mirror of
https://github.com/ZDoom/raze-gles.git
synced 2024-11-10 23:02:03 +00:00
- removed all dirty headers from third party code.
This code should be project independent and remain project independent
This commit is contained in:
parent
e4f55d4d90
commit
5b805af4e7
3 changed files with 35 additions and 39 deletions
35
source/thirdparty/include/fix16.h
vendored
35
source/thirdparty/include/fix16.h
vendored
|
@ -1,9 +1,6 @@
|
|||
#ifndef __libfixmath_fix16_h__
|
||||
#define __libfixmath_fix16_h__
|
||||
|
||||
#include "compat.h"
|
||||
|
||||
|
||||
|
||||
/* These options may let the optimizer to remove some calls to the functions.
|
||||
* Refer to http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html
|
||||
|
@ -24,18 +21,18 @@
|
|||
|
||||
typedef int32_t fix16_t;
|
||||
|
||||
static CONSTEXPR const fix16_t FIX16_MAX = 0x7FFFFFFF; /*!< the maximum value of fix16_t */
|
||||
static CONSTEXPR const fix16_t FIX16_MIN = 0x80000000; /*!< the minimum value of fix16_t */
|
||||
static CONSTEXPR const fix16_t FIX16_OVERFLOW = 0x80000000; /*!< the value used to indicate overflows when FIXMATH_NO_OVERFLOW is not specified */
|
||||
static const fix16_t FIX16_MAX = 0x7FFFFFFF; /*!< the maximum value of fix16_t */
|
||||
static const fix16_t FIX16_MIN = 0x80000000; /*!< the minimum value of fix16_t */
|
||||
static const fix16_t FIX16_OVERFLOW = 0x80000000; /*!< the value used to indicate overflows when FIXMATH_NO_OVERFLOW is not specified */
|
||||
|
||||
static CONSTEXPR const fix16_t fix16_one = 0x00010000; /*!< fix16_t value of 1 */
|
||||
static const fix16_t fix16_one = 0x00010000; /*!< fix16_t value of 1 */
|
||||
|
||||
/* Conversion functions between fix16_t and float/integer.
|
||||
* These are inlined to allow compiler to optimize away constant numbers
|
||||
*/
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_from_int(int a) { return a * fix16_one; }
|
||||
static FORCE_INLINE CONSTEXPR float fix16_to_float(fix16_t a) { return (float)a / fix16_one; }
|
||||
static FORCE_INLINE CONSTEXPR double fix16_to_dbl(fix16_t a) { return (double)a / fix16_one; }
|
||||
static inline constexpr fix16_t fix16_from_int(int a) { return a * fix16_one; }
|
||||
static inline constexpr float fix16_to_float(fix16_t a) { return (float)a / fix16_one; }
|
||||
static inline constexpr double fix16_to_dbl(fix16_t a) { return (double)a / fix16_one; }
|
||||
|
||||
static inline constexpr int fix16_to_int(fix16_t a)
|
||||
{
|
||||
|
@ -77,24 +74,24 @@ static inline fix16_t fix16_from_dbl(double a)
|
|||
*/
|
||||
#define F16(x) ((fix16_t)(((x) >= 0) ? ((x) * 65536.0 + 0.5) : ((x) * 65536.0 - 0.5)))
|
||||
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_abs(fix16_t x)
|
||||
static inline constexpr fix16_t fix16_abs(fix16_t x)
|
||||
{ return (x < 0 ? -x : x); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_floor(fix16_t x)
|
||||
static inline constexpr fix16_t fix16_floor(fix16_t x)
|
||||
{ return (x & 0xFFFF0000UL); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_ceil(fix16_t x)
|
||||
static inline constexpr fix16_t fix16_ceil(fix16_t x)
|
||||
{ return (x & 0xFFFF0000UL) + ((x & 0x0000FFFFUL) ? fix16_one : 0); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_min(fix16_t x, fix16_t y)
|
||||
static inline constexpr fix16_t fix16_min(fix16_t x, fix16_t y)
|
||||
{ return (x < y ? x : y); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_max(fix16_t x, fix16_t y)
|
||||
static inline constexpr fix16_t fix16_max(fix16_t x, fix16_t y)
|
||||
{ return (x > y ? x : y); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_clamp(fix16_t x, fix16_t lo, fix16_t hi)
|
||||
static inline constexpr fix16_t fix16_clamp(fix16_t x, fix16_t lo, fix16_t hi)
|
||||
{ return fix16_min(fix16_max(x, lo), hi); }
|
||||
|
||||
/* Subtraction and addition with (optional) overflow detection. */
|
||||
#ifdef FIXMATH_NO_OVERFLOW
|
||||
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_add(fix16_t inArg0, fix16_t inArg1) { return (inArg0 + inArg1); }
|
||||
static FORCE_INLINE CONSTEXPR fix16_t fix16_sub(fix16_t inArg0, fix16_t inArg1) { return (inArg0 - inArg1); }
|
||||
static inline constexpr fix16_t fix16_add(fix16_t inArg0, fix16_t inArg1) { return (inArg0 + inArg1); }
|
||||
static inline constexpr fix16_t fix16_sub(fix16_t inArg0, fix16_t inArg1) { return (inArg0 - inArg1); }
|
||||
|
||||
#else
|
||||
|
||||
|
@ -127,7 +124,7 @@ extern fix16_t fix16_sdiv(fix16_t inArg0, fix16_t inArg1) FIXMATH_FUNC_ATTRS;
|
|||
|
||||
/*! Divides the first given fix16_t by the second and returns the remainder.
|
||||
*/
|
||||
static FORCE_INLINE fix16_t fix16_mod(fix16_t x, fix16_t y) { return x %= y; }
|
||||
static inline fix16_t fix16_mod(fix16_t x, fix16_t y) { return x %= y; }
|
||||
|
||||
|
||||
|
||||
|
|
36
source/thirdparty/include/fix16_int64.h
vendored
36
source/thirdparty/include/fix16_int64.h
vendored
|
@ -1,28 +1,28 @@
|
|||
#ifndef __libfixmath_int64_h__
|
||||
#define __libfixmath_int64_h__
|
||||
|
||||
#include "compat.h"
|
||||
#include <stdint.h>
|
||||
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_const(int32_t hi, uint32_t lo) { return (((int64_t)hi << 32) | lo); }
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_from_int32(int32_t x) { return (int64_t)x; }
|
||||
static FORCE_INLINE CONSTEXPR int32_t int64_hi(int64_t x) { return (x >> 32); }
|
||||
static FORCE_INLINE CONSTEXPR uint32_t int64_lo(int64_t x) { return (x & ((1ULL << 32) - 1)); }
|
||||
static inline constexpr int64_t int64_const(int32_t hi, uint32_t lo) { return (((int64_t)hi << 32) | lo); }
|
||||
static inline constexpr int64_t int64_from_int32(int32_t x) { return (int64_t)x; }
|
||||
static inline constexpr int32_t int64_hi(int64_t x) { return (x >> 32); }
|
||||
static inline constexpr uint32_t int64_lo(int64_t x) { return (x & ((1ULL << 32) - 1)); }
|
||||
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_add(int64_t x, int64_t y) { return (x + y); }
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_neg(int64_t x) { return (-x); }
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_sub(int64_t x, int64_t y) { return (x - y); }
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_shift(int64_t x, int8_t y) { return (y < 0 ? (x >> -y) : (x << y)); }
|
||||
static inline constexpr int64_t int64_add(int64_t x, int64_t y) { return (x + y); }
|
||||
static inline constexpr int64_t int64_neg(int64_t x) { return (-x); }
|
||||
static inline constexpr int64_t int64_sub(int64_t x, int64_t y) { return (x - y); }
|
||||
static inline constexpr int64_t int64_shift(int64_t x, int8_t y) { return (y < 0 ? (x >> -y) : (x << y)); }
|
||||
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_mul_i32_i32(int32_t x, int32_t y) { return (x * y); }
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_mul_i64_i32(int64_t x, int32_t y) { return (x * y); }
|
||||
static inline constexpr int64_t int64_mul_i32_i32(int32_t x, int32_t y) { return (x * y); }
|
||||
static inline constexpr int64_t int64_mul_i64_i32(int64_t x, int32_t y) { return (x * y); }
|
||||
|
||||
static FORCE_INLINE CONSTEXPR int64_t int64_div_i64_i32(int64_t x, int32_t y) { return (x / y); }
|
||||
static inline constexpr int64_t int64_div_i64_i32(int64_t x, int32_t y) { return (x / y); }
|
||||
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_eq(int64_t x, int64_t y) { return (x == y); }
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_ne(int64_t x, int64_t y) { return (x != y); }
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_gt(int64_t x, int64_t y) { return (x > y); }
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_ge(int64_t x, int64_t y) { return (x >= y); }
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_lt(int64_t x, int64_t y) { return (x < y); }
|
||||
static FORCE_INLINE CONSTEXPR int int64_cmp_le(int64_t x, int64_t y) { return (x <= y); }
|
||||
static inline constexpr int int64_cmp_eq(int64_t x, int64_t y) { return (x == y); }
|
||||
static inline constexpr int int64_cmp_ne(int64_t x, int64_t y) { return (x != y); }
|
||||
static inline constexpr int int64_cmp_gt(int64_t x, int64_t y) { return (x > y); }
|
||||
static inline constexpr int int64_cmp_ge(int64_t x, int64_t y) { return (x >= y); }
|
||||
static inline constexpr int int64_cmp_lt(int64_t x, int64_t y) { return (x < y); }
|
||||
static inline constexpr int int64_cmp_le(int64_t x, int64_t y) { return (x <= y); }
|
||||
|
||||
#endif
|
||||
|
|
3
source/thirdparty/src/md4.cpp
vendored
3
source/thirdparty/src/md4.cpp
vendored
|
@ -30,7 +30,6 @@ typedef unsigned short UINT2;
|
|||
typedef unsigned int UINT4;
|
||||
|
||||
#include "md4.h"
|
||||
#include "compat.h"
|
||||
|
||||
/* Constants for MD4Transform routine.
|
||||
*/
|
||||
|
@ -50,7 +49,7 @@ typedef unsigned int UINT4;
|
|||
static void MD4Transform(UINT4 [4], const unsigned char [64]);
|
||||
static void Encode(unsigned char *, const UINT4 *, unsigned int);
|
||||
static void Decode(UINT4 *, const unsigned char *, unsigned int);
|
||||
#define MD4_memcpy Bmemcpy
|
||||
#define MD4_memcpy memcpy
|
||||
#define MD4_memset memset
|
||||
|
||||
static unsigned char PADDING[64] =
|
||||
|
|
Loading…
Reference in a new issue