shell bypass 403
UnknownSec Shell
:
/
usr
/
include
/
jasper
/ [
drwxr-xr-x
]
upload
mass deface
mass delete
console
info server
name :
jas_fix.h
/* * Copyright (c) 1999-2000 Image Power, Inc. and the University of * British Columbia. * Copyright (c) 2001-2002 Michael David Adams. * All rights reserved. */ /* __START_OF_JASPER_LICENSE__ * * JasPer License Version 2.0 * * Copyright (c) 2001-2006 Michael David Adams * Copyright (c) 1999-2000 Image Power, Inc. * Copyright (c) 1999-2000 The University of British Columbia * * All rights reserved. * * Permission is hereby granted, free of charge, to any person (the * "User") obtaining a copy of this software and associated documentation * files (the "Software"), to deal in the Software without restriction, * including without limitation the rights to use, copy, modify, merge, * publish, distribute, and/or sell copies of the Software, and to permit * persons to whom the Software is furnished to do so, subject to the * following conditions: * * 1. The above copyright notices and this permission notice (which * includes the disclaimer below) shall be included in all copies or * substantial portions of the Software. * * 2. The name of a copyright holder shall not be used to endorse or * promote products derived from the Software without specific prior * written permission. * * THIS DISCLAIMER OF WARRANTY CONSTITUTES AN ESSENTIAL PART OF THIS * LICENSE. NO USE OF THE SOFTWARE IS AUTHORIZED HEREUNDER EXCEPT UNDER * THIS DISCLAIMER. THE SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS * "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING * BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A * PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS. IN NO * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, OR ANY SPECIAL * INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY DAMAGES WHATSOEVER RESULTING * FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, * NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION * WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. NO ASSURANCES ARE * PROVIDED BY THE COPYRIGHT HOLDERS THAT THE SOFTWARE DOES NOT INFRINGE * THE PATENT OR OTHER INTELLECTUAL PROPERTY RIGHTS OF ANY OTHER ENTITY. * EACH COPYRIGHT HOLDER DISCLAIMS ANY LIABILITY TO THE USER FOR CLAIMS * BROUGHT BY ANY OTHER ENTITY BASED ON INFRINGEMENT OF INTELLECTUAL * PROPERTY RIGHTS OR OTHERWISE. AS A CONDITION TO EXERCISING THE RIGHTS * GRANTED HEREUNDER, EACH USER HEREBY ASSUMES SOLE RESPONSIBILITY TO SECURE * ANY OTHER INTELLECTUAL PROPERTY RIGHTS NEEDED, IF ANY. THE SOFTWARE * IS NOT FAULT-TOLERANT AND IS NOT INTENDED FOR USE IN MISSION-CRITICAL * SYSTEMS, SUCH AS THOSE USED IN THE OPERATION OF NUCLEAR FACILITIES, * AIRCRAFT NAVIGATION OR COMMUNICATION SYSTEMS, AIR TRAFFIC CONTROL * SYSTEMS, DIRECT LIFE SUPPORT MACHINES, OR WEAPONS SYSTEMS, IN WHICH * THE FAILURE OF THE SOFTWARE OR SYSTEM COULD LEAD DIRECTLY TO DEATH, * PERSONAL INJURY, OR SEVERE PHYSICAL OR ENVIRONMENTAL DAMAGE ("HIGH * RISK ACTIVITIES"). THE COPYRIGHT HOLDERS SPECIFICALLY DISCLAIM ANY * EXPRESS OR IMPLIED WARRANTY OF FITNESS FOR HIGH RISK ACTIVITIES. * * __END_OF_JASPER_LICENSE__ */ /* * Fixed-Point Number Class * * $Id$ */ #ifndef JAS_FIX_H #define JAS_FIX_H /******************************************************************************\ * Includes. \******************************************************************************/ /* The configuration header file should be included first. */ #include <jasper/jas_config.h> #include <stdio.h> #include <stdlib.h> #include <math.h> #include <jasper/jas_types.h> #include <jasper/jas_debug.h> #ifdef __cplusplus extern "C" { #endif /******************************************************************************\ * Constants. \******************************************************************************/ /* The representation of the value zero. */ #define JAS_FIX_ZERO(fix_t, fracbits) \ JAS_CAST(fix_t, 0) /* The representation of the value one. */ #define JAS_FIX_ONE(fix_t, fracbits) \ (JAS_CAST(fix_t, 1) << (fracbits)) /* The representation of the value one half. */ #define JAS_FIX_HALF(fix_t, fracbits) \ (JAS_CAST(fix_t, 1) << ((fracbits) - 1)) /******************************************************************************\ * Conversion operations. \******************************************************************************/ /* Convert an int to a fixed-point number. */ #define JAS_INTTOFIX(fix_t, fracbits, x) \ JAS_CAST(fix_t, (x) << (fracbits)) /* Convert a fixed-point number to an int. */ #define JAS_FIXTOINT(fix_t, fracbits, x) \ JAS_CAST(int, (x) >> (fracbits)) /* Convert a fixed-point number to a double. */ #define JAS_FIXTODBL(fix_t, fracbits, x) \ (JAS_CAST(double, x) / (JAS_CAST(fix_t, 1) << (fracbits))) /* Convert a double to a fixed-point number. */ #define JAS_DBLTOFIX(fix_t, fracbits, x) \ JAS_CAST(fix_t, ((x) * JAS_CAST(double, JAS_CAST(fix_t, 1) << (fracbits)))) /******************************************************************************\ * Basic arithmetic operations. * All other arithmetic operations are synthesized from these basic operations. * There are three macros for each type of arithmetic operation. * One macro always performs overflow/underflow checking, one never performs * overflow/underflow checking, and one is generic with its behavior * depending on compile-time flags. * Only the generic macros should be invoked directly by application code. \******************************************************************************/ /* Calculate the sum of two fixed-point numbers. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_ADD JAS_FIX_ADD_FAST #else #define JAS_FIX_ADD JAS_FIX_ADD_OFLOW #endif /* Calculate the sum of two fixed-point numbers without overflow checking. */ #define JAS_FIX_ADD_FAST(fix_t, fracbits, x, y) ((x) + (y)) /* Calculate the sum of two fixed-point numbers with overflow checking. */ #define JAS_FIX_ADD_OFLOW(fix_t, fracbits, x, y) \ ((x) >= 0) ? \ (((y) >= 0) ? ((x) + (y) >= 0 || JAS_FIX_OFLOW(), (x) + (y)) : \ ((x) + (y))) : \ (((y) >= 0) ? ((x) + (y)) : ((x) + (y) < 0 || JAS_FIX_OFLOW(), \ (x) + (y))) /* Calculate the product of two fixed-point numbers. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_MUL JAS_FIX_MUL_FAST #else #define JAS_FIX_MUL JAS_FIX_MUL_OFLOW #endif /* Calculate the product of two fixed-point numbers without overflow checking. */ #define JAS_FIX_MUL_FAST(fix_t, fracbits, bigfix_t, x, y) \ JAS_CAST(fix_t, (JAS_CAST(bigfix_t, x) * JAS_CAST(bigfix_t, y)) >> \ (fracbits)) /* Calculate the product of two fixed-point numbers with overflow checking. */ #define JAS_FIX_MUL_OFLOW(fix_t, fracbits, bigfix_t, x, y) \ ((JAS_CAST(bigfix_t, x) * JAS_CAST(bigfix_t, y) >> (fracbits)) == \ JAS_CAST(fix_t, (JAS_CAST(bigfix_t, x) * JAS_CAST(bigfix_t, y) >> \ (fracbits))) ? \ JAS_CAST(fix_t, (JAS_CAST(bigfix_t, x) * JAS_CAST(bigfix_t, y) >> \ (fracbits))) : JAS_FIX_OFLOW()) /* Calculate the product of a fixed-point number and an int. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_MULBYINT JAS_FIX_MULBYINT_FAST #else #define JAS_FIX_MULBYINT JAS_FIX_MULBYINT_OFLOW #endif /* Calculate the product of a fixed-point number and an int without overflow checking. */ #define JAS_FIX_MULBYINT_FAST(fix_t, fracbits, x, y) \ JAS_CAST(fix_t, ((x) * (y))) /* Calculate the product of a fixed-point number and an int with overflow checking. */ #define JAS_FIX_MULBYINT_OFLOW(fix_t, fracbits, x, y) \ JAS_FIX_MULBYINT_FAST(fix_t, fracbits, x, y) /* Calculate the quotient of two fixed-point numbers. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_DIV JAS_FIX_DIV_FAST #else #define JAS_FIX_DIV JAS_FIX_DIV_UFLOW #endif /* Calculate the quotient of two fixed-point numbers without underflow checking. */ #define JAS_FIX_DIV_FAST(fix_t, fracbits, bigfix_t, x, y) \ JAS_CAST(fix_t, (JAS_CAST(bigfix_t, x) << (fracbits)) / (y)) /* Calculate the quotient of two fixed-point numbers with underflow checking. */ #define JAS_FIX_DIV_UFLOW(fix_t, fracbits, bigfix_t, x, y) \ JAS_FIX_DIV_FAST(fix_t, fracbits, bigfix_t, x, y) /* Negate a fixed-point number. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_NEG JAS_FIX_NEG_FAST #else #define JAS_FIX_NEG JAS_FIX_NEG_OFLOW #endif /* Negate a fixed-point number without overflow checking. */ #define JAS_FIX_NEG_FAST(fix_t, fracbits, x) \ (-(x)) /* Negate a fixed-point number with overflow checking. */ /* Yes, overflow is actually possible for two's complement representations, although highly unlikely to occur. */ #define JAS_FIX_NEG_OFLOW(fix_t, fracbits, x) \ (((x) < 0) ? (-(x) > 0 || JAS_FIX_OFLOW(), -(x)) : (-(x))) /* Perform an arithmetic shift left of a fixed-point number. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_ASL JAS_FIX_ASL_FAST #else #define JAS_FIX_ASL JAS_FIX_ASL_OFLOW #endif /* Perform an arithmetic shift left of a fixed-point number without overflow checking. */ #define JAS_FIX_ASL_FAST(fix_t, fracbits, x, n) \ ((x) << (n)) /* Perform an arithmetic shift left of a fixed-point number with overflow checking. */ #define JAS_FIX_ASL_OFLOW(fix_t, fracbits, x, n) \ ((((x) << (n)) >> (n)) == (x) || JAS_FIX_OFLOW(), (x) << (n)) /* Perform an arithmetic shift right of a fixed-point number. */ #if !defined(DEBUG_OVERFLOW) #define JAS_FIX_ASR JAS_FIX_ASR_FAST #else #define JAS_FIX_ASR JAS_FIX_ASR_UFLOW #endif /* Perform an arithmetic shift right of a fixed-point number without underflow checking. */ #define JAS_FIX_ASR_FAST(fix_t, fracbits, x, n) \ ((x) >> (n)) /* Perform an arithmetic shift right of a fixed-point number with underflow checking. */ #define JAS_FIX_ASR_UFLOW(fix_t, fracbits, x, n) \ JAS_FIX_ASR_FAST(fix_t, fracbits, x, n) /******************************************************************************\ * Other basic arithmetic operations. \******************************************************************************/ /* Calculate the difference between two fixed-point numbers. */ #define JAS_FIX_SUB(fix_t, fracbits, x, y) \ JAS_FIX_ADD(fix_t, fracbits, x, JAS_FIX_NEG(fix_t, fracbits, y)) /* Add one fixed-point number to another. */ #define JAS_FIX_PLUSEQ(fix_t, fracbits, x, y) \ ((x) = JAS_FIX_ADD(fix_t, fracbits, x, y)) /* Subtract one fixed-point number from another. */ #define JAS_FIX_MINUSEQ(fix_t, fracbits, x, y) \ ((x) = JAS_FIX_SUB(fix_t, fracbits, x, y)) /* Multiply one fixed-point number by another. */ #define JAS_FIX_MULEQ(fix_t, fracbits, bigfix_t, x, y) \ ((x) = JAS_FIX_MUL(fix_t, fracbits, bigfix_t, x, y)) /******************************************************************************\ * Miscellaneous operations. \******************************************************************************/ /* Calculate the absolute value of a fixed-point number. */ #define JAS_FIX_ABS(fix_t, fracbits, x) \ (((x) >= 0) ? (x) : (JAS_FIX_NEG(fix_t, fracbits, x))) /* Is a fixed-point number an integer? */ #define JAS_FIX_ISINT(fix_t, fracbits, x) \ (JAS_FIX_FLOOR(fix_t, fracbits, x) == (x)) /* Get the sign of a fixed-point number. */ #define JAS_FIX_SGN(fix_t, fracbits, x) \ ((x) >= 0 ? 1 : (-1)) /******************************************************************************\ * Relational operations. \******************************************************************************/ /* Compare two fixed-point numbers. */ #define JAS_FIX_CMP(fix_t, fracbits, x, y) \ ((x) > (y) ? 1 : (((x) == (y)) ? 0 : (-1))) /* Less than. */ #define JAS_FIX_LT(fix_t, fracbits, x, y) \ ((x) < (y)) /* Less than or equal. */ #define JAS_FIX_LTE(fix_t, fracbits, x, y) \ ((x) <= (y)) /* Greater than. */ #define JAS_FIX_GT(fix_t, fracbits, x, y) \ ((x) > (y)) /* Greater than or equal. */ #define JAS_FIX_GTE(fix_t, fracbits, x, y) \ ((x) >= (y)) /******************************************************************************\ * Rounding functions. \******************************************************************************/ /* Round a fixed-point number to the nearest integer. */ #define JAS_FIX_ROUND(fix_t, fracbits, x) \ (((x) < 0) ? JAS_FIX_FLOOR(fix_t, fracbits, JAS_FIX_ADD(fix_t, fracbits, \ (x), JAS_FIX_HALF(fix_t, fracbits))) : \ JAS_FIX_NEG(fix_t, fracbits, JAS_FIX_FLOOR(fix_t, fracbits, \ JAS_FIX_ADD(fix_t, fracbits, (-(x)), JAS_FIX_HALF(fix_t, fracbits))))) /* Round a fixed-point number to the nearest integer in the direction of negative infinity (i.e., the floor function). */ #define JAS_FIX_FLOOR(fix_t, fracbits, x) \ ((x) & (~((JAS_CAST(fix_t, 1) << (fracbits)) - 1))) /* Round a fixed-point number to the nearest integer in the direction of zero. */ #define JAS_FIX_TRUNC(fix_t, fracbits, x) \ (((x) >= 0) ? JAS_FIX_FLOOR(fix_t, fracbits, x) : \ JAS_FIX_CEIL(fix_t, fracbits, x)) /******************************************************************************\ * The below macros are for internal library use only. Do not invoke them * directly in application code. \******************************************************************************/ /* Handle overflow. */ #define JAS_FIX_OFLOW() \ jas_eprintf("overflow error: file %s, line %d\n", __FILE__, __LINE__) /* Handle underflow. */ #define JAS_FIX_UFLOW() \ jas_eprintf("underflow error: file %s, line %d\n", __FILE__, __LINE__) #ifdef __cplusplus } #endif #endif
© 2024 UnknownSec