslogtreecommitdiff
path: root/src/glm/gtc/half_float.hpp
diff options
context:
space:
mode:
Diffstat (limited to 'src/glm/gtc/half_float.hpp')
-rw-r--r--src/glm/gtc/half_float.hpp401
1 files changed, 401 insertions, 0 deletions
diff --git a/src/glm/gtc/half_float.hpp b/src/glm/gtc/half_float.hpp
new file mode 100644
index 0000000..32f6d4b
--- /dev/null
+++ b/src/glm/gtc/half_float.hpp
@@ -0,0 +1,401 @@
+///////////////////////////////////////////////////////////////////////////////////////////////////
+// OpenGL Mathematics Copyright (c) 2005 - 2011 G-Truc Creation (www.g-truc.net)
+///////////////////////////////////////////////////////////////////////////////////////////////////
+// Created : 2009-04-29
+// Updated : 2010-02-07
+// Licence : This source is under MIT License
+// File : glm/gtc/half_float.hpp
+///////////////////////////////////////////////////////////////////////////////////////////////////
+
+#ifndef glm_gtc_half_float
+#define glm_gtc_half_float
+
+// Dependency:
+#include "../glm.hpp"
+
+#if(defined(GLM_MESSAGES) && !defined(glm_ext))
+# pragma message("GLM: GLM_GTC_half_float extension included")
+#endif
+
+namespace glm{
+namespace detail
+{
+#ifndef _MSC_EXTENSIONS
+ template <>
+ struct tvec2<thalf>
+ {
+ enum ctor{null};
+ typedef thalf value_type;
+ typedef std::size_t size_type;
+ static size_type value_size();
+
+ typedef tvec2<thalf> type;
+ typedef tvec2<bool> bool_type;
+
+ //////////////////////////////////////
+ // Data
+
+ thalf x, y;
+
+ //////////////////////////////////////
+ // Accesses
+
+ thalf & operator[](size_type i);
+ thalf const & operator[](size_type i) const;
+
+ //////////////////////////////////////
+ // Implicit basic constructors
+
+ tvec2();
+ tvec2(tvec2<thalf> const & v);
+
+ //////////////////////////////////////
+ // Explicit basic constructors
+
+ explicit tvec2(ctor);
+ explicit tvec2(
+ thalf const & s);
+ explicit tvec2(
+ thalf const & s1,
+ thalf const & s2);
+
+ //////////////////////////////////////
+ // Swizzle constructors
+
+ tvec2(tref2<thalf> const & r);
+
+ //////////////////////////////////////
+ // Convertion scalar constructors
+
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec2(U const & x);
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U, typename V>
+ explicit tvec2(U const & x, V const & y);
+
+ //////////////////////////////////////
+ // Convertion vector constructors
+
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec2(tvec2<U> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec2(tvec3<U> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec2(tvec4<U> const & v);
+
+ //////////////////////////////////////
+ // Unary arithmetic operators
+
+ tvec2<thalf>& operator= (tvec2<thalf> const & v);
+
+ tvec2<thalf>& operator+=(thalf const & s);
+ tvec2<thalf>& operator+=(tvec2<thalf> const & v);
+ tvec2<thalf>& operator-=(thalf const & s);
+ tvec2<thalf>& operator-=(tvec2<thalf> const & v);
+ tvec2<thalf>& operator*=(thalf const & s);
+ tvec2<thalf>& operator*=(tvec2<thalf> const & v);
+ tvec2<thalf>& operator/=(thalf const & s);
+ tvec2<thalf>& operator/=(tvec2<thalf> const & v);
+ tvec2<thalf>& operator++();
+ tvec2<thalf>& operator--();
+
+ //////////////////////////////////////
+ // Swizzle operators
+
+ thalf swizzle(comp X) const;
+ tvec2<thalf> swizzle(comp X, comp Y) const;
+ tvec3<thalf> swizzle(comp X, comp Y, comp Z) const;
+ tvec4<thalf> swizzle(comp X, comp Y, comp Z, comp W) const;
+ tref2<thalf> swizzle(comp X, comp Y);
+ };
+
+ template <>
+ struct tvec3<thalf>
+ {
+ enum ctor{null};
+ typedef thalf value_type;
+ typedef std::size_t size_type;
+ static size_type value_size();
+
+ typedef tvec3<thalf> type;
+ typedef tvec3<bool> bool_type;
+
+ //////////////////////////////////////
+ // Data
+
+ thalf x, y, z;
+
+ //////////////////////////////////////
+ // Accesses
+
+ thalf & operator[](size_type i);
+ thalf const & operator[](size_type i) const;
+
+ //////////////////////////////////////
+ // Implicit basic constructors
+
+ tvec3();
+ tvec3(tvec3<thalf> const & v);
+
+ //////////////////////////////////////
+ // Explicit basic constructors
+
+ explicit tvec3(ctor);
+ explicit tvec3(
+ thalf const & s);
+ explicit tvec3(
+ thalf const & s1,
+ thalf const & s2,
+ thalf const & s3);
+
+ //////////////////////////////////////
+ // Swizzle constructors
+
+ tvec3(tref3<thalf> const & r);
+
+ //////////////////////////////////////
+ // Convertion scalar constructors
+
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec3(U const & x);
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U, typename V, typename W>
+ explicit tvec3(U const & x, V const & y, W const & z);
+
+ //////////////////////////////////////
+ // Convertion vector constructors
+
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B>
+ explicit tvec3(tvec2<A> const & v, B const & s);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B>
+ explicit tvec3(A const & s, tvec2<B> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec3(tvec3<U> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec3(tvec4<U> const & v);
+
+ //////////////////////////////////////
+ // Unary arithmetic operators
+
+ tvec3<thalf>& operator= (tvec3<thalf> const & v);
+
+ tvec3<thalf>& operator+=(thalf const & s);
+ tvec3<thalf>& operator+=(tvec3<thalf> const & v);
+ tvec3<thalf>& operator-=(thalf const & s);
+ tvec3<thalf>& operator-=(tvec3<thalf> const & v);
+ tvec3<thalf>& operator*=(thalf const & s);
+ tvec3<thalf>& operator*=(tvec3<thalf> const & v);
+ tvec3<thalf>& operator/=(thalf const & s);
+ tvec3<thalf>& operator/=(tvec3<thalf> const & v);
+ tvec3<thalf>& operator++();
+ tvec3<thalf>& operator--();
+
+ //////////////////////////////////////
+ // Swizzle operators
+
+ thalf swizzle(comp X) const;
+ tvec2<thalf> swizzle(comp X, comp Y) const;
+ tvec3<thalf> swizzle(comp X, comp Y, comp Z) const;
+ tvec4<thalf> swizzle(comp X, comp Y, comp Z, comp W) const;
+ tref3<thalf> swizzle(comp X, comp Y, comp Z);
+ };
+
+ template <>
+ struct tvec4<thalf>
+ {
+ enum ctor{null};
+ typedef thalf value_type;
+ typedef std::size_t size_type;
+ static size_type value_size();
+
+ typedef tvec4<thalf> type;
+ typedef tvec4<bool> bool_type;
+
+ //////////////////////////////////////
+ // Data
+
+ thalf x, y, z, w;
+
+ //////////////////////////////////////
+ // Accesses
+
+ thalf & operator[](size_type i);
+ thalf const & operator[](size_type i) const;
+
+ //////////////////////////////////////
+ // Implicit basic constructors
+
+ tvec4();
+ tvec4(tvec4<thalf> const & v);
+
+ //////////////////////////////////////
+ // Explicit basic constructors
+
+ explicit tvec4(ctor);
+ explicit tvec4(
+ thalf const & s);
+ explicit tvec4(
+ thalf const & s0,
+ thalf const & s1,
+ thalf const & s2,
+ thalf const & s3);
+
+ //////////////////////////////////////
+ // Swizzle constructors
+
+ tvec4(tref4<thalf> const & r);
+
+ //////////////////////////////////////
+ // Convertion scalar constructors
+
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec4(U const & x);
+ //! Explicit converions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B, typename C, typename D>
+ explicit tvec4(A const & x, B const & y, C const & z, D const & w);
+
+ //////////////////////////////////////
+ // Convertion vector constructors
+
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B, typename C>
+ explicit tvec4(tvec2<A> const & v, B const & s1, C const & s2);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B, typename C>
+ explicit tvec4(A const & s1, tvec2<B> const & v, C const & s2);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B, typename C>
+ explicit tvec4(A const & s1, B const & s2, tvec2<C> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B>
+ explicit tvec4(tvec3<A> const & v, B const & s);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B>
+ explicit tvec4(A const & s, tvec3<B> const & v);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename A, typename B>
+ explicit tvec4(tvec2<A> const & v1, tvec2<B> const & v2);
+ //! Explicit conversions (From section 5.4.1 Conversion and scalar constructors of GLSL 1.30.08 specification)
+ template <typename U>
+ explicit tvec4(tvec4<U> const & v);
+
+ //////////////////////////////////////
+ // Unary arithmetic operators
+
+ tvec4<thalf>& operator= (tvec4<thalf> const & v);
+
+ tvec4<thalf>& operator+=(thalf const & s);
+ tvec4<thalf>& operator+=(tvec4<thalf> const & v);
+ tvec4<thalf>& operator-=(thalf const & s);
+ tvec4<thalf>& operator-=(tvec4<thalf> const & v);
+ tvec4<thalf>& operator*=(thalf const & s);
+ tvec4<thalf>& operator*=(tvec4<thalf> const & v);
+ tvec4<thalf>& operator/=(thalf const & s);
+ tvec4<thalf>& operator/=(tvec4<thalf> const & v);
+ tvec4<thalf>& operator++();
+ tvec4<thalf>& operator--();
+
+ //////////////////////////////////////
+ // Swizzle operators
+
+ thalf swizzle(comp X) const;
+ tvec2<thalf> swizzle(comp X, comp Y) const;
+ tvec3<thalf> swizzle(comp X, comp Y, comp Z) const;
+ tvec4<thalf> swizzle(comp X, comp Y, comp Z, comp W) const;
+ tref4<thalf> swizzle(comp X, comp Y, comp Z, comp W);
+ };
+#endif//_MSC_EXTENSIONS
+}
+//namespace detail
+
+namespace gtc{
+namespace half_float ///< GLM_GTC_half_float extension: Add support for half precision floating-point types
+{
+ /// \addtogroup gtc_half_float
+ ///@{
+
+ /// Type for half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::thalf half;
+
+ /// Vector of 2 half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tvec2<detail::thalf> hvec2;
+
+ /// Vector of 3 half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tvec3<detail::thalf> hvec3;
+
+ /// Vector of 4 half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tvec4<detail::thalf> hvec4;
+
+ /// 2 * 2 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat2x2<detail::thalf> hmat2;
+
+ /// 3 * 3 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat3x3<detail::thalf> hmat3;
+
+ /// 4 * 4 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat4x4<detail::thalf> hmat4;
+
+ /// 2 * 2 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat2x2<detail::thalf> hmat2x2;
+
+ /// 2 * 3 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat2x3<detail::thalf> hmat2x3;
+
+ /// 2 * 4 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat2x4<detail::thalf> hmat2x4;
+
+ /// 3 * 2 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat3x2<detail::thalf> hmat3x2;
+
+ /// 3 * 3 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat3x3<detail::thalf> hmat3x3;
+
+ /// 3 * 4 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat3x4<detail::thalf> hmat3x4;
+
+ /// 4 * 2 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat4x2<detail::thalf> hmat4x2;
+
+ /// 4 * 3 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat4x3<detail::thalf> hmat4x3;
+
+ /// 4 * 4 matrix of half-precision floating-point numbers.
+ /// From GLM_GTC_half_float extension.
+ typedef detail::tmat4x4<detail::thalf> hmat4x4;
+
+ /// @}
+
+}// namespace half_float
+}// namespace gtc
+}// namespace glm
+
+#include "half_float.inl"
+
+namespace glm{using namespace gtc::half_float;}
+
+#endif//glm_gtc_half_float