diff --git a/catcoon.png b/catcoon.png new file mode 100644 index 0000000..2c81022 Binary files /dev/null and b/catcoon.png differ diff --git a/demo/lenet.c b/demo/lenet.c new file mode 100644 index 0000000..cd86bc3 --- /dev/null +++ b/demo/lenet.c @@ -0,0 +1,136 @@ +#include +#include + +#include "catcoon.h" + +#define INPUT_W 28 +#define INPUT_H 28 + +const char *parameters_files[]={ + "conv1_w.bin", + "conv1_b.bin", + "conv2_w.bin", + "conv2_b.bin", + "fc1_w.bin", + "fc1_b.bin", + "fc2_w.bin", + "fc2_b.bin", +}; +const char *parameters_path = "/home/evi/catcoon-pytorch-model/lenet/dump"; + +int main(int argc, const char *argv[]) +{ + cc_int32 i, j; + cc_float32 v; + char filepath[256]; + /* Image */ + utim_image_t *img, *img_read; + /* Layers */ + cc_tensor_t *input,*l1, *l1_pool, *l2, *l2_pool, *l2_flat, *l3, *l4; + /* Parameters */ + cc_tensor_t *conv1_w, *conv1_b, + *conv2_w, *conv2_b, *fc1_w, *fc1_b, *fc2_w, *fc2_b; + cc_int32 shape_conv1_w[] = {5, 5, 1, 32, 0}; + cc_int32 shape_conv1_b[] = {32, 0}; + cc_int32 shape_conv2_w[] = {5, 5, 32, 64, 0}; + cc_int32 shape_conv2_b[] = {64, 0}; + cc_int32 shape_flat[] = {-1, 1, 1, 0}; + cc_int32 shape_fc1_w[] = {1, 1, 3136, 128, 0}; + cc_int32 shape_fc1_b[] = {128, 0}; + cc_int32 shape_fc2_w[] = {1, 1, 128, 10, 0}; + cc_int32 shape_fc2_b[] = {10, 0}; + + if (argc < 2) { + fprintf(stderr, "usage: lenet [filename]\n"); + exit(255); + } + + img_read = utim_read(argv[1]); + utim_img2gray(img_read); + img = utim_resize(img_read, INPUT_H, INPUT_W, 0); + input = cc_image2tensor(img, "input"); + utim_free_image(img); + utim_free_image(img_read); + input = cc_cast_tensor(input, CC_FLOAT32, "input"); + v = 255.; + input = cc_tensor_by_scalar(input, '/', &v, "input"); + + i = 0; + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + conv1_w = cc_load_bin(filepath, + shape_conv1_w, CC_FLOAT32, "conv1_w"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + conv1_b = cc_load_bin(filepath, + shape_conv1_b, CC_FLOAT32, "conv1_b"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + conv2_w = cc_load_bin(filepath, + shape_conv2_w, CC_FLOAT32, "conv2_w"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + conv2_b = cc_load_bin(filepath, + shape_conv2_b, CC_FLOAT32, "conv2_b"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + fc1_w = cc_load_bin(filepath, + shape_fc1_w, CC_FLOAT32, "fc1_w"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + fc1_b = cc_load_bin(filepath, + shape_fc1_b, CC_FLOAT32, "fc1_b"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + fc2_w = cc_load_bin(filepath, + shape_fc2_w, CC_FLOAT32, "fc2_w"); + + sprintf(filepath, "%s/%s", + parameters_path, parameters_files[i++]); + fc2_b = cc_load_bin(filepath, + shape_fc2_b, CC_FLOAT32, "fc2_b"); + + l1 = cc_conv2d(input, conv1_w, conv1_b, 1, 2, 0, "l1_conv"); + + l1 = cc_relu(l1, NULL); + + l1_pool = cc_max_pool2d(l1, 2, "l1_pool"); + + l2 = cc_conv2d(l1_pool, conv2_w, conv2_b, 1, 2, 0, "l2_conv"); + + l2 = cc_relu(l2, NULL); + + l2_pool = cc_max_pool2d(l2, 2, "l2_pool"); + + /* l2_flat = cc_fmap2d_flat(l2_pool, "l2_flat"); */ + l2_flat = cc_tensor_reshape(l2_pool, shape_flat); + + /* l3 = cc_conv2d(l2_flat, fc1_w, fc1_b, 1, 0, 0, "fc1"); */ + l3 = cc_fully_connected(l2_flat, fc1_w, fc1_b, "fc1"); + + l3 = cc_relu(l3, NULL); + + /* l4 = cc_conv2d(l3, fc2_w, fc2_b, 1, 0, 0, "fc2"); */ + l4 = cc_fully_connected(l3, fc2_w, fc2_b, "fc2"); + + l4 = cc_softmax(l4, NULL); + + v = 0.; + for (i = 0, j = 0; i < 10; ++i) { + if (*((cc_float32*)l4->data + i) > v) { + v = *((cc_float32*)l4->data + i); + j = i; + } + printf("[%d]: %f\n", i, *((cc_float32*)l4->data + i)); + } + printf("Result of \"%s\": [%d]\n", argv[1], j); + cc_tsrmgr_list(); + cc_clear(); + return 0; +} diff --git a/demo/simple.c b/demo/simple.c new file mode 100644 index 0000000..4816db1 --- /dev/null +++ b/demo/simple.c @@ -0,0 +1,11 @@ +#include "catcoon.h" + +int main(int argc, const char *argv[]) +{ + cc_tensor_t *tensor; + cc_int32 shape[] = {3, 3, 3, 0}; + tensor = cc_create_tensor(shape, CC_FLOAT32, "tensor0"); + cc_print_tensor_property(tensor); + cc_free_tensor(tensor); + return 0; +} diff --git a/makefile b/makefile new file mode 100644 index 0000000..f65047c --- /dev/null +++ b/makefile @@ -0,0 +1,89 @@ +cc = gcc + +# Debug flag, cc_assert +DFLAG += -DENABLE_CC_ASSERT +# AddressSanitizer for gcc/clang +DFLAG += # -g -fsanitize=address -fno-omit-frame-pointer + +CFLAG += # -std=c89 +CFLAG += -Wall # -Wpedantic + +OFLAG += -Os + +# Enable OpenMP +OFLAG += -DENABLE_OPENMP -fopenmp + +CFLAG += $(DFLAG) $(WFLAG) $(OFLAG) + +# Enable automatic tensor manager +CFLAG += -DAUTO_TSRMGR + +# 3rd party source/library configurations +# stb for jpg, png, tga format images, +# will only support bmp image if disabled +3RDSRC_CF += stb + +ifneq ($(findstring -std=c89, $(CFLAG)),) + CFLAG += -DCONFIG_STD_C89 + ifneq ($(findstring $(cc), gcc clang),) + WFLAG += -Wno-long-long + endif +endif + +LINK += -lm +INC += -I ./src/ +ALL_O = catcoon.o cc_tensor.o cc_dtype.o cc_image.o cc_array.o cc_tsrmgr.o \ + cc_basic.o cc_fmap2d.o cc_pad2d.o cc_conv2d.o cc_pool2d.o cc_cpufn.o \ + cc_fullycon.o global_fn_cfg.o util_rbt.o util_list.o util_log.o \ + util_image.o + +APP_NAMES = simple lenet + +ifeq ($(OS),Windows_NT) + RM = del + APPS = $(foreach v, $(APP_NAMES), $(v).exe) +else + RM = rm + APPS = $(APP_NAMES) +endif + +all: $(APPS) + +%.o: ./src/%.c + $(cc) -c -o $@ $< $(CFLAG) $(INC) +# APPS For Linux +%: ./demo/%.c $(ALL_O) + $(cc) -o $@ $< $(ALL_O) $(CFLAG) $(INC) $(LINK) +# APPS For Windows +%.exe: ./demo/%.c $(ALL_O) + $(cc) -o $@ $< $(ALL_O) $(CFLAG) $(INC) $(LINK) + +global_fn_cfg.o : $(patsubst %, ./src/%, global_fn_cfg.h global_fn_cfg.c) + +catcoon.o : $(patsubst %, ./src/%, catcoon.h catcoon.c) +cc_array.o : $(patsubst %, ./src/%, cc_array.h cc_array.c) +cc_basic.o : $(patsubst %, ./src/%, cc_basic.h cc_basic.c) +cc_conv2d.o : $(patsubst %, ./src/%, cc_conv2d.h cc_conv2d.c) +cc_cpufn.o : $(patsubst %, ./src/%, cc_cpufn.h cc_cpufn.c) +cc_dtype.o : $(patsubst %, ./src/%, cc_dtype.h cc_dtype.c) +cc_fmap2d.o : $(patsubst %, ./src/%, cc_fmap2d.h cc_fmap2d.c) +cc_fullycon.o : $(patsubst %, ./src/%, cc_fullycon.h cc_fullycon.c) +cc_image.o : $(patsubst %, ./src/%, cc_image.h cc_image.c) +cc_pad2d.o : $(patsubst %, ./src/%, cc_pad2d.h cc_pad2d.c) +cc_pool2d.o : $(patsubst %, ./src/%, cc_pool2d.h cc_pool2d.c) +cc_tsrmgr.o : $(patsubst %, ./src/%, cc_tsrmgr.h cc_tsrmgr.c) +cc_tensor.o : $(patsubst %, ./src/%, cc_tensor.h cc_tensor.c) + +util_log.o : $(patsubst %, ./src/%, util_log.h util_log.c) +util_rbt.o : $(patsubst %, ./src/%, util_rbt.h util_rbt.c) +util_list.o : ./src/util_list.h ./src/util_list.c + $(cc) -c -o $@ ./src/util_list.c $(CFLAG) -DENABLE_FOPS +UTIM_COND = +ifneq ($(findstring stb, $(3RDSRC_CF)),) + UTIM_COND += -DUSE3RD_STB_IMAGE -I ./src/3rd_party/stb/ +endif +util_image.o : ./src/util_image.h ./src/util_image.c + $(cc) -c -o $@ ./src/util_image.c $(CFLAG) $(UTIM_COND) + +clean: + $(RM) *.o && $(RM) $(APPS) \ No newline at end of file diff --git a/src/3rd_party/stb/stb_image.h b/src/3rd_party/stb/stb_image.h new file mode 100644 index 0000000..1499f6f --- /dev/null +++ b/src/3rd_party/stb/stb_image.h @@ -0,0 +1,6077 @@ +/* View upstream repo: https://github.com/nothings/stb +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person 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, sublicense, 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: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ + +#ifndef STBI_INCLUDE_STB_IMAGE_H +#define STBI_INCLUDE_STB_IMAGE_H +#ifndef STBI_NO_STDIO +#include +#endif +#define STBI_VERSION 1 +enum +{ + STBI_default = 0, + STBI_grey = 1, + STBI_grey_alpha = 2, + STBI_rgb = 3, + STBI_rgb_alpha = 4 +}; +#include +typedef unsigned char stbi_uc; +typedef unsigned short stbi_us; +#ifdef __cplusplus +extern "C" { +#endif +#ifndef STBIDEF +#ifdef STB_IMAGE_STATIC +#define STBIDEF static +#else +#define STBIDEF extern +#endif +#endif +typedef struct +{ + int (*read) (void *user,char *data,int size); + void (*skip) (void *user,int n); + int (*eof) (void *user); +} stbi_io_callbacks; +STBIDEF stbi_uc *stbi_load_from_memory (stbi_uc const *buffer, int len , int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk , void *user, int *x, int *y, int *channels_in_file, int desired_channels); +#ifndef STBI_NO_STDIO +STBIDEF stbi_uc *stbi_load (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_uc *stbi_load_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +#endif +#ifdef STBI_WINDOWS_UTF8 +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input); +#endif +STBIDEF stbi_us *stbi_load_16_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); +#ifndef STBI_NO_STDIO +STBIDEF stbi_us *stbi_load_16 (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); +STBIDEF stbi_us *stbi_load_from_file_16(FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); +#endif +#ifndef STBI_NO_LINEAR + STBIDEF float *stbi_loadf_from_memory (stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_callbacks (stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels); + #ifndef STBI_NO_STDIO + STBIDEF float *stbi_loadf (char const *filename, int *x, int *y, int *channels_in_file, int desired_channels); + STBIDEF float *stbi_loadf_from_file (FILE *f, int *x, int *y, int *channels_in_file, int desired_channels); + #endif +#endif +#ifndef STBI_NO_HDR + STBIDEF void stbi_hdr_to_ldr_gamma(float gamma); + STBIDEF void stbi_hdr_to_ldr_scale(float scale); +#endif +#ifndef STBI_NO_LINEAR + STBIDEF void stbi_ldr_to_hdr_gamma(float gamma); + STBIDEF void stbi_ldr_to_hdr_scale(float scale); +#endif +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user); +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len); +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename); +STBIDEF int stbi_is_hdr_from_file(FILE *f); +#endif +STBIDEF const char *stbi_failure_reason (void); +STBIDEF void stbi_image_free (void *retval_from_stbi_load); +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len); +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *clbk, void *user); +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info (char const *filename, int *x, int *y, int *comp); +STBIDEF int stbi_info_from_file (FILE *f, int *x, int *y, int *comp); +STBIDEF int stbi_is_16_bit (char const *filename); +STBIDEF int stbi_is_16_bit_from_file(FILE *f); +#endif +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply); +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert); +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip); +STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip); +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen); +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header); +STBIDEF char *stbi_zlib_decode_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); +STBIDEF char *stbi_zlib_decode_noheader_malloc(const char *buffer, int len, int *outlen); +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen); +#ifdef __cplusplus +} +#endif +#endif +#ifdef STB_IMAGE_IMPLEMENTATION +#if defined(STBI_ONLY_JPEG) || defined(STBI_ONLY_PNG) || defined(STBI_ONLY_BMP) \ + || defined(STBI_ONLY_TGA) || defined(STBI_ONLY_GIF) || defined(STBI_ONLY_PSD) \ + || defined(STBI_ONLY_HDR) || defined(STBI_ONLY_PIC) || defined(STBI_ONLY_PNM) \ + || defined(STBI_ONLY_ZLIB) + #ifndef STBI_ONLY_JPEG + #define STBI_NO_JPEG + #endif + #ifndef STBI_ONLY_PNG + #define STBI_NO_PNG + #endif + #ifndef STBI_ONLY_BMP + #define STBI_NO_BMP + #endif + #ifndef STBI_ONLY_PSD + #define STBI_NO_PSD + #endif + #ifndef STBI_ONLY_TGA + #define STBI_NO_TGA + #endif + #ifndef STBI_ONLY_GIF + #define STBI_NO_GIF + #endif + #ifndef STBI_ONLY_HDR + #define STBI_NO_HDR + #endif + #ifndef STBI_ONLY_PIC + #define STBI_NO_PIC + #endif + #ifndef STBI_ONLY_PNM + #define STBI_NO_PNM + #endif +#endif +#if defined(STBI_NO_PNG) && !defined(STBI_SUPPORT_ZLIB) && !defined(STBI_NO_ZLIB) +#define STBI_NO_ZLIB +#endif +#include +#include +#include +#include +#include +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +#include +#endif +#ifndef STBI_NO_STDIO +#include +#endif +#ifndef STBI_ASSERT +#include +#define STBI_ASSERT(x) assert(x) +#endif +#ifdef __cplusplus +#define STBI_EXTERN extern "C" +#else +#define STBI_EXTERN extern +#endif +#ifndef _MSC_VER + #ifdef __cplusplus + #define stbi_inline inline + #else + #define stbi_inline + #endif +#else + #define stbi_inline __forceinline +#endif +#ifndef STBI_NO_THREAD_LOCALS + #if defined(__cplusplus) && __cplusplus >= 201103L + #define STBI_THREAD_LOCAL thread_local + #elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 201112L + #define STBI_THREAD_LOCAL _Thread_local + #elif defined(__GNUC__) + #define STBI_THREAD_LOCAL __thread + #elif defined(_MSC_VER) + #define STBI_THREAD_LOCAL __declspec(thread) +#endif +#endif +#ifdef _MSC_VER +typedef unsigned short stbi__uint16; +typedef signed short stbi__int16; +typedef unsigned int stbi__uint32; +typedef signed int stbi__int32; +#else +#include +typedef uint16_t stbi__uint16; +typedef int16_t stbi__int16; +typedef uint32_t stbi__uint32; +typedef int32_t stbi__int32; +#endif +typedef unsigned char validate_uint32[sizeof(stbi__uint32)==4 ? 1 : -1]; +#ifdef _MSC_VER +#define STBI_NOTUSED(v) (void)(v) +#else +#define STBI_NOTUSED(v) (void)sizeof(v) +#endif +#ifdef _MSC_VER +#define STBI_HAS_LROTL +#endif +#ifdef STBI_HAS_LROTL + #define stbi_lrot(x,y) _lrotl(x,y) +#else + #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (32 - (y)))) +#endif +#if defined(STBI_MALLOC) && defined(STBI_FREE) && (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED)) +#elif !defined(STBI_MALLOC) && !defined(STBI_FREE) && !defined(STBI_REALLOC) && !defined(STBI_REALLOC_SIZED) +#else +#error "Must define all or none of STBI_MALLOC, STBI_FREE, and STBI_REALLOC (or STBI_REALLOC_SIZED)." +#endif +#ifndef STBI_MALLOC +#define STBI_MALLOC(sz) malloc(sz) +#define STBI_REALLOC(p,newsz) realloc(p,newsz) +#define STBI_FREE(p) free(p) +#endif +#ifndef STBI_REALLOC_SIZED +#define STBI_REALLOC_SIZED(p,oldsz,newsz) STBI_REALLOC(p,newsz) +#endif +#if defined(__x86_64__) || defined(_M_X64) +#define STBI__X64_TARGET +#elif defined(__i386) || defined(_M_IX86) +#define STBI__X86_TARGET +#endif +#if defined(__GNUC__) && defined(STBI__X86_TARGET) && !defined(__SSE2__) && !defined(STBI_NO_SIMD) +#define STBI_NO_SIMD +#endif +#if defined(__MINGW32__) && defined(STBI__X86_TARGET) && !defined(STBI_MINGW_ENABLE_SSE2) && !defined(STBI_NO_SIMD) +#define STBI_NO_SIMD +#endif +#if !defined(STBI_NO_SIMD) && (defined(STBI__X86_TARGET) || defined(STBI__X64_TARGET)) +#define STBI_SSE2 +#include +#ifdef _MSC_VER +#if _MSC_VER >= 1400 +#include +static int stbi__cpuid3(void) +{ + int info[4]; + __cpuid(info,1); + return info[3]; +} +#else +static int stbi__cpuid3(void) +{ + int res; + __asm { + mov eax,1 + cpuid + mov res,edx + } + return res; +} +#endif +#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + int info3 = stbi__cpuid3(); + return ((info3 >> 26) & 1) != 0; +} +#endif +#else +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) +#if !defined(STBI_NO_JPEG) && defined(STBI_SSE2) +static int stbi__sse2_available(void) +{ + + + + return 1; +} +#endif +#endif +#endif +#if defined(STBI_NO_SIMD) && defined(STBI_NEON) +#undef STBI_NEON +#endif +#ifdef STBI_NEON +#include +#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16))) +#endif +#ifndef STBI_SIMD_ALIGN +#define STBI_SIMD_ALIGN(type, name) type name +#endif +typedef struct +{ + stbi__uint32 img_x, img_y; + int img_n, img_out_n; + stbi_io_callbacks io; + void *io_user_data; + int read_from_callbacks; + int buflen; + stbi_uc buffer_start[128]; + stbi_uc *img_buffer, *img_buffer_end; + stbi_uc *img_buffer_original, *img_buffer_original_end; +} stbi__context; +static void stbi__refill_buffer(stbi__context *s); +static void stbi__start_mem(stbi__context *s, stbi_uc const *buffer, int len) +{ + s->io.read = NULL; + s->read_from_callbacks = 0; + s->img_buffer = s->img_buffer_original = (stbi_uc *) buffer; + s->img_buffer_end = s->img_buffer_original_end = (stbi_uc *) buffer+len; +} +static void stbi__start_callbacks(stbi__context *s, stbi_io_callbacks *c, void *user) +{ + s->io = *c; + s->io_user_data = user; + s->buflen = sizeof(s->buffer_start); + s->read_from_callbacks = 1; + s->img_buffer_original = s->buffer_start; + stbi__refill_buffer(s); + s->img_buffer_original_end = s->img_buffer_end; +} +#ifndef STBI_NO_STDIO +static int stbi__stdio_read(void *user, char *data, int size) +{ + return (int) fread(data,1,size,(FILE*) user); +} +static void stbi__stdio_skip(void *user, int n) +{ + fseek((FILE*) user, n, SEEK_CUR); +} +static int stbi__stdio_eof(void *user) +{ + return feof((FILE*) user); +} +static stbi_io_callbacks stbi__stdio_callbacks = +{ + stbi__stdio_read, + stbi__stdio_skip, + stbi__stdio_eof, +}; +static void stbi__start_file(stbi__context *s, FILE *f) +{ + stbi__start_callbacks(s, &stbi__stdio_callbacks, (void *) f); +} +#endif +static void stbi__rewind(stbi__context *s) +{ + + + + s->img_buffer = s->img_buffer_original; + s->img_buffer_end = s->img_buffer_original_end; +} +enum +{ + STBI_ORDER_RGB, + STBI_ORDER_BGR +}; +typedef struct +{ + int bits_per_channel; + int num_channels; + int channel_order; +} stbi__result_info; +#ifndef STBI_NO_JPEG +static int stbi__jpeg_test(stbi__context *s); +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_PNG +static int stbi__png_test(stbi__context *s); +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__png_is16(stbi__context *s); +#endif +#ifndef STBI_NO_BMP +static int stbi__bmp_test(stbi__context *s); +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_TGA +static int stbi__tga_test(stbi__context *s); +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s); +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc); +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp); +static int stbi__psd_is16(stbi__context *s); +#endif +#ifndef STBI_NO_HDR +static int stbi__hdr_test(stbi__context *s); +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_PIC +static int stbi__pic_test(stbi__context *s); +static void *stbi__pic_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_GIF +static int stbi__gif_test(stbi__context *s); +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp); +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp); +#endif +#ifndef STBI_NO_PNM +static int stbi__pnm_test(stbi__context *s); +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri); +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp); +#endif +static +#ifdef STBI_THREAD_LOCAL +STBI_THREAD_LOCAL +#endif +const char *stbi__g_failure_reason; +STBIDEF const char *stbi_failure_reason(void) +{ + return stbi__g_failure_reason; +} +#ifndef STBI_NO_FAILURE_STRINGS +static int stbi__err(const char *str) +{ + stbi__g_failure_reason = str; + return 0; +} +#endif +static void *stbi__malloc(size_t size) +{ + return STBI_MALLOC(size); +} +static int stbi__addsizes_valid(int a, int b) +{ + if (b < 0) return 0; + + + + + return a <= INT_MAX - b; +} +static int stbi__mul2sizes_valid(int a, int b) +{ + if (a < 0 || b < 0) return 0; + if (b == 0) return 1; + + return a <= INT_MAX/b; +} +#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR) +static int stbi__mad2sizes_valid(int a, int b, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a*b, add); +} +#endif +static int stbi__mad3sizes_valid(int a, int b, int c, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__addsizes_valid(a*b*c, add); +} +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static int stbi__mad4sizes_valid(int a, int b, int c, int d, int add) +{ + return stbi__mul2sizes_valid(a, b) && stbi__mul2sizes_valid(a*b, c) && + stbi__mul2sizes_valid(a*b*c, d) && stbi__addsizes_valid(a*b*c*d, add); +} +#endif +#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR) +static void *stbi__malloc_mad2(int a, int b, int add) +{ + if (!stbi__mad2sizes_valid(a, b, add)) return NULL; + return stbi__malloc(a*b + add); +} +#endif +static void *stbi__malloc_mad3(int a, int b, int c, int add) +{ + if (!stbi__mad3sizes_valid(a, b, c, add)) return NULL; + return stbi__malloc(a*b*c + add); +} +#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) +static void *stbi__malloc_mad4(int a, int b, int c, int d, int add) +{ + if (!stbi__mad4sizes_valid(a, b, c, d, add)) return NULL; + return stbi__malloc(a*b*c*d + add); +} +#endif +#ifdef STBI_NO_FAILURE_STRINGS + #define stbi__err(x,y) 0 +#elif defined(STBI_FAILURE_USERMSG) + #define stbi__err(x,y) stbi__err(y) +#else + #define stbi__err(x,y) stbi__err(x) +#endif +#define stbi__errpf(x,y) ((float *)(size_t) (stbi__err(x,y)?NULL:NULL)) +#define stbi__errpuc(x,y) ((unsigned char *)(size_t) (stbi__err(x,y)?NULL:NULL)) +STBIDEF void stbi_image_free(void *retval_from_stbi_load) +{ + STBI_FREE(retval_from_stbi_load); +} +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp); +#endif +#ifndef STBI_NO_HDR +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp); +#endif +static int stbi__vertically_flip_on_load_global = 0; +STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load_global = flag_true_if_should_flip; +} +#ifndef STBI_THREAD_LOCAL +#define stbi__vertically_flip_on_load stbi__vertically_flip_on_load_global +#else +static STBI_THREAD_LOCAL int stbi__vertically_flip_on_load_local, stbi__vertically_flip_on_load_set; +STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip) +{ + stbi__vertically_flip_on_load_local = flag_true_if_should_flip; + stbi__vertically_flip_on_load_set = 1; +} +#define stbi__vertically_flip_on_load (stbi__vertically_flip_on_load_set \ + ? stbi__vertically_flip_on_load_local \ + : stbi__vertically_flip_on_load_global) +#endif +static void *stbi__load_main(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + memset(ri, 0, sizeof(*ri)); + ri->bits_per_channel = 8; + ri->channel_order = STBI_ORDER_RGB; + ri->num_channels = 0; + #ifndef STBI_NO_JPEG + if (stbi__jpeg_test(s)) return stbi__jpeg_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNG + if (stbi__png_test(s)) return stbi__png_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_BMP + if (stbi__bmp_test(s)) return stbi__bmp_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_GIF + if (stbi__gif_test(s)) return stbi__gif_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_test(s)) return stbi__psd_load(s,x,y,comp,req_comp, ri, bpc); + #else + STBI_NOTUSED(bpc); + #endif + #ifndef STBI_NO_PIC + if (stbi__pic_test(s)) return stbi__pic_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_PNM + if (stbi__pnm_test(s)) return stbi__pnm_load(s,x,y,comp,req_comp, ri); + #endif + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + float *hdr = stbi__hdr_load(s, x,y,comp,req_comp, ri); + return stbi__hdr_to_ldr(hdr, *x, *y, req_comp ? req_comp : *comp); + } + #endif + #ifndef STBI_NO_TGA + + if (stbi__tga_test(s)) + return stbi__tga_load(s,x,y,comp,req_comp, ri); + #endif + return stbi__errpuc("unknown image type", "Image not of any known type, or corrupt"); +} +static stbi_uc *stbi__convert_16_to_8(stbi__uint16 *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi_uc *reduced; + reduced = (stbi_uc *) stbi__malloc(img_len); + if (reduced == NULL) return stbi__errpuc("outofmem", "Out of memory"); + for (i = 0; i < img_len; ++i) + reduced[i] = (stbi_uc)((orig[i] >> 8) & 0xFF); + STBI_FREE(orig); + return reduced; +} +static stbi__uint16 *stbi__convert_8_to_16(stbi_uc *orig, int w, int h, int channels) +{ + int i; + int img_len = w * h * channels; + stbi__uint16 *enlarged; + enlarged = (stbi__uint16 *) stbi__malloc(img_len*2); + if (enlarged == NULL) return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + for (i = 0; i < img_len; ++i) + enlarged[i] = (stbi__uint16)((orig[i] << 8) + orig[i]); + STBI_FREE(orig); + return enlarged; +} +static void stbi__vertical_flip(void *image, int w, int h, int bytes_per_pixel) +{ + int row; + size_t bytes_per_row = (size_t)w * bytes_per_pixel; + stbi_uc temp[2048]; + stbi_uc *bytes = (stbi_uc *)image; + for (row = 0; row < (h>>1); row++) { + stbi_uc *row0 = bytes + row*bytes_per_row; + stbi_uc *row1 = bytes + (h - row - 1)*bytes_per_row; + + size_t bytes_left = bytes_per_row; + while (bytes_left) { + size_t bytes_copy = (bytes_left < sizeof(temp)) ? bytes_left : sizeof(temp); + memcpy(temp, row0, bytes_copy); + memcpy(row0, row1, bytes_copy); + memcpy(row1, temp, bytes_copy); + row0 += bytes_copy; + row1 += bytes_copy; + bytes_left -= bytes_copy; + } + } +} +#ifndef STBI_NO_GIF +static void stbi__vertical_flip_slices(void *image, int w, int h, int z, int bytes_per_pixel) +{ + int slice; + int slice_size = w * h * bytes_per_pixel; + stbi_uc *bytes = (stbi_uc *)image; + for (slice = 0; slice < z; ++slice) { + stbi__vertical_flip(bytes, w, h, bytes_per_pixel); + bytes += slice_size; + } +} +#endif +static unsigned char *stbi__load_and_postprocess_8bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 8); + if (result == NULL) + return NULL; + if (ri.bits_per_channel != 8) { + STBI_ASSERT(ri.bits_per_channel == 16); + result = stbi__convert_16_to_8((stbi__uint16 *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 8; + } + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi_uc)); + } + return (unsigned char *) result; +} +static stbi__uint16 *stbi__load_and_postprocess_16bit(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + stbi__result_info ri; + void *result = stbi__load_main(s, x, y, comp, req_comp, &ri, 16); + if (result == NULL) + return NULL; + if (ri.bits_per_channel != 16) { + STBI_ASSERT(ri.bits_per_channel == 8); + result = stbi__convert_8_to_16((stbi_uc *) result, *x, *y, req_comp == 0 ? *comp : req_comp); + ri.bits_per_channel = 16; + } + + + if (stbi__vertically_flip_on_load) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(stbi__uint16)); + } + return (stbi__uint16 *) result; +} +#if !defined(STBI_NO_HDR) && !defined(STBI_NO_LINEAR) +static void stbi__float_postprocess(float *result, int *x, int *y, int *comp, int req_comp) +{ + if (stbi__vertically_flip_on_load && result != NULL) { + int channels = req_comp ? req_comp : *comp; + stbi__vertical_flip(result, *x, *y, channels * sizeof(float)); + } +} +#endif +#ifndef STBI_NO_STDIO +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +STBI_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, unsigned long flags, const char *str, int cbmb, wchar_t *widestr, int cchwide); +STBI_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, unsigned long flags, const wchar_t *widestr, int cchwide, char *str, int cbmb, const char *defchar, int *used_default); +#endif +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input) +{ + return WideCharToMultiByte(65001 , 0, input, -1, buffer, (int) bufferlen, NULL, NULL); +} +#endif +static FILE *stbi__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) + wchar_t wMode[64]; + wchar_t wFilename[1024]; + if (0 == MultiByteToWideChar(65001 , 0, filename, -1, wFilename, sizeof(wFilename))) + return 0; + if (0 == MultiByteToWideChar(65001 , 0, mode, -1, wMode, sizeof(wMode))) + return 0; +#if _MSC_VER >= 1400 + if (0 != _wfopen_s(&f, wFilename, wMode)) + f = 0; +#else + f = _wfopen(wFilename, wMode); +#endif +#elif defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} +STBIDEF stbi_uc *stbi_load(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + unsigned char *result; + if (!f) return stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} +STBIDEF stbi_uc *stbi_load_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); + if (result) { + + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} +STBIDEF stbi__uint16 *stbi_load_from_file_16(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__uint16 *result; + stbi__context s; + stbi__start_file(&s,f); + result = stbi__load_and_postprocess_16bit(&s,x,y,comp,req_comp); + if (result) { + + fseek(f, - (int) (s.img_buffer_end - s.img_buffer), SEEK_CUR); + } + return result; +} +STBIDEF stbi_us *stbi_load_16(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + stbi__uint16 *result; + if (!f) return (stbi_us *) stbi__errpuc("can't fopen", "Unable to open file"); + result = stbi_load_from_file_16(f,x,y,comp,req_comp); + fclose(f); + return result; +} +#endif +STBIDEF stbi_us *stbi_load_16_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} +STBIDEF stbi_us *stbi_load_16_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *channels_in_file, int desired_channels) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *)clbk, user); + return stbi__load_and_postprocess_16bit(&s,x,y,channels_in_file,desired_channels); +} +STBIDEF stbi_uc *stbi_load_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} +STBIDEF stbi_uc *stbi_load_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__load_and_postprocess_8bit(&s,x,y,comp,req_comp); +} +#ifndef STBI_NO_GIF +STBIDEF stbi_uc *stbi_load_gif_from_memory(stbi_uc const *buffer, int len, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + unsigned char *result; + stbi__context s; + stbi__start_mem(&s,buffer,len); + result = (unsigned char*) stbi__load_gif_main(&s, delays, x, y, z, comp, req_comp); + if (stbi__vertically_flip_on_load) { + stbi__vertical_flip_slices( result, *x, *y, *z, *comp ); + } + return result; +} +#endif +#ifndef STBI_NO_LINEAR +static float *stbi__loadf_main(stbi__context *s, int *x, int *y, int *comp, int req_comp) +{ + unsigned char *data; + #ifndef STBI_NO_HDR + if (stbi__hdr_test(s)) { + stbi__result_info ri; + float *hdr_data = stbi__hdr_load(s,x,y,comp,req_comp, &ri); + if (hdr_data) + stbi__float_postprocess(hdr_data,x,y,comp,req_comp); + return hdr_data; + } + #endif + data = stbi__load_and_postprocess_8bit(s, x, y, comp, req_comp); + if (data) + return stbi__ldr_to_hdr(data, *x, *y, req_comp ? req_comp : *comp); + return stbi__errpf("unknown image type", "Image not of any known type, or corrupt"); +} +STBIDEF float *stbi_loadf_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +STBIDEF float *stbi_loadf_from_callbacks(stbi_io_callbacks const *clbk, void *user, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +#ifndef STBI_NO_STDIO +STBIDEF float *stbi_loadf(char const *filename, int *x, int *y, int *comp, int req_comp) +{ + float *result; + FILE *f = stbi__fopen(filename, "rb"); + if (!f) return stbi__errpf("can't fopen", "Unable to open file"); + result = stbi_loadf_from_file(f,x,y,comp,req_comp); + fclose(f); + return result; +} +STBIDEF float *stbi_loadf_from_file(FILE *f, int *x, int *y, int *comp, int req_comp) +{ + stbi__context s; + stbi__start_file(&s,f); + return stbi__loadf_main(&s,x,y,comp,req_comp); +} +#endif +#endif +STBIDEF int stbi_is_hdr_from_memory(stbi_uc const *buffer, int len) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(buffer); + STBI_NOTUSED(len); + return 0; + #endif +} +#ifndef STBI_NO_STDIO +STBIDEF int stbi_is_hdr (char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result=0; + if (f) { + result = stbi_is_hdr_from_file(f); + fclose(f); + } + return result; +} +STBIDEF int stbi_is_hdr_from_file(FILE *f) +{ + #ifndef STBI_NO_HDR + long pos = ftell(f); + int res; + stbi__context s; + stbi__start_file(&s,f); + res = stbi__hdr_test(&s); + fseek(f, pos, SEEK_SET); + return res; + #else + STBI_NOTUSED(f); + return 0; + #endif +} +#endif +STBIDEF int stbi_is_hdr_from_callbacks(stbi_io_callbacks const *clbk, void *user) +{ + #ifndef STBI_NO_HDR + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) clbk, user); + return stbi__hdr_test(&s); + #else + STBI_NOTUSED(clbk); + STBI_NOTUSED(user); + return 0; + #endif +} +#ifndef STBI_NO_LINEAR +static float stbi__l2h_gamma=2.2f, stbi__l2h_scale=1.0f; +STBIDEF void stbi_ldr_to_hdr_gamma(float gamma) { stbi__l2h_gamma = gamma; } +STBIDEF void stbi_ldr_to_hdr_scale(float scale) { stbi__l2h_scale = scale; } +#endif +static float stbi__h2l_gamma_i=1.0f/2.2f, stbi__h2l_scale_i=1.0f; +STBIDEF void stbi_hdr_to_ldr_gamma(float gamma) { stbi__h2l_gamma_i = 1/gamma; } +STBIDEF void stbi_hdr_to_ldr_scale(float scale) { stbi__h2l_scale_i = 1/scale; } +enum +{ + STBI__SCAN_load=0, + STBI__SCAN_type, + STBI__SCAN_header +}; +static void stbi__refill_buffer(stbi__context *s) +{ + int n = (s->io.read)(s->io_user_data,(char*)s->buffer_start,s->buflen); + if (n == 0) { + + + s->read_from_callbacks = 0; + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start+1; + *s->img_buffer = 0; + } else { + s->img_buffer = s->buffer_start; + s->img_buffer_end = s->buffer_start + n; + } +} +stbi_inline static stbi_uc stbi__get8(stbi__context *s) +{ + if (s->img_buffer < s->img_buffer_end) + return *s->img_buffer++; + if (s->read_from_callbacks) { + stbi__refill_buffer(s); + return *s->img_buffer++; + } + return 0; +} +#if defined(STBI_NO_JPEG) && defined(STBI_NO_HDR) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +#else +stbi_inline static int stbi__at_eof(stbi__context *s) +{ + if (s->io.read) { + if (!(s->io.eof)(s->io_user_data)) return 0; + + + if (s->read_from_callbacks == 0) return 1; + } + return s->img_buffer >= s->img_buffer_end; +} +#endif +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) +#else +static void stbi__skip(stbi__context *s, int n) +{ + if (n < 0) { + s->img_buffer = s->img_buffer_end; + return; + } + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + s->img_buffer = s->img_buffer_end; + (s->io.skip)(s->io_user_data, n - blen); + return; + } + } + s->img_buffer += n; +} +#endif +#if defined(STBI_NO_PNG) && defined(STBI_NO_TGA) && defined(STBI_NO_HDR) && defined(STBI_NO_PNM) +#else +static int stbi__getn(stbi__context *s, stbi_uc *buffer, int n) +{ + if (s->io.read) { + int blen = (int) (s->img_buffer_end - s->img_buffer); + if (blen < n) { + int res, count; + memcpy(buffer, s->img_buffer, blen); + count = (s->io.read)(s->io_user_data, (char*) buffer + blen, n - blen); + res = (count == (n-blen)); + s->img_buffer = s->img_buffer_end; + return res; + } + } + if (s->img_buffer+n <= s->img_buffer_end) { + memcpy(buffer, s->img_buffer, n); + s->img_buffer += n; + return 1; + } else + return 0; +} +#endif +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) +#else +static int stbi__get16be(stbi__context *s) +{ + int z = stbi__get8(s); + return (z << 8) + stbi__get8(s); +} +#endif +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC) +#else +static stbi__uint32 stbi__get32be(stbi__context *s) +{ + stbi__uint32 z = stbi__get16be(s); + return (z << 16) + stbi__get16be(s); +} +#endif +#if defined(STBI_NO_BMP) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) +#else +static int stbi__get16le(stbi__context *s) +{ + int z = stbi__get8(s); + return z + (stbi__get8(s) << 8); +} +#endif +#ifndef STBI_NO_BMP +static stbi__uint32 stbi__get32le(stbi__context *s) +{ + stbi__uint32 z = stbi__get16le(s); + return z + (stbi__get16le(s) << 16); +} +#endif +#define STBI__BYTECAST(x) ((stbi_uc) ((x) & 255)) +#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +#else +static stbi_uc stbi__compute_y(int r, int g, int b) +{ + return (stbi_uc) (((r*77) + (g*150) + (29*b)) >> 8); +} +#endif +#if defined(STBI_NO_PNG) && defined(STBI_NO_BMP) && defined(STBI_NO_PSD) && defined(STBI_NO_TGA) && defined(STBI_NO_GIF) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM) +#else +static unsigned char *stbi__convert_format(unsigned char *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + unsigned char *good; + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + good = (unsigned char *) stbi__malloc_mad3(req_comp, x, y, 0); + if (good == NULL) { + STBI_FREE(data); + return stbi__errpuc("outofmem", "Out of memory"); + } + for (j=0; j < (int) y; ++j) { + unsigned char *src = data + j * x * img_n ; + unsigned char *dest = good + j * x * req_comp; + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + + + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=255; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=255; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=255; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = 255; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + STBI_FREE(data); + return good; +} +#endif +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) +#else +static stbi__uint16 stbi__compute_y_16(int r, int g, int b) +{ + return (stbi__uint16) (((r*77) + (g*150) + (29*b)) >> 8); +} +#endif +#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD) +#else +static stbi__uint16 *stbi__convert_format16(stbi__uint16 *data, int img_n, int req_comp, unsigned int x, unsigned int y) +{ + int i,j; + stbi__uint16 *good; + if (req_comp == img_n) return data; + STBI_ASSERT(req_comp >= 1 && req_comp <= 4); + good = (stbi__uint16 *) stbi__malloc(req_comp * x * y * 2); + if (good == NULL) { + STBI_FREE(data); + return (stbi__uint16 *) stbi__errpuc("outofmem", "Out of memory"); + } + for (j=0; j < (int) y; ++j) { + stbi__uint16 *src = data + j * x * img_n ; + stbi__uint16 *dest = good + j * x * req_comp; + #define STBI__COMBO(a,b) ((a)*8+(b)) + #define STBI__CASE(a,b) case STBI__COMBO(a,b): for(i=x-1; i >= 0; --i, src += a, dest += b) + + + switch (STBI__COMBO(img_n, req_comp)) { + STBI__CASE(1,2) { dest[0]=src[0]; dest[1]=0xffff; } break; + STBI__CASE(1,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(1,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=0xffff; } break; + STBI__CASE(2,1) { dest[0]=src[0]; } break; + STBI__CASE(2,3) { dest[0]=dest[1]=dest[2]=src[0]; } break; + STBI__CASE(2,4) { dest[0]=dest[1]=dest[2]=src[0]; dest[3]=src[1]; } break; + STBI__CASE(3,4) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2];dest[3]=0xffff; } break; + STBI__CASE(3,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(3,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = 0xffff; } break; + STBI__CASE(4,1) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); } break; + STBI__CASE(4,2) { dest[0]=stbi__compute_y_16(src[0],src[1],src[2]); dest[1] = src[3]; } break; + STBI__CASE(4,3) { dest[0]=src[0];dest[1]=src[1];dest[2]=src[2]; } break; + default: STBI_ASSERT(0); + } + #undef STBI__CASE + } + STBI_FREE(data); + return good; +} +#endif +#ifndef STBI_NO_LINEAR +static float *stbi__ldr_to_hdr(stbi_uc *data, int x, int y, int comp) +{ + int i,k,n; + float *output; + if (!data) return NULL; + output = (float *) stbi__malloc_mad4(x, y, comp, sizeof(float), 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpf("outofmem", "Out of memory"); } + + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + output[i*comp + k] = (float) (pow(data[i*comp+k]/255.0f, stbi__l2h_gamma) * stbi__l2h_scale); + } + } + if (n < comp) { + for (i=0; i < x*y; ++i) { + output[i*comp + n] = data[i*comp + n]/255.0f; + } + } + STBI_FREE(data); + return output; +} +#endif +#ifndef STBI_NO_HDR +#define stbi__float2int(x) ((int) (x)) +static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp) +{ + int i,k,n; + stbi_uc *output; + if (!data) return NULL; + output = (stbi_uc *) stbi__malloc_mad3(x, y, comp, 0); + if (output == NULL) { STBI_FREE(data); return stbi__errpuc("outofmem", "Out of memory"); } + + if (comp & 1) n = comp; else n = comp-1; + for (i=0; i < x*y; ++i) { + for (k=0; k < n; ++k) { + float z = (float) pow(data[i*comp+k]*stbi__h2l_scale_i, stbi__h2l_gamma_i) * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + if (k < comp) { + float z = data[i*comp+k] * 255 + 0.5f; + if (z < 0) z = 0; + if (z > 255) z = 255; + output[i*comp + k] = (stbi_uc) stbi__float2int(z); + } + } + STBI_FREE(data); + return output; +} +#endif +#ifndef STBI_NO_JPEG +#define FAST_BITS 9 +typedef struct +{ + stbi_uc fast[1 << FAST_BITS]; + + stbi__uint16 code[256]; + stbi_uc values[256]; + stbi_uc size[257]; + unsigned int maxcode[18]; + int delta[17]; +} stbi__huffman; +typedef struct +{ + stbi__context *s; + stbi__huffman huff_dc[4]; + stbi__huffman huff_ac[4]; + stbi__uint16 dequant[4][64]; + stbi__int16 fast_ac[4][1 << FAST_BITS]; + int img_h_max, img_v_max; + int img_mcu_x, img_mcu_y; + int img_mcu_w, img_mcu_h; + struct + { + int id; + int h,v; + int tq; + int hd,ha; + int dc_pred; + int x,y,w2,h2; + stbi_uc *data; + void *raw_data, *raw_coeff; + stbi_uc *linebuf; + short *coeff; + int coeff_w, coeff_h; + } img_comp[4]; + stbi__uint32 code_buffer; + int code_bits; + unsigned char marker; + int nomore; + int progressive; + int spec_start; + int spec_end; + int succ_high; + int succ_low; + int eob_run; + int jfif; + int app14_color_transform; + int rgb; + int scan_n, order[4]; + int restart_interval, todo; + void (*idct_block_kernel)(stbi_uc *out, int out_stride, short data[64]); + void (*YCbCr_to_RGB_kernel)(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step); + stbi_uc *(*resample_row_hv_2_kernel)(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs); +} stbi__jpeg; +static int stbi__build_huffman(stbi__huffman *h, int *count) +{ + int i,j,k=0; + unsigned int code; + + for (i=0; i < 16; ++i) + for (j=0; j < count[i]; ++j) + h->size[k++] = (stbi_uc) (i+1); + h->size[k] = 0; + + code = 0; + k = 0; + for(j=1; j <= 16; ++j) { + + h->delta[j] = k - code; + if (h->size[k] == j) { + while (h->size[k] == j) + h->code[k++] = (stbi__uint16) (code++); + if (code-1 >= (1u << j)) return stbi__err("bad code lengths","Corrupt JPEG"); + } + + h->maxcode[j] = code << (16-j); + code <<= 1; + } + h->maxcode[j] = 0xffffffff; + + memset(h->fast, 255, 1 << FAST_BITS); + for (i=0; i < k; ++i) { + int s = h->size[i]; + if (s <= FAST_BITS) { + int c = h->code[i] << (FAST_BITS-s); + int m = 1 << (FAST_BITS-s); + for (j=0; j < m; ++j) { + h->fast[c+j] = (stbi_uc) i; + } + } + } + return 1; +} +static void stbi__build_fast_ac(stbi__int16 *fast_ac, stbi__huffman *h) +{ + int i; + for (i=0; i < (1 << FAST_BITS); ++i) { + stbi_uc fast = h->fast[i]; + fast_ac[i] = 0; + if (fast < 255) { + int rs = h->values[fast]; + int run = (rs >> 4) & 15; + int magbits = rs & 15; + int len = h->size[fast]; + if (magbits && len + magbits <= FAST_BITS) { + + int k = ((i << len) & ((1 << FAST_BITS) - 1)) >> (FAST_BITS - magbits); + int m = 1 << (magbits - 1); + if (k < m) k += (~0U << magbits) + 1; + + if (k >= -128 && k <= 127) + fast_ac[i] = (stbi__int16) ((k * 256) + (run * 16) + (len + magbits)); + } + } + } +} +static void stbi__grow_buffer_unsafe(stbi__jpeg *j) +{ + do { + unsigned int b = j->nomore ? 0 : stbi__get8(j->s); + if (b == 0xff) { + int c = stbi__get8(j->s); + while (c == 0xff) c = stbi__get8(j->s); + if (c != 0) { + j->marker = (unsigned char) c; + j->nomore = 1; + return; + } + } + j->code_buffer |= b << (24 - j->code_bits); + j->code_bits += 8; + } while (j->code_bits <= 24); +} +static const stbi__uint32 stbi__bmask[17]={0,1,3,7,15,31,63,127,255,511,1023,2047,4095,8191,16383,32767,65535}; +stbi_inline static int stbi__jpeg_huff_decode(stbi__jpeg *j, stbi__huffman *h) +{ + unsigned int temp; + int c,k; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + + + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + k = h->fast[c]; + if (k < 255) { + int s = h->size[k]; + if (s > j->code_bits) + return -1; + j->code_buffer <<= s; + j->code_bits -= s; + return h->values[k]; + } + + + + + + + temp = j->code_buffer >> 16; + for (k=FAST_BITS+1 ; ; ++k) + if (temp < h->maxcode[k]) + break; + if (k == 17) { + + j->code_bits -= 16; + return -1; + } + if (k > j->code_bits) + return -1; + + c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k]; + STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]); + + j->code_bits -= k; + j->code_buffer <<= k; + return h->values[c]; +} +static const int stbi__jbias[16] = {0,-1,-3,-7,-15,-31,-63,-127,-255,-511,-1023,-2047,-4095,-8191,-16383,-32767}; +stbi_inline static int stbi__extend_receive(stbi__jpeg *j, int n) +{ + unsigned int k; + int sgn; + if (j->code_bits < n) stbi__grow_buffer_unsafe(j); + sgn = (stbi__int32)j->code_buffer >> 31; + k = stbi_lrot(j->code_buffer, n); + STBI_ASSERT(n >= 0 && n < (int) (sizeof(stbi__bmask)/sizeof(*stbi__bmask))); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k + (stbi__jbias[n] & ~sgn); +} +stbi_inline static int stbi__jpeg_get_bits(stbi__jpeg *j, int n) +{ + unsigned int k; + if (j->code_bits < n) stbi__grow_buffer_unsafe(j); + k = stbi_lrot(j->code_buffer, n); + j->code_buffer = k & ~stbi__bmask[n]; + k &= stbi__bmask[n]; + j->code_bits -= n; + return k; +} +stbi_inline static int stbi__jpeg_get_bit(stbi__jpeg *j) +{ + unsigned int k; + if (j->code_bits < 1) stbi__grow_buffer_unsafe(j); + k = j->code_buffer; + j->code_buffer <<= 1; + --j->code_bits; + return k & 0x80000000; +} +static const stbi_uc stbi__jpeg_dezigzag[64+15] = +{ + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 7, 14, 21, 28, + 35, 42, 49, 56, 57, 50, 43, 36, + 29, 22, 15, 23, 30, 37, 44, 51, + 58, 59, 52, 45, 38, 31, 39, 46, + 53, 60, 61, 54, 47, 55, 62, 63, + + 63, 63, 63, 63, 63, 63, 63, 63, + 63, 63, 63, 63, 63, 63, 63 +}; +static int stbi__jpeg_decode_block(stbi__jpeg *j, short data[64], stbi__huffman *hdc, stbi__huffman *hac, stbi__int16 *fac, int b, stbi__uint16 *dequant) +{ + int diff,dc,k; + int t; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + t = stbi__jpeg_huff_decode(j, hdc); + if (t < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + + memset(data,0,64*sizeof(data[0])); + diff = t ? stbi__extend_receive(j, t) : 0; + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc * dequant[0]); + + k = 1; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { + k += (r >> 4) & 15; + s = r & 15; + j->code_buffer <<= s; + j->code_bits -= s; + + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) * dequant[zig]); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (rs != 0xf0) break; + k += 16; + } else { + k += r; + + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) * dequant[zig]); + } + } + } while (k < 64); + return 1; +} +static int stbi__jpeg_decode_block_prog_dc(stbi__jpeg *j, short data[64], stbi__huffman *hdc, int b) +{ + int diff,dc; + int t; + if (j->spec_end != 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + if (j->succ_high == 0) { + + memset(data,0,64*sizeof(data[0])); + t = stbi__jpeg_huff_decode(j, hdc); + diff = t ? stbi__extend_receive(j, t) : 0; + dc = j->img_comp[b].dc_pred + diff; + j->img_comp[b].dc_pred = dc; + data[0] = (short) (dc << j->succ_low); + } else { + + if (stbi__jpeg_get_bit(j)) + data[0] += (short) (1 << j->succ_low); + } + return 1; +} +static int stbi__jpeg_decode_block_prog_ac(stbi__jpeg *j, short data[64], stbi__huffman *hac, stbi__int16 *fac) +{ + int k; + if (j->spec_start == 0) return stbi__err("can't merge dc and ac", "Corrupt JPEG"); + if (j->succ_high == 0) { + int shift = j->succ_low; + if (j->eob_run) { + --j->eob_run; + return 1; + } + k = j->spec_start; + do { + unsigned int zig; + int c,r,s; + if (j->code_bits < 16) stbi__grow_buffer_unsafe(j); + c = (j->code_buffer >> (32 - FAST_BITS)) & ((1 << FAST_BITS)-1); + r = fac[c]; + if (r) { + k += (r >> 4) & 15; + s = r & 15; + j->code_buffer <<= s; + j->code_bits -= s; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) ((r >> 8) << shift); + } else { + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r); + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + --j->eob_run; + break; + } + k += 16; + } else { + k += r; + zig = stbi__jpeg_dezigzag[k++]; + data[zig] = (short) (stbi__extend_receive(j,s) << shift); + } + } + } while (k <= j->spec_end); + } else { + + short bit = (short) (1 << j->succ_low); + if (j->eob_run) { + --j->eob_run; + for (k = j->spec_start; k <= j->spec_end; ++k) { + short *p = &data[stbi__jpeg_dezigzag[k]]; + if (*p != 0) + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } + } else { + k = j->spec_start; + do { + int r,s; + int rs = stbi__jpeg_huff_decode(j, hac); + if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG"); + s = rs & 15; + r = rs >> 4; + if (s == 0) { + if (r < 15) { + j->eob_run = (1 << r) - 1; + if (r) + j->eob_run += stbi__jpeg_get_bits(j, r); + r = 64; + } else { + + + + } + } else { + if (s != 1) return stbi__err("bad huffman code", "Corrupt JPEG"); + + if (stbi__jpeg_get_bit(j)) + s = bit; + else + s = -bit; + } + + while (k <= j->spec_end) { + short *p = &data[stbi__jpeg_dezigzag[k++]]; + if (*p != 0) { + if (stbi__jpeg_get_bit(j)) + if ((*p & bit)==0) { + if (*p > 0) + *p += bit; + else + *p -= bit; + } + } else { + if (r == 0) { + *p = (short) s; + break; + } + --r; + } + } + } while (k <= j->spec_end); + } + } + return 1; +} +stbi_inline static stbi_uc stbi__clamp(int x) +{ + + if ((unsigned int) x > 255) { + if (x < 0) return 0; + if (x > 255) return 255; + } + return (stbi_uc) x; +} +#define stbi__f2f(x) ((int) (((x) * 4096 + 0.5))) +#define stbi__fsh(x) ((x) * 4096) +#define STBI__IDCT_1D(s0,s1,s2,s3,s4,s5,s6,s7) \ + int t0,t1,t2,t3,p1,p2,p3,p4,p5,x0,x1,x2,x3; \ + p2 = s2; \ + p3 = s6; \ + p1 = (p2+p3) * stbi__f2f(0.5411961f); \ + t2 = p1 + p3*stbi__f2f(-1.847759065f); \ + t3 = p1 + p2*stbi__f2f( 0.765366865f); \ + p2 = s0; \ + p3 = s4; \ + t0 = stbi__fsh(p2+p3); \ + t1 = stbi__fsh(p2-p3); \ + x0 = t0+t3; \ + x3 = t0-t3; \ + x1 = t1+t2; \ + x2 = t1-t2; \ + t0 = s7; \ + t1 = s5; \ + t2 = s3; \ + t3 = s1; \ + p3 = t0+t2; \ + p4 = t1+t3; \ + p1 = t0+t3; \ + p2 = t1+t2; \ + p5 = (p3+p4)*stbi__f2f( 1.175875602f); \ + t0 = t0*stbi__f2f( 0.298631336f); \ + t1 = t1*stbi__f2f( 2.053119869f); \ + t2 = t2*stbi__f2f( 3.072711026f); \ + t3 = t3*stbi__f2f( 1.501321110f); \ + p1 = p5 + p1*stbi__f2f(-0.899976223f); \ + p2 = p5 + p2*stbi__f2f(-2.562915447f); \ + p3 = p3*stbi__f2f(-1.961570560f); \ + p4 = p4*stbi__f2f(-0.390180644f); \ + t3 += p1+p4; \ + t2 += p2+p3; \ + t1 += p2+p4; \ + t0 += p1+p3; +static void stbi__idct_block(stbi_uc *out, int out_stride, short data[64]) +{ + int i,val[64],*v=val; + stbi_uc *o; + short *d = data; + + for (i=0; i < 8; ++i,++d, ++v) { + + if (d[ 8]==0 && d[16]==0 && d[24]==0 && d[32]==0 + && d[40]==0 && d[48]==0 && d[56]==0) { + + + + + int dcterm = d[0]*4; + v[0] = v[8] = v[16] = v[24] = v[32] = v[40] = v[48] = v[56] = dcterm; + } else { + STBI__IDCT_1D(d[ 0],d[ 8],d[16],d[24],d[32],d[40],d[48],d[56]) + + + x0 += 512; x1 += 512; x2 += 512; x3 += 512; + v[ 0] = (x0+t3) >> 10; + v[56] = (x0-t3) >> 10; + v[ 8] = (x1+t2) >> 10; + v[48] = (x1-t2) >> 10; + v[16] = (x2+t1) >> 10; + v[40] = (x2-t1) >> 10; + v[24] = (x3+t0) >> 10; + v[32] = (x3-t0) >> 10; + } + } + for (i=0, v=val, o=out; i < 8; ++i,v+=8,o+=out_stride) { + + STBI__IDCT_1D(v[0],v[1],v[2],v[3],v[4],v[5],v[6],v[7]) + + + + + + + x0 += 65536 + (128<<17); + x1 += 65536 + (128<<17); + x2 += 65536 + (128<<17); + x3 += 65536 + (128<<17); + + + o[0] = stbi__clamp((x0+t3) >> 17); + o[7] = stbi__clamp((x0-t3) >> 17); + o[1] = stbi__clamp((x1+t2) >> 17); + o[6] = stbi__clamp((x1-t2) >> 17); + o[2] = stbi__clamp((x2+t1) >> 17); + o[5] = stbi__clamp((x2-t1) >> 17); + o[3] = stbi__clamp((x3+t0) >> 17); + o[4] = stbi__clamp((x3-t0) >> 17); + } +} +#ifdef STBI_SSE2 +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + + __m128i row0, row1, row2, row3, row4, row5, row6, row7; + __m128i tmp; + + #define dct_const(x,y) _mm_setr_epi16((x),(y),(x),(y),(x),(y),(x),(y)) + + + #define dct_rot(out0,out1, x,y,c0,c1) \ + __m128i c0##lo = _mm_unpacklo_epi16((x),(y)); \ + __m128i c0##hi = _mm_unpackhi_epi16((x),(y)); \ + __m128i out0##_l = _mm_madd_epi16(c0##lo, c0); \ + __m128i out0##_h = _mm_madd_epi16(c0##hi, c0); \ + __m128i out1##_l = _mm_madd_epi16(c0##lo, c1); \ + __m128i out1##_h = _mm_madd_epi16(c0##hi, c1) + + #define dct_widen(out, in) \ + __m128i out##_l = _mm_srai_epi32(_mm_unpacklo_epi16(_mm_setzero_si128(), (in)), 4); \ + __m128i out##_h = _mm_srai_epi32(_mm_unpackhi_epi16(_mm_setzero_si128(), (in)), 4) + + #define dct_wadd(out, a, b) \ + __m128i out##_l = _mm_add_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_add_epi32(a##_h, b##_h) + + #define dct_wsub(out, a, b) \ + __m128i out##_l = _mm_sub_epi32(a##_l, b##_l); \ + __m128i out##_h = _mm_sub_epi32(a##_h, b##_h) + + #define dct_bfly32o(out0, out1, a,b,bias,s) \ + { \ + __m128i abiased_l = _mm_add_epi32(a##_l, bias); \ + __m128i abiased_h = _mm_add_epi32(a##_h, bias); \ + dct_wadd(sum, abiased, b); \ + dct_wsub(dif, abiased, b); \ + out0 = _mm_packs_epi32(_mm_srai_epi32(sum_l, s), _mm_srai_epi32(sum_h, s)); \ + out1 = _mm_packs_epi32(_mm_srai_epi32(dif_l, s), _mm_srai_epi32(dif_h, s)); \ + } + + #define dct_interleave8(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi8(a, b); \ + b = _mm_unpackhi_epi8(tmp, b) + + #define dct_interleave16(a, b) \ + tmp = a; \ + a = _mm_unpacklo_epi16(a, b); \ + b = _mm_unpackhi_epi16(tmp, b) + #define dct_pass(bias,shift) \ + { \ + \ + dct_rot(t2e,t3e, row2,row6, rot0_0,rot0_1); \ + __m128i sum04 = _mm_add_epi16(row0, row4); \ + __m128i dif04 = _mm_sub_epi16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + \ + dct_rot(y0o,y2o, row7,row3, rot2_0,rot2_1); \ + dct_rot(y1o,y3o, row5,row1, rot3_0,rot3_1); \ + __m128i sum17 = _mm_add_epi16(row1, row7); \ + __m128i sum35 = _mm_add_epi16(row3, row5); \ + dct_rot(y4o,y5o, sum17,sum35, rot1_0,rot1_1); \ + dct_wadd(x4, y0o, y4o); \ + dct_wadd(x5, y1o, y5o); \ + dct_wadd(x6, y2o, y5o); \ + dct_wadd(x7, y3o, y4o); \ + dct_bfly32o(row0,row7, x0,x7,bias,shift); \ + dct_bfly32o(row1,row6, x1,x6,bias,shift); \ + dct_bfly32o(row2,row5, x2,x5,bias,shift); \ + dct_bfly32o(row3,row4, x3,x4,bias,shift); \ + } + __m128i rot0_0 = dct_const(stbi__f2f(0.5411961f), stbi__f2f(0.5411961f) + stbi__f2f(-1.847759065f)); + __m128i rot0_1 = dct_const(stbi__f2f(0.5411961f) + stbi__f2f( 0.765366865f), stbi__f2f(0.5411961f)); + __m128i rot1_0 = dct_const(stbi__f2f(1.175875602f) + stbi__f2f(-0.899976223f), stbi__f2f(1.175875602f)); + __m128i rot1_1 = dct_const(stbi__f2f(1.175875602f), stbi__f2f(1.175875602f) + stbi__f2f(-2.562915447f)); + __m128i rot2_0 = dct_const(stbi__f2f(-1.961570560f) + stbi__f2f( 0.298631336f), stbi__f2f(-1.961570560f)); + __m128i rot2_1 = dct_const(stbi__f2f(-1.961570560f), stbi__f2f(-1.961570560f) + stbi__f2f( 3.072711026f)); + __m128i rot3_0 = dct_const(stbi__f2f(-0.390180644f) + stbi__f2f( 2.053119869f), stbi__f2f(-0.390180644f)); + __m128i rot3_1 = dct_const(stbi__f2f(-0.390180644f), stbi__f2f(-0.390180644f) + stbi__f2f( 1.501321110f)); + + __m128i bias_0 = _mm_set1_epi32(512); + __m128i bias_1 = _mm_set1_epi32(65536 + (128<<17)); + + row0 = _mm_load_si128((const __m128i *) (data + 0*8)); + row1 = _mm_load_si128((const __m128i *) (data + 1*8)); + row2 = _mm_load_si128((const __m128i *) (data + 2*8)); + row3 = _mm_load_si128((const __m128i *) (data + 3*8)); + row4 = _mm_load_si128((const __m128i *) (data + 4*8)); + row5 = _mm_load_si128((const __m128i *) (data + 5*8)); + row6 = _mm_load_si128((const __m128i *) (data + 6*8)); + row7 = _mm_load_si128((const __m128i *) (data + 7*8)); + + dct_pass(bias_0, 10); + { + + dct_interleave16(row0, row4); + dct_interleave16(row1, row5); + dct_interleave16(row2, row6); + dct_interleave16(row3, row7); + + dct_interleave16(row0, row2); + dct_interleave16(row1, row3); + dct_interleave16(row4, row6); + dct_interleave16(row5, row7); + + dct_interleave16(row0, row1); + dct_interleave16(row2, row3); + dct_interleave16(row4, row5); + dct_interleave16(row6, row7); + } + + dct_pass(bias_1, 17); + { + + __m128i p0 = _mm_packus_epi16(row0, row1); + __m128i p1 = _mm_packus_epi16(row2, row3); + __m128i p2 = _mm_packus_epi16(row4, row5); + __m128i p3 = _mm_packus_epi16(row6, row7); + + dct_interleave8(p0, p2); + dct_interleave8(p1, p3); + + dct_interleave8(p0, p1); + dct_interleave8(p2, p3); + + dct_interleave8(p0, p2); + dct_interleave8(p1, p3); + + _mm_storel_epi64((__m128i *) out, p0); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p0, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p2); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p2, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p1); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p1, 0x4e)); out += out_stride; + _mm_storel_epi64((__m128i *) out, p3); out += out_stride; + _mm_storel_epi64((__m128i *) out, _mm_shuffle_epi32(p3, 0x4e)); + } +#undef dct_const +#undef dct_rot +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_interleave8 +#undef dct_interleave16 +#undef dct_pass +} +#endif +#ifdef STBI_NEON +static void stbi__idct_simd(stbi_uc *out, int out_stride, short data[64]) +{ + int16x8_t row0, row1, row2, row3, row4, row5, row6, row7; + int16x4_t rot0_0 = vdup_n_s16(stbi__f2f(0.5411961f)); + int16x4_t rot0_1 = vdup_n_s16(stbi__f2f(-1.847759065f)); + int16x4_t rot0_2 = vdup_n_s16(stbi__f2f( 0.765366865f)); + int16x4_t rot1_0 = vdup_n_s16(stbi__f2f( 1.175875602f)); + int16x4_t rot1_1 = vdup_n_s16(stbi__f2f(-0.899976223f)); + int16x4_t rot1_2 = vdup_n_s16(stbi__f2f(-2.562915447f)); + int16x4_t rot2_0 = vdup_n_s16(stbi__f2f(-1.961570560f)); + int16x4_t rot2_1 = vdup_n_s16(stbi__f2f(-0.390180644f)); + int16x4_t rot3_0 = vdup_n_s16(stbi__f2f( 0.298631336f)); + int16x4_t rot3_1 = vdup_n_s16(stbi__f2f( 2.053119869f)); + int16x4_t rot3_2 = vdup_n_s16(stbi__f2f( 3.072711026f)); + int16x4_t rot3_3 = vdup_n_s16(stbi__f2f( 1.501321110f)); +#define dct_long_mul(out, inq, coeff) \ + int32x4_t out##_l = vmull_s16(vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmull_s16(vget_high_s16(inq), coeff) +#define dct_long_mac(out, acc, inq, coeff) \ + int32x4_t out##_l = vmlal_s16(acc##_l, vget_low_s16(inq), coeff); \ + int32x4_t out##_h = vmlal_s16(acc##_h, vget_high_s16(inq), coeff) +#define dct_widen(out, inq) \ + int32x4_t out##_l = vshll_n_s16(vget_low_s16(inq), 12); \ + int32x4_t out##_h = vshll_n_s16(vget_high_s16(inq), 12) +#define dct_wadd(out, a, b) \ + int32x4_t out##_l = vaddq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vaddq_s32(a##_h, b##_h) +#define dct_wsub(out, a, b) \ + int32x4_t out##_l = vsubq_s32(a##_l, b##_l); \ + int32x4_t out##_h = vsubq_s32(a##_h, b##_h) +#define dct_bfly32o(out0,out1, a,b,shiftop,s) \ + { \ + dct_wadd(sum, a, b); \ + dct_wsub(dif, a, b); \ + out0 = vcombine_s16(shiftop(sum_l, s), shiftop(sum_h, s)); \ + out1 = vcombine_s16(shiftop(dif_l, s), shiftop(dif_h, s)); \ + } +#define dct_pass(shiftop, shift) \ + { \ + \ + int16x8_t sum26 = vaddq_s16(row2, row6); \ + dct_long_mul(p1e, sum26, rot0_0); \ + dct_long_mac(t2e, p1e, row6, rot0_1); \ + dct_long_mac(t3e, p1e, row2, rot0_2); \ + int16x8_t sum04 = vaddq_s16(row0, row4); \ + int16x8_t dif04 = vsubq_s16(row0, row4); \ + dct_widen(t0e, sum04); \ + dct_widen(t1e, dif04); \ + dct_wadd(x0, t0e, t3e); \ + dct_wsub(x3, t0e, t3e); \ + dct_wadd(x1, t1e, t2e); \ + dct_wsub(x2, t1e, t2e); \ + \ + int16x8_t sum15 = vaddq_s16(row1, row5); \ + int16x8_t sum17 = vaddq_s16(row1, row7); \ + int16x8_t sum35 = vaddq_s16(row3, row5); \ + int16x8_t sum37 = vaddq_s16(row3, row7); \ + int16x8_t sumodd = vaddq_s16(sum17, sum35); \ + dct_long_mul(p5o, sumodd, rot1_0); \ + dct_long_mac(p1o, p5o, sum17, rot1_1); \ + dct_long_mac(p2o, p5o, sum35, rot1_2); \ + dct_long_mul(p3o, sum37, rot2_0); \ + dct_long_mul(p4o, sum15, rot2_1); \ + dct_wadd(sump13o, p1o, p3o); \ + dct_wadd(sump24o, p2o, p4o); \ + dct_wadd(sump23o, p2o, p3o); \ + dct_wadd(sump14o, p1o, p4o); \ + dct_long_mac(x4, sump13o, row7, rot3_0); \ + dct_long_mac(x5, sump24o, row5, rot3_1); \ + dct_long_mac(x6, sump23o, row3, rot3_2); \ + dct_long_mac(x7, sump14o, row1, rot3_3); \ + dct_bfly32o(row0,row7, x0,x7,shiftop,shift); \ + dct_bfly32o(row1,row6, x1,x6,shiftop,shift); \ + dct_bfly32o(row2,row5, x2,x5,shiftop,shift); \ + dct_bfly32o(row3,row4, x3,x4,shiftop,shift); \ + } + + row0 = vld1q_s16(data + 0*8); + row1 = vld1q_s16(data + 1*8); + row2 = vld1q_s16(data + 2*8); + row3 = vld1q_s16(data + 3*8); + row4 = vld1q_s16(data + 4*8); + row5 = vld1q_s16(data + 5*8); + row6 = vld1q_s16(data + 6*8); + row7 = vld1q_s16(data + 7*8); + + row0 = vaddq_s16(row0, vsetq_lane_s16(1024, vdupq_n_s16(0), 0)); + + dct_pass(vrshrn_n_s32, 10); + + { +#define dct_trn16(x, y) { int16x8x2_t t = vtrnq_s16(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn32(x, y) { int32x4x2_t t = vtrnq_s32(vreinterpretq_s32_s16(x), vreinterpretq_s32_s16(y)); x = vreinterpretq_s16_s32(t.val[0]); y = vreinterpretq_s16_s32(t.val[1]); } +#define dct_trn64(x, y) { int16x8_t x0 = x; int16x8_t y0 = y; x = vcombine_s16(vget_low_s16(x0), vget_low_s16(y0)); y = vcombine_s16(vget_high_s16(x0), vget_high_s16(y0)); } + + dct_trn16(row0, row1); + dct_trn16(row2, row3); + dct_trn16(row4, row5); + dct_trn16(row6, row7); + + dct_trn32(row0, row2); + dct_trn32(row1, row3); + dct_trn32(row4, row6); + dct_trn32(row5, row7); + + dct_trn64(row0, row4); + dct_trn64(row1, row5); + dct_trn64(row2, row6); + dct_trn64(row3, row7); +#undef dct_trn16 +#undef dct_trn32 +#undef dct_trn64 + } + + + + + dct_pass(vshrn_n_s32, 16); + { + + uint8x8_t p0 = vqrshrun_n_s16(row0, 1); + uint8x8_t p1 = vqrshrun_n_s16(row1, 1); + uint8x8_t p2 = vqrshrun_n_s16(row2, 1); + uint8x8_t p3 = vqrshrun_n_s16(row3, 1); + uint8x8_t p4 = vqrshrun_n_s16(row4, 1); + uint8x8_t p5 = vqrshrun_n_s16(row5, 1); + uint8x8_t p6 = vqrshrun_n_s16(row6, 1); + uint8x8_t p7 = vqrshrun_n_s16(row7, 1); + +#define dct_trn8_8(x, y) { uint8x8x2_t t = vtrn_u8(x, y); x = t.val[0]; y = t.val[1]; } +#define dct_trn8_16(x, y) { uint16x4x2_t t = vtrn_u16(vreinterpret_u16_u8(x), vreinterpret_u16_u8(y)); x = vreinterpret_u8_u16(t.val[0]); y = vreinterpret_u8_u16(t.val[1]); } +#define dct_trn8_32(x, y) { uint32x2x2_t t = vtrn_u32(vreinterpret_u32_u8(x), vreinterpret_u32_u8(y)); x = vreinterpret_u8_u32(t.val[0]); y = vreinterpret_u8_u32(t.val[1]); } + + + + dct_trn8_8(p0, p1); + dct_trn8_8(p2, p3); + dct_trn8_8(p4, p5); + dct_trn8_8(p6, p7); + + dct_trn8_16(p0, p2); + dct_trn8_16(p1, p3); + dct_trn8_16(p4, p6); + dct_trn8_16(p5, p7); + + dct_trn8_32(p0, p4); + dct_trn8_32(p1, p5); + dct_trn8_32(p2, p6); + dct_trn8_32(p3, p7); + + vst1_u8(out, p0); out += out_stride; + vst1_u8(out, p1); out += out_stride; + vst1_u8(out, p2); out += out_stride; + vst1_u8(out, p3); out += out_stride; + vst1_u8(out, p4); out += out_stride; + vst1_u8(out, p5); out += out_stride; + vst1_u8(out, p6); out += out_stride; + vst1_u8(out, p7); +#undef dct_trn8_8 +#undef dct_trn8_16 +#undef dct_trn8_32 + } +#undef dct_long_mul +#undef dct_long_mac +#undef dct_widen +#undef dct_wadd +#undef dct_wsub +#undef dct_bfly32o +#undef dct_pass +} +#endif +#define STBI__MARKER_none 0xff +static stbi_uc stbi__get_marker(stbi__jpeg *j) +{ + stbi_uc x; + if (j->marker != STBI__MARKER_none) { x = j->marker; j->marker = STBI__MARKER_none; return x; } + x = stbi__get8(j->s); + if (x != 0xff) return STBI__MARKER_none; + while (x == 0xff) + x = stbi__get8(j->s); + return x; +} +#define STBI__RESTART(x) ((x) >= 0xd0 && (x) <= 0xd7) +static void stbi__jpeg_reset(stbi__jpeg *j) +{ + j->code_bits = 0; + j->code_buffer = 0; + j->nomore = 0; + j->img_comp[0].dc_pred = j->img_comp[1].dc_pred = j->img_comp[2].dc_pred = j->img_comp[3].dc_pred = 0; + j->marker = STBI__MARKER_none; + j->todo = j->restart_interval ? j->restart_interval : 0x7fffffff; + j->eob_run = 0; + + +} +static int stbi__parse_entropy_coded_data(stbi__jpeg *z) +{ + stbi__jpeg_reset(z); + if (!z->progressive) { + if (z->scan_n == 1) { + int i,j; + STBI_SIMD_ALIGN(short, data[64]); + int n = z->order[0]; + + + + + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + + + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { + int i,j,k,x,y; + STBI_SIMD_ALIGN(short, data[64]); + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + + + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x)*8; + int y2 = (j*z->img_comp[n].v + y)*8; + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block(z, data, z->huff_dc+z->img_comp[n].hd, z->huff_ac+ha, z->fast_ac[ha], n, z->dequant[z->img_comp[n].tq])) return 0; + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*y2+x2, z->img_comp[n].w2, data); + } + } + } + + + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } else { + if (z->scan_n == 1) { + int i,j; + int n = z->order[0]; + + + + + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + if (z->spec_start == 0) { + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } else { + int ha = z->img_comp[n].ha; + if (!stbi__jpeg_decode_block_prog_ac(z, data, &z->huff_ac[ha], z->fast_ac[ha])) + return 0; + } + + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } else { + int i,j,k,x,y; + for (j=0; j < z->img_mcu_y; ++j) { + for (i=0; i < z->img_mcu_x; ++i) { + + for (k=0; k < z->scan_n; ++k) { + int n = z->order[k]; + + + for (y=0; y < z->img_comp[n].v; ++y) { + for (x=0; x < z->img_comp[n].h; ++x) { + int x2 = (i*z->img_comp[n].h + x); + int y2 = (j*z->img_comp[n].v + y); + short *data = z->img_comp[n].coeff + 64 * (x2 + y2 * z->img_comp[n].coeff_w); + if (!stbi__jpeg_decode_block_prog_dc(z, data, &z->huff_dc[z->img_comp[n].hd], n)) + return 0; + } + } + } + + + if (--z->todo <= 0) { + if (z->code_bits < 24) stbi__grow_buffer_unsafe(z); + if (!STBI__RESTART(z->marker)) return 1; + stbi__jpeg_reset(z); + } + } + } + return 1; + } + } +} +static void stbi__jpeg_dequantize(short *data, stbi__uint16 *dequant) +{ + int i; + for (i=0; i < 64; ++i) + data[i] *= dequant[i]; +} +static void stbi__jpeg_finish(stbi__jpeg *z) +{ + if (z->progressive) { + + int i,j,n; + for (n=0; n < z->s->img_n; ++n) { + int w = (z->img_comp[n].x+7) >> 3; + int h = (z->img_comp[n].y+7) >> 3; + for (j=0; j < h; ++j) { + for (i=0; i < w; ++i) { + short *data = z->img_comp[n].coeff + 64 * (i + j * z->img_comp[n].coeff_w); + stbi__jpeg_dequantize(data, z->dequant[z->img_comp[n].tq]); + z->idct_block_kernel(z->img_comp[n].data+z->img_comp[n].w2*j*8+i*8, z->img_comp[n].w2, data); + } + } + } + } +} +static int stbi__process_marker(stbi__jpeg *z, int m) +{ + int L; + switch (m) { + case STBI__MARKER_none: + return stbi__err("expected marker","Corrupt JPEG"); + case 0xDD: + if (stbi__get16be(z->s) != 4) return stbi__err("bad DRI len","Corrupt JPEG"); + z->restart_interval = stbi__get16be(z->s); + return 1; + case 0xDB: + L = stbi__get16be(z->s)-2; + while (L > 0) { + int q = stbi__get8(z->s); + int p = q >> 4, sixteen = (p != 0); + int t = q & 15,i; + if (p != 0 && p != 1) return stbi__err("bad DQT type","Corrupt JPEG"); + if (t > 3) return stbi__err("bad DQT table","Corrupt JPEG"); + for (i=0; i < 64; ++i) + z->dequant[t][stbi__jpeg_dezigzag[i]] = (stbi__uint16)(sixteen ? stbi__get16be(z->s) : stbi__get8(z->s)); + L -= (sixteen ? 129 : 65); + } + return L==0; + case 0xC4: + L = stbi__get16be(z->s)-2; + while (L > 0) { + stbi_uc *v; + int sizes[16],i,n=0; + int q = stbi__get8(z->s); + int tc = q >> 4; + int th = q & 15; + if (tc > 1 || th > 3) return stbi__err("bad DHT header","Corrupt JPEG"); + for (i=0; i < 16; ++i) { + sizes[i] = stbi__get8(z->s); + n += sizes[i]; + } + L -= 17; + if (tc == 0) { + if (!stbi__build_huffman(z->huff_dc+th, sizes)) return 0; + v = z->huff_dc[th].values; + } else { + if (!stbi__build_huffman(z->huff_ac+th, sizes)) return 0; + v = z->huff_ac[th].values; + } + for (i=0; i < n; ++i) + v[i] = stbi__get8(z->s); + if (tc != 0) + stbi__build_fast_ac(z->fast_ac[th], z->huff_ac + th); + L -= n; + } + return L==0; + } + + if ((m >= 0xE0 && m <= 0xEF) || m == 0xFE) { + L = stbi__get16be(z->s); + if (L < 2) { + if (m == 0xFE) + return stbi__err("bad COM len","Corrupt JPEG"); + else + return stbi__err("bad APP len","Corrupt JPEG"); + } + L -= 2; + if (m == 0xE0 && L >= 5) { + static const unsigned char tag[5] = {'J','F','I','F','\0'}; + int ok = 1; + int i; + for (i=0; i < 5; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 5; + if (ok) + z->jfif = 1; + } else if (m == 0xEE && L >= 12) { + static const unsigned char tag[6] = {'A','d','o','b','e','\0'}; + int ok = 1; + int i; + for (i=0; i < 6; ++i) + if (stbi__get8(z->s) != tag[i]) + ok = 0; + L -= 6; + if (ok) { + stbi__get8(z->s); + stbi__get16be(z->s); + stbi__get16be(z->s); + z->app14_color_transform = stbi__get8(z->s); + L -= 6; + } + } + stbi__skip(z->s, L); + return 1; + } + return stbi__err("unknown marker","Corrupt JPEG"); +} +static int stbi__process_scan_header(stbi__jpeg *z) +{ + int i; + int Ls = stbi__get16be(z->s); + z->scan_n = stbi__get8(z->s); + if (z->scan_n < 1 || z->scan_n > 4 || z->scan_n > (int) z->s->img_n) return stbi__err("bad SOS component count","Corrupt JPEG"); + if (Ls != 6+2*z->scan_n) return stbi__err("bad SOS len","Corrupt JPEG"); + for (i=0; i < z->scan_n; ++i) { + int id = stbi__get8(z->s), which; + int q = stbi__get8(z->s); + for (which = 0; which < z->s->img_n; ++which) + if (z->img_comp[which].id == id) + break; + if (which == z->s->img_n) return 0; + z->img_comp[which].hd = q >> 4; if (z->img_comp[which].hd > 3) return stbi__err("bad DC huff","Corrupt JPEG"); + z->img_comp[which].ha = q & 15; if (z->img_comp[which].ha > 3) return stbi__err("bad AC huff","Corrupt JPEG"); + z->order[i] = which; + } + { + int aa; + z->spec_start = stbi__get8(z->s); + z->spec_end = stbi__get8(z->s); + aa = stbi__get8(z->s); + z->succ_high = (aa >> 4); + z->succ_low = (aa & 15); + if (z->progressive) { + if (z->spec_start > 63 || z->spec_end > 63 || z->spec_start > z->spec_end || z->succ_high > 13 || z->succ_low > 13) + return stbi__err("bad SOS", "Corrupt JPEG"); + } else { + if (z->spec_start != 0) return stbi__err("bad SOS","Corrupt JPEG"); + if (z->succ_high != 0 || z->succ_low != 0) return stbi__err("bad SOS","Corrupt JPEG"); + z->spec_end = 63; + } + } + return 1; +} +static int stbi__free_jpeg_components(stbi__jpeg *z, int ncomp, int why) +{ + int i; + for (i=0; i < ncomp; ++i) { + if (z->img_comp[i].raw_data) { + STBI_FREE(z->img_comp[i].raw_data); + z->img_comp[i].raw_data = NULL; + z->img_comp[i].data = NULL; + } + if (z->img_comp[i].raw_coeff) { + STBI_FREE(z->img_comp[i].raw_coeff); + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].coeff = 0; + } + if (z->img_comp[i].linebuf) { + STBI_FREE(z->img_comp[i].linebuf); + z->img_comp[i].linebuf = NULL; + } + } + return why; +} +static int stbi__process_frame_header(stbi__jpeg *z, int scan) +{ + stbi__context *s = z->s; + int Lf,p,i,q, h_max=1,v_max=1,c; + Lf = stbi__get16be(s); if (Lf < 11) return stbi__err("bad SOF len","Corrupt JPEG"); + p = stbi__get8(s); if (p != 8) return stbi__err("only 8-bit","JPEG format not supported: 8-bit only"); + s->img_y = stbi__get16be(s); if (s->img_y == 0) return stbi__err("no header height", "JPEG format not supported: delayed height"); + s->img_x = stbi__get16be(s); if (s->img_x == 0) return stbi__err("0 width","Corrupt JPEG"); + c = stbi__get8(s); + if (c != 3 && c != 1 && c != 4) return stbi__err("bad component count","Corrupt JPEG"); + s->img_n = c; + for (i=0; i < c; ++i) { + z->img_comp[i].data = NULL; + z->img_comp[i].linebuf = NULL; + } + if (Lf != 8+3*s->img_n) return stbi__err("bad SOF len","Corrupt JPEG"); + z->rgb = 0; + for (i=0; i < s->img_n; ++i) { + static const unsigned char rgb[3] = { 'R', 'G', 'B' }; + z->img_comp[i].id = stbi__get8(s); + if (s->img_n == 3 && z->img_comp[i].id == rgb[i]) + ++z->rgb; + q = stbi__get8(s); + z->img_comp[i].h = (q >> 4); if (!z->img_comp[i].h || z->img_comp[i].h > 4) return stbi__err("bad H","Corrupt JPEG"); + z->img_comp[i].v = q & 15; if (!z->img_comp[i].v || z->img_comp[i].v > 4) return stbi__err("bad V","Corrupt JPEG"); + z->img_comp[i].tq = stbi__get8(s); if (z->img_comp[i].tq > 3) return stbi__err("bad TQ","Corrupt JPEG"); + } + if (scan != STBI__SCAN_load) return 1; + if (!stbi__mad3sizes_valid(s->img_x, s->img_y, s->img_n, 0)) return stbi__err("too large", "Image too large to decode"); + for (i=0; i < s->img_n; ++i) { + if (z->img_comp[i].h > h_max) h_max = z->img_comp[i].h; + if (z->img_comp[i].v > v_max) v_max = z->img_comp[i].v; + } + + z->img_h_max = h_max; + z->img_v_max = v_max; + z->img_mcu_w = h_max * 8; + z->img_mcu_h = v_max * 8; + + z->img_mcu_x = (s->img_x + z->img_mcu_w-1) / z->img_mcu_w; + z->img_mcu_y = (s->img_y + z->img_mcu_h-1) / z->img_mcu_h; + for (i=0; i < s->img_n; ++i) { + + z->img_comp[i].x = (s->img_x * z->img_comp[i].h + h_max-1) / h_max; + z->img_comp[i].y = (s->img_y * z->img_comp[i].v + v_max-1) / v_max; + + + + + + + + z->img_comp[i].w2 = z->img_mcu_x * z->img_comp[i].h * 8; + z->img_comp[i].h2 = z->img_mcu_y * z->img_comp[i].v * 8; + z->img_comp[i].coeff = 0; + z->img_comp[i].raw_coeff = 0; + z->img_comp[i].linebuf = NULL; + z->img_comp[i].raw_data = stbi__malloc_mad2(z->img_comp[i].w2, z->img_comp[i].h2, 15); + if (z->img_comp[i].raw_data == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + + z->img_comp[i].data = (stbi_uc*) (((size_t) z->img_comp[i].raw_data + 15) & ~15); + if (z->progressive) { + + z->img_comp[i].coeff_w = z->img_comp[i].w2 / 8; + z->img_comp[i].coeff_h = z->img_comp[i].h2 / 8; + z->img_comp[i].raw_coeff = stbi__malloc_mad3(z->img_comp[i].w2, z->img_comp[i].h2, sizeof(short), 15); + if (z->img_comp[i].raw_coeff == NULL) + return stbi__free_jpeg_components(z, i+1, stbi__err("outofmem", "Out of memory")); + z->img_comp[i].coeff = (short*) (((size_t) z->img_comp[i].raw_coeff + 15) & ~15); + } + } + return 1; +} +#define stbi__DNL(x) ((x) == 0xdc) +#define stbi__SOI(x) ((x) == 0xd8) +#define stbi__EOI(x) ((x) == 0xd9) +#define stbi__SOF(x) ((x) == 0xc0 || (x) == 0xc1 || (x) == 0xc2) +#define stbi__SOS(x) ((x) == 0xda) +#define stbi__SOF_progressive(x) ((x) == 0xc2) +static int stbi__decode_jpeg_header(stbi__jpeg *z, int scan) +{ + int m; + z->jfif = 0; + z->app14_color_transform = -1; + z->marker = STBI__MARKER_none; + m = stbi__get_marker(z); + if (!stbi__SOI(m)) return stbi__err("no SOI","Corrupt JPEG"); + if (scan == STBI__SCAN_type) return 1; + m = stbi__get_marker(z); + while (!stbi__SOF(m)) { + if (!stbi__process_marker(z,m)) return 0; + m = stbi__get_marker(z); + while (m == STBI__MARKER_none) { + + if (stbi__at_eof(z->s)) return stbi__err("no SOF", "Corrupt JPEG"); + m = stbi__get_marker(z); + } + } + z->progressive = stbi__SOF_progressive(m); + if (!stbi__process_frame_header(z, scan)) return 0; + return 1; +} +static int stbi__decode_jpeg_image(stbi__jpeg *j) +{ + int m; + for (m = 0; m < 4; m++) { + j->img_comp[m].raw_data = NULL; + j->img_comp[m].raw_coeff = NULL; + } + j->restart_interval = 0; + if (!stbi__decode_jpeg_header(j, STBI__SCAN_load)) return 0; + m = stbi__get_marker(j); + while (!stbi__EOI(m)) { + if (stbi__SOS(m)) { + if (!stbi__process_scan_header(j)) return 0; + if (!stbi__parse_entropy_coded_data(j)) return 0; + if (j->marker == STBI__MARKER_none ) { + + while (!stbi__at_eof(j->s)) { + int x = stbi__get8(j->s); + if (x == 255) { + j->marker = stbi__get8(j->s); + break; + } + } + + } + } else if (stbi__DNL(m)) { + int Ld = stbi__get16be(j->s); + stbi__uint32 NL = stbi__get16be(j->s); + if (Ld != 4) return stbi__err("bad DNL len", "Corrupt JPEG"); + if (NL != j->s->img_y) return stbi__err("bad DNL height", "Corrupt JPEG"); + } else { + if (!stbi__process_marker(j, m)) return 0; + } + m = stbi__get_marker(j); + } + if (j->progressive) + stbi__jpeg_finish(j); + return 1; +} +typedef stbi_uc *(*resample_row_func)(stbi_uc *out, stbi_uc *in0, stbi_uc *in1, + int w, int hs); +#define stbi__div4(x) ((stbi_uc) ((x) >> 2)) +static stbi_uc *resample_row_1(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + STBI_NOTUSED(out); + STBI_NOTUSED(in_far); + STBI_NOTUSED(w); + STBI_NOTUSED(hs); + return in_near; +} +static stbi_uc* stbi__resample_row_v_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + + int i; + STBI_NOTUSED(hs); + for (i=0; i < w; ++i) + out[i] = stbi__div4(3*in_near[i] + in_far[i] + 2); + return out; +} +static stbi_uc* stbi__resample_row_h_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + + int i; + stbi_uc *input = in_near; + if (w == 1) { + + out[0] = out[1] = input[0]; + return out; + } + out[0] = input[0]; + out[1] = stbi__div4(input[0]*3 + input[1] + 2); + for (i=1; i < w-1; ++i) { + int n = 3*input[i]+2; + out[i*2+0] = stbi__div4(n+input[i-1]); + out[i*2+1] = stbi__div4(n+input[i+1]); + } + out[i*2+0] = stbi__div4(input[w-2]*3 + input[w-1] + 2); + out[i*2+1] = input[w-1]; + STBI_NOTUSED(in_far); + STBI_NOTUSED(hs); + return out; +} +#define stbi__div16(x) ((stbi_uc) ((x) >> 4)) +static stbi_uc *stbi__resample_row_hv_2(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + + int i,t0,t1; + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + t1 = 3*in_near[0] + in_far[0]; + out[0] = stbi__div4(t1+2); + for (i=1; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + STBI_NOTUSED(hs); + return out; +} +#if defined(STBI_SSE2) || defined(STBI_NEON) +static stbi_uc *stbi__resample_row_hv_2_simd(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + + int i=0,t0,t1; + if (w == 1) { + out[0] = out[1] = stbi__div4(3*in_near[0] + in_far[0] + 2); + return out; + } + t1 = 3*in_near[0] + in_far[0]; + + + + for (; i < ((w-1) & ~7); i += 8) { +#if defined(STBI_SSE2) + + + __m128i zero = _mm_setzero_si128(); + __m128i farb = _mm_loadl_epi64((__m128i *) (in_far + i)); + __m128i nearb = _mm_loadl_epi64((__m128i *) (in_near + i)); + __m128i farw = _mm_unpacklo_epi8(farb, zero); + __m128i nearw = _mm_unpacklo_epi8(nearb, zero); + __m128i diff = _mm_sub_epi16(farw, nearw); + __m128i nears = _mm_slli_epi16(nearw, 2); + __m128i curr = _mm_add_epi16(nears, diff); + + + + + + __m128i prv0 = _mm_slli_si128(curr, 2); + __m128i nxt0 = _mm_srli_si128(curr, 2); + __m128i prev = _mm_insert_epi16(prv0, t1, 0); + __m128i next = _mm_insert_epi16(nxt0, 3*in_near[i+8] + in_far[i+8], 7); + + + + + __m128i bias = _mm_set1_epi16(8); + __m128i curs = _mm_slli_epi16(curr, 2); + __m128i prvd = _mm_sub_epi16(prev, curr); + __m128i nxtd = _mm_sub_epi16(next, curr); + __m128i curb = _mm_add_epi16(curs, bias); + __m128i even = _mm_add_epi16(prvd, curb); + __m128i odd = _mm_add_epi16(nxtd, curb); + + __m128i int0 = _mm_unpacklo_epi16(even, odd); + __m128i int1 = _mm_unpackhi_epi16(even, odd); + __m128i de0 = _mm_srli_epi16(int0, 4); + __m128i de1 = _mm_srli_epi16(int1, 4); + + __m128i outv = _mm_packus_epi16(de0, de1); + _mm_storeu_si128((__m128i *) (out + i*2), outv); +#elif defined(STBI_NEON) + + + uint8x8_t farb = vld1_u8(in_far + i); + uint8x8_t nearb = vld1_u8(in_near + i); + int16x8_t diff = vreinterpretq_s16_u16(vsubl_u8(farb, nearb)); + int16x8_t nears = vreinterpretq_s16_u16(vshll_n_u8(nearb, 2)); + int16x8_t curr = vaddq_s16(nears, diff); + + + + + + int16x8_t prv0 = vextq_s16(curr, curr, 7); + int16x8_t nxt0 = vextq_s16(curr, curr, 1); + int16x8_t prev = vsetq_lane_s16(t1, prv0, 0); + int16x8_t next = vsetq_lane_s16(3*in_near[i+8] + in_far[i+8], nxt0, 7); + + + + + int16x8_t curs = vshlq_n_s16(curr, 2); + int16x8_t prvd = vsubq_s16(prev, curr); + int16x8_t nxtd = vsubq_s16(next, curr); + int16x8_t even = vaddq_s16(curs, prvd); + int16x8_t odd = vaddq_s16(curs, nxtd); + + uint8x8x2_t o; + o.val[0] = vqrshrun_n_s16(even, 4); + o.val[1] = vqrshrun_n_s16(odd, 4); + vst2_u8(out + i*2, o); +#endif + + t1 = 3*in_near[i+7] + in_far[i+7]; + } + t0 = t1; + t1 = 3*in_near[i] + in_far[i]; + out[i*2] = stbi__div16(3*t1 + t0 + 8); + for (++i; i < w; ++i) { + t0 = t1; + t1 = 3*in_near[i]+in_far[i]; + out[i*2-1] = stbi__div16(3*t0 + t1 + 8); + out[i*2 ] = stbi__div16(3*t1 + t0 + 8); + } + out[w*2-1] = stbi__div4(t1+2); + STBI_NOTUSED(hs); + return out; +} +#endif +static stbi_uc *stbi__resample_row_generic(stbi_uc *out, stbi_uc *in_near, stbi_uc *in_far, int w, int hs) +{ + + int i,j; + STBI_NOTUSED(in_far); + for (i=0; i < w; ++i) + for (j=0; j < hs; ++j) + out[i*hs+j] = in_near[i]; + return out; +} +#define stbi__float2fixed(x) (((int) ((x) * 4096.0f + 0.5f)) << 8) +static void stbi__YCbCr_to_RGB_row(stbi_uc *out, const stbi_uc *y, const stbi_uc *pcb, const stbi_uc *pcr, int count, int step) +{ + int i; + for (i=0; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + (cr*-stbi__float2fixed(0.71414f)) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} +#if defined(STBI_SSE2) || defined(STBI_NEON) +static void stbi__YCbCr_to_RGB_simd(stbi_uc *out, stbi_uc const *y, stbi_uc const *pcb, stbi_uc const *pcr, int count, int step) +{ + int i = 0; +#ifdef STBI_SSE2 + + + + if (step == 4) { + + __m128i signflip = _mm_set1_epi8(-0x80); + __m128i cr_const0 = _mm_set1_epi16( (short) ( 1.40200f*4096.0f+0.5f)); + __m128i cr_const1 = _mm_set1_epi16( - (short) ( 0.71414f*4096.0f+0.5f)); + __m128i cb_const0 = _mm_set1_epi16( - (short) ( 0.34414f*4096.0f+0.5f)); + __m128i cb_const1 = _mm_set1_epi16( (short) ( 1.77200f*4096.0f+0.5f)); + __m128i y_bias = _mm_set1_epi8((char) (unsigned char) 128); + __m128i xw = _mm_set1_epi16(255); + for (; i+7 < count; i += 8) { + + __m128i y_bytes = _mm_loadl_epi64((__m128i *) (y+i)); + __m128i cr_bytes = _mm_loadl_epi64((__m128i *) (pcr+i)); + __m128i cb_bytes = _mm_loadl_epi64((__m128i *) (pcb+i)); + __m128i cr_biased = _mm_xor_si128(cr_bytes, signflip); + __m128i cb_biased = _mm_xor_si128(cb_bytes, signflip); + + __m128i yw = _mm_unpacklo_epi8(y_bias, y_bytes); + __m128i crw = _mm_unpacklo_epi8(_mm_setzero_si128(), cr_biased); + __m128i cbw = _mm_unpacklo_epi8(_mm_setzero_si128(), cb_biased); + + __m128i yws = _mm_srli_epi16(yw, 4); + __m128i cr0 = _mm_mulhi_epi16(cr_const0, crw); + __m128i cb0 = _mm_mulhi_epi16(cb_const0, cbw); + __m128i cb1 = _mm_mulhi_epi16(cbw, cb_const1); + __m128i cr1 = _mm_mulhi_epi16(crw, cr_const1); + __m128i rws = _mm_add_epi16(cr0, yws); + __m128i gwt = _mm_add_epi16(cb0, yws); + __m128i bws = _mm_add_epi16(yws, cb1); + __m128i gws = _mm_add_epi16(gwt, cr1); + + __m128i rw = _mm_srai_epi16(rws, 4); + __m128i bw = _mm_srai_epi16(bws, 4); + __m128i gw = _mm_srai_epi16(gws, 4); + + __m128i brb = _mm_packus_epi16(rw, bw); + __m128i gxb = _mm_packus_epi16(gw, xw); + + __m128i t0 = _mm_unpacklo_epi8(brb, gxb); + __m128i t1 = _mm_unpackhi_epi8(brb, gxb); + __m128i o0 = _mm_unpacklo_epi16(t0, t1); + __m128i o1 = _mm_unpackhi_epi16(t0, t1); + + _mm_storeu_si128((__m128i *) (out + 0), o0); + _mm_storeu_si128((__m128i *) (out + 16), o1); + out += 32; + } + } +#endif +#ifdef STBI_NEON + + if (step == 4) { + + uint8x8_t signflip = vdup_n_u8(0x80); + int16x8_t cr_const0 = vdupq_n_s16( (short) ( 1.40200f*4096.0f+0.5f)); + int16x8_t cr_const1 = vdupq_n_s16( - (short) ( 0.71414f*4096.0f+0.5f)); + int16x8_t cb_const0 = vdupq_n_s16( - (short) ( 0.34414f*4096.0f+0.5f)); + int16x8_t cb_const1 = vdupq_n_s16( (short) ( 1.77200f*4096.0f+0.5f)); + for (; i+7 < count; i += 8) { + + uint8x8_t y_bytes = vld1_u8(y + i); + uint8x8_t cr_bytes = vld1_u8(pcr + i); + uint8x8_t cb_bytes = vld1_u8(pcb + i); + int8x8_t cr_biased = vreinterpret_s8_u8(vsub_u8(cr_bytes, signflip)); + int8x8_t cb_biased = vreinterpret_s8_u8(vsub_u8(cb_bytes, signflip)); + + int16x8_t yws = vreinterpretq_s16_u16(vshll_n_u8(y_bytes, 4)); + int16x8_t crw = vshll_n_s8(cr_biased, 7); + int16x8_t cbw = vshll_n_s8(cb_biased, 7); + + int16x8_t cr0 = vqdmulhq_s16(crw, cr_const0); + int16x8_t cb0 = vqdmulhq_s16(cbw, cb_const0); + int16x8_t cr1 = vqdmulhq_s16(crw, cr_const1); + int16x8_t cb1 = vqdmulhq_s16(cbw, cb_const1); + int16x8_t rws = vaddq_s16(yws, cr0); + int16x8_t gws = vaddq_s16(vaddq_s16(yws, cb0), cr1); + int16x8_t bws = vaddq_s16(yws, cb1); + + uint8x8x4_t o; + o.val[0] = vqrshrun_n_s16(rws, 4); + o.val[1] = vqrshrun_n_s16(gws, 4); + o.val[2] = vqrshrun_n_s16(bws, 4); + o.val[3] = vdup_n_u8(255); + + vst4_u8(out, o); + out += 8*4; + } + } +#endif + for (; i < count; ++i) { + int y_fixed = (y[i] << 20) + (1<<19); + int r,g,b; + int cr = pcr[i] - 128; + int cb = pcb[i] - 128; + r = y_fixed + cr* stbi__float2fixed(1.40200f); + g = y_fixed + cr*-stbi__float2fixed(0.71414f) + ((cb*-stbi__float2fixed(0.34414f)) & 0xffff0000); + b = y_fixed + cb* stbi__float2fixed(1.77200f); + r >>= 20; + g >>= 20; + b >>= 20; + if ((unsigned) r > 255) { if (r < 0) r = 0; else r = 255; } + if ((unsigned) g > 255) { if (g < 0) g = 0; else g = 255; } + if ((unsigned) b > 255) { if (b < 0) b = 0; else b = 255; } + out[0] = (stbi_uc)r; + out[1] = (stbi_uc)g; + out[2] = (stbi_uc)b; + out[3] = 255; + out += step; + } +} +#endif +static void stbi__setup_jpeg(stbi__jpeg *j) +{ + j->idct_block_kernel = stbi__idct_block; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_row; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2; +#ifdef STBI_SSE2 + if (stbi__sse2_available()) { + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; + } +#endif +#ifdef STBI_NEON + j->idct_block_kernel = stbi__idct_simd; + j->YCbCr_to_RGB_kernel = stbi__YCbCr_to_RGB_simd; + j->resample_row_hv_2_kernel = stbi__resample_row_hv_2_simd; +#endif +} +static void stbi__cleanup_jpeg(stbi__jpeg *j) +{ + stbi__free_jpeg_components(j, j->s->img_n, 0); +} +typedef struct +{ + resample_row_func resample; + stbi_uc *line0,*line1; + int hs,vs; + int w_lores; + int ystep; + int ypos; +} stbi__resample; +static stbi_uc stbi__blinn_8x8(stbi_uc x, stbi_uc y) +{ + unsigned int t = x*y + 128; + return (stbi_uc) ((t + (t >>8)) >> 8); +} +static stbi_uc *load_jpeg_image(stbi__jpeg *z, int *out_x, int *out_y, int *comp, int req_comp) +{ + int n, decode_n, is_rgb; + z->s->img_n = 0; + + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + + if (!stbi__decode_jpeg_image(z)) { stbi__cleanup_jpeg(z); return NULL; } + + n = req_comp ? req_comp : z->s->img_n >= 3 ? 3 : 1; + is_rgb = z->s->img_n == 3 && (z->rgb == 3 || (z->app14_color_transform == 0 && !z->jfif)); + if (z->s->img_n == 3 && n < 3 && !is_rgb) + decode_n = 1; + else + decode_n = z->s->img_n; + + { + int k; + unsigned int i,j; + stbi_uc *output; + stbi_uc *coutput[4] = { NULL, NULL, NULL, NULL }; + stbi__resample res_comp[4]; + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + + + z->img_comp[k].linebuf = (stbi_uc *) stbi__malloc(z->s->img_x + 3); + if (!z->img_comp[k].linebuf) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + r->hs = z->img_h_max / z->img_comp[k].h; + r->vs = z->img_v_max / z->img_comp[k].v; + r->ystep = r->vs >> 1; + r->w_lores = (z->s->img_x + r->hs-1) / r->hs; + r->ypos = 0; + r->line0 = r->line1 = z->img_comp[k].data; + if (r->hs == 1 && r->vs == 1) r->resample = resample_row_1; + else if (r->hs == 1 && r->vs == 2) r->resample = stbi__resample_row_v_2; + else if (r->hs == 2 && r->vs == 1) r->resample = stbi__resample_row_h_2; + else if (r->hs == 2 && r->vs == 2) r->resample = z->resample_row_hv_2_kernel; + else r->resample = stbi__resample_row_generic; + } + + output = (stbi_uc *) stbi__malloc_mad3(n, z->s->img_x, z->s->img_y, 1); + if (!output) { stbi__cleanup_jpeg(z); return stbi__errpuc("outofmem", "Out of memory"); } + + for (j=0; j < z->s->img_y; ++j) { + stbi_uc *out = output + n * z->s->img_x * j; + for (k=0; k < decode_n; ++k) { + stbi__resample *r = &res_comp[k]; + int y_bot = r->ystep >= (r->vs >> 1); + coutput[k] = r->resample(z->img_comp[k].linebuf, + y_bot ? r->line1 : r->line0, + y_bot ? r->line0 : r->line1, + r->w_lores, r->hs); + if (++r->ystep >= r->vs) { + r->ystep = 0; + r->line0 = r->line1; + if (++r->ypos < z->img_comp[k].y) + r->line1 += z->img_comp[k].w2; + } + } + if (n >= 3) { + stbi_uc *y = coutput[0]; + if (z->s->img_n == 3) { + if (is_rgb) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = y[i]; + out[1] = coutput[1][i]; + out[2] = coutput[2][i]; + out[3] = 255; + out += n; + } + } else { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else if (z->s->img_n == 4) { + if (z->app14_color_transform == 0) { + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(coutput[0][i], m); + out[1] = stbi__blinn_8x8(coutput[1][i], m); + out[2] = stbi__blinn_8x8(coutput[2][i], m); + out[3] = 255; + out += n; + } + } else if (z->app14_color_transform == 2) { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + out[0] = stbi__blinn_8x8(255 - out[0], m); + out[1] = stbi__blinn_8x8(255 - out[1], m); + out[2] = stbi__blinn_8x8(255 - out[2], m); + out += n; + } + } else { + z->YCbCr_to_RGB_kernel(out, y, coutput[1], coutput[2], z->s->img_x, n); + } + } else + for (i=0; i < z->s->img_x; ++i) { + out[0] = out[1] = out[2] = y[i]; + out[3] = 255; + out += n; + } + } else { + if (is_rgb) { + if (n == 1) + for (i=0; i < z->s->img_x; ++i) + *out++ = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + else { + for (i=0; i < z->s->img_x; ++i, out += 2) { + out[0] = stbi__compute_y(coutput[0][i], coutput[1][i], coutput[2][i]); + out[1] = 255; + } + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 0) { + for (i=0; i < z->s->img_x; ++i) { + stbi_uc m = coutput[3][i]; + stbi_uc r = stbi__blinn_8x8(coutput[0][i], m); + stbi_uc g = stbi__blinn_8x8(coutput[1][i], m); + stbi_uc b = stbi__blinn_8x8(coutput[2][i], m); + out[0] = stbi__compute_y(r, g, b); + out[1] = 255; + out += n; + } + } else if (z->s->img_n == 4 && z->app14_color_transform == 2) { + for (i=0; i < z->s->img_x; ++i) { + out[0] = stbi__blinn_8x8(255 - coutput[0][i], coutput[3][i]); + out[1] = 255; + out += n; + } + } else { + stbi_uc *y = coutput[0]; + if (n == 1) + for (i=0; i < z->s->img_x; ++i) out[i] = y[i]; + else + for (i=0; i < z->s->img_x; ++i) { *out++ = y[i]; *out++ = 255; } + } + } + } + stbi__cleanup_jpeg(z); + *out_x = z->s->img_x; + *out_y = z->s->img_y; + if (comp) *comp = z->s->img_n >= 3 ? 3 : 1; + return output; + } +} +static void *stbi__jpeg_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + unsigned char* result; + stbi__jpeg* j = (stbi__jpeg*) stbi__malloc(sizeof(stbi__jpeg)); + STBI_NOTUSED(ri); + j->s = s; + stbi__setup_jpeg(j); + result = load_jpeg_image(j, x,y,comp,req_comp); + STBI_FREE(j); + return result; +} +static int stbi__jpeg_test(stbi__context *s) +{ + int r; + stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg)); + j->s = s; + stbi__setup_jpeg(j); + r = stbi__decode_jpeg_header(j, STBI__SCAN_type); + stbi__rewind(s); + STBI_FREE(j); + return r; +} +static int stbi__jpeg_info_raw(stbi__jpeg *j, int *x, int *y, int *comp) +{ + if (!stbi__decode_jpeg_header(j, STBI__SCAN_header)) { + stbi__rewind( j->s ); + return 0; + } + if (x) *x = j->s->img_x; + if (y) *y = j->s->img_y; + if (comp) *comp = j->s->img_n >= 3 ? 3 : 1; + return 1; +} +static int stbi__jpeg_info(stbi__context *s, int *x, int *y, int *comp) +{ + int result; + stbi__jpeg* j = (stbi__jpeg*) (stbi__malloc(sizeof(stbi__jpeg))); + j->s = s; + result = stbi__jpeg_info_raw(j, x, y, comp); + STBI_FREE(j); + return result; +} +#endif +#ifndef STBI_NO_ZLIB +#define STBI__ZFAST_BITS 9 +#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1) +typedef struct +{ + stbi__uint16 fast[1 << STBI__ZFAST_BITS]; + stbi__uint16 firstcode[16]; + int maxcode[17]; + stbi__uint16 firstsymbol[16]; + stbi_uc size[288]; + stbi__uint16 value[288]; +} stbi__zhuffman; +stbi_inline static int stbi__bitreverse16(int n) +{ + n = ((n & 0xAAAA) >> 1) | ((n & 0x5555) << 1); + n = ((n & 0xCCCC) >> 2) | ((n & 0x3333) << 2); + n = ((n & 0xF0F0) >> 4) | ((n & 0x0F0F) << 4); + n = ((n & 0xFF00) >> 8) | ((n & 0x00FF) << 8); + return n; +} +stbi_inline static int stbi__bit_reverse(int v, int bits) +{ + STBI_ASSERT(bits <= 16); + + + return stbi__bitreverse16(v) >> (16-bits); +} +static int stbi__zbuild_huffman(stbi__zhuffman *z, const stbi_uc *sizelist, int num) +{ + int i,k=0; + int code, next_code[16], sizes[17]; + + memset(sizes, 0, sizeof(sizes)); + memset(z->fast, 0, sizeof(z->fast)); + for (i=0; i < num; ++i) + ++sizes[sizelist[i]]; + sizes[0] = 0; + for (i=1; i < 16; ++i) + if (sizes[i] > (1 << i)) + return stbi__err("bad sizes", "Corrupt PNG"); + code = 0; + for (i=1; i < 16; ++i) { + next_code[i] = code; + z->firstcode[i] = (stbi__uint16) code; + z->firstsymbol[i] = (stbi__uint16) k; + code = (code + sizes[i]); + if (sizes[i]) + if (code-1 >= (1 << i)) return stbi__err("bad codelengths","Corrupt PNG"); + z->maxcode[i] = code << (16-i); + code <<= 1; + k += sizes[i]; + } + z->maxcode[16] = 0x10000; + for (i=0; i < num; ++i) { + int s = sizelist[i]; + if (s) { + int c = next_code[s] - z->firstcode[s] + z->firstsymbol[s]; + stbi__uint16 fastv = (stbi__uint16) ((s << 9) | i); + z->size [c] = (stbi_uc ) s; + z->value[c] = (stbi__uint16) i; + if (s <= STBI__ZFAST_BITS) { + int j = stbi__bit_reverse(next_code[s],s); + while (j < (1 << STBI__ZFAST_BITS)) { + z->fast[j] = fastv; + j += (1 << s); + } + } + ++next_code[s]; + } + } + return 1; +} +typedef struct +{ + stbi_uc *zbuffer, *zbuffer_end; + int num_bits; + stbi__uint32 code_buffer; + char *zout; + char *zout_start; + char *zout_end; + int z_expandable; + stbi__zhuffman z_length, z_distance; +} stbi__zbuf; +stbi_inline static stbi_uc stbi__zget8(stbi__zbuf *z) +{ + if (z->zbuffer >= z->zbuffer_end) return 0; + return *z->zbuffer++; +} +static void stbi__fill_bits(stbi__zbuf *z) +{ + do { + STBI_ASSERT(z->code_buffer < (1U << z->num_bits)); + z->code_buffer |= (unsigned int) stbi__zget8(z) << z->num_bits; + z->num_bits += 8; + } while (z->num_bits <= 24); +} +stbi_inline static unsigned int stbi__zreceive(stbi__zbuf *z, int n) +{ + unsigned int k; + if (z->num_bits < n) stbi__fill_bits(z); + k = z->code_buffer & ((1 << n) - 1); + z->code_buffer >>= n; + z->num_bits -= n; + return k; +} +static int stbi__zhuffman_decode_slowpath(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s,k; + + + k = stbi__bit_reverse(a->code_buffer, 16); + for (s=STBI__ZFAST_BITS+1; ; ++s) + if (k < z->maxcode[s]) + break; + if (s == 16) return -1; + + b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s]; + STBI_ASSERT(z->size[b] == s); + a->code_buffer >>= s; + a->num_bits -= s; + return z->value[b]; +} +stbi_inline static int stbi__zhuffman_decode(stbi__zbuf *a, stbi__zhuffman *z) +{ + int b,s; + if (a->num_bits < 16) stbi__fill_bits(a); + b = z->fast[a->code_buffer & STBI__ZFAST_MASK]; + if (b) { + s = b >> 9; + a->code_buffer >>= s; + a->num_bits -= s; + return b & 511; + } + return stbi__zhuffman_decode_slowpath(a, z); +} +static int stbi__zexpand(stbi__zbuf *z, char *zout, int n) +{ + char *q; + int cur, limit, old_limit; + z->zout = zout; + if (!z->z_expandable) return stbi__err("output buffer limit","Corrupt PNG"); + cur = (int) (z->zout - z->zout_start); + limit = old_limit = (int) (z->zout_end - z->zout_start); + while (cur + n > limit) + limit *= 2; + q = (char *) STBI_REALLOC_SIZED(z->zout_start, old_limit, limit); + STBI_NOTUSED(old_limit); + if (q == NULL) return stbi__err("outofmem", "Out of memory"); + z->zout_start = q; + z->zout = q + cur; + z->zout_end = q + limit; + return 1; +} +static const int stbi__zlength_base[31] = { + 3,4,5,6,7,8,9,10,11,13, + 15,17,19,23,27,31,35,43,51,59, + 67,83,99,115,131,163,195,227,258,0,0 }; +static const int stbi__zlength_extra[31]= +{ 0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0,0,0 }; +static const int stbi__zdist_base[32] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193, +257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577,0,0}; +static const int stbi__zdist_extra[32] = +{ 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13}; +static int stbi__parse_huffman_block(stbi__zbuf *a) +{ + char *zout = a->zout; + for(;;) { + int z = stbi__zhuffman_decode(a, &a->z_length); + if (z < 256) { + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); + if (zout >= a->zout_end) { + if (!stbi__zexpand(a, zout, 1)) return 0; + zout = a->zout; + } + *zout++ = (char) z; + } else { + stbi_uc *p; + int len,dist; + if (z == 256) { + a->zout = zout; + return 1; + } + z -= 257; + len = stbi__zlength_base[z]; + if (stbi__zlength_extra[z]) len += stbi__zreceive(a, stbi__zlength_extra[z]); + z = stbi__zhuffman_decode(a, &a->z_distance); + if (z < 0) return stbi__err("bad huffman code","Corrupt PNG"); + dist = stbi__zdist_base[z]; + if (stbi__zdist_extra[z]) dist += stbi__zreceive(a, stbi__zdist_extra[z]); + if (zout - a->zout_start < dist) return stbi__err("bad dist","Corrupt PNG"); + if (zout + len > a->zout_end) { + if (!stbi__zexpand(a, zout, len)) return 0; + zout = a->zout; + } + p = (stbi_uc *) (zout - dist); + if (dist == 1) { + stbi_uc v = *p; + if (len) { do *zout++ = v; while (--len); } + } else { + if (len) { do *zout++ = *p++; while (--len); } + } + } + } +} +static int stbi__compute_huffman_codes(stbi__zbuf *a) +{ + static const stbi_uc length_dezigzag[19] = { 16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15 }; + stbi__zhuffman z_codelength; + stbi_uc lencodes[286+32+137]; + stbi_uc codelength_sizes[19]; + int i,n; + int hlit = stbi__zreceive(a,5) + 257; + int hdist = stbi__zreceive(a,5) + 1; + int hclen = stbi__zreceive(a,4) + 4; + int ntot = hlit + hdist; + memset(codelength_sizes, 0, sizeof(codelength_sizes)); + for (i=0; i < hclen; ++i) { + int s = stbi__zreceive(a,3); + codelength_sizes[length_dezigzag[i]] = (stbi_uc) s; + } + if (!stbi__zbuild_huffman(&z_codelength, codelength_sizes, 19)) return 0; + n = 0; + while (n < ntot) { + int c = stbi__zhuffman_decode(a, &z_codelength); + if (c < 0 || c >= 19) return stbi__err("bad codelengths", "Corrupt PNG"); + if (c < 16) + lencodes[n++] = (stbi_uc) c; + else { + stbi_uc fill = 0; + if (c == 16) { + c = stbi__zreceive(a,2)+3; + if (n == 0) return stbi__err("bad codelengths", "Corrupt PNG"); + fill = lencodes[n-1]; + } else if (c == 17) + c = stbi__zreceive(a,3)+3; + else { + STBI_ASSERT(c == 18); + c = stbi__zreceive(a,7)+11; + } + if (ntot - n < c) return stbi__err("bad codelengths", "Corrupt PNG"); + memset(lencodes+n, fill, c); + n += c; + } + } + if (n != ntot) return stbi__err("bad codelengths","Corrupt PNG"); + if (!stbi__zbuild_huffman(&a->z_length, lencodes, hlit)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, lencodes+hlit, hdist)) return 0; + return 1; +} +static int stbi__parse_uncompressed_block(stbi__zbuf *a) +{ + stbi_uc header[4]; + int len,nlen,k; + if (a->num_bits & 7) + stbi__zreceive(a, a->num_bits & 7); + + k = 0; + while (a->num_bits > 0) { + header[k++] = (stbi_uc) (a->code_buffer & 255); + a->code_buffer >>= 8; + a->num_bits -= 8; + } + STBI_ASSERT(a->num_bits == 0); + + while (k < 4) + header[k++] = stbi__zget8(a); + len = header[1] * 256 + header[0]; + nlen = header[3] * 256 + header[2]; + if (nlen != (len ^ 0xffff)) return stbi__err("zlib corrupt","Corrupt PNG"); + if (a->zbuffer + len > a->zbuffer_end) return stbi__err("read past buffer","Corrupt PNG"); + if (a->zout + len > a->zout_end) + if (!stbi__zexpand(a, a->zout, len)) return 0; + memcpy(a->zout, a->zbuffer, len); + a->zbuffer += len; + a->zout += len; + return 1; +} +static int stbi__parse_zlib_header(stbi__zbuf *a) +{ + int cmf = stbi__zget8(a); + int cm = cmf & 15; + + int flg = stbi__zget8(a); + if ((cmf*256+flg) % 31 != 0) return stbi__err("bad zlib header","Corrupt PNG"); + if (flg & 32) return stbi__err("no preset dict","Corrupt PNG"); + if (cm != 8) return stbi__err("bad compression","Corrupt PNG"); + + return 1; +} +static const stbi_uc stbi__zdefault_length[288] = +{ + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, + 8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, 9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, 7,7,7,7,7,7,7,7,8,8,8,8,8,8,8,8 +}; +static const stbi_uc stbi__zdefault_distance[32] = +{ + 5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5 +}; +static int stbi__parse_zlib(stbi__zbuf *a, int parse_header) +{ + int final, type; + if (parse_header) + if (!stbi__parse_zlib_header(a)) return 0; + a->num_bits = 0; + a->code_buffer = 0; + do { + final = stbi__zreceive(a,1); + type = stbi__zreceive(a,2); + if (type == 0) { + if (!stbi__parse_uncompressed_block(a)) return 0; + } else if (type == 3) { + return 0; + } else { + if (type == 1) { + + if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , 288)) return 0; + if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) return 0; + } else { + if (!stbi__compute_huffman_codes(a)) return 0; + } + if (!stbi__parse_huffman_block(a)) return 0; + } + } while (!final); + return 1; +} +static int stbi__do_zlib(stbi__zbuf *a, char *obuf, int olen, int exp, int parse_header) +{ + a->zout_start = obuf; + a->zout = obuf; + a->zout_end = obuf + olen; + a->z_expandable = exp; + return stbi__parse_zlib(a, parse_header); +} +STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, 1)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} +STBIDEF char *stbi_zlib_decode_malloc(char const *buffer, int len, int *outlen) +{ + return stbi_zlib_decode_malloc_guesssize(buffer, len, 16384, outlen); +} +STBIDEF char *stbi_zlib_decode_malloc_guesssize_headerflag(const char *buffer, int len, int initial_size, int *outlen, int parse_header) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(initial_size); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer + len; + if (stbi__do_zlib(&a, p, initial_size, 1, parse_header)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} +STBIDEF int stbi_zlib_decode_buffer(char *obuffer, int olen, char const *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 1)) + return (int) (a.zout - a.zout_start); + else + return -1; +} +STBIDEF char *stbi_zlib_decode_noheader_malloc(char const *buffer, int len, int *outlen) +{ + stbi__zbuf a; + char *p = (char *) stbi__malloc(16384); + if (p == NULL) return NULL; + a.zbuffer = (stbi_uc *) buffer; + a.zbuffer_end = (stbi_uc *) buffer+len; + if (stbi__do_zlib(&a, p, 16384, 1, 0)) { + if (outlen) *outlen = (int) (a.zout - a.zout_start); + return a.zout_start; + } else { + STBI_FREE(a.zout_start); + return NULL; + } +} +STBIDEF int stbi_zlib_decode_noheader_buffer(char *obuffer, int olen, const char *ibuffer, int ilen) +{ + stbi__zbuf a; + a.zbuffer = (stbi_uc *) ibuffer; + a.zbuffer_end = (stbi_uc *) ibuffer + ilen; + if (stbi__do_zlib(&a, obuffer, olen, 0, 0)) + return (int) (a.zout - a.zout_start); + else + return -1; +} +#endif +#ifndef STBI_NO_PNG +typedef struct +{ + stbi__uint32 length; + stbi__uint32 type; +} stbi__pngchunk; +static stbi__pngchunk stbi__get_chunk_header(stbi__context *s) +{ + stbi__pngchunk c; + c.length = stbi__get32be(s); + c.type = stbi__get32be(s); + return c; +} +static int stbi__check_png_header(stbi__context *s) +{ + static const stbi_uc png_sig[8] = { 137,80,78,71,13,10,26,10 }; + int i; + for (i=0; i < 8; ++i) + if (stbi__get8(s) != png_sig[i]) return stbi__err("bad png sig","Not a PNG"); + return 1; +} +typedef struct +{ + stbi__context *s; + stbi_uc *idata, *expanded, *out; + int depth; +} stbi__png; +enum { + STBI__F_none=0, + STBI__F_sub=1, + STBI__F_up=2, + STBI__F_avg=3, + STBI__F_paeth=4, + + STBI__F_avg_first, + STBI__F_paeth_first +}; +static stbi_uc first_row_filter[5] = +{ + STBI__F_none, + STBI__F_sub, + STBI__F_none, + STBI__F_avg_first, + STBI__F_paeth_first +}; +static int stbi__paeth(int a, int b, int c) +{ + int p = a + b - c; + int pa = abs(p-a); + int pb = abs(p-b); + int pc = abs(p-c); + if (pa <= pb && pa <= pc) return a; + if (pb <= pc) return b; + return c; +} +static const stbi_uc stbi__depth_scale_table[9] = { 0, 0xff, 0x55, 0, 0x11, 0,0,0, 0x01 }; +static int stbi__create_png_image_raw(stbi__png *a, stbi_uc *raw, stbi__uint32 raw_len, int out_n, stbi__uint32 x, stbi__uint32 y, int depth, int color) +{ + int bytes = (depth == 16? 2 : 1); + stbi__context *s = a->s; + stbi__uint32 i,j,stride = x*out_n*bytes; + stbi__uint32 img_len, img_width_bytes; + int k; + int img_n = s->img_n; + int output_bytes = out_n*bytes; + int filter_bytes = img_n*bytes; + int width = x; + STBI_ASSERT(out_n == s->img_n || out_n == s->img_n+1); + a->out = (stbi_uc *) stbi__malloc_mad3(x, y, output_bytes, 0); + if (!a->out) return stbi__err("outofmem", "Out of memory"); + if (!stbi__mad3sizes_valid(img_n, x, depth, 7)) return stbi__err("too large", "Corrupt PNG"); + img_width_bytes = (((img_n * x * depth) + 7) >> 3); + img_len = (img_width_bytes + 1) * y; + + + + if (raw_len < img_len) return stbi__err("not enough pixels","Corrupt PNG"); + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *prior; + int filter = *raw++; + if (filter > 4) + return stbi__err("invalid filter","Corrupt PNG"); + if (depth < 8) { + STBI_ASSERT(img_width_bytes <= x); + cur += x*out_n - img_width_bytes; + filter_bytes = 1; + width = img_width_bytes; + } + prior = cur - stride; + + if (j == 0) filter = first_row_filter[filter]; + + for (k=0; k < filter_bytes; ++k) { + switch (filter) { + case STBI__F_none : cur[k] = raw[k]; break; + case STBI__F_sub : cur[k] = raw[k]; break; + case STBI__F_up : cur[k] = STBI__BYTECAST(raw[k] + prior[k]); break; + case STBI__F_avg : cur[k] = STBI__BYTECAST(raw[k] + (prior[k]>>1)); break; + case STBI__F_paeth : cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(0,prior[k],0)); break; + case STBI__F_avg_first : cur[k] = raw[k]; break; + case STBI__F_paeth_first: cur[k] = raw[k]; break; + } + } + if (depth == 8) { + if (img_n != out_n) + cur[img_n] = 255; + raw += img_n; + cur += out_n; + prior += out_n; + } else if (depth == 16) { + if (img_n != out_n) { + cur[filter_bytes] = 255; + cur[filter_bytes+1] = 255; + } + raw += filter_bytes; + cur += output_bytes; + prior += output_bytes; + } else { + raw += 1; + cur += 1; + prior += 1; + } + + if (depth < 8 || img_n == out_n) { + int nk = (width - 1)*filter_bytes; + #define STBI__CASE(f) \ + case f: \ + for (k=0; k < nk; ++k) + switch (filter) { + + case STBI__F_none: memcpy(cur, raw, nk); break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k-filter_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k-filter_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],prior[k],prior[k-filter_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k-filter_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k-filter_bytes],0,0)); } break; + } + #undef STBI__CASE + raw += nk; + } else { + STBI_ASSERT(img_n+1 == out_n); + #define STBI__CASE(f) \ + case f: \ + for (i=x-1; i >= 1; --i, cur[filter_bytes]=255,raw+=filter_bytes,cur+=output_bytes,prior+=output_bytes) \ + for (k=0; k < filter_bytes; ++k) + switch (filter) { + STBI__CASE(STBI__F_none) { cur[k] = raw[k]; } break; + STBI__CASE(STBI__F_sub) { cur[k] = STBI__BYTECAST(raw[k] + cur[k- output_bytes]); } break; + STBI__CASE(STBI__F_up) { cur[k] = STBI__BYTECAST(raw[k] + prior[k]); } break; + STBI__CASE(STBI__F_avg) { cur[k] = STBI__BYTECAST(raw[k] + ((prior[k] + cur[k- output_bytes])>>1)); } break; + STBI__CASE(STBI__F_paeth) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],prior[k],prior[k- output_bytes])); } break; + STBI__CASE(STBI__F_avg_first) { cur[k] = STBI__BYTECAST(raw[k] + (cur[k- output_bytes] >> 1)); } break; + STBI__CASE(STBI__F_paeth_first) { cur[k] = STBI__BYTECAST(raw[k] + stbi__paeth(cur[k- output_bytes],0,0)); } break; + } + #undef STBI__CASE + + + if (depth == 16) { + cur = a->out + stride*j; + for (i=0; i < x; ++i,cur+=output_bytes) { + cur[filter_bytes+1] = 255; + } + } + } + } + + + + if (depth < 8) { + for (j=0; j < y; ++j) { + stbi_uc *cur = a->out + stride*j; + stbi_uc *in = a->out + stride*j + x*out_n - img_width_bytes; + + + stbi_uc scale = (color == 0) ? stbi__depth_scale_table[depth] : 1; + + + + + + if (depth == 4) { + for (k=x*img_n; k >= 2; k-=2, ++in) { + *cur++ = scale * ((*in >> 4) ); + *cur++ = scale * ((*in ) & 0x0f); + } + if (k > 0) *cur++ = scale * ((*in >> 4) ); + } else if (depth == 2) { + for (k=x*img_n; k >= 4; k-=4, ++in) { + *cur++ = scale * ((*in >> 6) ); + *cur++ = scale * ((*in >> 4) & 0x03); + *cur++ = scale * ((*in >> 2) & 0x03); + *cur++ = scale * ((*in ) & 0x03); + } + if (k > 0) *cur++ = scale * ((*in >> 6) ); + if (k > 1) *cur++ = scale * ((*in >> 4) & 0x03); + if (k > 2) *cur++ = scale * ((*in >> 2) & 0x03); + } else if (depth == 1) { + for (k=x*img_n; k >= 8; k-=8, ++in) { + *cur++ = scale * ((*in >> 7) ); + *cur++ = scale * ((*in >> 6) & 0x01); + *cur++ = scale * ((*in >> 5) & 0x01); + *cur++ = scale * ((*in >> 4) & 0x01); + *cur++ = scale * ((*in >> 3) & 0x01); + *cur++ = scale * ((*in >> 2) & 0x01); + *cur++ = scale * ((*in >> 1) & 0x01); + *cur++ = scale * ((*in ) & 0x01); + } + if (k > 0) *cur++ = scale * ((*in >> 7) ); + if (k > 1) *cur++ = scale * ((*in >> 6) & 0x01); + if (k > 2) *cur++ = scale * ((*in >> 5) & 0x01); + if (k > 3) *cur++ = scale * ((*in >> 4) & 0x01); + if (k > 4) *cur++ = scale * ((*in >> 3) & 0x01); + if (k > 5) *cur++ = scale * ((*in >> 2) & 0x01); + if (k > 6) *cur++ = scale * ((*in >> 1) & 0x01); + } + if (img_n != out_n) { + int q; + + cur = a->out + stride*j; + if (img_n == 1) { + for (q=x-1; q >= 0; --q) { + cur[q*2+1] = 255; + cur[q*2+0] = cur[q]; + } + } else { + STBI_ASSERT(img_n == 3); + for (q=x-1; q >= 0; --q) { + cur[q*4+3] = 255; + cur[q*4+2] = cur[q*3+2]; + cur[q*4+1] = cur[q*3+1]; + cur[q*4+0] = cur[q*3+0]; + } + } + } + } + } else if (depth == 16) { + + + + + stbi_uc *cur = a->out; + stbi__uint16 *cur16 = (stbi__uint16*)cur; + for(i=0; i < x*y*out_n; ++i,cur16++,cur+=2) { + *cur16 = (cur[0] << 8) | cur[1]; + } + } + return 1; +} +static int stbi__create_png_image(stbi__png *a, stbi_uc *image_data, stbi__uint32 image_data_len, int out_n, int depth, int color, int interlaced) +{ + int bytes = (depth == 16 ? 2 : 1); + int out_bytes = out_n * bytes; + stbi_uc *final; + int p; + if (!interlaced) + return stbi__create_png_image_raw(a, image_data, image_data_len, out_n, a->s->img_x, a->s->img_y, depth, color); + + final = (stbi_uc *) stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0); + for (p=0; p < 7; ++p) { + int xorig[] = { 0,4,0,2,0,1,0 }; + int yorig[] = { 0,0,4,0,2,0,1 }; + int xspc[] = { 8,8,4,4,2,2,1 }; + int yspc[] = { 8,8,8,4,4,2,2 }; + int i,j,x,y; + + x = (a->s->img_x - xorig[p] + xspc[p]-1) / xspc[p]; + y = (a->s->img_y - yorig[p] + yspc[p]-1) / yspc[p]; + if (x && y) { + stbi__uint32 img_len = ((((a->s->img_n * x * depth) + 7) >> 3) + 1) * y; + if (!stbi__create_png_image_raw(a, image_data, image_data_len, out_n, x, y, depth, color)) { + STBI_FREE(final); + return 0; + } + for (j=0; j < y; ++j) { + for (i=0; i < x; ++i) { + int out_y = j*yspc[p]+yorig[p]; + int out_x = i*xspc[p]+xorig[p]; + memcpy(final + out_y*a->s->img_x*out_bytes + out_x*out_bytes, + a->out + (j*x+i)*out_bytes, out_bytes); + } + } + STBI_FREE(a->out); + image_data += img_len; + image_data_len -= img_len; + } + } + a->out = final; + return 1; +} +static int stbi__compute_transparency(stbi__png *z, stbi_uc tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + + + STBI_ASSERT(out_n == 2 || out_n == 4); + if (out_n == 2) { + for (i=0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 255); + p += 2; + } + } else { + for (i=0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} +static int stbi__compute_transparency16(stbi__png *z, stbi__uint16 tc[3], int out_n) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi__uint16 *p = (stbi__uint16*) z->out; + + + STBI_ASSERT(out_n == 2 || out_n == 4); + if (out_n == 2) { + for (i = 0; i < pixel_count; ++i) { + p[1] = (p[0] == tc[0] ? 0 : 65535); + p += 2; + } + } else { + for (i = 0; i < pixel_count; ++i) { + if (p[0] == tc[0] && p[1] == tc[1] && p[2] == tc[2]) + p[3] = 0; + p += 4; + } + } + return 1; +} +static int stbi__expand_png_palette(stbi__png *a, stbi_uc *palette, int len, int pal_img_n) +{ + stbi__uint32 i, pixel_count = a->s->img_x * a->s->img_y; + stbi_uc *p, *temp_out, *orig = a->out; + p = (stbi_uc *) stbi__malloc_mad2(pixel_count, pal_img_n, 0); + if (p == NULL) return stbi__err("outofmem", "Out of memory"); + + temp_out = p; + if (pal_img_n == 3) { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p += 3; + } + } else { + for (i=0; i < pixel_count; ++i) { + int n = orig[i]*4; + p[0] = palette[n ]; + p[1] = palette[n+1]; + p[2] = palette[n+2]; + p[3] = palette[n+3]; + p += 4; + } + } + STBI_FREE(a->out); + a->out = temp_out; + STBI_NOTUSED(len); + return 1; +} +static int stbi__unpremultiply_on_load = 0; +static int stbi__de_iphone_flag = 0; +STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply) +{ + stbi__unpremultiply_on_load = flag_true_if_should_unpremultiply; +} +STBIDEF void stbi_convert_iphone_png_to_rgb(int flag_true_if_should_convert) +{ + stbi__de_iphone_flag = flag_true_if_should_convert; +} +static void stbi__de_iphone(stbi__png *z) +{ + stbi__context *s = z->s; + stbi__uint32 i, pixel_count = s->img_x * s->img_y; + stbi_uc *p = z->out; + if (s->img_out_n == 3) { + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 3; + } + } else { + STBI_ASSERT(s->img_out_n == 4); + if (stbi__unpremultiply_on_load) { + + for (i=0; i < pixel_count; ++i) { + stbi_uc a = p[3]; + stbi_uc t = p[0]; + if (a) { + stbi_uc half = a / 2; + p[0] = (p[2] * 255 + half) / a; + p[1] = (p[1] * 255 + half) / a; + p[2] = ( t * 255 + half) / a; + } else { + p[0] = p[2]; + p[2] = t; + } + p += 4; + } + } else { + + for (i=0; i < pixel_count; ++i) { + stbi_uc t = p[0]; + p[0] = p[2]; + p[2] = t; + p += 4; + } + } + } +} +#define STBI__PNG_TYPE(a,b,c,d) (((unsigned) (a) << 24) + ((unsigned) (b) << 16) + ((unsigned) (c) << 8) + (unsigned) (d)) +static int stbi__parse_png_file(stbi__png *z, int scan, int req_comp) +{ + stbi_uc palette[1024], pal_img_n=0; + stbi_uc has_trans=0, tc[3]={0}; + stbi__uint16 tc16[3]; + stbi__uint32 ioff=0, idata_limit=0, i, pal_len=0; + int first=1,k,interlace=0, color=0, is_iphone=0; + stbi__context *s = z->s; + z->expanded = NULL; + z->idata = NULL; + z->out = NULL; + if (!stbi__check_png_header(s)) return 0; + if (scan == STBI__SCAN_type) return 1; + for (;;) { + stbi__pngchunk c = stbi__get_chunk_header(s); + switch (c.type) { + case STBI__PNG_TYPE('C','g','B','I'): + is_iphone = 1; + stbi__skip(s, c.length); + break; + case STBI__PNG_TYPE('I','H','D','R'): { + int comp,filter; + if (!first) return stbi__err("multiple IHDR","Corrupt PNG"); + first = 0; + if (c.length != 13) return stbi__err("bad IHDR len","Corrupt PNG"); + s->img_x = stbi__get32be(s); if (s->img_x > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + s->img_y = stbi__get32be(s); if (s->img_y > (1 << 24)) return stbi__err("too large","Very large image (corrupt?)"); + z->depth = stbi__get8(s); if (z->depth != 1 && z->depth != 2 && z->depth != 4 && z->depth != 8 && z->depth != 16) return stbi__err("1/2/4/8/16-bit only","PNG not supported: 1/2/4/8/16-bit only"); + color = stbi__get8(s); if (color > 6) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3 && z->depth == 16) return stbi__err("bad ctype","Corrupt PNG"); + if (color == 3) pal_img_n = 3; else if (color & 1) return stbi__err("bad ctype","Corrupt PNG"); + comp = stbi__get8(s); if (comp) return stbi__err("bad comp method","Corrupt PNG"); + filter= stbi__get8(s); if (filter) return stbi__err("bad filter method","Corrupt PNG"); + interlace = stbi__get8(s); if (interlace>1) return stbi__err("bad interlace method","Corrupt PNG"); + if (!s->img_x || !s->img_y) return stbi__err("0-pixel image","Corrupt PNG"); + if (!pal_img_n) { + s->img_n = (color & 2 ? 3 : 1) + (color & 4 ? 1 : 0); + if ((1 << 30) / s->img_x / s->img_n < s->img_y) return stbi__err("too large", "Image too large to decode"); + if (scan == STBI__SCAN_header) return 1; + } else { + + + s->img_n = 1; + if ((1 << 30) / s->img_x / 4 < s->img_y) return stbi__err("too large","Corrupt PNG"); + + } + break; + } + case STBI__PNG_TYPE('P','L','T','E'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (c.length > 256*3) return stbi__err("invalid PLTE","Corrupt PNG"); + pal_len = c.length / 3; + if (pal_len * 3 != c.length) return stbi__err("invalid PLTE","Corrupt PNG"); + for (i=0; i < pal_len; ++i) { + palette[i*4+0] = stbi__get8(s); + palette[i*4+1] = stbi__get8(s); + palette[i*4+2] = stbi__get8(s); + palette[i*4+3] = 255; + } + break; + } + case STBI__PNG_TYPE('t','R','N','S'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (z->idata) return stbi__err("tRNS after IDAT","Corrupt PNG"); + if (pal_img_n) { + if (scan == STBI__SCAN_header) { s->img_n = 4; return 1; } + if (pal_len == 0) return stbi__err("tRNS before PLTE","Corrupt PNG"); + if (c.length > pal_len) return stbi__err("bad tRNS len","Corrupt PNG"); + pal_img_n = 4; + for (i=0; i < c.length; ++i) + palette[i*4+3] = stbi__get8(s); + } else { + if (!(s->img_n & 1)) return stbi__err("tRNS with alpha","Corrupt PNG"); + if (c.length != (stbi__uint32) s->img_n*2) return stbi__err("bad tRNS len","Corrupt PNG"); + has_trans = 1; + if (z->depth == 16) { + for (k = 0; k < s->img_n; ++k) tc16[k] = (stbi__uint16)stbi__get16be(s); + } else { + for (k = 0; k < s->img_n; ++k) tc[k] = (stbi_uc)(stbi__get16be(s) & 255) * stbi__depth_scale_table[z->depth]; + } + } + break; + } + case STBI__PNG_TYPE('I','D','A','T'): { + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (pal_img_n && !pal_len) return stbi__err("no PLTE","Corrupt PNG"); + if (scan == STBI__SCAN_header) { s->img_n = pal_img_n; return 1; } + if ((int)(ioff + c.length) < (int)ioff) return 0; + if (ioff + c.length > idata_limit) { + stbi__uint32 idata_limit_old = idata_limit; + stbi_uc *p; + if (idata_limit == 0) idata_limit = c.length > 4096 ? c.length : 4096; + while (ioff + c.length > idata_limit) + idata_limit *= 2; + STBI_NOTUSED(idata_limit_old); + p = (stbi_uc *) STBI_REALLOC_SIZED(z->idata, idata_limit_old, idata_limit); if (p == NULL) return stbi__err("outofmem", "Out of memory"); + z->idata = p; + } + if (!stbi__getn(s, z->idata+ioff,c.length)) return stbi__err("outofdata","Corrupt PNG"); + ioff += c.length; + break; + } + case STBI__PNG_TYPE('I','E','N','D'): { + stbi__uint32 raw_len, bpl; + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if (scan != STBI__SCAN_load) return 1; + if (z->idata == NULL) return stbi__err("no IDAT","Corrupt PNG"); + + bpl = (s->img_x * z->depth + 7) / 8; + raw_len = bpl * s->img_y * s->img_n + s->img_y ; + z->expanded = (stbi_uc *) stbi_zlib_decode_malloc_guesssize_headerflag((char *) z->idata, ioff, raw_len, (int *) &raw_len, !is_iphone); + if (z->expanded == NULL) return 0; + STBI_FREE(z->idata); z->idata = NULL; + if ((req_comp == s->img_n+1 && req_comp != 3 && !pal_img_n) || has_trans) + s->img_out_n = s->img_n+1; + else + s->img_out_n = s->img_n; + if (!stbi__create_png_image(z, z->expanded, raw_len, s->img_out_n, z->depth, color, interlace)) return 0; + if (has_trans) { + if (z->depth == 16) { + if (!stbi__compute_transparency16(z, tc16, s->img_out_n)) return 0; + } else { + if (!stbi__compute_transparency(z, tc, s->img_out_n)) return 0; + } + } + if (is_iphone && stbi__de_iphone_flag && s->img_out_n > 2) + stbi__de_iphone(z); + if (pal_img_n) { + + s->img_n = pal_img_n; + s->img_out_n = pal_img_n; + if (req_comp >= 3) s->img_out_n = req_comp; + if (!stbi__expand_png_palette(z, palette, pal_len, s->img_out_n)) + return 0; + } else if (has_trans) { + + ++s->img_n; + } + STBI_FREE(z->expanded); z->expanded = NULL; + + stbi__get32be(s); + return 1; + } + default: + + if (first) return stbi__err("first not IHDR", "Corrupt PNG"); + if ((c.type & (1 << 29)) == 0) { + #ifndef STBI_NO_FAILURE_STRINGS + + static char invalid_chunk[] = "XXXX PNG chunk not known"; + invalid_chunk[0] = STBI__BYTECAST(c.type >> 24); + invalid_chunk[1] = STBI__BYTECAST(c.type >> 16); + invalid_chunk[2] = STBI__BYTECAST(c.type >> 8); + invalid_chunk[3] = STBI__BYTECAST(c.type >> 0); + #endif + return stbi__err(invalid_chunk, "PNG not supported: unknown PNG chunk type"); + } + stbi__skip(s, c.length); + break; + } + + stbi__get32be(s); + } +} +static void *stbi__do_png(stbi__png *p, int *x, int *y, int *n, int req_comp, stbi__result_info *ri) +{ + void *result=NULL; + if (req_comp < 0 || req_comp > 4) return stbi__errpuc("bad req_comp", "Internal error"); + if (stbi__parse_png_file(p, STBI__SCAN_load, req_comp)) { + if (p->depth < 8) + ri->bits_per_channel = 8; + else + ri->bits_per_channel = p->depth; + result = p->out; + p->out = NULL; + if (req_comp && req_comp != p->s->img_out_n) { + if (ri->bits_per_channel == 8) + result = stbi__convert_format((unsigned char *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + else + result = stbi__convert_format16((stbi__uint16 *) result, p->s->img_out_n, req_comp, p->s->img_x, p->s->img_y); + p->s->img_out_n = req_comp; + if (result == NULL) return result; + } + *x = p->s->img_x; + *y = p->s->img_y; + if (n) *n = p->s->img_n; + } + STBI_FREE(p->out); p->out = NULL; + STBI_FREE(p->expanded); p->expanded = NULL; + STBI_FREE(p->idata); p->idata = NULL; + return result; +} +static void *stbi__png_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi__png p; + p.s = s; + return stbi__do_png(&p, x,y,comp,req_comp, ri); +} +static int stbi__png_test(stbi__context *s) +{ + int r; + r = stbi__check_png_header(s); + stbi__rewind(s); + return r; +} +static int stbi__png_info_raw(stbi__png *p, int *x, int *y, int *comp) +{ + if (!stbi__parse_png_file(p, STBI__SCAN_header, 0)) { + stbi__rewind( p->s ); + return 0; + } + if (x) *x = p->s->img_x; + if (y) *y = p->s->img_y; + if (comp) *comp = p->s->img_n; + return 1; +} +static int stbi__png_info(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__png p; + p.s = s; + return stbi__png_info_raw(&p, x, y, comp); +} +static int stbi__png_is16(stbi__context *s) +{ + stbi__png p; + p.s = s; + if (!stbi__png_info_raw(&p, NULL, NULL, NULL)) + return 0; + if (p.depth != 16) { + stbi__rewind(p.s); + return 0; + } + return 1; +} +#endif +#ifndef STBI_NO_BMP +static int stbi__bmp_test_raw(stbi__context *s) +{ + int r; + int sz; + if (stbi__get8(s) != 'B') return 0; + if (stbi__get8(s) != 'M') return 0; + stbi__get32le(s); + stbi__get16le(s); + stbi__get16le(s); + stbi__get32le(s); + sz = stbi__get32le(s); + r = (sz == 12 || sz == 40 || sz == 56 || sz == 108 || sz == 124); + return r; +} +static int stbi__bmp_test(stbi__context *s) +{ + int r = stbi__bmp_test_raw(s); + stbi__rewind(s); + return r; +} +static int stbi__high_bit(unsigned int z) +{ + int n=0; + if (z == 0) return -1; + if (z >= 0x10000) { n += 16; z >>= 16; } + if (z >= 0x00100) { n += 8; z >>= 8; } + if (z >= 0x00010) { n += 4; z >>= 4; } + if (z >= 0x00004) { n += 2; z >>= 2; } + if (z >= 0x00002) { n += 1; } + return n; +} +static int stbi__bitcount(unsigned int a) +{ + a = (a & 0x55555555) + ((a >> 1) & 0x55555555); + a = (a & 0x33333333) + ((a >> 2) & 0x33333333); + a = (a + (a >> 4)) & 0x0f0f0f0f; + a = (a + (a >> 8)); + a = (a + (a >> 16)); + return a & 0xff; +} +static int stbi__shiftsigned(unsigned int v, int shift, int bits) +{ + static unsigned int mul_table[9] = { + 0, + 0xff, 0x55, 0x49, 0x11, + 0x21, 0x41, 0x81, 0x01, + }; + static unsigned int shift_table[9] = { + 0, 0,0,1,0,2,4,6,0, + }; + if (shift < 0) + v <<= -shift; + else + v >>= shift; + STBI_ASSERT(v < 256); + v >>= (8-bits); + STBI_ASSERT(bits >= 0 && bits <= 8); + return (int) ((unsigned) v * mul_table[bits]) >> shift_table[bits]; +} +typedef struct +{ + int bpp, offset, hsz; + unsigned int mr,mg,mb,ma, all_a; + int extra_read; +} stbi__bmp_data; +static void *stbi__bmp_parse_header(stbi__context *s, stbi__bmp_data *info) +{ + int hsz; + if (stbi__get8(s) != 'B' || stbi__get8(s) != 'M') return stbi__errpuc("not BMP", "Corrupt BMP"); + stbi__get32le(s); + stbi__get16le(s); + stbi__get16le(s); + info->offset = stbi__get32le(s); + info->hsz = hsz = stbi__get32le(s); + info->mr = info->mg = info->mb = info->ma = 0; + info->extra_read = 14; + if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) return stbi__errpuc("unknown BMP", "BMP type not supported: unknown"); + if (hsz == 12) { + s->img_x = stbi__get16le(s); + s->img_y = stbi__get16le(s); + } else { + s->img_x = stbi__get32le(s); + s->img_y = stbi__get32le(s); + } + if (stbi__get16le(s) != 1) return stbi__errpuc("bad BMP", "bad BMP"); + info->bpp = stbi__get16le(s); + if (hsz != 12) { + int compress = stbi__get32le(s); + if (compress == 1 || compress == 2) return stbi__errpuc("BMP RLE", "BMP type not supported: RLE"); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + if (hsz == 40 || hsz == 56) { + if (hsz == 56) { + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + } + if (info->bpp == 16 || info->bpp == 32) { + if (compress == 0) { + if (info->bpp == 32) { + info->mr = 0xffu << 16; + info->mg = 0xffu << 8; + info->mb = 0xffu << 0; + info->ma = 0xffu << 24; + info->all_a = 0; + } else { + info->mr = 31u << 10; + info->mg = 31u << 5; + info->mb = 31u << 0; + } + } else if (compress == 3) { + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->extra_read += 12; + + if (info->mr == info->mg && info->mg == info->mb) { + + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else + return stbi__errpuc("bad BMP", "bad BMP"); + } + } else { + int i; + if (hsz != 108 && hsz != 124) + return stbi__errpuc("bad BMP", "bad BMP"); + info->mr = stbi__get32le(s); + info->mg = stbi__get32le(s); + info->mb = stbi__get32le(s); + info->ma = stbi__get32le(s); + stbi__get32le(s); + for (i=0; i < 12; ++i) + stbi__get32le(s); + if (hsz == 124) { + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + stbi__get32le(s); + } + } + } + return (void *) 1; +} +static void *stbi__bmp_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + unsigned int mr=0,mg=0,mb=0,ma=0, all_a; + stbi_uc pal[256][4]; + int psize=0,i,j,width; + int flip_vertically, pad, target; + stbi__bmp_data info; + STBI_NOTUSED(ri); + info.all_a = 255; + if (stbi__bmp_parse_header(s, &info) == NULL) + return NULL; + flip_vertically = ((int) s->img_y) > 0; + s->img_y = abs((int) s->img_y); + mr = info.mr; + mg = info.mg; + mb = info.mb; + ma = info.ma; + all_a = info.all_a; + if (info.hsz == 12) { + if (info.bpp < 24) + psize = (info.offset - info.extra_read - 24) / 3; + } else { + if (info.bpp < 16) + psize = (info.offset - info.extra_read - info.hsz) >> 2; + } + if (psize == 0) { + STBI_ASSERT(info.offset == (s->img_buffer - s->buffer_start)); + } + if (info.bpp == 24 && ma == 0xff000000) + s->img_n = 3; + else + s->img_n = ma ? 4 : 3; + if (req_comp && req_comp >= 3) + target = req_comp; + else + target = s->img_n; + + if (!stbi__mad3sizes_valid(target, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "Corrupt BMP"); + out = (stbi_uc *) stbi__malloc_mad3(target, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + if (info.bpp < 16) { + int z=0; + if (psize == 0 || psize > 256) { STBI_FREE(out); return stbi__errpuc("invalid", "Corrupt BMP"); } + for (i=0; i < psize; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + if (info.hsz != 12) stbi__get8(s); + pal[i][3] = 255; + } + stbi__skip(s, info.offset - info.extra_read - info.hsz - psize * (info.hsz == 12 ? 3 : 4)); + if (info.bpp == 1) width = (s->img_x + 7) >> 3; + else if (info.bpp == 4) width = (s->img_x + 1) >> 1; + else if (info.bpp == 8) width = s->img_x; + else { STBI_FREE(out); return stbi__errpuc("bad bpp", "Corrupt BMP"); } + pad = (-width)&3; + if (info.bpp == 1) { + for (j=0; j < (int) s->img_y; ++j) { + int bit_offset = 7, v = stbi__get8(s); + for (i=0; i < (int) s->img_x; ++i) { + int color = (v>>bit_offset)&0x1; + out[z++] = pal[color][0]; + out[z++] = pal[color][1]; + out[z++] = pal[color][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + if((--bit_offset) < 0) { + bit_offset = 7; + v = stbi__get8(s); + } + } + stbi__skip(s, pad); + } + } else { + for (j=0; j < (int) s->img_y; ++j) { + for (i=0; i < (int) s->img_x; i += 2) { + int v=stbi__get8(s),v2=0; + if (info.bpp == 4) { + v2 = v & 15; + v >>= 4; + } + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + if (i+1 == (int) s->img_x) break; + v = (info.bpp == 8) ? stbi__get8(s) : v2; + out[z++] = pal[v][0]; + out[z++] = pal[v][1]; + out[z++] = pal[v][2]; + if (target == 4) out[z++] = 255; + } + stbi__skip(s, pad); + } + } + } else { + int rshift=0,gshift=0,bshift=0,ashift=0,rcount=0,gcount=0,bcount=0,acount=0; + int z = 0; + int easy=0; + stbi__skip(s, info.offset - info.extra_read - info.hsz); + if (info.bpp == 24) width = 3 * s->img_x; + else if (info.bpp == 16) width = 2*s->img_x; + else width=0; + pad = (-width) & 3; + if (info.bpp == 24) { + easy = 1; + } else if (info.bpp == 32) { + if (mb == 0xff && mg == 0xff00 && mr == 0x00ff0000 && ma == 0xff000000) + easy = 2; + } + if (!easy) { + if (!mr || !mg || !mb) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); } + + rshift = stbi__high_bit(mr)-7; rcount = stbi__bitcount(mr); + gshift = stbi__high_bit(mg)-7; gcount = stbi__bitcount(mg); + bshift = stbi__high_bit(mb)-7; bcount = stbi__bitcount(mb); + ashift = stbi__high_bit(ma)-7; acount = stbi__bitcount(ma); + } + for (j=0; j < (int) s->img_y; ++j) { + if (easy) { + for (i=0; i < (int) s->img_x; ++i) { + unsigned char a; + out[z+2] = stbi__get8(s); + out[z+1] = stbi__get8(s); + out[z+0] = stbi__get8(s); + z += 3; + a = (easy == 2 ? stbi__get8(s) : 255); + all_a |= a; + if (target == 4) out[z++] = a; + } + } else { + int bpp = info.bpp; + for (i=0; i < (int) s->img_x; ++i) { + stbi__uint32 v = (bpp == 16 ? (stbi__uint32) stbi__get16le(s) : stbi__get32le(s)); + unsigned int a; + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mr, rshift, rcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mg, gshift, gcount)); + out[z++] = STBI__BYTECAST(stbi__shiftsigned(v & mb, bshift, bcount)); + a = (ma ? stbi__shiftsigned(v & ma, ashift, acount) : 255); + all_a |= a; + if (target == 4) out[z++] = STBI__BYTECAST(a); + } + } + stbi__skip(s, pad); + } + } + + if (target == 4 && all_a == 0) + for (i=4*s->img_x*s->img_y-1; i >= 0; i -= 4) + out[i] = 255; + if (flip_vertically) { + stbi_uc t; + for (j=0; j < (int) s->img_y>>1; ++j) { + stbi_uc *p1 = out + j *s->img_x*target; + stbi_uc *p2 = out + (s->img_y-1-j)*s->img_x*target; + for (i=0; i < (int) s->img_x*target; ++i) { + t = p1[i]; p1[i] = p2[i]; p2[i] = t; + } + } + } + if (req_comp && req_comp != target) { + out = stbi__convert_format(out, target, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; + } + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + return out; +} +#endif +#ifndef STBI_NO_TGA +static int stbi__tga_get_comp(int bits_per_pixel, int is_grey, int* is_rgb16) +{ + + if (is_rgb16) *is_rgb16 = 0; + switch(bits_per_pixel) { + case 8: return STBI_grey; + case 16: if(is_grey) return STBI_grey_alpha; + + case 15: if(is_rgb16) *is_rgb16 = 1; + return STBI_rgb; + case 24: + case 32: return bits_per_pixel/8; + default: return 0; + } +} +static int stbi__tga_info(stbi__context *s, int *x, int *y, int *comp) +{ + int tga_w, tga_h, tga_comp, tga_image_type, tga_bits_per_pixel, tga_colormap_bpp; + int sz, tga_colormap_type; + stbi__get8(s); + tga_colormap_type = stbi__get8(s); + if( tga_colormap_type > 1 ) { + stbi__rewind(s); + return 0; + } + tga_image_type = stbi__get8(s); + if ( tga_colormap_type == 1 ) { + if (tga_image_type != 1 && tga_image_type != 9) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); + sz = stbi__get8(s); + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,4); + tga_colormap_bpp = sz; + } else { + if ( (tga_image_type != 2) && (tga_image_type != 3) && (tga_image_type != 10) && (tga_image_type != 11) ) { + stbi__rewind(s); + return 0; + } + stbi__skip(s,9); + tga_colormap_bpp = 0; + } + tga_w = stbi__get16le(s); + if( tga_w < 1 ) { + stbi__rewind(s); + return 0; + } + tga_h = stbi__get16le(s); + if( tga_h < 1 ) { + stbi__rewind(s); + return 0; + } + tga_bits_per_pixel = stbi__get8(s); + stbi__get8(s); + if (tga_colormap_bpp != 0) { + if((tga_bits_per_pixel != 8) && (tga_bits_per_pixel != 16)) { + + + stbi__rewind(s); + return 0; + } + tga_comp = stbi__tga_get_comp(tga_colormap_bpp, 0, NULL); + } else { + tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3) || (tga_image_type == 11), NULL); + } + if(!tga_comp) { + stbi__rewind(s); + return 0; + } + if (x) *x = tga_w; + if (y) *y = tga_h; + if (comp) *comp = tga_comp; + return 1; +} +static int stbi__tga_test(stbi__context *s) +{ + int res = 0; + int sz, tga_color_type; + stbi__get8(s); + tga_color_type = stbi__get8(s); + if ( tga_color_type > 1 ) goto errorEnd; + sz = stbi__get8(s); + if ( tga_color_type == 1 ) { + if (sz != 1 && sz != 9) goto errorEnd; + stbi__skip(s,4); + sz = stbi__get8(s); + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + stbi__skip(s,4); + } else { + if ( (sz != 2) && (sz != 3) && (sz != 10) && (sz != 11) ) goto errorEnd; + stbi__skip(s,9); + } + if ( stbi__get16le(s) < 1 ) goto errorEnd; + if ( stbi__get16le(s) < 1 ) goto errorEnd; + sz = stbi__get8(s); + if ( (tga_color_type == 1) && (sz != 8) && (sz != 16) ) goto errorEnd; + if ( (sz != 8) && (sz != 15) && (sz != 16) && (sz != 24) && (sz != 32) ) goto errorEnd; + res = 1; +errorEnd: + stbi__rewind(s); + return res; +} +static void stbi__tga_read_rgb16(stbi__context *s, stbi_uc* out) +{ + stbi__uint16 px = (stbi__uint16)stbi__get16le(s); + stbi__uint16 fiveBitMask = 31; + + int r = (px >> 10) & fiveBitMask; + int g = (px >> 5) & fiveBitMask; + int b = px & fiveBitMask; + + out[0] = (stbi_uc)((r * 255)/31); + out[1] = (stbi_uc)((g * 255)/31); + out[2] = (stbi_uc)((b * 255)/31); + + + + +} +static void *stbi__tga_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + + int tga_offset = stbi__get8(s); + int tga_indexed = stbi__get8(s); + int tga_image_type = stbi__get8(s); + int tga_is_RLE = 0; + int tga_palette_start = stbi__get16le(s); + int tga_palette_len = stbi__get16le(s); + int tga_palette_bits = stbi__get8(s); + int tga_x_origin = stbi__get16le(s); + int tga_y_origin = stbi__get16le(s); + int tga_width = stbi__get16le(s); + int tga_height = stbi__get16le(s); + int tga_bits_per_pixel = stbi__get8(s); + int tga_comp, tga_rgb16=0; + int tga_inverted = stbi__get8(s); + + + unsigned char *tga_data; + unsigned char *tga_palette = NULL; + int i, j; + unsigned char raw_data[4] = {0}; + int RLE_count = 0; + int RLE_repeating = 0; + int read_next_pixel = 1; + STBI_NOTUSED(ri); + STBI_NOTUSED(tga_x_origin); + STBI_NOTUSED(tga_y_origin); + + if ( tga_image_type >= 8 ) + { + tga_image_type -= 8; + tga_is_RLE = 1; + } + tga_inverted = 1 - ((tga_inverted >> 5) & 1); + + if ( tga_indexed ) tga_comp = stbi__tga_get_comp(tga_palette_bits, 0, &tga_rgb16); + else tga_comp = stbi__tga_get_comp(tga_bits_per_pixel, (tga_image_type == 3), &tga_rgb16); + if(!tga_comp) + return stbi__errpuc("bad format", "Can't find out TGA pixelformat"); + + *x = tga_width; + *y = tga_height; + if (comp) *comp = tga_comp; + if (!stbi__mad3sizes_valid(tga_width, tga_height, tga_comp, 0)) + return stbi__errpuc("too large", "Corrupt TGA"); + tga_data = (unsigned char*)stbi__malloc_mad3(tga_width, tga_height, tga_comp, 0); + if (!tga_data) return stbi__errpuc("outofmem", "Out of memory"); + + stbi__skip(s, tga_offset ); + if ( !tga_indexed && !tga_is_RLE && !tga_rgb16 ) { + for (i=0; i < tga_height; ++i) { + int row = tga_inverted ? tga_height -i - 1 : i; + stbi_uc *tga_row = tga_data + row*tga_width*tga_comp; + stbi__getn(s, tga_row, tga_width * tga_comp); + } + } else { + + if ( tga_indexed) + { + + stbi__skip(s, tga_palette_start ); + + tga_palette = (unsigned char*)stbi__malloc_mad2(tga_palette_len, tga_comp, 0); + if (!tga_palette) { + STBI_FREE(tga_data); + return stbi__errpuc("outofmem", "Out of memory"); + } + if (tga_rgb16) { + stbi_uc *pal_entry = tga_palette; + STBI_ASSERT(tga_comp == STBI_rgb); + for (i=0; i < tga_palette_len; ++i) { + stbi__tga_read_rgb16(s, pal_entry); + pal_entry += tga_comp; + } + } else if (!stbi__getn(s, tga_palette, tga_palette_len * tga_comp)) { + STBI_FREE(tga_data); + STBI_FREE(tga_palette); + return stbi__errpuc("bad palette", "Corrupt TGA"); + } + } + + for (i=0; i < tga_width * tga_height; ++i) + { + + if ( tga_is_RLE ) + { + if ( RLE_count == 0 ) + { + + int RLE_cmd = stbi__get8(s); + RLE_count = 1 + (RLE_cmd & 127); + RLE_repeating = RLE_cmd >> 7; + read_next_pixel = 1; + } else if ( !RLE_repeating ) + { + read_next_pixel = 1; + } + } else + { + read_next_pixel = 1; + } + + if ( read_next_pixel ) + { + + if ( tga_indexed ) + { + + int pal_idx = (tga_bits_per_pixel == 8) ? stbi__get8(s) : stbi__get16le(s); + if ( pal_idx >= tga_palette_len ) { + + pal_idx = 0; + } + pal_idx *= tga_comp; + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = tga_palette[pal_idx+j]; + } + } else if(tga_rgb16) { + STBI_ASSERT(tga_comp == STBI_rgb); + stbi__tga_read_rgb16(s, raw_data); + } else { + + for (j = 0; j < tga_comp; ++j) { + raw_data[j] = stbi__get8(s); + } + } + + read_next_pixel = 0; + } + + for (j = 0; j < tga_comp; ++j) + tga_data[i*tga_comp+j] = raw_data[j]; + + --RLE_count; + } + + if ( tga_inverted ) + { + for (j = 0; j*2 < tga_height; ++j) + { + int index1 = j * tga_width * tga_comp; + int index2 = (tga_height - 1 - j) * tga_width * tga_comp; + for (i = tga_width * tga_comp; i > 0; --i) + { + unsigned char temp = tga_data[index1]; + tga_data[index1] = tga_data[index2]; + tga_data[index2] = temp; + ++index1; + ++index2; + } + } + } + + if ( tga_palette != NULL ) + { + STBI_FREE( tga_palette ); + } + } + + if (tga_comp >= 3 && !tga_rgb16) + { + unsigned char* tga_pixel = tga_data; + for (i=0; i < tga_width * tga_height; ++i) + { + unsigned char temp = tga_pixel[0]; + tga_pixel[0] = tga_pixel[2]; + tga_pixel[2] = temp; + tga_pixel += tga_comp; + } + } + + if (req_comp && req_comp != tga_comp) + tga_data = stbi__convert_format(tga_data, tga_comp, req_comp, tga_width, tga_height); + + + tga_palette_start = tga_palette_len = tga_palette_bits = + tga_x_origin = tga_y_origin = 0; + STBI_NOTUSED(tga_palette_start); + + return tga_data; +} +#endif +#ifndef STBI_NO_PSD +static int stbi__psd_test(stbi__context *s) +{ + int r = (stbi__get32be(s) == 0x38425053); + stbi__rewind(s); + return r; +} +static int stbi__psd_decode_rle(stbi__context *s, stbi_uc *p, int pixelCount) +{ + int count, nleft, len; + count = 0; + while ((nleft = pixelCount - count) > 0) { + len = stbi__get8(s); + if (len == 128) { + + } else if (len < 128) { + + len++; + if (len > nleft) return 0; + count += len; + while (len) { + *p = stbi__get8(s); + p += 4; + len--; + } + } else if (len > 128) { + stbi_uc val; + + + len = 257 - len; + if (len > nleft) return 0; + val = stbi__get8(s); + count += len; + while (len) { + *p = val; + p += 4; + len--; + } + } + } + return 1; +} +static void *stbi__psd_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri, int bpc) +{ + int pixelCount; + int channelCount, compression; + int channel, i; + int bitdepth; + int w,h; + stbi_uc *out; + STBI_NOTUSED(ri); + + if (stbi__get32be(s) != 0x38425053) + return stbi__errpuc("not PSD", "Corrupt PSD image"); + + if (stbi__get16be(s) != 1) + return stbi__errpuc("wrong version", "Unsupported version of PSD image"); + + stbi__skip(s, 6 ); + + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) + return stbi__errpuc("wrong channel count", "Unsupported number of channels in PSD image"); + + h = stbi__get32be(s); + w = stbi__get32be(s); + + bitdepth = stbi__get16be(s); + if (bitdepth != 8 && bitdepth != 16) + return stbi__errpuc("unsupported bit depth", "PSD bit depth is not 8 or 16 bit"); + + + + + + + + + + + if (stbi__get16be(s) != 3) + return stbi__errpuc("wrong color format", "PSD is not in RGB color format"); + + stbi__skip(s,stbi__get32be(s) ); + + stbi__skip(s, stbi__get32be(s) ); + + stbi__skip(s, stbi__get32be(s) ); + + + + + compression = stbi__get16be(s); + if (compression > 1) + return stbi__errpuc("bad compression", "PSD has an unknown compression format"); + + if (!stbi__mad3sizes_valid(4, w, h, 0)) + return stbi__errpuc("too large", "Corrupt PSD"); + + if (!compression && bitdepth == 16 && bpc == 16) { + out = (stbi_uc *) stbi__malloc_mad3(8, w, h, 0); + ri->bits_per_channel = 16; + } else + out = (stbi_uc *) stbi__malloc(4 * w*h); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + pixelCount = w*h; + + + + if (compression) { + + + + + + + + + + stbi__skip(s, h * channelCount * 2 ); + + for (channel = 0; channel < 4; channel++) { + stbi_uc *p; + p = out+channel; + if (channel >= channelCount) { + + for (i = 0; i < pixelCount; i++, p += 4) + *p = (channel == 3 ? 255 : 0); + } else { + + if (!stbi__psd_decode_rle(s, p, pixelCount)) { + STBI_FREE(out); + return stbi__errpuc("corrupt", "bad RLE data"); + } + } + } + } else { + + + + for (channel = 0; channel < 4; channel++) { + if (channel >= channelCount) { + + if (bitdepth == 16 && bpc == 16) { + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + stbi__uint16 val = channel == 3 ? 65535 : 0; + for (i = 0; i < pixelCount; i++, q += 4) + *q = val; + } else { + stbi_uc *p = out+channel; + stbi_uc val = channel == 3 ? 255 : 0; + for (i = 0; i < pixelCount; i++, p += 4) + *p = val; + } + } else { + if (ri->bits_per_channel == 16) { + stbi__uint16 *q = ((stbi__uint16 *) out) + channel; + for (i = 0; i < pixelCount; i++, q += 4) + *q = (stbi__uint16) stbi__get16be(s); + } else { + stbi_uc *p = out+channel; + if (bitdepth == 16) { + for (i = 0; i < pixelCount; i++, p += 4) + *p = (stbi_uc) (stbi__get16be(s) >> 8); + } else { + for (i = 0; i < pixelCount; i++, p += 4) + *p = stbi__get8(s); + } + } + } + } + } + + if (channelCount >= 4) { + if (ri->bits_per_channel == 16) { + for (i=0; i < w*h; ++i) { + stbi__uint16 *pixel = (stbi__uint16 *) out + 4*i; + if (pixel[3] != 0 && pixel[3] != 65535) { + float a = pixel[3] / 65535.0f; + float ra = 1.0f / a; + float inv_a = 65535.0f * (1 - ra); + pixel[0] = (stbi__uint16) (pixel[0]*ra + inv_a); + pixel[1] = (stbi__uint16) (pixel[1]*ra + inv_a); + pixel[2] = (stbi__uint16) (pixel[2]*ra + inv_a); + } + } + } else { + for (i=0; i < w*h; ++i) { + unsigned char *pixel = out + 4*i; + if (pixel[3] != 0 && pixel[3] != 255) { + float a = pixel[3] / 255.0f; + float ra = 1.0f / a; + float inv_a = 255.0f * (1 - ra); + pixel[0] = (unsigned char) (pixel[0]*ra + inv_a); + pixel[1] = (unsigned char) (pixel[1]*ra + inv_a); + pixel[2] = (unsigned char) (pixel[2]*ra + inv_a); + } + } + } + } + + if (req_comp && req_comp != 4) { + if (ri->bits_per_channel == 16) + out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, 4, req_comp, w, h); + else + out = stbi__convert_format(out, 4, req_comp, w, h); + if (out == NULL) return out; + } + if (comp) *comp = 4; + *y = h; + *x = w; + return out; +} +#endif +#ifndef STBI_NO_PIC +static int stbi__pic_is4(stbi__context *s,const char *str) +{ + int i; + for (i=0; i<4; ++i) + if (stbi__get8(s) != (stbi_uc)str[i]) + return 0; + return 1; +} +static int stbi__pic_test_core(stbi__context *s) +{ + int i; + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) + return 0; + for(i=0;i<84;++i) + stbi__get8(s); + if (!stbi__pic_is4(s,"PICT")) + return 0; + return 1; +} +typedef struct +{ + stbi_uc size,type,channel; +} stbi__pic_packet; +static stbi_uc *stbi__readval(stbi__context *s, int channel, stbi_uc *dest) +{ + int mask=0x80, i; + for (i=0; i<4; ++i, mask>>=1) { + if (channel & mask) { + if (stbi__at_eof(s)) return stbi__errpuc("bad file","PIC file too short"); + dest[i]=stbi__get8(s); + } + } + return dest; +} +static void stbi__copyval(int channel,stbi_uc *dest,const stbi_uc *src) +{ + int mask=0x80,i; + for (i=0;i<4; ++i, mask>>=1) + if (channel&mask) + dest[i]=src[i]; +} +static stbi_uc *stbi__pic_load_core(stbi__context *s,int width,int height,int *comp, stbi_uc *result) +{ + int act_comp=0,num_packets=0,y,chained; + stbi__pic_packet packets[10]; + + + do { + stbi__pic_packet *packet; + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return stbi__errpuc("bad format","too many packets"); + packet = &packets[num_packets++]; + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + act_comp |= packet->channel; + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (reading packets)"); + if (packet->size != 8) return stbi__errpuc("bad format","packet isn't 8bpp"); + } while (chained); + *comp = (act_comp & 0x10 ? 4 : 3); + for(y=0; ytype) { + default: + return stbi__errpuc("bad format","packet has bad compression type"); + case 0: { + int x; + for(x=0;xchannel,dest)) + return 0; + break; + } + case 1: + { + int left=width, i; + while (left>0) { + stbi_uc count,value[4]; + count=stbi__get8(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pure read count)"); + if (count > left) + count = (stbi_uc) left; + if (!stbi__readval(s,packet->channel,value)) return 0; + for(i=0; ichannel,dest,value); + left -= count; + } + } + break; + case 2: { + int left=width; + while (left>0) { + int count = stbi__get8(s), i; + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (mixed read count)"); + if (count >= 128) { + stbi_uc value[4]; + if (count==128) + count = stbi__get16be(s); + else + count -= 127; + if (count > left) + return stbi__errpuc("bad file","scanline overrun"); + if (!stbi__readval(s,packet->channel,value)) + return 0; + for(i=0;ichannel,dest,value); + } else { + ++count; + if (count>left) return stbi__errpuc("bad file","scanline overrun"); + for(i=0;ichannel,dest)) + return 0; + } + left-=count; + } + break; + } + } + } + } + return result; +} +static void *stbi__pic_load(stbi__context *s,int *px,int *py,int *comp,int req_comp, stbi__result_info *ri) +{ + stbi_uc *result; + int i, x,y, internal_comp; + STBI_NOTUSED(ri); + if (!comp) comp = &internal_comp; + for (i=0; i<92; ++i) + stbi__get8(s); + x = stbi__get16be(s); + y = stbi__get16be(s); + if (stbi__at_eof(s)) return stbi__errpuc("bad file","file too short (pic header)"); + if (!stbi__mad3sizes_valid(x, y, 4, 0)) return stbi__errpuc("too large", "PIC image too large to decode"); + stbi__get32be(s); + stbi__get16be(s); + stbi__get16be(s); + + result = (stbi_uc *) stbi__malloc_mad3(x, y, 4, 0); + memset(result, 0xff, x*y*4); + if (!stbi__pic_load_core(s,x,y,comp, result)) { + STBI_FREE(result); + result=0; + } + *px = x; + *py = y; + if (req_comp == 0) req_comp = *comp; + result=stbi__convert_format(result,4,req_comp,x,y); + return result; +} +static int stbi__pic_test(stbi__context *s) +{ + int r = stbi__pic_test_core(s); + stbi__rewind(s); + return r; +} +#endif +#ifndef STBI_NO_GIF +typedef struct +{ + stbi__int16 prefix; + stbi_uc first; + stbi_uc suffix; +} stbi__gif_lzw; +typedef struct +{ + int w,h; + stbi_uc *out; + stbi_uc *background; + stbi_uc *history; + int flags, bgindex, ratio, transparent, eflags; + stbi_uc pal[256][4]; + stbi_uc lpal[256][4]; + stbi__gif_lzw codes[8192]; + stbi_uc *color_table; + int parse, step; + int lflags; + int start_x, start_y; + int max_x, max_y; + int cur_x, cur_y; + int line_size; + int delay; +} stbi__gif; +static int stbi__gif_test_raw(stbi__context *s) +{ + int sz; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') return 0; + sz = stbi__get8(s); + if (sz != '9' && sz != '7') return 0; + if (stbi__get8(s) != 'a') return 0; + return 1; +} +static int stbi__gif_test(stbi__context *s) +{ + int r = stbi__gif_test_raw(s); + stbi__rewind(s); + return r; +} +static void stbi__gif_parse_colortable(stbi__context *s, stbi_uc pal[256][4], int num_entries, int transp) +{ + int i; + for (i=0; i < num_entries; ++i) { + pal[i][2] = stbi__get8(s); + pal[i][1] = stbi__get8(s); + pal[i][0] = stbi__get8(s); + pal[i][3] = transp == i ? 0 : 255; + } +} +static int stbi__gif_header(stbi__context *s, stbi__gif *g, int *comp, int is_info) +{ + stbi_uc version; + if (stbi__get8(s) != 'G' || stbi__get8(s) != 'I' || stbi__get8(s) != 'F' || stbi__get8(s) != '8') + return stbi__err("not GIF", "Corrupt GIF"); + version = stbi__get8(s); + if (version != '7' && version != '9') return stbi__err("not GIF", "Corrupt GIF"); + if (stbi__get8(s) != 'a') return stbi__err("not GIF", "Corrupt GIF"); + stbi__g_failure_reason = ""; + g->w = stbi__get16le(s); + g->h = stbi__get16le(s); + g->flags = stbi__get8(s); + g->bgindex = stbi__get8(s); + g->ratio = stbi__get8(s); + g->transparent = -1; + if (comp != 0) *comp = 4; + if (is_info) return 1; + if (g->flags & 0x80) + stbi__gif_parse_colortable(s,g->pal, 2 << (g->flags & 7), -1); + return 1; +} +static int stbi__gif_info_raw(stbi__context *s, int *x, int *y, int *comp) +{ + stbi__gif* g = (stbi__gif*) stbi__malloc(sizeof(stbi__gif)); + if (!stbi__gif_header(s, g, comp, 1)) { + STBI_FREE(g); + stbi__rewind( s ); + return 0; + } + if (x) *x = g->w; + if (y) *y = g->h; + STBI_FREE(g); + return 1; +} +static void stbi__out_gif_code(stbi__gif *g, stbi__uint16 code) +{ + stbi_uc *p, *c; + int idx; + + + if (g->codes[code].prefix >= 0) + stbi__out_gif_code(g, g->codes[code].prefix); + if (g->cur_y >= g->max_y) return; + idx = g->cur_x + g->cur_y; + p = &g->out[idx]; + g->history[idx / 4] = 1; + c = &g->color_table[g->codes[code].suffix * 4]; + if (c[3] > 128) { + p[0] = c[2]; + p[1] = c[1]; + p[2] = c[0]; + p[3] = c[3]; + } + g->cur_x += 4; + if (g->cur_x >= g->max_x) { + g->cur_x = g->start_x; + g->cur_y += g->step; + while (g->cur_y >= g->max_y && g->parse > 0) { + g->step = (1 << g->parse) * g->line_size; + g->cur_y = g->start_y + (g->step >> 1); + --g->parse; + } + } +} +static stbi_uc *stbi__process_gif_raster(stbi__context *s, stbi__gif *g) +{ + stbi_uc lzw_cs; + stbi__int32 len, init_code; + stbi__uint32 first; + stbi__int32 codesize, codemask, avail, oldcode, bits, valid_bits, clear; + stbi__gif_lzw *p; + lzw_cs = stbi__get8(s); + if (lzw_cs > 12) return NULL; + clear = 1 << lzw_cs; + first = 1; + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + bits = 0; + valid_bits = 0; + for (init_code = 0; init_code < clear; init_code++) { + g->codes[init_code].prefix = -1; + g->codes[init_code].first = (stbi_uc) init_code; + g->codes[init_code].suffix = (stbi_uc) init_code; + } + + avail = clear+2; + oldcode = -1; + len = 0; + for(;;) { + if (valid_bits < codesize) { + if (len == 0) { + len = stbi__get8(s); + if (len == 0) + return g->out; + } + --len; + bits |= (stbi__int32) stbi__get8(s) << valid_bits; + valid_bits += 8; + } else { + stbi__int32 code = bits & codemask; + bits >>= codesize; + valid_bits -= codesize; + + if (code == clear) { + codesize = lzw_cs + 1; + codemask = (1 << codesize) - 1; + avail = clear + 2; + oldcode = -1; + first = 0; + } else if (code == clear + 1) { + stbi__skip(s, len); + while ((len = stbi__get8(s)) > 0) + stbi__skip(s,len); + return g->out; + } else if (code <= avail) { + if (first) { + return stbi__errpuc("no clear code", "Corrupt GIF"); + } + if (oldcode >= 0) { + p = &g->codes[avail++]; + if (avail > 8192) { + return stbi__errpuc("too many codes", "Corrupt GIF"); + } + p->prefix = (stbi__int16) oldcode; + p->first = g->codes[oldcode].first; + p->suffix = (code == avail) ? p->first : g->codes[code].first; + } else if (code == avail) + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + stbi__out_gif_code(g, (stbi__uint16) code); + if ((avail & codemask) == 0 && avail <= 0x0FFF) { + codesize++; + codemask = (1 << codesize) - 1; + } + oldcode = code; + } else { + return stbi__errpuc("illegal code in raster", "Corrupt GIF"); + } + } + } +} +static stbi_uc *stbi__gif_load_next(stbi__context *s, stbi__gif *g, int *comp, int req_comp, stbi_uc *two_back) +{ + int dispose; + int first_frame; + int pi; + int pcount; + STBI_NOTUSED(req_comp); + + first_frame = 0; + if (g->out == 0) { + if (!stbi__gif_header(s, g, comp,0)) return 0; + if (!stbi__mad3sizes_valid(4, g->w, g->h, 0)) + return stbi__errpuc("too large", "GIF image is too large"); + pcount = g->w * g->h; + g->out = (stbi_uc *) stbi__malloc(4 * pcount); + g->background = (stbi_uc *) stbi__malloc(4 * pcount); + g->history = (stbi_uc *) stbi__malloc(pcount); + if (!g->out || !g->background || !g->history) + return stbi__errpuc("outofmem", "Out of memory"); + + + + memset(g->out, 0x00, 4 * pcount); + memset(g->background, 0x00, 4 * pcount); + memset(g->history, 0x00, pcount); + first_frame = 1; + } else { + + dispose = (g->eflags & 0x1C) >> 2; + pcount = g->w * g->h; + if ((dispose == 3) && (two_back == 0)) { + dispose = 2; + } + if (dispose == 3) { + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 ); + } + } + } else if (dispose == 2) { + + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi]) { + memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 ); + } + } + } else { + + + + + } + + memcpy( g->background, g->out, 4 * g->w * g->h ); + } + + memset( g->history, 0x00, g->w * g->h ); + for (;;) { + int tag = stbi__get8(s); + switch (tag) { + case 0x2C: + { + stbi__int32 x, y, w, h; + stbi_uc *o; + x = stbi__get16le(s); + y = stbi__get16le(s); + w = stbi__get16le(s); + h = stbi__get16le(s); + if (((x + w) > (g->w)) || ((y + h) > (g->h))) + return stbi__errpuc("bad Image Descriptor", "Corrupt GIF"); + g->line_size = g->w * 4; + g->start_x = x * 4; + g->start_y = y * g->line_size; + g->max_x = g->start_x + w * 4; + g->max_y = g->start_y + h * g->line_size; + g->cur_x = g->start_x; + g->cur_y = g->start_y; + + + + + if (w == 0) + g->cur_y = g->max_y; + g->lflags = stbi__get8(s); + if (g->lflags & 0x40) { + g->step = 8 * g->line_size; + g->parse = 3; + } else { + g->step = g->line_size; + g->parse = 0; + } + if (g->lflags & 0x80) { + stbi__gif_parse_colortable(s,g->lpal, 2 << (g->lflags & 7), g->eflags & 0x01 ? g->transparent : -1); + g->color_table = (stbi_uc *) g->lpal; + } else if (g->flags & 0x80) { + g->color_table = (stbi_uc *) g->pal; + } else + return stbi__errpuc("missing color table", "Corrupt GIF"); + o = stbi__process_gif_raster(s, g); + if (!o) return NULL; + + pcount = g->w * g->h; + if (first_frame && (g->bgindex > 0)) { + + for (pi = 0; pi < pcount; ++pi) { + if (g->history[pi] == 0) { + g->pal[g->bgindex][3] = 255; + memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 ); + } + } + } + return o; + } + case 0x21: + { + int len; + int ext = stbi__get8(s); + if (ext == 0xF9) { + len = stbi__get8(s); + if (len == 4) { + g->eflags = stbi__get8(s); + g->delay = 10 * stbi__get16le(s); + + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 255; + } + if (g->eflags & 0x01) { + g->transparent = stbi__get8(s); + if (g->transparent >= 0) { + g->pal[g->transparent][3] = 0; + } + } else { + + stbi__skip(s, 1); + g->transparent = -1; + } + } else { + stbi__skip(s, len); + break; + } + } + while ((len = stbi__get8(s)) != 0) { + stbi__skip(s, len); + } + break; + } + case 0x3B: + return (stbi_uc *) s; + default: + return stbi__errpuc("unknown code", "Corrupt GIF"); + } + } +} +static void *stbi__load_gif_main(stbi__context *s, int **delays, int *x, int *y, int *z, int *comp, int req_comp) +{ + if (stbi__gif_test(s)) { + int layers = 0; + stbi_uc *u = 0; + stbi_uc *out = 0; + stbi_uc *two_back = 0; + stbi__gif g; + int stride; + memset(&g, 0, sizeof(g)); + if (delays) { + *delays = 0; + } + do { + u = stbi__gif_load_next(s, &g, comp, req_comp, two_back); + if (u == (stbi_uc *) s) u = 0; + if (u) { + *x = g.w; + *y = g.h; + ++layers; + stride = g.w * g.h * 4; + if (out) { + void *tmp = (stbi_uc*) STBI_REALLOC( out, layers * stride ); + if (NULL == tmp) { + STBI_FREE(g.out); + STBI_FREE(g.history); + STBI_FREE(g.background); + return stbi__errpuc("outofmem", "Out of memory"); + } + else + out = (stbi_uc*) tmp; + if (delays) { + *delays = (int*) STBI_REALLOC( *delays, sizeof(int) * layers ); + } + } else { + out = (stbi_uc*)stbi__malloc( layers * stride ); + if (delays) { + *delays = (int*) stbi__malloc( layers * sizeof(int) ); + } + } + memcpy( out + ((layers - 1) * stride), u, stride ); + if (layers >= 2) { + two_back = out - 2 * stride; + } + if (delays) { + (*delays)[layers - 1U] = g.delay; + } + } + } while (u != 0); + + STBI_FREE(g.out); + STBI_FREE(g.history); + STBI_FREE(g.background); + + if (req_comp && req_comp != 4) + out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h); + *z = layers; + return out; + } else { + return stbi__errpuc("not GIF", "Image was not as a gif type."); + } +} +static void *stbi__gif_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *u = 0; + stbi__gif g; + memset(&g, 0, sizeof(g)); + STBI_NOTUSED(ri); + u = stbi__gif_load_next(s, &g, comp, req_comp, 0); + if (u == (stbi_uc *) s) u = 0; + if (u) { + *x = g.w; + *y = g.h; + + + if (req_comp && req_comp != 4) + u = stbi__convert_format(u, 4, req_comp, g.w, g.h); + } else if (g.out) { + + STBI_FREE(g.out); + } + + STBI_FREE(g.history); + STBI_FREE(g.background); + return u; +} +static int stbi__gif_info(stbi__context *s, int *x, int *y, int *comp) +{ + return stbi__gif_info_raw(s,x,y,comp); +} +#endif +#ifndef STBI_NO_HDR +static int stbi__hdr_test_core(stbi__context *s, const char *signature) +{ + int i; + for (i=0; signature[i]; ++i) + if (stbi__get8(s) != signature[i]) + return 0; + stbi__rewind(s); + return 1; +} +static int stbi__hdr_test(stbi__context* s) +{ + int r = stbi__hdr_test_core(s, "#?RADIANCE\n"); + stbi__rewind(s); + if(!r) { + r = stbi__hdr_test_core(s, "#?RGBE\n"); + stbi__rewind(s); + } + return r; +} +#define STBI__HDR_BUFLEN 1024 +static char *stbi__hdr_gettoken(stbi__context *z, char *buffer) +{ + int len=0; + char c = '\0'; + c = (char) stbi__get8(z); + while (!stbi__at_eof(z) && c != '\n') { + buffer[len++] = c; + if (len == STBI__HDR_BUFLEN-1) { + + while (!stbi__at_eof(z) && stbi__get8(z) != '\n') + ; + break; + } + c = (char) stbi__get8(z); + } + buffer[len] = 0; + return buffer; +} +static void stbi__hdr_convert(float *output, stbi_uc *input, int req_comp) +{ + if ( input[3] != 0 ) { + float f1; + + f1 = (float) ldexp(1.0f, input[3] - (int)(128 + 8)); + if (req_comp <= 2) + output[0] = (input[0] + input[1] + input[2]) * f1 / 3; + else { + output[0] = input[0] * f1; + output[1] = input[1] * f1; + output[2] = input[2] * f1; + } + if (req_comp == 2) output[1] = 1; + if (req_comp == 4) output[3] = 1; + } else { + switch (req_comp) { + case 4: output[3] = 1; + case 3: output[0] = output[1] = output[2] = 0; + break; + case 2: output[1] = 1; + case 1: output[0] = 0; + break; + } + } +} +static float *stbi__hdr_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int width, height; + stbi_uc *scanline; + float *hdr_data; + int len; + unsigned char count, value; + int i, j, k, c1,c2, z; + const char *headerToken; + STBI_NOTUSED(ri); + + headerToken = stbi__hdr_gettoken(s,buffer); + if (strcmp(headerToken, "#?RADIANCE") != 0 && strcmp(headerToken, "#?RGBE") != 0) + return stbi__errpf("not HDR", "Corrupt HDR image"); + + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + if (!valid) return stbi__errpf("unsupported format", "Unsupported HDR format"); + + + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + height = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) return stbi__errpf("unsupported data layout", "Unsupported HDR format"); + token += 3; + width = (int) strtol(token, NULL, 10); + *x = width; + *y = height; + if (comp) *comp = 3; + if (req_comp == 0) req_comp = 3; + if (!stbi__mad4sizes_valid(width, height, req_comp, sizeof(float), 0)) + return stbi__errpf("too large", "HDR image is too large"); + + hdr_data = (float *) stbi__malloc_mad4(width, height, req_comp, sizeof(float), 0); + if (!hdr_data) + return stbi__errpf("outofmem", "Out of memory"); + + + if ( width < 8 || width >= 32768) { + + for (j=0; j < height; ++j) { + for (i=0; i < width; ++i) { + stbi_uc rgbe[4]; + main_decode_loop: + stbi__getn(s, rgbe, 4); + stbi__hdr_convert(hdr_data + j * width * req_comp + i * req_comp, rgbe, req_comp); + } + } + } else { + + scanline = NULL; + for (j = 0; j < height; ++j) { + c1 = stbi__get8(s); + c2 = stbi__get8(s); + len = stbi__get8(s); + if (c1 != 2 || c2 != 2 || (len & 0x80)) { + + + stbi_uc rgbe[4]; + rgbe[0] = (stbi_uc) c1; + rgbe[1] = (stbi_uc) c2; + rgbe[2] = (stbi_uc) len; + rgbe[3] = (stbi_uc) stbi__get8(s); + stbi__hdr_convert(hdr_data, rgbe, req_comp); + i = 1; + j = 0; + STBI_FREE(scanline); + goto main_decode_loop; + } + len <<= 8; + len |= stbi__get8(s); + if (len != width) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("invalid decoded scanline length", "corrupt HDR"); } + if (scanline == NULL) { + scanline = (stbi_uc *) stbi__malloc_mad2(width, 4, 0); + if (!scanline) { + STBI_FREE(hdr_data); + return stbi__errpf("outofmem", "Out of memory"); + } + } + for (k = 0; k < 4; ++k) { + int nleft; + i = 0; + while ((nleft = width - i) > 0) { + count = stbi__get8(s); + if (count > 128) { + + value = stbi__get8(s); + count -= 128; + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = value; + } else { + + if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); } + for (z = 0; z < count; ++z) + scanline[i++ * 4 + k] = stbi__get8(s); + } + } + } + for (i=0; i < width; ++i) + stbi__hdr_convert(hdr_data+(j*width + i)*req_comp, scanline + i*4, req_comp); + } + if (scanline) + STBI_FREE(scanline); + } + return hdr_data; +} +static int stbi__hdr_info(stbi__context *s, int *x, int *y, int *comp) +{ + char buffer[STBI__HDR_BUFLEN]; + char *token; + int valid = 0; + int dummy; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (stbi__hdr_test(s) == 0) { + stbi__rewind( s ); + return 0; + } + for(;;) { + token = stbi__hdr_gettoken(s,buffer); + if (token[0] == 0) break; + if (strcmp(token, "FORMAT=32-bit_rle_rgbe") == 0) valid = 1; + } + if (!valid) { + stbi__rewind( s ); + return 0; + } + token = stbi__hdr_gettoken(s,buffer); + if (strncmp(token, "-Y ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *y = (int) strtol(token, &token, 10); + while (*token == ' ') ++token; + if (strncmp(token, "+X ", 3)) { + stbi__rewind( s ); + return 0; + } + token += 3; + *x = (int) strtol(token, NULL, 10); + *comp = 3; + return 1; +} +#endif +#ifndef STBI_NO_BMP +static int stbi__bmp_info(stbi__context *s, int *x, int *y, int *comp) +{ + void *p; + stbi__bmp_data info; + info.all_a = 255; + p = stbi__bmp_parse_header(s, &info); + stbi__rewind( s ); + if (p == NULL) + return 0; + if (x) *x = s->img_x; + if (y) *y = s->img_y; + if (comp) { + if (info.bpp == 24 && info.ma == 0xff000000) + *comp = 3; + else + *comp = info.ma ? 4 : 3; + } + return 1; +} +#endif +#ifndef STBI_NO_PSD +static int stbi__psd_info(stbi__context *s, int *x, int *y, int *comp) +{ + int channelCount, dummy, depth; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + *y = stbi__get32be(s); + *x = stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 8 && depth != 16) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 3) { + stbi__rewind( s ); + return 0; + } + *comp = 4; + return 1; +} +static int stbi__psd_is16(stbi__context *s) +{ + int channelCount, depth; + if (stbi__get32be(s) != 0x38425053) { + stbi__rewind( s ); + return 0; + } + if (stbi__get16be(s) != 1) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 6); + channelCount = stbi__get16be(s); + if (channelCount < 0 || channelCount > 16) { + stbi__rewind( s ); + return 0; + } + (void) stbi__get32be(s); + (void) stbi__get32be(s); + depth = stbi__get16be(s); + if (depth != 16) { + stbi__rewind( s ); + return 0; + } + return 1; +} +#endif +#ifndef STBI_NO_PIC +static int stbi__pic_info(stbi__context *s, int *x, int *y, int *comp) +{ + int act_comp=0,num_packets=0,chained,dummy; + stbi__pic_packet packets[10]; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + if (!stbi__pic_is4(s,"\x53\x80\xF6\x34")) { + stbi__rewind(s); + return 0; + } + stbi__skip(s, 88); + *x = stbi__get16be(s); + *y = stbi__get16be(s); + if (stbi__at_eof(s)) { + stbi__rewind( s); + return 0; + } + if ( (*x) != 0 && (1 << 28) / (*x) < (*y)) { + stbi__rewind( s ); + return 0; + } + stbi__skip(s, 8); + do { + stbi__pic_packet *packet; + if (num_packets==sizeof(packets)/sizeof(packets[0])) + return 0; + packet = &packets[num_packets++]; + chained = stbi__get8(s); + packet->size = stbi__get8(s); + packet->type = stbi__get8(s); + packet->channel = stbi__get8(s); + act_comp |= packet->channel; + if (stbi__at_eof(s)) { + stbi__rewind( s ); + return 0; + } + if (packet->size != 8) { + stbi__rewind( s ); + return 0; + } + } while (chained); + *comp = (act_comp & 0x10 ? 4 : 3); + return 1; +} +#endif +#ifndef STBI_NO_PNM +static int stbi__pnm_test(stbi__context *s) +{ + char p, t; + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind( s ); + return 0; + } + return 1; +} +static void *stbi__pnm_load(stbi__context *s, int *x, int *y, int *comp, int req_comp, stbi__result_info *ri) +{ + stbi_uc *out; + STBI_NOTUSED(ri); + if (!stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n)) + return 0; + *x = s->img_x; + *y = s->img_y; + if (comp) *comp = s->img_n; + if (!stbi__mad3sizes_valid(s->img_n, s->img_x, s->img_y, 0)) + return stbi__errpuc("too large", "PNM too large"); + out = (stbi_uc *) stbi__malloc_mad3(s->img_n, s->img_x, s->img_y, 0); + if (!out) return stbi__errpuc("outofmem", "Out of memory"); + stbi__getn(s, out, s->img_n * s->img_x * s->img_y); + if (req_comp && req_comp != s->img_n) { + out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y); + if (out == NULL) return out; + } + return out; +} +static int stbi__pnm_isspace(char c) +{ + return c == ' ' || c == '\t' || c == '\n' || c == '\v' || c == '\f' || c == '\r'; +} +static void stbi__pnm_skip_whitespace(stbi__context *s, char *c) +{ + for (;;) { + while (!stbi__at_eof(s) && stbi__pnm_isspace(*c)) + *c = (char) stbi__get8(s); + if (stbi__at_eof(s) || *c != '#') + break; + while (!stbi__at_eof(s) && *c != '\n' && *c != '\r' ) + *c = (char) stbi__get8(s); + } +} +static int stbi__pnm_isdigit(char c) +{ + return c >= '0' && c <= '9'; +} +static int stbi__pnm_getinteger(stbi__context *s, char *c) +{ + int value = 0; + while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) { + value = value*10 + (*c - '0'); + *c = (char) stbi__get8(s); + } + return value; +} +static int stbi__pnm_info(stbi__context *s, int *x, int *y, int *comp) +{ + int maxv, dummy; + char c, p, t; + if (!x) x = &dummy; + if (!y) y = &dummy; + if (!comp) comp = &dummy; + stbi__rewind(s); + + p = (char) stbi__get8(s); + t = (char) stbi__get8(s); + if (p != 'P' || (t != '5' && t != '6')) { + stbi__rewind(s); + return 0; + } + *comp = (t == '6') ? 3 : 1; + c = (char) stbi__get8(s); + stbi__pnm_skip_whitespace(s, &c); + *x = stbi__pnm_getinteger(s, &c); + stbi__pnm_skip_whitespace(s, &c); + *y = stbi__pnm_getinteger(s, &c); + stbi__pnm_skip_whitespace(s, &c); + maxv = stbi__pnm_getinteger(s, &c); + if (maxv > 255) + return stbi__err("max value > 255", "PPM image not 8-bit"); + else + return 1; +} +#endif +static int stbi__info_main(stbi__context *s, int *x, int *y, int *comp) +{ + #ifndef STBI_NO_JPEG + if (stbi__jpeg_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_PNG + if (stbi__png_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_GIF + if (stbi__gif_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_BMP + if (stbi__bmp_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_PIC + if (stbi__pic_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_PNM + if (stbi__pnm_info(s, x, y, comp)) return 1; + #endif + #ifndef STBI_NO_HDR + if (stbi__hdr_info(s, x, y, comp)) return 1; + #endif + + #ifndef STBI_NO_TGA + if (stbi__tga_info(s, x, y, comp)) + return 1; + #endif + return stbi__err("unknown image type", "Image not of any known type, or corrupt"); +} +static int stbi__is_16_main(stbi__context *s) +{ + #ifndef STBI_NO_PNG + if (stbi__png_is16(s)) return 1; + #endif + #ifndef STBI_NO_PSD + if (stbi__psd_is16(s)) return 1; + #endif + return 0; +} +#ifndef STBI_NO_STDIO +STBIDEF int stbi_info(char const *filename, int *x, int *y, int *comp) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_info_from_file(f, x, y, comp); + fclose(f); + return result; +} +STBIDEF int stbi_info_from_file(FILE *f, int *x, int *y, int *comp) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__info_main(&s,x,y,comp); + fseek(f,pos,SEEK_SET); + return r; +} +STBIDEF int stbi_is_16_bit(char const *filename) +{ + FILE *f = stbi__fopen(filename, "rb"); + int result; + if (!f) return stbi__err("can't fopen", "Unable to open file"); + result = stbi_is_16_bit_from_file(f); + fclose(f); + return result; +} +STBIDEF int stbi_is_16_bit_from_file(FILE *f) +{ + int r; + stbi__context s; + long pos = ftell(f); + stbi__start_file(&s, f); + r = stbi__is_16_main(&s); + fseek(f,pos,SEEK_SET); + return r; +} +#endif +STBIDEF int stbi_info_from_memory(stbi_uc const *buffer, int len, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__info_main(&s,x,y,comp); +} +STBIDEF int stbi_info_from_callbacks(stbi_io_callbacks const *c, void *user, int *x, int *y, int *comp) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__info_main(&s,x,y,comp); +} +STBIDEF int stbi_is_16_bit_from_memory(stbi_uc const *buffer, int len) +{ + stbi__context s; + stbi__start_mem(&s,buffer,len); + return stbi__is_16_main(&s); +} +STBIDEF int stbi_is_16_bit_from_callbacks(stbi_io_callbacks const *c, void *user) +{ + stbi__context s; + stbi__start_callbacks(&s, (stbi_io_callbacks *) c, user); + return stbi__is_16_main(&s); +} +#endif diff --git a/src/3rd_party/stb/stb_image_resize.h b/src/3rd_party/stb/stb_image_resize.h new file mode 100644 index 0000000..e354a63 --- /dev/null +++ b/src/3rd_party/stb/stb_image_resize.h @@ -0,0 +1,1967 @@ +/* +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person 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, sublicense, 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: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ + +#ifndef STBIR_INCLUDE_STB_IMAGE_RESIZE_H +#define STBIR_INCLUDE_STB_IMAGE_RESIZE_H +#ifdef _MSC_VER +typedef unsigned char stbir_uint8; +typedef unsigned short stbir_uint16; +typedef unsigned int stbir_uint32; +#else +#include +typedef uint8_t stbir_uint8; +typedef uint16_t stbir_uint16; +typedef uint32_t stbir_uint32; +#endif +#ifndef STBIRDEF +#ifdef STB_IMAGE_RESIZE_STATIC +#define STBIRDEF static +#else +#ifdef __cplusplus +#define STBIRDEF extern "C" +#else +#define STBIRDEF extern +#endif +#endif +#endif +STBIRDEF int stbir_resize_uint8( const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels); +STBIRDEF int stbir_resize_float( const float *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + float *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels); +#define STBIR_ALPHA_CHANNEL_NONE -1 +#define STBIR_FLAG_ALPHA_PREMULTIPLIED (1 << 0) +#define STBIR_FLAG_ALPHA_USES_COLORSPACE (1 << 1) +STBIRDEF int stbir_resize_uint8_srgb(const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags); +typedef enum +{ + STBIR_EDGE_CLAMP = 1, + STBIR_EDGE_REFLECT = 2, + STBIR_EDGE_WRAP = 3, + STBIR_EDGE_ZERO = 4, +} stbir_edge; +STBIRDEF int stbir_resize_uint8_srgb_edgemode(const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode); +typedef enum +{ + STBIR_FILTER_DEFAULT = 0, + STBIR_FILTER_BOX = 1, + STBIR_FILTER_TRIANGLE = 2, + STBIR_FILTER_CUBICBSPLINE = 3, + STBIR_FILTER_CATMULLROM = 4, + STBIR_FILTER_MITCHELL = 5, +} stbir_filter; +typedef enum +{ + STBIR_COLORSPACE_LINEAR, + STBIR_COLORSPACE_SRGB, + STBIR_MAX_COLORSPACES, +} stbir_colorspace; +STBIRDEF int stbir_resize_uint8_generic( const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context); +STBIRDEF int stbir_resize_uint16_generic(const stbir_uint16 *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + stbir_uint16 *output_pixels , int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context); +STBIRDEF int stbir_resize_float_generic( const float *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + float *output_pixels , int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context); +typedef enum +{ + STBIR_TYPE_UINT8 , + STBIR_TYPE_UINT16, + STBIR_TYPE_UINT32, + STBIR_TYPE_FLOAT , + STBIR_MAX_TYPES +} stbir_datatype; +STBIRDEF int stbir_resize( const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context); +STBIRDEF int stbir_resize_subpixel(const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context, + float x_scale, float y_scale, + float x_offset, float y_offset); +STBIRDEF int stbir_resize_region( const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context, + float s0, float t0, float s1, float t1); +#endif +#ifdef STB_IMAGE_RESIZE_IMPLEMENTATION +#ifndef STBIR_ASSERT +#include +#define STBIR_ASSERT(x) assert(x) +#endif +#include +#include +#ifndef STBIR_MALLOC +#include +#define STBIR_MALLOC(size,c) ((void)(c), malloc(size)) +#define STBIR_FREE(ptr,c) ((void)(c), free(ptr)) +#endif +#ifndef _MSC_VER +#ifdef __cplusplus +#define stbir__inline inline +#else +#define stbir__inline +#endif +#else +#define stbir__inline __forceinline +#endif +typedef unsigned char stbir__validate_uint32[sizeof(stbir_uint32) == 4 ? 1 : -1]; +#ifdef _MSC_VER +#define STBIR__NOTUSED(v) (void)(v) +#else +#define STBIR__NOTUSED(v) (void)sizeof(v) +#endif +#define STBIR__ARRAY_SIZE(a) (sizeof((a))/sizeof((a)[0])) +#ifndef STBIR_DEFAULT_FILTER_UPSAMPLE +#define STBIR_DEFAULT_FILTER_UPSAMPLE STBIR_FILTER_CATMULLROM +#endif +#ifndef STBIR_DEFAULT_FILTER_DOWNSAMPLE +#define STBIR_DEFAULT_FILTER_DOWNSAMPLE STBIR_FILTER_MITCHELL +#endif +#ifndef STBIR_PROGRESS_REPORT +#define STBIR_PROGRESS_REPORT(float_0_to_1) +#endif +#ifndef STBIR_MAX_CHANNELS +#define STBIR_MAX_CHANNELS 64 +#endif +#if STBIR_MAX_CHANNELS > 65536 +#error "Too many channels; STBIR_MAX_CHANNELS must be no more than 65536." +#endif +#ifndef STBIR_ALPHA_EPSILON +#define STBIR_ALPHA_EPSILON ((float)1 / (1 << 20) / (1 << 20) / (1 << 20) / (1 << 20)) +#endif +#ifdef _MSC_VER +#define STBIR__UNUSED_PARAM(v) (void)(v) +#else +#define STBIR__UNUSED_PARAM(v) (void)sizeof(v) +#endif +static unsigned char stbir__type_size[] = { + 1, + 2, + 4, + 4, +}; +typedef float (stbir__kernel_fn)(float x, float scale); +typedef float (stbir__support_fn)(float scale); +typedef struct +{ + stbir__kernel_fn* kernel; + stbir__support_fn* support; +} stbir__filter_info; +typedef struct +{ + int n0; + int n1; +} stbir__contributors; +typedef struct +{ + const void* input_data; + int input_w; + int input_h; + int input_stride_bytes; + void* output_data; + int output_w; + int output_h; + int output_stride_bytes; + float s0, t0, s1, t1; + float horizontal_shift; + float vertical_shift; + float horizontal_scale; + float vertical_scale; + int channels; + int alpha_channel; + stbir_uint32 flags; + stbir_datatype type; + stbir_filter horizontal_filter; + stbir_filter vertical_filter; + stbir_edge edge_horizontal; + stbir_edge edge_vertical; + stbir_colorspace colorspace; + stbir__contributors* horizontal_contributors; + float* horizontal_coefficients; + stbir__contributors* vertical_contributors; + float* vertical_coefficients; + int decode_buffer_pixels; + float* decode_buffer; + float* horizontal_buffer; + + int horizontal_coefficient_width; + int vertical_coefficient_width; + int horizontal_filter_pixel_width; + int vertical_filter_pixel_width; + int horizontal_filter_pixel_margin; + int vertical_filter_pixel_margin; + int horizontal_num_contributors; + int vertical_num_contributors; + int ring_buffer_length_bytes; + int ring_buffer_num_entries; + int ring_buffer_first_scanline; + int ring_buffer_last_scanline; + int ring_buffer_begin_index; + float* ring_buffer; + float* encode_buffer; + int horizontal_contributors_size; + int horizontal_coefficients_size; + int vertical_contributors_size; + int vertical_coefficients_size; + int decode_buffer_size; + int horizontal_buffer_size; + int ring_buffer_size; + int encode_buffer_size; +} stbir__info; +static const float stbir__max_uint8_as_float = 255.0f; +static const float stbir__max_uint16_as_float = 65535.0f; +static const double stbir__max_uint32_as_float = 4294967295.0; +static stbir__inline int stbir__min(int a, int b) +{ + return a < b ? a : b; +} +static stbir__inline float stbir__saturate(float x) +{ + if (x < 0) + return 0; + if (x > 1) + return 1; + return x; +} +#ifdef STBIR_SATURATE_INT +static stbir__inline stbir_uint8 stbir__saturate8(int x) +{ + if ((unsigned int) x <= 255) + return x; + if (x < 0) + return 0; + return 255; +} +static stbir__inline stbir_uint16 stbir__saturate16(int x) +{ + if ((unsigned int) x <= 65535) + return x; + if (x < 0) + return 0; + return 65535; +} +#endif +static float stbir__srgb_uchar_to_linear_float[256] = { + 0.000000f, 0.000304f, 0.000607f, 0.000911f, 0.001214f, 0.001518f, 0.001821f, 0.002125f, 0.002428f, 0.002732f, 0.003035f, + 0.003347f, 0.003677f, 0.004025f, 0.004391f, 0.004777f, 0.005182f, 0.005605f, 0.006049f, 0.006512f, 0.006995f, 0.007499f, + 0.008023f, 0.008568f, 0.009134f, 0.009721f, 0.010330f, 0.010960f, 0.011612f, 0.012286f, 0.012983f, 0.013702f, 0.014444f, + 0.015209f, 0.015996f, 0.016807f, 0.017642f, 0.018500f, 0.019382f, 0.020289f, 0.021219f, 0.022174f, 0.023153f, 0.024158f, + 0.025187f, 0.026241f, 0.027321f, 0.028426f, 0.029557f, 0.030713f, 0.031896f, 0.033105f, 0.034340f, 0.035601f, 0.036889f, + 0.038204f, 0.039546f, 0.040915f, 0.042311f, 0.043735f, 0.045186f, 0.046665f, 0.048172f, 0.049707f, 0.051269f, 0.052861f, + 0.054480f, 0.056128f, 0.057805f, 0.059511f, 0.061246f, 0.063010f, 0.064803f, 0.066626f, 0.068478f, 0.070360f, 0.072272f, + 0.074214f, 0.076185f, 0.078187f, 0.080220f, 0.082283f, 0.084376f, 0.086500f, 0.088656f, 0.090842f, 0.093059f, 0.095307f, + 0.097587f, 0.099899f, 0.102242f, 0.104616f, 0.107023f, 0.109462f, 0.111932f, 0.114435f, 0.116971f, 0.119538f, 0.122139f, + 0.124772f, 0.127438f, 0.130136f, 0.132868f, 0.135633f, 0.138432f, 0.141263f, 0.144128f, 0.147027f, 0.149960f, 0.152926f, + 0.155926f, 0.158961f, 0.162029f, 0.165132f, 0.168269f, 0.171441f, 0.174647f, 0.177888f, 0.181164f, 0.184475f, 0.187821f, + 0.191202f, 0.194618f, 0.198069f, 0.201556f, 0.205079f, 0.208637f, 0.212231f, 0.215861f, 0.219526f, 0.223228f, 0.226966f, + 0.230740f, 0.234551f, 0.238398f, 0.242281f, 0.246201f, 0.250158f, 0.254152f, 0.258183f, 0.262251f, 0.266356f, 0.270498f, + 0.274677f, 0.278894f, 0.283149f, 0.287441f, 0.291771f, 0.296138f, 0.300544f, 0.304987f, 0.309469f, 0.313989f, 0.318547f, + 0.323143f, 0.327778f, 0.332452f, 0.337164f, 0.341914f, 0.346704f, 0.351533f, 0.356400f, 0.361307f, 0.366253f, 0.371238f, + 0.376262f, 0.381326f, 0.386430f, 0.391573f, 0.396755f, 0.401978f, 0.407240f, 0.412543f, 0.417885f, 0.423268f, 0.428691f, + 0.434154f, 0.439657f, 0.445201f, 0.450786f, 0.456411f, 0.462077f, 0.467784f, 0.473532f, 0.479320f, 0.485150f, 0.491021f, + 0.496933f, 0.502887f, 0.508881f, 0.514918f, 0.520996f, 0.527115f, 0.533276f, 0.539480f, 0.545725f, 0.552011f, 0.558340f, + 0.564712f, 0.571125f, 0.577581f, 0.584078f, 0.590619f, 0.597202f, 0.603827f, 0.610496f, 0.617207f, 0.623960f, 0.630757f, + 0.637597f, 0.644480f, 0.651406f, 0.658375f, 0.665387f, 0.672443f, 0.679543f, 0.686685f, 0.693872f, 0.701102f, 0.708376f, + 0.715694f, 0.723055f, 0.730461f, 0.737911f, 0.745404f, 0.752942f, 0.760525f, 0.768151f, 0.775822f, 0.783538f, 0.791298f, + 0.799103f, 0.806952f, 0.814847f, 0.822786f, 0.830770f, 0.838799f, 0.846873f, 0.854993f, 0.863157f, 0.871367f, 0.879622f, + 0.887923f, 0.896269f, 0.904661f, 0.913099f, 0.921582f, 0.930111f, 0.938686f, 0.947307f, 0.955974f, 0.964686f, 0.973445f, + 0.982251f, 0.991102f, 1.0f +}; +static float stbir__srgb_to_linear(float f) +{ + if (f <= 0.04045f) + return f / 12.92f; + else + return (float)pow((f + 0.055f) / 1.055f, 2.4f); +} +static float stbir__linear_to_srgb(float f) +{ + if (f <= 0.0031308f) + return f * 12.92f; + else + return 1.055f * (float)pow(f, 1 / 2.4f) - 0.055f; +} +#ifndef STBIR_NON_IEEE_FLOAT +typedef union +{ + stbir_uint32 u; + float f; +} stbir__FP32; +static const stbir_uint32 fp32_to_srgb8_tab4[104] = { + 0x0073000d, 0x007a000d, 0x0080000d, 0x0087000d, 0x008d000d, 0x0094000d, 0x009a000d, 0x00a1000d, + 0x00a7001a, 0x00b4001a, 0x00c1001a, 0x00ce001a, 0x00da001a, 0x00e7001a, 0x00f4001a, 0x0101001a, + 0x010e0033, 0x01280033, 0x01410033, 0x015b0033, 0x01750033, 0x018f0033, 0x01a80033, 0x01c20033, + 0x01dc0067, 0x020f0067, 0x02430067, 0x02760067, 0x02aa0067, 0x02dd0067, 0x03110067, 0x03440067, + 0x037800ce, 0x03df00ce, 0x044600ce, 0x04ad00ce, 0x051400ce, 0x057b00c5, 0x05dd00bc, 0x063b00b5, + 0x06970158, 0x07420142, 0x07e30130, 0x087b0120, 0x090b0112, 0x09940106, 0x0a1700fc, 0x0a9500f2, + 0x0b0f01cb, 0x0bf401ae, 0x0ccb0195, 0x0d950180, 0x0e56016e, 0x0f0d015e, 0x0fbc0150, 0x10630143, + 0x11070264, 0x1238023e, 0x1357021d, 0x14660201, 0x156601e9, 0x165a01d3, 0x174401c0, 0x182401af, + 0x18fe0331, 0x1a9602fe, 0x1c1502d2, 0x1d7e02ad, 0x1ed4028d, 0x201a0270, 0x21520256, 0x227d0240, + 0x239f0443, 0x25c003fe, 0x27bf03c4, 0x29a10392, 0x2b6a0367, 0x2d1d0341, 0x2ebe031f, 0x304d0300, + 0x31d105b0, 0x34a80555, 0x37520507, 0x39d504c5, 0x3c37048b, 0x3e7c0458, 0x40a8042a, 0x42bd0401, + 0x44c20798, 0x488e071e, 0x4c1c06b6, 0x4f76065d, 0x52a50610, 0x55ac05cc, 0x5892058f, 0x5b590559, + 0x5e0c0a23, 0x631c0980, 0x67db08f6, 0x6c55087f, 0x70940818, 0x74a007bd, 0x787d076c, 0x7c330723, +}; +static stbir_uint8 stbir__linear_to_srgb_uchar(float in) +{ + static const stbir__FP32 almostone = { 0x3f7fffff }; + static const stbir__FP32 minval = { (127-13) << 23 }; + stbir_uint32 tab,bias,scale,t; + stbir__FP32 f; + + + + if (!(in > minval.f)) + in = minval.f; + if (in > almostone.f) + in = almostone.f; + + f.f = in; + tab = fp32_to_srgb8_tab4[(f.u - minval.u) >> 20]; + bias = (tab >> 16) << 9; + scale = tab & 0xffff; + + t = (f.u >> 12) & 0xff; + return (unsigned char) ((bias + scale*t) >> 16); +} +#else +static int stbir__srgb_offset_to_linear_scaled[256] = +{ + 0, 40738, 122216, 203693, 285170, 366648, 448125, 529603, + 611080, 692557, 774035, 855852, 942009, 1033024, 1128971, 1229926, + 1335959, 1447142, 1563542, 1685229, 1812268, 1944725, 2082664, 2226148, + 2375238, 2529996, 2690481, 2856753, 3028870, 3206888, 3390865, 3580856, + 3776916, 3979100, 4187460, 4402049, 4622919, 4850123, 5083710, 5323731, + 5570236, 5823273, 6082892, 6349140, 6622065, 6901714, 7188133, 7481369, + 7781466, 8088471, 8402427, 8723380, 9051372, 9386448, 9728650, 10078021, + 10434603, 10798439, 11169569, 11548036, 11933879, 12327139, 12727857, 13136073, + 13551826, 13975156, 14406100, 14844697, 15290987, 15745007, 16206795, 16676389, + 17153826, 17639142, 18132374, 18633560, 19142734, 19659934, 20185196, 20718552, + 21260042, 21809696, 22367554, 22933648, 23508010, 24090680, 24681686, 25281066, + 25888850, 26505076, 27129772, 27762974, 28404716, 29055026, 29713942, 30381490, + 31057708, 31742624, 32436272, 33138682, 33849884, 34569912, 35298800, 36036568, + 36783260, 37538896, 38303512, 39077136, 39859796, 40651528, 41452360, 42262316, + 43081432, 43909732, 44747252, 45594016, 46450052, 47315392, 48190064, 49074096, + 49967516, 50870356, 51782636, 52704392, 53635648, 54576432, 55526772, 56486700, + 57456236, 58435408, 59424248, 60422780, 61431036, 62449032, 63476804, 64514376, + 65561776, 66619028, 67686160, 68763192, 69850160, 70947088, 72053992, 73170912, + 74297864, 75434880, 76581976, 77739184, 78906536, 80084040, 81271736, 82469648, + 83677792, 84896192, 86124888, 87363888, 88613232, 89872928, 91143016, 92423512, + 93714432, 95015816, 96327688, 97650056, 98982952, 100326408, 101680440, 103045072, + 104420320, 105806224, 107202800, 108610064, 110028048, 111456776, 112896264, 114346544, + 115807632, 117279552, 118762328, 120255976, 121760536, 123276016, 124802440, 126339832, + 127888216, 129447616, 131018048, 132599544, 134192112, 135795792, 137410592, 139036528, + 140673648, 142321952, 143981456, 145652208, 147334208, 149027488, 150732064, 152447968, + 154175200, 155913792, 157663776, 159425168, 161197984, 162982240, 164777968, 166585184, + 168403904, 170234160, 172075968, 173929344, 175794320, 177670896, 179559120, 181458992, + 183370528, 185293776, 187228736, 189175424, 191133888, 193104112, 195086128, 197079968, + 199085648, 201103184, 203132592, 205173888, 207227120, 209292272, 211369392, 213458480, + 215559568, 217672656, 219797792, 221934976, 224084240, 226245600, 228419056, 230604656, + 232802400, 235012320, 237234432, 239468736, 241715280, 243974080, 246245120, 248528464, + 250824112, 253132064, 255452368, 257785040, 260130080, 262487520, 264857376, 267239664, +}; +static stbir_uint8 stbir__linear_to_srgb_uchar(float f) +{ + int x = (int) (f * (1 << 28)); + int v = 0; + int i; + + i = v + 128; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 64; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 32; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 16; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 8; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 4; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 2; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + i = v + 1; if (x >= stbir__srgb_offset_to_linear_scaled[i]) v = i; + return (stbir_uint8) v; +} +#endif +static float stbir__filter_trapezoid(float x, float scale) +{ + float halfscale = scale / 2; + float t = 0.5f + halfscale; + STBIR_ASSERT(scale <= 1); + x = (float)fabs(x); + if (x >= t) + return 0; + else + { + float r = 0.5f - halfscale; + if (x <= r) + return 1; + else + return (t - x) / scale; + } +} +static float stbir__support_trapezoid(float scale) +{ + STBIR_ASSERT(scale <= 1); + return 0.5f + scale / 2; +} +static float stbir__filter_triangle(float x, float s) +{ + STBIR__UNUSED_PARAM(s); + x = (float)fabs(x); + if (x <= 1.0f) + return 1 - x; + else + return 0; +} +static float stbir__filter_cubic(float x, float s) +{ + STBIR__UNUSED_PARAM(s); + x = (float)fabs(x); + if (x < 1.0f) + return (4 + x*x*(3*x - 6))/6; + else if (x < 2.0f) + return (8 + x*(-12 + x*(6 - x)))/6; + return (0.0f); +} +static float stbir__filter_catmullrom(float x, float s) +{ + STBIR__UNUSED_PARAM(s); + x = (float)fabs(x); + if (x < 1.0f) + return 1 - x*x*(2.5f - 1.5f*x); + else if (x < 2.0f) + return 2 - x*(4 + x*(0.5f*x - 2.5f)); + return (0.0f); +} +static float stbir__filter_mitchell(float x, float s) +{ + STBIR__UNUSED_PARAM(s); + x = (float)fabs(x); + if (x < 1.0f) + return (16 + x*x*(21 * x - 36))/18; + else if (x < 2.0f) + return (32 + x*(-60 + x*(36 - 7*x)))/18; + return (0.0f); +} +static float stbir__support_zero(float s) +{ + STBIR__UNUSED_PARAM(s); + return 0; +} +static float stbir__support_one(float s) +{ + STBIR__UNUSED_PARAM(s); + return 1; +} +static float stbir__support_two(float s) +{ + STBIR__UNUSED_PARAM(s); + return 2; +} +static stbir__filter_info stbir__filter_info_table[] = { + { NULL, stbir__support_zero }, + { stbir__filter_trapezoid, stbir__support_trapezoid }, + { stbir__filter_triangle, stbir__support_one }, + { stbir__filter_cubic, stbir__support_two }, + { stbir__filter_catmullrom, stbir__support_two }, + { stbir__filter_mitchell, stbir__support_two }, +}; +stbir__inline static int stbir__use_upsampling(float ratio) +{ + return ratio > 1; +} +stbir__inline static int stbir__use_width_upsampling(stbir__info* stbir_info) +{ + return stbir__use_upsampling(stbir_info->horizontal_scale); +} +stbir__inline static int stbir__use_height_upsampling(stbir__info* stbir_info) +{ + return stbir__use_upsampling(stbir_info->vertical_scale); +} +static int stbir__get_filter_pixel_width(stbir_filter filter, float scale) +{ + STBIR_ASSERT(filter != 0); + STBIR_ASSERT(filter < STBIR__ARRAY_SIZE(stbir__filter_info_table)); + if (stbir__use_upsampling(scale)) + return (int)ceil(stbir__filter_info_table[filter].support(1/scale) * 2); + else + return (int)ceil(stbir__filter_info_table[filter].support(scale) * 2 / scale); +} +static int stbir__get_filter_pixel_margin(stbir_filter filter, float scale) +{ + return stbir__get_filter_pixel_width(filter, scale) / 2; +} +static int stbir__get_coefficient_width(stbir_filter filter, float scale) +{ + if (stbir__use_upsampling(scale)) + return (int)ceil(stbir__filter_info_table[filter].support(1 / scale) * 2); + else + return (int)ceil(stbir__filter_info_table[filter].support(scale) * 2); +} +static int stbir__get_contributors(float scale, stbir_filter filter, int input_size, int output_size) +{ + if (stbir__use_upsampling(scale)) + return output_size; + else + return (input_size + stbir__get_filter_pixel_margin(filter, scale) * 2); +} +static int stbir__get_total_horizontal_coefficients(stbir__info* info) +{ + return info->horizontal_num_contributors + * stbir__get_coefficient_width (info->horizontal_filter, info->horizontal_scale); +} +static int stbir__get_total_vertical_coefficients(stbir__info* info) +{ + return info->vertical_num_contributors + * stbir__get_coefficient_width (info->vertical_filter, info->vertical_scale); +} +static stbir__contributors* stbir__get_contributor(stbir__contributors* contributors, int n) +{ + return &contributors[n]; +} +static float* stbir__get_coefficient(float* coefficients, stbir_filter filter, float scale, int n, int c) +{ + int width = stbir__get_coefficient_width(filter, scale); + return &coefficients[width*n + c]; +} +static int stbir__edge_wrap_slow(stbir_edge edge, int n, int max) +{ + switch (edge) + { + case STBIR_EDGE_ZERO: + return 0; + case STBIR_EDGE_CLAMP: + if (n < 0) + return 0; + if (n >= max) + return max - 1; + return n; + case STBIR_EDGE_REFLECT: + { + if (n < 0) + { + if (n < max) + return -n; + else + return max - 1; + } + if (n >= max) + { + int max2 = max * 2; + if (n >= max2) + return 0; + else + return max2 - n - 1; + } + return n; + } + case STBIR_EDGE_WRAP: + if (n >= 0) + return (n % max); + else + { + int m = (-n) % max; + if (m != 0) + m = max - m; + return (m); + } + + default: + STBIR_ASSERT(!"Unimplemented edge type"); + return 0; + } +} +stbir__inline static int stbir__edge_wrap(stbir_edge edge, int n, int max) +{ + + if (n >= 0 && n < max) + return n; + return stbir__edge_wrap_slow(edge, n, max); +} +static void stbir__calculate_sample_range_upsample(int n, float out_filter_radius, float scale_ratio, float out_shift, int* in_first_pixel, int* in_last_pixel, float* in_center_of_out) +{ + float out_pixel_center = (float)n + 0.5f; + float out_pixel_influence_lowerbound = out_pixel_center - out_filter_radius; + float out_pixel_influence_upperbound = out_pixel_center + out_filter_radius; + float in_pixel_influence_lowerbound = (out_pixel_influence_lowerbound + out_shift) / scale_ratio; + float in_pixel_influence_upperbound = (out_pixel_influence_upperbound + out_shift) / scale_ratio; + *in_center_of_out = (out_pixel_center + out_shift) / scale_ratio; + *in_first_pixel = (int)(floor(in_pixel_influence_lowerbound + 0.5)); + *in_last_pixel = (int)(floor(in_pixel_influence_upperbound - 0.5)); +} +static void stbir__calculate_sample_range_downsample(int n, float in_pixels_radius, float scale_ratio, float out_shift, int* out_first_pixel, int* out_last_pixel, float* out_center_of_in) +{ + float in_pixel_center = (float)n + 0.5f; + float in_pixel_influence_lowerbound = in_pixel_center - in_pixels_radius; + float in_pixel_influence_upperbound = in_pixel_center + in_pixels_radius; + float out_pixel_influence_lowerbound = in_pixel_influence_lowerbound * scale_ratio - out_shift; + float out_pixel_influence_upperbound = in_pixel_influence_upperbound * scale_ratio - out_shift; + *out_center_of_in = in_pixel_center * scale_ratio - out_shift; + *out_first_pixel = (int)(floor(out_pixel_influence_lowerbound + 0.5)); + *out_last_pixel = (int)(floor(out_pixel_influence_upperbound - 0.5)); +} +static void stbir__calculate_coefficients_upsample(stbir_filter filter, float scale, int in_first_pixel, int in_last_pixel, float in_center_of_out, stbir__contributors* contributor, float* coefficient_group) +{ + int i; + float total_filter = 0; + float filter_scale; + STBIR_ASSERT(in_last_pixel - in_first_pixel <= (int)ceil(stbir__filter_info_table[filter].support(1/scale) * 2)); + contributor->n0 = in_first_pixel; + contributor->n1 = in_last_pixel; + STBIR_ASSERT(contributor->n1 >= contributor->n0); + for (i = 0; i <= in_last_pixel - in_first_pixel; i++) + { + float in_pixel_center = (float)(i + in_first_pixel) + 0.5f; + coefficient_group[i] = stbir__filter_info_table[filter].kernel(in_center_of_out - in_pixel_center, 1 / scale); + + if (i == 0 && !coefficient_group[i]) + { + contributor->n0 = ++in_first_pixel; + i--; + continue; + } + total_filter += coefficient_group[i]; + } + STBIR_ASSERT(stbir__filter_info_table[filter].kernel((float)(in_last_pixel + 1) + 0.5f - in_center_of_out, 1/scale) == 0); + STBIR_ASSERT(total_filter > 0.9); + STBIR_ASSERT(total_filter < 1.1f); + + filter_scale = 1 / total_filter; + for (i = 0; i <= in_last_pixel - in_first_pixel; i++) + coefficient_group[i] *= filter_scale; + for (i = in_last_pixel - in_first_pixel; i >= 0; i--) + { + if (coefficient_group[i]) + break; + + contributor->n1 = contributor->n0 + i - 1; + } +} +static void stbir__calculate_coefficients_downsample(stbir_filter filter, float scale_ratio, int out_first_pixel, int out_last_pixel, float out_center_of_in, stbir__contributors* contributor, float* coefficient_group) +{ + int i; + STBIR_ASSERT(out_last_pixel - out_first_pixel <= (int)ceil(stbir__filter_info_table[filter].support(scale_ratio) * 2)); + contributor->n0 = out_first_pixel; + contributor->n1 = out_last_pixel; + STBIR_ASSERT(contributor->n1 >= contributor->n0); + for (i = 0; i <= out_last_pixel - out_first_pixel; i++) + { + float out_pixel_center = (float)(i + out_first_pixel) + 0.5f; + float x = out_pixel_center - out_center_of_in; + coefficient_group[i] = stbir__filter_info_table[filter].kernel(x, scale_ratio) * scale_ratio; + } + STBIR_ASSERT(stbir__filter_info_table[filter].kernel((float)(out_last_pixel + 1) + 0.5f - out_center_of_in, scale_ratio) == 0); + for (i = out_last_pixel - out_first_pixel; i >= 0; i--) + { + if (coefficient_group[i]) + break; + + contributor->n1 = contributor->n0 + i - 1; + } +} +static void stbir__normalize_downsample_coefficients(stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, int input_size, int output_size) +{ + int num_contributors = stbir__get_contributors(scale_ratio, filter, input_size, output_size); + int num_coefficients = stbir__get_coefficient_width(filter, scale_ratio); + int i, j; + int skip; + for (i = 0; i < output_size; i++) + { + float scale; + float total = 0; + for (j = 0; j < num_contributors; j++) + { + if (i >= contributors[j].n0 && i <= contributors[j].n1) + { + float coefficient = *stbir__get_coefficient(coefficients, filter, scale_ratio, j, i - contributors[j].n0); + total += coefficient; + } + else if (i < contributors[j].n0) + break; + } + STBIR_ASSERT(total > 0.9f); + STBIR_ASSERT(total < 1.1f); + scale = 1 / total; + for (j = 0; j < num_contributors; j++) + { + if (i >= contributors[j].n0 && i <= contributors[j].n1) + *stbir__get_coefficient(coefficients, filter, scale_ratio, j, i - contributors[j].n0) *= scale; + else if (i < contributors[j].n0) + break; + } + } + + + for (j = 0; j < num_contributors; j++) + { + int range, max, width; + skip = 0; + while (*stbir__get_coefficient(coefficients, filter, scale_ratio, j, skip) == 0) + skip++; + contributors[j].n0 += skip; + while (contributors[j].n0 < 0) + { + contributors[j].n0++; + skip++; + } + range = contributors[j].n1 - contributors[j].n0 + 1; + max = stbir__min(num_coefficients, range); + width = stbir__get_coefficient_width(filter, scale_ratio); + for (i = 0; i < max; i++) + { + if (i + skip >= width) + break; + *stbir__get_coefficient(coefficients, filter, scale_ratio, j, i) = *stbir__get_coefficient(coefficients, filter, scale_ratio, j, i + skip); + } + continue; + } + + for (i = 0; i < num_contributors; i++) + contributors[i].n1 = stbir__min(contributors[i].n1, output_size - 1); +} +static void stbir__calculate_filters(stbir__contributors* contributors, float* coefficients, stbir_filter filter, float scale_ratio, float shift, int input_size, int output_size) +{ + int n; + int total_contributors = stbir__get_contributors(scale_ratio, filter, input_size, output_size); + if (stbir__use_upsampling(scale_ratio)) + { + float out_pixels_radius = stbir__filter_info_table[filter].support(1 / scale_ratio) * scale_ratio; + + for (n = 0; n < total_contributors; n++) + { + float in_center_of_out; + int in_first_pixel, in_last_pixel; + stbir__calculate_sample_range_upsample(n, out_pixels_radius, scale_ratio, shift, &in_first_pixel, &in_last_pixel, &in_center_of_out); + stbir__calculate_coefficients_upsample(filter, scale_ratio, in_first_pixel, in_last_pixel, in_center_of_out, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0)); + } + } + else + { + float in_pixels_radius = stbir__filter_info_table[filter].support(scale_ratio) / scale_ratio; + + for (n = 0; n < total_contributors; n++) + { + float out_center_of_in; + int out_first_pixel, out_last_pixel; + int n_adjusted = n - stbir__get_filter_pixel_margin(filter, scale_ratio); + stbir__calculate_sample_range_downsample(n_adjusted, in_pixels_radius, scale_ratio, shift, &out_first_pixel, &out_last_pixel, &out_center_of_in); + stbir__calculate_coefficients_downsample(filter, scale_ratio, out_first_pixel, out_last_pixel, out_center_of_in, stbir__get_contributor(contributors, n), stbir__get_coefficient(coefficients, filter, scale_ratio, n, 0)); + } + stbir__normalize_downsample_coefficients(contributors, coefficients, filter, scale_ratio, input_size, output_size); + } +} +static float* stbir__get_decode_buffer(stbir__info* stbir_info) +{ + + + return &stbir_info->decode_buffer[stbir_info->horizontal_filter_pixel_margin * stbir_info->channels]; +} +#define STBIR__DECODE(type, colorspace) ((int)(type) * (STBIR_MAX_COLORSPACES) + (int)(colorspace)) +static void stbir__decode_scanline(stbir__info* stbir_info, int n) +{ + int c; + int channels = stbir_info->channels; + int alpha_channel = stbir_info->alpha_channel; + int type = stbir_info->type; + int colorspace = stbir_info->colorspace; + int input_w = stbir_info->input_w; + size_t input_stride_bytes = stbir_info->input_stride_bytes; + float* decode_buffer = stbir__get_decode_buffer(stbir_info); + stbir_edge edge_horizontal = stbir_info->edge_horizontal; + stbir_edge edge_vertical = stbir_info->edge_vertical; + size_t in_buffer_row_offset = stbir__edge_wrap(edge_vertical, n, stbir_info->input_h) * input_stride_bytes; + const void* input_data = (char *) stbir_info->input_data + in_buffer_row_offset; + int max_x = input_w + stbir_info->horizontal_filter_pixel_margin; + int decode = STBIR__DECODE(type, colorspace); + int x = -stbir_info->horizontal_filter_pixel_margin; + + + if (edge_vertical == STBIR_EDGE_ZERO && (n < 0 || n >= stbir_info->input_h)) + { + for (; x < max_x; x++) + for (c = 0; c < channels; c++) + decode_buffer[x*channels + c] = 0; + return; + } + switch (decode) + { + case STBIR__DECODE(STBIR_TYPE_UINT8, STBIR_COLORSPACE_LINEAR): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = ((float)((const unsigned char*)input_data)[input_pixel_index + c]) / stbir__max_uint8_as_float; + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT8, STBIR_COLORSPACE_SRGB): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = stbir__srgb_uchar_to_linear_float[((const unsigned char*)input_data)[input_pixel_index + c]]; + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned char*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint8_as_float; + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT16, STBIR_COLORSPACE_LINEAR): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = ((float)((const unsigned short*)input_data)[input_pixel_index + c]) / stbir__max_uint16_as_float; + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT16, STBIR_COLORSPACE_SRGB): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear(((float)((const unsigned short*)input_data)[input_pixel_index + c]) / stbir__max_uint16_as_float); + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + decode_buffer[decode_pixel_index + alpha_channel] = ((float)((const unsigned short*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint16_as_float; + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT32, STBIR_COLORSPACE_LINEAR): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / stbir__max_uint32_as_float); + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT32, STBIR_COLORSPACE_SRGB): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear((float)(((double)((const unsigned int*)input_data)[input_pixel_index + c]) / stbir__max_uint32_as_float)); + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + decode_buffer[decode_pixel_index + alpha_channel] = (float)(((double)((const unsigned int*)input_data)[input_pixel_index + alpha_channel]) / stbir__max_uint32_as_float); + } + break; + case STBIR__DECODE(STBIR_TYPE_FLOAT, STBIR_COLORSPACE_LINEAR): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = ((const float*)input_data)[input_pixel_index + c]; + } + break; + case STBIR__DECODE(STBIR_TYPE_FLOAT, STBIR_COLORSPACE_SRGB): + for (; x < max_x; x++) + { + int decode_pixel_index = x * channels; + int input_pixel_index = stbir__edge_wrap(edge_horizontal, x, input_w) * channels; + for (c = 0; c < channels; c++) + decode_buffer[decode_pixel_index + c] = stbir__srgb_to_linear(((const float*)input_data)[input_pixel_index + c]); + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + decode_buffer[decode_pixel_index + alpha_channel] = ((const float*)input_data)[input_pixel_index + alpha_channel]; + } + break; + default: + STBIR_ASSERT(!"Unknown type/colorspace/channels combination."); + break; + } + if (!(stbir_info->flags & STBIR_FLAG_ALPHA_PREMULTIPLIED)) + { + for (x = -stbir_info->horizontal_filter_pixel_margin; x < max_x; x++) + { + int decode_pixel_index = x * channels; + + float alpha = decode_buffer[decode_pixel_index + alpha_channel]; +#ifndef STBIR_NO_ALPHA_EPSILON + if (stbir_info->type != STBIR_TYPE_FLOAT) { + alpha += STBIR_ALPHA_EPSILON; + decode_buffer[decode_pixel_index + alpha_channel] = alpha; + } +#endif + for (c = 0; c < channels; c++) + { + if (c == alpha_channel) + continue; + decode_buffer[decode_pixel_index + c] *= alpha; + } + } + } + if (edge_horizontal == STBIR_EDGE_ZERO) + { + for (x = -stbir_info->horizontal_filter_pixel_margin; x < 0; x++) + { + for (c = 0; c < channels; c++) + decode_buffer[x*channels + c] = 0; + } + for (x = input_w; x < max_x; x++) + { + for (c = 0; c < channels; c++) + decode_buffer[x*channels + c] = 0; + } + } +} +static float* stbir__get_ring_buffer_entry(float* ring_buffer, int index, int ring_buffer_length) +{ + return &ring_buffer[index * ring_buffer_length]; +} +static float* stbir__add_empty_ring_buffer_entry(stbir__info* stbir_info, int n) +{ + int ring_buffer_index; + float* ring_buffer; + stbir_info->ring_buffer_last_scanline = n; + if (stbir_info->ring_buffer_begin_index < 0) + { + ring_buffer_index = stbir_info->ring_buffer_begin_index = 0; + stbir_info->ring_buffer_first_scanline = n; + } + else + { + ring_buffer_index = (stbir_info->ring_buffer_begin_index + (stbir_info->ring_buffer_last_scanline - stbir_info->ring_buffer_first_scanline)) % stbir_info->ring_buffer_num_entries; + STBIR_ASSERT(ring_buffer_index != stbir_info->ring_buffer_begin_index); + } + ring_buffer = stbir__get_ring_buffer_entry(stbir_info->ring_buffer, ring_buffer_index, stbir_info->ring_buffer_length_bytes / sizeof(float)); + memset(ring_buffer, 0, stbir_info->ring_buffer_length_bytes); + return ring_buffer; +} +static void stbir__resample_horizontal_upsample(stbir__info* stbir_info, float* output_buffer) +{ + int x, k; + int output_w = stbir_info->output_w; + int channels = stbir_info->channels; + float* decode_buffer = stbir__get_decode_buffer(stbir_info); + stbir__contributors* horizontal_contributors = stbir_info->horizontal_contributors; + float* horizontal_coefficients = stbir_info->horizontal_coefficients; + int coefficient_width = stbir_info->horizontal_coefficient_width; + for (x = 0; x < output_w; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int out_pixel_index = x * channels; + int coefficient_group = coefficient_width * x; + int coefficient_counter = 0; + STBIR_ASSERT(n1 >= n0); + STBIR_ASSERT(n0 >= -stbir_info->horizontal_filter_pixel_margin); + STBIR_ASSERT(n1 >= -stbir_info->horizontal_filter_pixel_margin); + STBIR_ASSERT(n0 < stbir_info->input_w + stbir_info->horizontal_filter_pixel_margin); + STBIR_ASSERT(n1 < stbir_info->input_w + stbir_info->horizontal_filter_pixel_margin); + switch (channels) { + case 1: + for (k = n0; k <= n1; k++) + { + int in_pixel_index = k * 1; + float coefficient = horizontal_coefficients[coefficient_group + coefficient_counter++]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + } + break; + case 2: + for (k = n0; k <= n1; k++) + { + int in_pixel_index = k * 2; + float coefficient = horizontal_coefficients[coefficient_group + coefficient_counter++]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + } + break; + case 3: + for (k = n0; k <= n1; k++) + { + int in_pixel_index = k * 3; + float coefficient = horizontal_coefficients[coefficient_group + coefficient_counter++]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + output_buffer[out_pixel_index + 2] += decode_buffer[in_pixel_index + 2] * coefficient; + } + break; + case 4: + for (k = n0; k <= n1; k++) + { + int in_pixel_index = k * 4; + float coefficient = horizontal_coefficients[coefficient_group + coefficient_counter++]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + output_buffer[out_pixel_index + 2] += decode_buffer[in_pixel_index + 2] * coefficient; + output_buffer[out_pixel_index + 3] += decode_buffer[in_pixel_index + 3] * coefficient; + } + break; + default: + for (k = n0; k <= n1; k++) + { + int in_pixel_index = k * channels; + float coefficient = horizontal_coefficients[coefficient_group + coefficient_counter++]; + int c; + STBIR_ASSERT(coefficient != 0); + for (c = 0; c < channels; c++) + output_buffer[out_pixel_index + c] += decode_buffer[in_pixel_index + c] * coefficient; + } + break; + } + } +} +static void stbir__resample_horizontal_downsample(stbir__info* stbir_info, float* output_buffer) +{ + int x, k; + int input_w = stbir_info->input_w; + int channels = stbir_info->channels; + float* decode_buffer = stbir__get_decode_buffer(stbir_info); + stbir__contributors* horizontal_contributors = stbir_info->horizontal_contributors; + float* horizontal_coefficients = stbir_info->horizontal_coefficients; + int coefficient_width = stbir_info->horizontal_coefficient_width; + int filter_pixel_margin = stbir_info->horizontal_filter_pixel_margin; + int max_x = input_w + filter_pixel_margin * 2; + STBIR_ASSERT(!stbir__use_width_upsampling(stbir_info)); + switch (channels) { + case 1: + for (x = 0; x < max_x; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int in_x = x - filter_pixel_margin; + int in_pixel_index = in_x * 1; + int max_n = n1; + int coefficient_group = coefficient_width * x; + for (k = n0; k <= max_n; k++) + { + int out_pixel_index = k * 1; + float coefficient = horizontal_coefficients[coefficient_group + k - n0]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + } + } + break; + case 2: + for (x = 0; x < max_x; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int in_x = x - filter_pixel_margin; + int in_pixel_index = in_x * 2; + int max_n = n1; + int coefficient_group = coefficient_width * x; + for (k = n0; k <= max_n; k++) + { + int out_pixel_index = k * 2; + float coefficient = horizontal_coefficients[coefficient_group + k - n0]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + } + } + break; + case 3: + for (x = 0; x < max_x; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int in_x = x - filter_pixel_margin; + int in_pixel_index = in_x * 3; + int max_n = n1; + int coefficient_group = coefficient_width * x; + for (k = n0; k <= max_n; k++) + { + int out_pixel_index = k * 3; + float coefficient = horizontal_coefficients[coefficient_group + k - n0]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + output_buffer[out_pixel_index + 2] += decode_buffer[in_pixel_index + 2] * coefficient; + } + } + break; + case 4: + for (x = 0; x < max_x; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int in_x = x - filter_pixel_margin; + int in_pixel_index = in_x * 4; + int max_n = n1; + int coefficient_group = coefficient_width * x; + for (k = n0; k <= max_n; k++) + { + int out_pixel_index = k * 4; + float coefficient = horizontal_coefficients[coefficient_group + k - n0]; + STBIR_ASSERT(coefficient != 0); + output_buffer[out_pixel_index + 0] += decode_buffer[in_pixel_index + 0] * coefficient; + output_buffer[out_pixel_index + 1] += decode_buffer[in_pixel_index + 1] * coefficient; + output_buffer[out_pixel_index + 2] += decode_buffer[in_pixel_index + 2] * coefficient; + output_buffer[out_pixel_index + 3] += decode_buffer[in_pixel_index + 3] * coefficient; + } + } + break; + default: + for (x = 0; x < max_x; x++) + { + int n0 = horizontal_contributors[x].n0; + int n1 = horizontal_contributors[x].n1; + int in_x = x - filter_pixel_margin; + int in_pixel_index = in_x * channels; + int max_n = n1; + int coefficient_group = coefficient_width * x; + for (k = n0; k <= max_n; k++) + { + int c; + int out_pixel_index = k * channels; + float coefficient = horizontal_coefficients[coefficient_group + k - n0]; + STBIR_ASSERT(coefficient != 0); + for (c = 0; c < channels; c++) + output_buffer[out_pixel_index + c] += decode_buffer[in_pixel_index + c] * coefficient; + } + } + break; + } +} +static void stbir__decode_and_resample_upsample(stbir__info* stbir_info, int n) +{ + + stbir__decode_scanline(stbir_info, n); + + if (stbir__use_width_upsampling(stbir_info)) + stbir__resample_horizontal_upsample(stbir_info, stbir__add_empty_ring_buffer_entry(stbir_info, n)); + else + stbir__resample_horizontal_downsample(stbir_info, stbir__add_empty_ring_buffer_entry(stbir_info, n)); + +} +static void stbir__decode_and_resample_downsample(stbir__info* stbir_info, int n) +{ + + stbir__decode_scanline(stbir_info, n); + memset(stbir_info->horizontal_buffer, 0, stbir_info->output_w * stbir_info->channels * sizeof(float)); + + if (stbir__use_width_upsampling(stbir_info)) + stbir__resample_horizontal_upsample(stbir_info, stbir_info->horizontal_buffer); + else + stbir__resample_horizontal_downsample(stbir_info, stbir_info->horizontal_buffer); + +} +static float* stbir__get_ring_buffer_scanline(int get_scanline, float* ring_buffer, int begin_index, int first_scanline, int ring_buffer_num_entries, int ring_buffer_length) +{ + int ring_buffer_index = (begin_index + (get_scanline - first_scanline)) % ring_buffer_num_entries; + return stbir__get_ring_buffer_entry(ring_buffer, ring_buffer_index, ring_buffer_length); +} +static void stbir__encode_scanline(stbir__info* stbir_info, int num_pixels, void *output_buffer, float *encode_buffer, int channels, int alpha_channel, int decode) +{ + int x; + int n; + int num_nonalpha; + stbir_uint16 nonalpha[STBIR_MAX_CHANNELS]; + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_PREMULTIPLIED)) + { + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + float alpha = encode_buffer[pixel_index + alpha_channel]; + float reciprocal_alpha = alpha ? 1.0f / alpha : 0; + + for (n = 0; n < channels; n++) + if (n != alpha_channel) + encode_buffer[pixel_index + n] *= reciprocal_alpha; + + + + + } + } + + + for (x = 0, num_nonalpha = 0; x < channels; ++x) + { + if (x != alpha_channel || (stbir_info->flags & STBIR_FLAG_ALPHA_USES_COLORSPACE)) + { + nonalpha[num_nonalpha++] = (stbir_uint16)x; + } + } + #define STBIR__ROUND_INT(f) ((int) ((f)+0.5)) + #define STBIR__ROUND_UINT(f) ((stbir_uint32) ((f)+0.5)) + #ifdef STBIR__SATURATE_INT + #define STBIR__ENCODE_LINEAR8(f) stbir__saturate8 (STBIR__ROUND_INT((f) * stbir__max_uint8_as_float )) + #define STBIR__ENCODE_LINEAR16(f) stbir__saturate16(STBIR__ROUND_INT((f) * stbir__max_uint16_as_float)) + #else + #define STBIR__ENCODE_LINEAR8(f) (unsigned char ) STBIR__ROUND_INT(stbir__saturate(f) * stbir__max_uint8_as_float ) + #define STBIR__ENCODE_LINEAR16(f) (unsigned short) STBIR__ROUND_INT(stbir__saturate(f) * stbir__max_uint16_as_float) + #endif + switch (decode) + { + case STBIR__DECODE(STBIR_TYPE_UINT8, STBIR_COLORSPACE_LINEAR): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < channels; n++) + { + int index = pixel_index + n; + ((unsigned char*)output_buffer)[index] = STBIR__ENCODE_LINEAR8(encode_buffer[index]); + } + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT8, STBIR_COLORSPACE_SRGB): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < num_nonalpha; n++) + { + int index = pixel_index + nonalpha[n]; + ((unsigned char*)output_buffer)[index] = stbir__linear_to_srgb_uchar(encode_buffer[index]); + } + if (!(stbir_info->flags & STBIR_FLAG_ALPHA_USES_COLORSPACE)) + ((unsigned char *)output_buffer)[pixel_index + alpha_channel] = STBIR__ENCODE_LINEAR8(encode_buffer[pixel_index+alpha_channel]); + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT16, STBIR_COLORSPACE_LINEAR): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < channels; n++) + { + int index = pixel_index + n; + ((unsigned short*)output_buffer)[index] = STBIR__ENCODE_LINEAR16(encode_buffer[index]); + } + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT16, STBIR_COLORSPACE_SRGB): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < num_nonalpha; n++) + { + int index = pixel_index + nonalpha[n]; + ((unsigned short*)output_buffer)[index] = (unsigned short)STBIR__ROUND_INT(stbir__linear_to_srgb(stbir__saturate(encode_buffer[index])) * stbir__max_uint16_as_float); + } + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + ((unsigned short*)output_buffer)[pixel_index + alpha_channel] = STBIR__ENCODE_LINEAR16(encode_buffer[pixel_index + alpha_channel]); + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT32, STBIR_COLORSPACE_LINEAR): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < channels; n++) + { + int index = pixel_index + n; + ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__saturate(encode_buffer[index])) * stbir__max_uint32_as_float); + } + } + break; + case STBIR__DECODE(STBIR_TYPE_UINT32, STBIR_COLORSPACE_SRGB): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < num_nonalpha; n++) + { + int index = pixel_index + nonalpha[n]; + ((unsigned int*)output_buffer)[index] = (unsigned int)STBIR__ROUND_UINT(((double)stbir__linear_to_srgb(stbir__saturate(encode_buffer[index]))) * stbir__max_uint32_as_float); + } + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + ((unsigned int*)output_buffer)[pixel_index + alpha_channel] = (unsigned int)STBIR__ROUND_INT(((double)stbir__saturate(encode_buffer[pixel_index + alpha_channel])) * stbir__max_uint32_as_float); + } + break; + case STBIR__DECODE(STBIR_TYPE_FLOAT, STBIR_COLORSPACE_LINEAR): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < channels; n++) + { + int index = pixel_index + n; + ((float*)output_buffer)[index] = encode_buffer[index]; + } + } + break; + case STBIR__DECODE(STBIR_TYPE_FLOAT, STBIR_COLORSPACE_SRGB): + for (x=0; x < num_pixels; ++x) + { + int pixel_index = x*channels; + for (n = 0; n < num_nonalpha; n++) + { + int index = pixel_index + nonalpha[n]; + ((float*)output_buffer)[index] = stbir__linear_to_srgb(encode_buffer[index]); + } + if (!(stbir_info->flags&STBIR_FLAG_ALPHA_USES_COLORSPACE)) + ((float*)output_buffer)[pixel_index + alpha_channel] = encode_buffer[pixel_index + alpha_channel]; + } + break; + default: + STBIR_ASSERT(!"Unknown type/colorspace/channels combination."); + break; + } +} +static void stbir__resample_vertical_upsample(stbir__info* stbir_info, int n) +{ + int x, k; + int output_w = stbir_info->output_w; + stbir__contributors* vertical_contributors = stbir_info->vertical_contributors; + float* vertical_coefficients = stbir_info->vertical_coefficients; + int channels = stbir_info->channels; + int alpha_channel = stbir_info->alpha_channel; + int type = stbir_info->type; + int colorspace = stbir_info->colorspace; + int ring_buffer_entries = stbir_info->ring_buffer_num_entries; + void* output_data = stbir_info->output_data; + float* encode_buffer = stbir_info->encode_buffer; + int decode = STBIR__DECODE(type, colorspace); + int coefficient_width = stbir_info->vertical_coefficient_width; + int coefficient_counter; + int contributor = n; + float* ring_buffer = stbir_info->ring_buffer; + int ring_buffer_begin_index = stbir_info->ring_buffer_begin_index; + int ring_buffer_first_scanline = stbir_info->ring_buffer_first_scanline; + int ring_buffer_length = stbir_info->ring_buffer_length_bytes/sizeof(float); + int n0,n1, output_row_start; + int coefficient_group = coefficient_width * contributor; + n0 = vertical_contributors[contributor].n0; + n1 = vertical_contributors[contributor].n1; + output_row_start = n * stbir_info->output_stride_bytes; + STBIR_ASSERT(stbir__use_height_upsampling(stbir_info)); + memset(encode_buffer, 0, output_w * sizeof(float) * channels); + + + coefficient_counter = 0; + switch (channels) { + case 1: + for (k = n0; k <= n1; k++) + { + int coefficient_index = coefficient_counter++; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + for (x = 0; x < output_w; ++x) + { + int in_pixel_index = x * 1; + encode_buffer[in_pixel_index + 0] += ring_buffer_entry[in_pixel_index + 0] * coefficient; + } + } + break; + case 2: + for (k = n0; k <= n1; k++) + { + int coefficient_index = coefficient_counter++; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + for (x = 0; x < output_w; ++x) + { + int in_pixel_index = x * 2; + encode_buffer[in_pixel_index + 0] += ring_buffer_entry[in_pixel_index + 0] * coefficient; + encode_buffer[in_pixel_index + 1] += ring_buffer_entry[in_pixel_index + 1] * coefficient; + } + } + break; + case 3: + for (k = n0; k <= n1; k++) + { + int coefficient_index = coefficient_counter++; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + for (x = 0; x < output_w; ++x) + { + int in_pixel_index = x * 3; + encode_buffer[in_pixel_index + 0] += ring_buffer_entry[in_pixel_index + 0] * coefficient; + encode_buffer[in_pixel_index + 1] += ring_buffer_entry[in_pixel_index + 1] * coefficient; + encode_buffer[in_pixel_index + 2] += ring_buffer_entry[in_pixel_index + 2] * coefficient; + } + } + break; + case 4: + for (k = n0; k <= n1; k++) + { + int coefficient_index = coefficient_counter++; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + for (x = 0; x < output_w; ++x) + { + int in_pixel_index = x * 4; + encode_buffer[in_pixel_index + 0] += ring_buffer_entry[in_pixel_index + 0] * coefficient; + encode_buffer[in_pixel_index + 1] += ring_buffer_entry[in_pixel_index + 1] * coefficient; + encode_buffer[in_pixel_index + 2] += ring_buffer_entry[in_pixel_index + 2] * coefficient; + encode_buffer[in_pixel_index + 3] += ring_buffer_entry[in_pixel_index + 3] * coefficient; + } + } + break; + default: + for (k = n0; k <= n1; k++) + { + int coefficient_index = coefficient_counter++; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + for (x = 0; x < output_w; ++x) + { + int in_pixel_index = x * channels; + int c; + for (c = 0; c < channels; c++) + encode_buffer[in_pixel_index + c] += ring_buffer_entry[in_pixel_index + c] * coefficient; + } + } + break; + } + stbir__encode_scanline(stbir_info, output_w, (char *) output_data + output_row_start, encode_buffer, channels, alpha_channel, decode); +} +static void stbir__resample_vertical_downsample(stbir__info* stbir_info, int n) +{ + int x, k; + int output_w = stbir_info->output_w; + stbir__contributors* vertical_contributors = stbir_info->vertical_contributors; + float* vertical_coefficients = stbir_info->vertical_coefficients; + int channels = stbir_info->channels; + int ring_buffer_entries = stbir_info->ring_buffer_num_entries; + float* horizontal_buffer = stbir_info->horizontal_buffer; + int coefficient_width = stbir_info->vertical_coefficient_width; + int contributor = n + stbir_info->vertical_filter_pixel_margin; + float* ring_buffer = stbir_info->ring_buffer; + int ring_buffer_begin_index = stbir_info->ring_buffer_begin_index; + int ring_buffer_first_scanline = stbir_info->ring_buffer_first_scanline; + int ring_buffer_length = stbir_info->ring_buffer_length_bytes/sizeof(float); + int n0,n1; + n0 = vertical_contributors[contributor].n0; + n1 = vertical_contributors[contributor].n1; + STBIR_ASSERT(!stbir__use_height_upsampling(stbir_info)); + for (k = n0; k <= n1; k++) + { + int coefficient_index = k - n0; + int coefficient_group = coefficient_width * contributor; + float coefficient = vertical_coefficients[coefficient_group + coefficient_index]; + float* ring_buffer_entry = stbir__get_ring_buffer_scanline(k, ring_buffer, ring_buffer_begin_index, ring_buffer_first_scanline, ring_buffer_entries, ring_buffer_length); + switch (channels) { + case 1: + for (x = 0; x < output_w; x++) + { + int in_pixel_index = x * 1; + ring_buffer_entry[in_pixel_index + 0] += horizontal_buffer[in_pixel_index + 0] * coefficient; + } + break; + case 2: + for (x = 0; x < output_w; x++) + { + int in_pixel_index = x * 2; + ring_buffer_entry[in_pixel_index + 0] += horizontal_buffer[in_pixel_index + 0] * coefficient; + ring_buffer_entry[in_pixel_index + 1] += horizontal_buffer[in_pixel_index + 1] * coefficient; + } + break; + case 3: + for (x = 0; x < output_w; x++) + { + int in_pixel_index = x * 3; + ring_buffer_entry[in_pixel_index + 0] += horizontal_buffer[in_pixel_index + 0] * coefficient; + ring_buffer_entry[in_pixel_index + 1] += horizontal_buffer[in_pixel_index + 1] * coefficient; + ring_buffer_entry[in_pixel_index + 2] += horizontal_buffer[in_pixel_index + 2] * coefficient; + } + break; + case 4: + for (x = 0; x < output_w; x++) + { + int in_pixel_index = x * 4; + ring_buffer_entry[in_pixel_index + 0] += horizontal_buffer[in_pixel_index + 0] * coefficient; + ring_buffer_entry[in_pixel_index + 1] += horizontal_buffer[in_pixel_index + 1] * coefficient; + ring_buffer_entry[in_pixel_index + 2] += horizontal_buffer[in_pixel_index + 2] * coefficient; + ring_buffer_entry[in_pixel_index + 3] += horizontal_buffer[in_pixel_index + 3] * coefficient; + } + break; + default: + for (x = 0; x < output_w; x++) + { + int in_pixel_index = x * channels; + int c; + for (c = 0; c < channels; c++) + ring_buffer_entry[in_pixel_index + c] += horizontal_buffer[in_pixel_index + c] * coefficient; + } + break; + } + } +} +static void stbir__buffer_loop_upsample(stbir__info* stbir_info) +{ + int y; + float scale_ratio = stbir_info->vertical_scale; + float out_scanlines_radius = stbir__filter_info_table[stbir_info->vertical_filter].support(1/scale_ratio) * scale_ratio; + STBIR_ASSERT(stbir__use_height_upsampling(stbir_info)); + for (y = 0; y < stbir_info->output_h; y++) + { + float in_center_of_out = 0; + int in_first_scanline = 0, in_last_scanline = 0; + stbir__calculate_sample_range_upsample(y, out_scanlines_radius, scale_ratio, stbir_info->vertical_shift, &in_first_scanline, &in_last_scanline, &in_center_of_out); + STBIR_ASSERT(in_last_scanline - in_first_scanline + 1 <= stbir_info->ring_buffer_num_entries); + if (stbir_info->ring_buffer_begin_index >= 0) + { + + while (in_first_scanline > stbir_info->ring_buffer_first_scanline) + { + if (stbir_info->ring_buffer_first_scanline == stbir_info->ring_buffer_last_scanline) + { + + + stbir_info->ring_buffer_begin_index = -1; + stbir_info->ring_buffer_first_scanline = 0; + stbir_info->ring_buffer_last_scanline = 0; + break; + } + else + { + stbir_info->ring_buffer_first_scanline++; + stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->ring_buffer_num_entries; + } + } + } + + if (stbir_info->ring_buffer_begin_index < 0) + stbir__decode_and_resample_upsample(stbir_info, in_first_scanline); + while (in_last_scanline > stbir_info->ring_buffer_last_scanline) + stbir__decode_and_resample_upsample(stbir_info, stbir_info->ring_buffer_last_scanline + 1); + + stbir__resample_vertical_upsample(stbir_info, y); + STBIR_PROGRESS_REPORT((float)y / stbir_info->output_h); + } +} +static void stbir__empty_ring_buffer(stbir__info* stbir_info, int first_necessary_scanline) +{ + int output_stride_bytes = stbir_info->output_stride_bytes; + int channels = stbir_info->channels; + int alpha_channel = stbir_info->alpha_channel; + int type = stbir_info->type; + int colorspace = stbir_info->colorspace; + int output_w = stbir_info->output_w; + void* output_data = stbir_info->output_data; + int decode = STBIR__DECODE(type, colorspace); + float* ring_buffer = stbir_info->ring_buffer; + int ring_buffer_length = stbir_info->ring_buffer_length_bytes/sizeof(float); + if (stbir_info->ring_buffer_begin_index >= 0) + { + + while (first_necessary_scanline > stbir_info->ring_buffer_first_scanline) + { + if (stbir_info->ring_buffer_first_scanline >= 0 && stbir_info->ring_buffer_first_scanline < stbir_info->output_h) + { + int output_row_start = stbir_info->ring_buffer_first_scanline * output_stride_bytes; + float* ring_buffer_entry = stbir__get_ring_buffer_entry(ring_buffer, stbir_info->ring_buffer_begin_index, ring_buffer_length); + stbir__encode_scanline(stbir_info, output_w, (char *) output_data + output_row_start, ring_buffer_entry, channels, alpha_channel, decode); + STBIR_PROGRESS_REPORT((float)stbir_info->ring_buffer_first_scanline / stbir_info->output_h); + } + if (stbir_info->ring_buffer_first_scanline == stbir_info->ring_buffer_last_scanline) + { + + + stbir_info->ring_buffer_begin_index = -1; + stbir_info->ring_buffer_first_scanline = 0; + stbir_info->ring_buffer_last_scanline = 0; + break; + } + else + { + stbir_info->ring_buffer_first_scanline++; + stbir_info->ring_buffer_begin_index = (stbir_info->ring_buffer_begin_index + 1) % stbir_info->ring_buffer_num_entries; + } + } + } +} +static void stbir__buffer_loop_downsample(stbir__info* stbir_info) +{ + int y; + float scale_ratio = stbir_info->vertical_scale; + int output_h = stbir_info->output_h; + float in_pixels_radius = stbir__filter_info_table[stbir_info->vertical_filter].support(scale_ratio) / scale_ratio; + int pixel_margin = stbir_info->vertical_filter_pixel_margin; + int max_y = stbir_info->input_h + pixel_margin; + STBIR_ASSERT(!stbir__use_height_upsampling(stbir_info)); + for (y = -pixel_margin; y < max_y; y++) + { + float out_center_of_in; + int out_first_scanline, out_last_scanline; + stbir__calculate_sample_range_downsample(y, in_pixels_radius, scale_ratio, stbir_info->vertical_shift, &out_first_scanline, &out_last_scanline, &out_center_of_in); + STBIR_ASSERT(out_last_scanline - out_first_scanline + 1 <= stbir_info->ring_buffer_num_entries); + if (out_last_scanline < 0 || out_first_scanline >= output_h) + continue; + stbir__empty_ring_buffer(stbir_info, out_first_scanline); + stbir__decode_and_resample_downsample(stbir_info, y); + + if (stbir_info->ring_buffer_begin_index < 0) + stbir__add_empty_ring_buffer_entry(stbir_info, out_first_scanline); + while (out_last_scanline > stbir_info->ring_buffer_last_scanline) + stbir__add_empty_ring_buffer_entry(stbir_info, stbir_info->ring_buffer_last_scanline + 1); + + stbir__resample_vertical_downsample(stbir_info, y); + } + stbir__empty_ring_buffer(stbir_info, stbir_info->output_h); +} +static void stbir__setup(stbir__info *info, int input_w, int input_h, int output_w, int output_h, int channels) +{ + info->input_w = input_w; + info->input_h = input_h; + info->output_w = output_w; + info->output_h = output_h; + info->channels = channels; +} +static void stbir__calculate_transform(stbir__info *info, float s0, float t0, float s1, float t1, float *transform) +{ + info->s0 = s0; + info->t0 = t0; + info->s1 = s1; + info->t1 = t1; + if (transform) + { + info->horizontal_scale = transform[0]; + info->vertical_scale = transform[1]; + info->horizontal_shift = transform[2]; + info->vertical_shift = transform[3]; + } + else + { + info->horizontal_scale = ((float)info->output_w / info->input_w) / (s1 - s0); + info->vertical_scale = ((float)info->output_h / info->input_h) / (t1 - t0); + info->horizontal_shift = s0 * info->output_w / (s1 - s0); + info->vertical_shift = t0 * info->output_h / (t1 - t0); + } +} +static void stbir__choose_filter(stbir__info *info, stbir_filter h_filter, stbir_filter v_filter) +{ + if (h_filter == 0) + h_filter = stbir__use_upsampling(info->horizontal_scale) ? STBIR_DEFAULT_FILTER_UPSAMPLE : STBIR_DEFAULT_FILTER_DOWNSAMPLE; + if (v_filter == 0) + v_filter = stbir__use_upsampling(info->vertical_scale) ? STBIR_DEFAULT_FILTER_UPSAMPLE : STBIR_DEFAULT_FILTER_DOWNSAMPLE; + info->horizontal_filter = h_filter; + info->vertical_filter = v_filter; +} +static stbir_uint32 stbir__calculate_memory(stbir__info *info) +{ + int pixel_margin = stbir__get_filter_pixel_margin(info->horizontal_filter, info->horizontal_scale); + int filter_height = stbir__get_filter_pixel_width(info->vertical_filter, info->vertical_scale); + info->horizontal_num_contributors = stbir__get_contributors(info->horizontal_scale, info->horizontal_filter, info->input_w, info->output_w); + info->vertical_num_contributors = stbir__get_contributors(info->vertical_scale , info->vertical_filter , info->input_h, info->output_h); + + info->ring_buffer_num_entries = filter_height + 1; + info->horizontal_contributors_size = info->horizontal_num_contributors * sizeof(stbir__contributors); + info->horizontal_coefficients_size = stbir__get_total_horizontal_coefficients(info) * sizeof(float); + info->vertical_contributors_size = info->vertical_num_contributors * sizeof(stbir__contributors); + info->vertical_coefficients_size = stbir__get_total_vertical_coefficients(info) * sizeof(float); + info->decode_buffer_size = (info->input_w + pixel_margin * 2) * info->channels * sizeof(float); + info->horizontal_buffer_size = info->output_w * info->channels * sizeof(float); + info->ring_buffer_size = info->output_w * info->channels * info->ring_buffer_num_entries * sizeof(float); + info->encode_buffer_size = info->output_w * info->channels * sizeof(float); + STBIR_ASSERT(info->horizontal_filter != 0); + STBIR_ASSERT(info->horizontal_filter < STBIR__ARRAY_SIZE(stbir__filter_info_table)); + STBIR_ASSERT(info->vertical_filter != 0); + STBIR_ASSERT(info->vertical_filter < STBIR__ARRAY_SIZE(stbir__filter_info_table)); + if (stbir__use_height_upsampling(info)) + + + + info->horizontal_buffer_size = 0; + else + + + info->encode_buffer_size = 0; + return info->horizontal_contributors_size + info->horizontal_coefficients_size + + info->vertical_contributors_size + info->vertical_coefficients_size + + info->decode_buffer_size + info->horizontal_buffer_size + + info->ring_buffer_size + info->encode_buffer_size; +} +static int stbir__resize_allocated(stbir__info *info, + const void* input_data, int input_stride_in_bytes, + void* output_data, int output_stride_in_bytes, + int alpha_channel, stbir_uint32 flags, stbir_datatype type, + stbir_edge edge_horizontal, stbir_edge edge_vertical, stbir_colorspace colorspace, + void* tempmem, size_t tempmem_size_in_bytes) +{ + size_t memory_required = stbir__calculate_memory(info); + int width_stride_input = input_stride_in_bytes ? input_stride_in_bytes : info->channels * info->input_w * stbir__type_size[type]; + int width_stride_output = output_stride_in_bytes ? output_stride_in_bytes : info->channels * info->output_w * stbir__type_size[type]; +#ifdef STBIR_DEBUG_OVERWRITE_TEST +#define OVERWRITE_ARRAY_SIZE 8 + unsigned char overwrite_output_before_pre[OVERWRITE_ARRAY_SIZE]; + unsigned char overwrite_tempmem_before_pre[OVERWRITE_ARRAY_SIZE]; + unsigned char overwrite_output_after_pre[OVERWRITE_ARRAY_SIZE]; + unsigned char overwrite_tempmem_after_pre[OVERWRITE_ARRAY_SIZE]; + size_t begin_forbidden = width_stride_output * (info->output_h - 1) + info->output_w * info->channels * stbir__type_size[type]; + memcpy(overwrite_output_before_pre, &((unsigned char*)output_data)[-OVERWRITE_ARRAY_SIZE], OVERWRITE_ARRAY_SIZE); + memcpy(overwrite_output_after_pre, &((unsigned char*)output_data)[begin_forbidden], OVERWRITE_ARRAY_SIZE); + memcpy(overwrite_tempmem_before_pre, &((unsigned char*)tempmem)[-OVERWRITE_ARRAY_SIZE], OVERWRITE_ARRAY_SIZE); + memcpy(overwrite_tempmem_after_pre, &((unsigned char*)tempmem)[tempmem_size_in_bytes], OVERWRITE_ARRAY_SIZE); +#endif + STBIR_ASSERT(info->channels >= 0); + STBIR_ASSERT(info->channels <= STBIR_MAX_CHANNELS); + if (info->channels < 0 || info->channels > STBIR_MAX_CHANNELS) + return 0; + STBIR_ASSERT(info->horizontal_filter < STBIR__ARRAY_SIZE(stbir__filter_info_table)); + STBIR_ASSERT(info->vertical_filter < STBIR__ARRAY_SIZE(stbir__filter_info_table)); + if (info->horizontal_filter >= STBIR__ARRAY_SIZE(stbir__filter_info_table)) + return 0; + if (info->vertical_filter >= STBIR__ARRAY_SIZE(stbir__filter_info_table)) + return 0; + if (alpha_channel < 0) + flags |= STBIR_FLAG_ALPHA_USES_COLORSPACE | STBIR_FLAG_ALPHA_PREMULTIPLIED; + if (!(flags&STBIR_FLAG_ALPHA_USES_COLORSPACE) || !(flags&STBIR_FLAG_ALPHA_PREMULTIPLIED)) { + STBIR_ASSERT(alpha_channel >= 0 && alpha_channel < info->channels); + } + if (alpha_channel >= info->channels) + return 0; + STBIR_ASSERT(tempmem); + if (!tempmem) + return 0; + STBIR_ASSERT(tempmem_size_in_bytes >= memory_required); + if (tempmem_size_in_bytes < memory_required) + return 0; + memset(tempmem, 0, tempmem_size_in_bytes); + info->input_data = input_data; + info->input_stride_bytes = width_stride_input; + info->output_data = output_data; + info->output_stride_bytes = width_stride_output; + info->alpha_channel = alpha_channel; + info->flags = flags; + info->type = type; + info->edge_horizontal = edge_horizontal; + info->edge_vertical = edge_vertical; + info->colorspace = colorspace; + info->horizontal_coefficient_width = stbir__get_coefficient_width (info->horizontal_filter, info->horizontal_scale); + info->vertical_coefficient_width = stbir__get_coefficient_width (info->vertical_filter , info->vertical_scale ); + info->horizontal_filter_pixel_width = stbir__get_filter_pixel_width (info->horizontal_filter, info->horizontal_scale); + info->vertical_filter_pixel_width = stbir__get_filter_pixel_width (info->vertical_filter , info->vertical_scale ); + info->horizontal_filter_pixel_margin = stbir__get_filter_pixel_margin(info->horizontal_filter, info->horizontal_scale); + info->vertical_filter_pixel_margin = stbir__get_filter_pixel_margin(info->vertical_filter , info->vertical_scale ); + info->ring_buffer_length_bytes = info->output_w * info->channels * sizeof(float); + info->decode_buffer_pixels = info->input_w + info->horizontal_filter_pixel_margin * 2; +#define STBIR__NEXT_MEMPTR(current, newtype) (newtype*)(((unsigned char*)current) + current##_size) + info->horizontal_contributors = (stbir__contributors *) tempmem; + info->horizontal_coefficients = STBIR__NEXT_MEMPTR(info->horizontal_contributors, float); + info->vertical_contributors = STBIR__NEXT_MEMPTR(info->horizontal_coefficients, stbir__contributors); + info->vertical_coefficients = STBIR__NEXT_MEMPTR(info->vertical_contributors, float); + info->decode_buffer = STBIR__NEXT_MEMPTR(info->vertical_coefficients, float); + if (stbir__use_height_upsampling(info)) + { + info->horizontal_buffer = NULL; + info->ring_buffer = STBIR__NEXT_MEMPTR(info->decode_buffer, float); + info->encode_buffer = STBIR__NEXT_MEMPTR(info->ring_buffer, float); + STBIR_ASSERT((size_t)STBIR__NEXT_MEMPTR(info->encode_buffer, unsigned char) == (size_t)tempmem + tempmem_size_in_bytes); + } + else + { + info->horizontal_buffer = STBIR__NEXT_MEMPTR(info->decode_buffer, float); + info->ring_buffer = STBIR__NEXT_MEMPTR(info->horizontal_buffer, float); + info->encode_buffer = NULL; + STBIR_ASSERT((size_t)STBIR__NEXT_MEMPTR(info->ring_buffer, unsigned char) == (size_t)tempmem + tempmem_size_in_bytes); + } +#undef STBIR__NEXT_MEMPTR + + info->ring_buffer_begin_index = -1; + stbir__calculate_filters(info->horizontal_contributors, info->horizontal_coefficients, info->horizontal_filter, info->horizontal_scale, info->horizontal_shift, info->input_w, info->output_w); + stbir__calculate_filters(info->vertical_contributors, info->vertical_coefficients, info->vertical_filter, info->vertical_scale, info->vertical_shift, info->input_h, info->output_h); + STBIR_PROGRESS_REPORT(0); + if (stbir__use_height_upsampling(info)) + stbir__buffer_loop_upsample(info); + else + stbir__buffer_loop_downsample(info); + STBIR_PROGRESS_REPORT(1); +#ifdef STBIR_DEBUG_OVERWRITE_TEST + STBIR_ASSERT(memcmp(overwrite_output_before_pre, &((unsigned char*)output_data)[-OVERWRITE_ARRAY_SIZE], OVERWRITE_ARRAY_SIZE) == 0); + STBIR_ASSERT(memcmp(overwrite_output_after_pre, &((unsigned char*)output_data)[begin_forbidden], OVERWRITE_ARRAY_SIZE) == 0); + STBIR_ASSERT(memcmp(overwrite_tempmem_before_pre, &((unsigned char*)tempmem)[-OVERWRITE_ARRAY_SIZE], OVERWRITE_ARRAY_SIZE) == 0); + STBIR_ASSERT(memcmp(overwrite_tempmem_after_pre, &((unsigned char*)tempmem)[tempmem_size_in_bytes], OVERWRITE_ARRAY_SIZE) == 0); +#endif + return 1; +} +static int stbir__resize_arbitrary( + void *alloc_context, + const void* input_data, int input_w, int input_h, int input_stride_in_bytes, + void* output_data, int output_w, int output_h, int output_stride_in_bytes, + float s0, float t0, float s1, float t1, float *transform, + int channels, int alpha_channel, stbir_uint32 flags, stbir_datatype type, + stbir_filter h_filter, stbir_filter v_filter, + stbir_edge edge_horizontal, stbir_edge edge_vertical, stbir_colorspace colorspace) +{ + stbir__info info; + int result; + size_t memory_required; + void* extra_memory; + stbir__setup(&info, input_w, input_h, output_w, output_h, channels); + stbir__calculate_transform(&info, s0,t0,s1,t1,transform); + stbir__choose_filter(&info, h_filter, v_filter); + memory_required = stbir__calculate_memory(&info); + extra_memory = STBIR_MALLOC(memory_required, alloc_context); + if (!extra_memory) + return 0; + result = stbir__resize_allocated(&info, input_data, input_stride_in_bytes, + output_data, output_stride_in_bytes, + alpha_channel, flags, type, + edge_horizontal, edge_vertical, + colorspace, extra_memory, memory_required); + STBIR_FREE(extra_memory, alloc_context); + return result; +} +STBIRDEF int stbir_resize_uint8( const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels) +{ + return stbir__resize_arbitrary(NULL, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,-1,0, STBIR_TYPE_UINT8, STBIR_FILTER_DEFAULT, STBIR_FILTER_DEFAULT, + STBIR_EDGE_CLAMP, STBIR_EDGE_CLAMP, STBIR_COLORSPACE_LINEAR); +} +STBIRDEF int stbir_resize_float( const float *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + float *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels) +{ + return stbir__resize_arbitrary(NULL, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,-1,0, STBIR_TYPE_FLOAT, STBIR_FILTER_DEFAULT, STBIR_FILTER_DEFAULT, + STBIR_EDGE_CLAMP, STBIR_EDGE_CLAMP, STBIR_COLORSPACE_LINEAR); +} +STBIRDEF int stbir_resize_uint8_srgb(const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags) +{ + return stbir__resize_arbitrary(NULL, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, STBIR_TYPE_UINT8, STBIR_FILTER_DEFAULT, STBIR_FILTER_DEFAULT, + STBIR_EDGE_CLAMP, STBIR_EDGE_CLAMP, STBIR_COLORSPACE_SRGB); +} +STBIRDEF int stbir_resize_uint8_srgb_edgemode(const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode) +{ + return stbir__resize_arbitrary(NULL, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, STBIR_TYPE_UINT8, STBIR_FILTER_DEFAULT, STBIR_FILTER_DEFAULT, + edge_wrap_mode, edge_wrap_mode, STBIR_COLORSPACE_SRGB); +} +STBIRDEF int stbir_resize_uint8_generic( const unsigned char *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + unsigned char *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context) +{ + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, STBIR_TYPE_UINT8, filter, filter, + edge_wrap_mode, edge_wrap_mode, space); +} +STBIRDEF int stbir_resize_uint16_generic(const stbir_uint16 *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + stbir_uint16 *output_pixels , int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context) +{ + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, STBIR_TYPE_UINT16, filter, filter, + edge_wrap_mode, edge_wrap_mode, space); +} +STBIRDEF int stbir_resize_float_generic( const float *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + float *output_pixels , int output_w, int output_h, int output_stride_in_bytes, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_wrap_mode, stbir_filter filter, stbir_colorspace space, + void *alloc_context) +{ + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, STBIR_TYPE_FLOAT, filter, filter, + edge_wrap_mode, edge_wrap_mode, space); +} +STBIRDEF int stbir_resize( const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context) +{ + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,NULL,num_channels,alpha_channel,flags, datatype, filter_horizontal, filter_vertical, + edge_mode_horizontal, edge_mode_vertical, space); +} +STBIRDEF int stbir_resize_subpixel(const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context, + float x_scale, float y_scale, + float x_offset, float y_offset) +{ + float transform[4]; + transform[0] = x_scale; + transform[1] = y_scale; + transform[2] = x_offset; + transform[3] = y_offset; + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + 0,0,1,1,transform,num_channels,alpha_channel,flags, datatype, filter_horizontal, filter_vertical, + edge_mode_horizontal, edge_mode_vertical, space); +} +STBIRDEF int stbir_resize_region( const void *input_pixels , int input_w , int input_h , int input_stride_in_bytes, + void *output_pixels, int output_w, int output_h, int output_stride_in_bytes, + stbir_datatype datatype, + int num_channels, int alpha_channel, int flags, + stbir_edge edge_mode_horizontal, stbir_edge edge_mode_vertical, + stbir_filter filter_horizontal, stbir_filter filter_vertical, + stbir_colorspace space, void *alloc_context, + float s0, float t0, float s1, float t1) +{ + return stbir__resize_arbitrary(alloc_context, input_pixels, input_w, input_h, input_stride_in_bytes, + output_pixels, output_w, output_h, output_stride_in_bytes, + s0,t0,s1,t1,NULL,num_channels,alpha_channel,flags, datatype, filter_horizontal, filter_vertical, + edge_mode_horizontal, edge_mode_vertical, space); +} +#endif diff --git a/src/3rd_party/stb/stb_image_write.h b/src/3rd_party/stb/stb_image_write.h new file mode 100644 index 0000000..9872d12 --- /dev/null +++ b/src/3rd_party/stb/stb_image_write.h @@ -0,0 +1,1286 @@ +/* View upstream repo: https://github.com/nothings/stb +------------------------------------------------------------------------------ +This software is available under 2 licenses -- choose whichever you prefer. +------------------------------------------------------------------------------ +ALTERNATIVE A - MIT License +Copyright (c) 2017 Sean Barrett +Permission is hereby granted, free of charge, to any person 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, sublicense, 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: +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. +------------------------------------------------------------------------------ +ALTERNATIVE B - Public Domain (www.unlicense.org) +This is free and unencumbered software released into the public domain. +Anyone is free to copy, modify, publish, use, compile, sell, or distribute this +software, either in source code form or as a compiled binary, for any purpose, +commercial or non-commercial, and by any means. +In jurisdictions that recognize copyright laws, the author or authors of this +software dedicate any and all copyright interest in the software to the public +domain. We make this dedication for the benefit of the public at large and to +the detriment of our heirs and successors. We intend this dedication to be an +overt act of relinquishment in perpetuity of all present and future rights to +this software under copyright law. +THE SOFTWARE IS PROVIDED "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. IN NO EVENT SHALL THE +AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN +ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION +WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +------------------------------------------------------------------------------ +*/ + +#ifndef INCLUDE_STB_IMAGE_WRITE_H +#define INCLUDE_STB_IMAGE_WRITE_H +#include +#ifndef STBIWDEF +#ifdef STB_IMAGE_WRITE_STATIC +#define STBIWDEF static +#else +#ifdef __cplusplus +#define STBIWDEF extern "C" +#else +#define STBIWDEF extern +#endif +#endif +#endif +#ifndef STB_IMAGE_WRITE_STATIC +extern int stbi_write_tga_with_rle; +extern int stbi_write_png_compression_level; +extern int stbi_write_force_png_filter; +#endif +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_png(char const *filename, int w, int h, int comp, const void *data, int stride_in_bytes); +STBIWDEF int stbi_write_bmp(char const *filename, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_tga(char const *filename, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_hdr(char const *filename, int w, int h, int comp, const float *data); +STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void *data, int quality); +#ifdef STBI_WINDOWS_UTF8 +STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input); +#endif +#endif +typedef void stbi_write_func(void *context, void *data, int size); +STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data, int stride_in_bytes); +STBIWDEF int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const void *data); +STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int w, int h, int comp, const float *data); +STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality); +STBIWDEF void stbi_flip_vertically_on_write(int flip_boolean); +#endif +#ifdef STB_IMAGE_WRITE_IMPLEMENTATION +#ifdef _WIN32 + #ifndef _CRT_SECURE_NO_WARNINGS + #define _CRT_SECURE_NO_WARNINGS + #endif + #ifndef _CRT_NONSTDC_NO_DEPRECATE + #define _CRT_NONSTDC_NO_DEPRECATE + #endif +#endif +#ifndef STBI_WRITE_NO_STDIO +#include +#endif +#include +#include +#include +#include +#if defined(STBIW_MALLOC) && defined(STBIW_FREE) && (defined(STBIW_REALLOC) || defined(STBIW_REALLOC_SIZED)) +#elif !defined(STBIW_MALLOC) && !defined(STBIW_FREE) && !defined(STBIW_REALLOC) && !defined(STBIW_REALLOC_SIZED) +#else +#error "Must define all or none of STBIW_MALLOC, STBIW_FREE, and STBIW_REALLOC (or STBIW_REALLOC_SIZED)." +#endif +#ifndef STBIW_MALLOC +#define STBIW_MALLOC(sz) malloc(sz) +#define STBIW_REALLOC(p,newsz) realloc(p,newsz) +#define STBIW_FREE(p) free(p) +#endif +#ifndef STBIW_REALLOC_SIZED +#define STBIW_REALLOC_SIZED(p,oldsz,newsz) STBIW_REALLOC(p,newsz) +#endif +#ifndef STBIW_MEMMOVE +#define STBIW_MEMMOVE(a,b,sz) memmove(a,b,sz) +#endif +#ifndef STBIW_ASSERT +#include +#define STBIW_ASSERT(x) assert(x) +#endif +#define STBIW_UCHAR(x) (unsigned char) ((x) & 0xff) +#ifdef STB_IMAGE_WRITE_STATIC +static int stbi_write_png_compression_level = 8; +static int stbi_write_tga_with_rle = 1; +static int stbi_write_force_png_filter = -1; +#else +int stbi_write_png_compression_level = 8; +int stbi_write_tga_with_rle = 1; +int stbi_write_force_png_filter = -1; +#endif +static int stbi__flip_vertically_on_write = 0; +STBIWDEF void stbi_flip_vertically_on_write(int flag) +{ + stbi__flip_vertically_on_write = flag; +} +typedef struct +{ + stbi_write_func *func; + void *context; +} stbi__write_context; +static void stbi__start_write_callbacks(stbi__write_context *s, stbi_write_func *c, void *context) +{ + s->func = c; + s->context = context; +} +#ifndef STBI_WRITE_NO_STDIO +static void stbi__stdio_write(void *context, void *data, int size) +{ + fwrite(data,1,size,(FILE*) context); +} +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) +#ifdef __cplusplus +#define STBIW_EXTERN extern "C" +#else +#define STBIW_EXTERN extern +#endif +STBIW_EXTERN __declspec(dllimport) int __stdcall MultiByteToWideChar(unsigned int cp, unsigned long flags, const char *str, int cbmb, wchar_t *widestr, int cchwide); +STBIW_EXTERN __declspec(dllimport) int __stdcall WideCharToMultiByte(unsigned int cp, unsigned long flags, const wchar_t *widestr, int cchwide, char *str, int cbmb, const char *defchar, int *used_default); +STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input) +{ + return WideCharToMultiByte(65001 , 0, input, -1, buffer, (int) bufferlen, NULL, NULL); +} +#endif +static FILE *stbiw__fopen(char const *filename, char const *mode) +{ + FILE *f; +#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8) + wchar_t wMode[64]; + wchar_t wFilename[1024]; + if (0 == MultiByteToWideChar(65001 , 0, filename, -1, wFilename, sizeof(wFilename))) + return 0; + if (0 == MultiByteToWideChar(65001 , 0, mode, -1, wMode, sizeof(wMode))) + return 0; +#if _MSC_VER >= 1400 + if (0 != _wfopen_s(&f, wFilename, wMode)) + f = 0; +#else + f = _wfopen(wFilename, wMode); +#endif +#elif defined(_MSC_VER) && _MSC_VER >= 1400 + if (0 != fopen_s(&f, filename, mode)) + f=0; +#else + f = fopen(filename, mode); +#endif + return f; +} +static int stbi__start_write_file(stbi__write_context *s, const char *filename) +{ + FILE *f = stbiw__fopen(filename, "wb"); + stbi__start_write_callbacks(s, stbi__stdio_write, (void *) f); + return f != NULL; +} +static void stbi__end_write_file(stbi__write_context *s) +{ + fclose((FILE *)s->context); +} +#endif +typedef unsigned int stbiw_uint32; +typedef int stb_image_write_test[sizeof(stbiw_uint32)==4 ? 1 : -1]; +static void stbiw__writefv(stbi__write_context *s, const char *fmt, va_list v) +{ + while (*fmt) { + switch (*fmt++) { + case ' ': break; + case '1': { unsigned char x = STBIW_UCHAR(va_arg(v, int)); + s->func(s->context,&x,1); + break; } + case '2': { int x = va_arg(v,int); + unsigned char b[2]; + b[0] = STBIW_UCHAR(x); + b[1] = STBIW_UCHAR(x>>8); + s->func(s->context,b,2); + break; } + case '4': { stbiw_uint32 x = va_arg(v,int); + unsigned char b[4]; + b[0]=STBIW_UCHAR(x); + b[1]=STBIW_UCHAR(x>>8); + b[2]=STBIW_UCHAR(x>>16); + b[3]=STBIW_UCHAR(x>>24); + s->func(s->context,b,4); + break; } + default: + STBIW_ASSERT(0); + return; + } + } +} +static void stbiw__writef(stbi__write_context *s, const char *fmt, ...) +{ + va_list v; + va_start(v, fmt); + stbiw__writefv(s, fmt, v); + va_end(v); +} +static void stbiw__putc(stbi__write_context *s, unsigned char c) +{ + s->func(s->context, &c, 1); +} +static void stbiw__write3(stbi__write_context *s, unsigned char a, unsigned char b, unsigned char c) +{ + unsigned char arr[3]; + arr[0] = a; arr[1] = b; arr[2] = c; + s->func(s->context, arr, 3); +} +static void stbiw__write_pixel(stbi__write_context *s, int rgb_dir, int comp, int write_alpha, int expand_mono, unsigned char *d) +{ + unsigned char bg[3] = { 255, 0, 255}, px[3]; + int k; + if (write_alpha < 0) + s->func(s->context, &d[comp - 1], 1); + switch (comp) { + case 2: + case 1: + if (expand_mono) + stbiw__write3(s, d[0], d[0], d[0]); + else + s->func(s->context, d, 1); + break; + case 4: + if (!write_alpha) { + + for (k = 0; k < 3; ++k) + px[k] = bg[k] + ((d[k] - bg[k]) * d[3]) / 255; + stbiw__write3(s, px[1 - rgb_dir], px[1], px[1 + rgb_dir]); + break; + } + + case 3: + stbiw__write3(s, d[1 - rgb_dir], d[1], d[1 + rgb_dir]); + break; + } + if (write_alpha > 0) + s->func(s->context, &d[comp - 1], 1); +} +static void stbiw__write_pixels(stbi__write_context *s, int rgb_dir, int vdir, int x, int y, int comp, void *data, int write_alpha, int scanline_pad, int expand_mono) +{ + stbiw_uint32 zero = 0; + int i,j, j_end; + if (y <= 0) + return; + if (stbi__flip_vertically_on_write) + vdir *= -1; + if (vdir < 0) { + j_end = -1; j = y-1; + } else { + j_end = y; j = 0; + } + for (; j != j_end; j += vdir) { + for (i=0; i < x; ++i) { + unsigned char *d = (unsigned char *) data + (j*x+i)*comp; + stbiw__write_pixel(s, rgb_dir, comp, write_alpha, expand_mono, d); + } + s->func(s->context, &zero, scanline_pad); + } +} +static int stbiw__outfile(stbi__write_context *s, int rgb_dir, int vdir, int x, int y, int comp, int expand_mono, void *data, int alpha, int pad, const char *fmt, ...) +{ + if (y < 0 || x < 0) { + return 0; + } else { + va_list v; + va_start(v, fmt); + stbiw__writefv(s, fmt, v); + va_end(v); + stbiw__write_pixels(s,rgb_dir,vdir,x,y,comp,data,alpha,pad, expand_mono); + return 1; + } +} +static int stbi_write_bmp_core(stbi__write_context *s, int x, int y, int comp, const void *data) +{ + int pad = (-x*3) & 3; + return stbiw__outfile(s,-1,-1,x,y,comp,1,(void *) data,0,pad, + "11 4 22 4" "4 44 22 444444", + 'B', 'M', 14+40+(x*3+pad)*y, 0,0, 14+40, + 40, x,y, 1,24, 0,0,0,0,0,0); +} +STBIWDEF int stbi_write_bmp_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_bmp_core(&s, x, y, comp, data); +} +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_bmp(char const *filename, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_bmp_core(&s, x, y, comp, data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif +static int stbi_write_tga_core(stbi__write_context *s, int x, int y, int comp, void *data) +{ + int has_alpha = (comp == 2 || comp == 4); + int colorbytes = has_alpha ? comp-1 : comp; + int format = colorbytes < 2 ? 3 : 2; + if (y < 0 || x < 0) + return 0; + if (!stbi_write_tga_with_rle) { + return stbiw__outfile(s, -1, -1, x, y, comp, 0, (void *) data, has_alpha, 0, + "111 221 2222 11", 0, 0, format, 0, 0, 0, 0, 0, x, y, (colorbytes + has_alpha) * 8, has_alpha * 8); + } else { + int i,j,k; + int jend, jdir; + stbiw__writef(s, "111 221 2222 11", 0,0,format+8, 0,0,0, 0,0,x,y, (colorbytes + has_alpha) * 8, has_alpha * 8); + if (stbi__flip_vertically_on_write) { + j = 0; + jend = y; + jdir = 1; + } else { + j = y-1; + jend = -1; + jdir = -1; + } + for (; j != jend; j += jdir) { + unsigned char *row = (unsigned char *) data + j * x * comp; + int len; + for (i = 0; i < x; i += len) { + unsigned char *begin = row + i * comp; + int diff = 1; + len = 1; + if (i < x - 1) { + ++len; + diff = memcmp(begin, row + (i + 1) * comp, comp); + if (diff) { + const unsigned char *prev = begin; + for (k = i + 2; k < x && len < 128; ++k) { + if (memcmp(prev, row + k * comp, comp)) { + prev += comp; + ++len; + } else { + --len; + break; + } + } + } else { + for (k = i + 2; k < x && len < 128; ++k) { + if (!memcmp(begin, row + k * comp, comp)) { + ++len; + } else { + break; + } + } + } + } + if (diff) { + unsigned char header = STBIW_UCHAR(len - 1); + s->func(s->context, &header, 1); + for (k = 0; k < len; ++k) { + stbiw__write_pixel(s, -1, comp, has_alpha, 0, begin + k * comp); + } + } else { + unsigned char header = STBIW_UCHAR(len - 129); + s->func(s->context, &header, 1); + stbiw__write_pixel(s, -1, comp, has_alpha, 0, begin); + } + } + } + } + return 1; +} +STBIWDEF int stbi_write_tga_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_tga_core(&s, x, y, comp, (void *) data); +} +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_tga(char const *filename, int x, int y, int comp, const void *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_tga_core(&s, x, y, comp, (void *) data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif +#define stbiw__max(a, b) ((a) > (b) ? (a) : (b)) +static void stbiw__linear_to_rgbe(unsigned char *rgbe, float *linear) +{ + int exponent; + float maxcomp = stbiw__max(linear[0], stbiw__max(linear[1], linear[2])); + if (maxcomp < 1e-32f) { + rgbe[0] = rgbe[1] = rgbe[2] = rgbe[3] = 0; + } else { + float normalize = (float) frexp(maxcomp, &exponent) * 256.0f/maxcomp; + rgbe[0] = (unsigned char)(linear[0] * normalize); + rgbe[1] = (unsigned char)(linear[1] * normalize); + rgbe[2] = (unsigned char)(linear[2] * normalize); + rgbe[3] = (unsigned char)(exponent + 128); + } +} +static void stbiw__write_run_data(stbi__write_context *s, int length, unsigned char databyte) +{ + unsigned char lengthbyte = STBIW_UCHAR(length+128); + STBIW_ASSERT(length+128 <= 255); + s->func(s->context, &lengthbyte, 1); + s->func(s->context, &databyte, 1); +} +static void stbiw__write_dump_data(stbi__write_context *s, int length, unsigned char *data) +{ + unsigned char lengthbyte = STBIW_UCHAR(length); + STBIW_ASSERT(length <= 128); + s->func(s->context, &lengthbyte, 1); + s->func(s->context, data, length); +} +static void stbiw__write_hdr_scanline(stbi__write_context *s, int width, int ncomp, unsigned char *scratch, float *scanline) +{ + unsigned char scanlineheader[4] = { 2, 2, 0, 0 }; + unsigned char rgbe[4]; + float linear[3]; + int x; + scanlineheader[2] = (width&0xff00)>>8; + scanlineheader[3] = (width&0x00ff); + + if (width < 8 || width >= 32768) { + for (x=0; x < width; x++) { + switch (ncomp) { + case 4: + case 3: linear[2] = scanline[x*ncomp + 2]; + linear[1] = scanline[x*ncomp + 1]; + linear[0] = scanline[x*ncomp + 0]; + break; + default: + linear[0] = linear[1] = linear[2] = scanline[x*ncomp + 0]; + break; + } + stbiw__linear_to_rgbe(rgbe, linear); + s->func(s->context, rgbe, 4); + } + } else { + int c,r; + + for (x=0; x < width; x++) { + switch(ncomp) { + case 4: + case 3: linear[2] = scanline[x*ncomp + 2]; + linear[1] = scanline[x*ncomp + 1]; + linear[0] = scanline[x*ncomp + 0]; + break; + default: + linear[0] = linear[1] = linear[2] = scanline[x*ncomp + 0]; + break; + } + stbiw__linear_to_rgbe(rgbe, linear); + scratch[x + width*0] = rgbe[0]; + scratch[x + width*1] = rgbe[1]; + scratch[x + width*2] = rgbe[2]; + scratch[x + width*3] = rgbe[3]; + } + s->func(s->context, scanlineheader, 4); + + for (c=0; c < 4; c++) { + unsigned char *comp = &scratch[width*c]; + x = 0; + while (x < width) { + + r = x; + while (r+2 < width) { + if (comp[r] == comp[r+1] && comp[r] == comp[r+2]) + break; + ++r; + } + if (r+2 >= width) + r = width; + + while (x < r) { + int len = r-x; + if (len > 128) len = 128; + stbiw__write_dump_data(s, len, &comp[x]); + x += len; + } + + if (r+2 < width) { + + while (r < width && comp[r] == comp[x]) + ++r; + + while (x < r) { + int len = r-x; + if (len > 127) len = 127; + stbiw__write_run_data(s, len, comp[x]); + x += len; + } + } + } + } + } +} +static int stbi_write_hdr_core(stbi__write_context *s, int x, int y, int comp, float *data) +{ + if (y <= 0 || x <= 0 || data == NULL) + return 0; + else { + + unsigned char *scratch = (unsigned char *) STBIW_MALLOC(x*4); + int i, len; + char buffer[128]; + char header[] = "#?RADIANCE\n# Written by stb_image_write.h\nFORMAT=32-bit_rle_rgbe\n"; + s->func(s->context, header, sizeof(header)-1); +#ifdef __STDC_WANT_SECURE_LIB__ + len = sprintf_s(buffer, sizeof(buffer), "EXPOSURE= 1.0000000000000\n\n-Y %d +X %d\n", y, x); +#else + len = sprintf(buffer, "EXPOSURE= 1.0000000000000\n\n-Y %d +X %d\n", y, x); +#endif + s->func(s->context, buffer, len); + for(i=0; i < y; i++) + stbiw__write_hdr_scanline(s, x, comp, scratch, data + comp*x*(stbi__flip_vertically_on_write ? y-1-i : i)); + STBIW_FREE(scratch); + return 1; + } +} +STBIWDEF int stbi_write_hdr_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const float *data) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_hdr_core(&s, x, y, comp, (float *) data); +} +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_hdr(char const *filename, int x, int y, int comp, const float *data) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_hdr_core(&s, x, y, comp, (float *) data); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif +#ifndef STBIW_ZLIB_COMPRESS +#define stbiw__sbraw(a) ((int *) (void *) (a) - 2) +#define stbiw__sbm(a) stbiw__sbraw(a)[0] +#define stbiw__sbn(a) stbiw__sbraw(a)[1] +#define stbiw__sbneedgrow(a,n) ((a)==0 || stbiw__sbn(a)+n >= stbiw__sbm(a)) +#define stbiw__sbmaybegrow(a,n) (stbiw__sbneedgrow(a,(n)) ? stbiw__sbgrow(a,n) : 0) +#define stbiw__sbgrow(a,n) stbiw__sbgrowf((void **) &(a), (n), sizeof(*(a))) +#define stbiw__sbpush(a, v) (stbiw__sbmaybegrow(a,1), (a)[stbiw__sbn(a)++] = (v)) +#define stbiw__sbcount(a) ((a) ? stbiw__sbn(a) : 0) +#define stbiw__sbfree(a) ((a) ? STBIW_FREE(stbiw__sbraw(a)),0 : 0) +static void *stbiw__sbgrowf(void **arr, int increment, int itemsize) +{ + int m = *arr ? 2*stbiw__sbm(*arr)+increment : increment+1; + void *p = STBIW_REALLOC_SIZED(*arr ? stbiw__sbraw(*arr) : 0, *arr ? (stbiw__sbm(*arr)*itemsize + sizeof(int)*2) : 0, itemsize * m + sizeof(int)*2); + STBIW_ASSERT(p); + if (p) { + if (!*arr) ((int *) p)[1] = 0; + *arr = (void *) ((int *) p + 2); + stbiw__sbm(*arr) = m; + } + return *arr; +} +static unsigned char *stbiw__zlib_flushf(unsigned char *data, unsigned int *bitbuffer, int *bitcount) +{ + while (*bitcount >= 8) { + stbiw__sbpush(data, STBIW_UCHAR(*bitbuffer)); + *bitbuffer >>= 8; + *bitcount -= 8; + } + return data; +} +static int stbiw__zlib_bitrev(int code, int codebits) +{ + int res=0; + while (codebits--) { + res = (res << 1) | (code & 1); + code >>= 1; + } + return res; +} +static unsigned int stbiw__zlib_countm(unsigned char *a, unsigned char *b, int limit) +{ + int i; + for (i=0; i < limit && i < 258; ++i) + if (a[i] != b[i]) break; + return i; +} +static unsigned int stbiw__zhash(unsigned char *data) +{ + stbiw_uint32 hash = data[0] + (data[1] << 8) + (data[2] << 16); + hash ^= hash << 3; + hash += hash >> 5; + hash ^= hash << 4; + hash += hash >> 17; + hash ^= hash << 25; + hash += hash >> 6; + return hash; +} +#define stbiw__zlib_flush() (out = stbiw__zlib_flushf(out, &bitbuf, &bitcount)) +#define stbiw__zlib_add(code,codebits) \ + (bitbuf |= (code) << bitcount, bitcount += (codebits), stbiw__zlib_flush()) +#define stbiw__zlib_huffa(b,c) stbiw__zlib_add(stbiw__zlib_bitrev(b,c),c) +#define stbiw__zlib_huff1(n) stbiw__zlib_huffa(0x30 + (n), 8) +#define stbiw__zlib_huff2(n) stbiw__zlib_huffa(0x190 + (n)-144, 9) +#define stbiw__zlib_huff3(n) stbiw__zlib_huffa(0 + (n)-256,7) +#define stbiw__zlib_huff4(n) stbiw__zlib_huffa(0xc0 + (n)-280,8) +#define stbiw__zlib_huff(n) ((n) <= 143 ? stbiw__zlib_huff1(n) : (n) <= 255 ? stbiw__zlib_huff2(n) : (n) <= 279 ? stbiw__zlib_huff3(n) : stbiw__zlib_huff4(n)) +#define stbiw__zlib_huffb(n) ((n) <= 143 ? stbiw__zlib_huff1(n) : stbiw__zlib_huff2(n)) +#define stbiw__ZHASH 16384 +#endif +STBIWDEF unsigned char * stbi_zlib_compress(unsigned char *data, int data_len, int *out_len, int quality) +{ +#ifdef STBIW_ZLIB_COMPRESS + + return STBIW_ZLIB_COMPRESS(data, data_len, out_len, quality); +#else + static unsigned short lengthc[] = { 3,4,5,6,7,8,9,10,11,13,15,17,19,23,27,31,35,43,51,59,67,83,99,115,131,163,195,227,258, 259 }; + static unsigned char lengtheb[]= { 0,0,0,0,0,0,0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0 }; + static unsigned short distc[] = { 1,2,3,4,5,7,9,13,17,25,33,49,65,97,129,193,257,385,513,769,1025,1537,2049,3073,4097,6145,8193,12289,16385,24577, 32768 }; + static unsigned char disteb[] = { 0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13 }; + unsigned int bitbuf=0; + int i,j, bitcount=0; + unsigned char *out = NULL; + unsigned char ***hash_table = (unsigned char***) STBIW_MALLOC(stbiw__ZHASH * sizeof(unsigned char**)); + if (hash_table == NULL) + return NULL; + if (quality < 5) quality = 5; + stbiw__sbpush(out, 0x78); + stbiw__sbpush(out, 0x5e); + stbiw__zlib_add(1,1); + stbiw__zlib_add(1,2); + for (i=0; i < stbiw__ZHASH; ++i) + hash_table[i] = NULL; + i=0; + while (i < data_len-3) { + + int h = stbiw__zhash(data+i)&(stbiw__ZHASH-1), best=3; + unsigned char *bestloc = 0; + unsigned char **hlist = hash_table[h]; + int n = stbiw__sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32768) { + int d = stbiw__zlib_countm(hlist[j], data+i, data_len-i); + if (d >= best) { best=d; bestloc=hlist[j]; } + } + } + + if (hash_table[h] && stbiw__sbn(hash_table[h]) == 2*quality) { + STBIW_MEMMOVE(hash_table[h], hash_table[h]+quality, sizeof(hash_table[h][0])*quality); + stbiw__sbn(hash_table[h]) = quality; + } + stbiw__sbpush(hash_table[h],data+i); + if (bestloc) { + + h = stbiw__zhash(data+i+1)&(stbiw__ZHASH-1); + hlist = hash_table[h]; + n = stbiw__sbcount(hlist); + for (j=0; j < n; ++j) { + if (hlist[j]-data > i-32767) { + int e = stbiw__zlib_countm(hlist[j], data+i+1, data_len-i-1); + if (e > best) { + bestloc = NULL; + break; + } + } + } + } + if (bestloc) { + int d = (int) (data+i - bestloc); + STBIW_ASSERT(d <= 32767 && best <= 258); + for (j=0; best > lengthc[j+1]-1; ++j); + stbiw__zlib_huff(j+257); + if (lengtheb[j]) stbiw__zlib_add(best - lengthc[j], lengtheb[j]); + for (j=0; d > distc[j+1]-1; ++j); + stbiw__zlib_add(stbiw__zlib_bitrev(j,5),5); + if (disteb[j]) stbiw__zlib_add(d - distc[j], disteb[j]); + i += best; + } else { + stbiw__zlib_huffb(data[i]); + ++i; + } + } + + for (;i < data_len; ++i) + stbiw__zlib_huffb(data[i]); + stbiw__zlib_huff(256); + + while (bitcount) + stbiw__zlib_add(0,1); + for (i=0; i < stbiw__ZHASH; ++i) + (void) stbiw__sbfree(hash_table[i]); + STBIW_FREE(hash_table); + { + + unsigned int s1=1, s2=0; + int blocklen = (int) (data_len % 5552); + j=0; + while (j < data_len) { + for (i=0; i < blocklen; ++i) { s1 += data[j+i]; s2 += s1; } + s1 %= 65521; s2 %= 65521; + j += blocklen; + blocklen = 5552; + } + stbiw__sbpush(out, STBIW_UCHAR(s2 >> 8)); + stbiw__sbpush(out, STBIW_UCHAR(s2)); + stbiw__sbpush(out, STBIW_UCHAR(s1 >> 8)); + stbiw__sbpush(out, STBIW_UCHAR(s1)); + } + *out_len = stbiw__sbn(out); + + STBIW_MEMMOVE(stbiw__sbraw(out), out, *out_len); + return (unsigned char *) stbiw__sbraw(out); +#endif +} +static unsigned int stbiw__crc32(unsigned char *buffer, int len) +{ +#ifdef STBIW_CRC32 + return STBIW_CRC32(buffer, len); +#else + static unsigned int crc_table[256] = + { + 0x00000000, 0x77073096, 0xEE0E612C, 0x990951BA, 0x076DC419, 0x706AF48F, 0xE963A535, 0x9E6495A3, + 0x0eDB8832, 0x79DCB8A4, 0xE0D5E91E, 0x97D2D988, 0x09B64C2B, 0x7EB17CBD, 0xE7B82D07, 0x90BF1D91, + 0x1DB71064, 0x6AB020F2, 0xF3B97148, 0x84BE41DE, 0x1ADAD47D, 0x6DDDE4EB, 0xF4D4B551, 0x83D385C7, + 0x136C9856, 0x646BA8C0, 0xFD62F97A, 0x8A65C9EC, 0x14015C4F, 0x63066CD9, 0xFA0F3D63, 0x8D080DF5, + 0x3B6E20C8, 0x4C69105E, 0xD56041E4, 0xA2677172, 0x3C03E4D1, 0x4B04D447, 0xD20D85FD, 0xA50AB56B, + 0x35B5A8FA, 0x42B2986C, 0xDBBBC9D6, 0xACBCF940, 0x32D86CE3, 0x45DF5C75, 0xDCD60DCF, 0xABD13D59, + 0x26D930AC, 0x51DE003A, 0xC8D75180, 0xBFD06116, 0x21B4F4B5, 0x56B3C423, 0xCFBA9599, 0xB8BDA50F, + 0x2802B89E, 0x5F058808, 0xC60CD9B2, 0xB10BE924, 0x2F6F7C87, 0x58684C11, 0xC1611DAB, 0xB6662D3D, + 0x76DC4190, 0x01DB7106, 0x98D220BC, 0xEFD5102A, 0x71B18589, 0x06B6B51F, 0x9FBFE4A5, 0xE8B8D433, + 0x7807C9A2, 0x0F00F934, 0x9609A88E, 0xE10E9818, 0x7F6A0DBB, 0x086D3D2D, 0x91646C97, 0xE6635C01, + 0x6B6B51F4, 0x1C6C6162, 0x856530D8, 0xF262004E, 0x6C0695ED, 0x1B01A57B, 0x8208F4C1, 0xF50FC457, + 0x65B0D9C6, 0x12B7E950, 0x8BBEB8EA, 0xFCB9887C, 0x62DD1DDF, 0x15DA2D49, 0x8CD37CF3, 0xFBD44C65, + 0x4DB26158, 0x3AB551CE, 0xA3BC0074, 0xD4BB30E2, 0x4ADFA541, 0x3DD895D7, 0xA4D1C46D, 0xD3D6F4FB, + 0x4369E96A, 0x346ED9FC, 0xAD678846, 0xDA60B8D0, 0x44042D73, 0x33031DE5, 0xAA0A4C5F, 0xDD0D7CC9, + 0x5005713C, 0x270241AA, 0xBE0B1010, 0xC90C2086, 0x5768B525, 0x206F85B3, 0xB966D409, 0xCE61E49F, + 0x5EDEF90E, 0x29D9C998, 0xB0D09822, 0xC7D7A8B4, 0x59B33D17, 0x2EB40D81, 0xB7BD5C3B, 0xC0BA6CAD, + 0xEDB88320, 0x9ABFB3B6, 0x03B6E20C, 0x74B1D29A, 0xEAD54739, 0x9DD277AF, 0x04DB2615, 0x73DC1683, + 0xE3630B12, 0x94643B84, 0x0D6D6A3E, 0x7A6A5AA8, 0xE40ECF0B, 0x9309FF9D, 0x0A00AE27, 0x7D079EB1, + 0xF00F9344, 0x8708A3D2, 0x1E01F268, 0x6906C2FE, 0xF762575D, 0x806567CB, 0x196C3671, 0x6E6B06E7, + 0xFED41B76, 0x89D32BE0, 0x10DA7A5A, 0x67DD4ACC, 0xF9B9DF6F, 0x8EBEEFF9, 0x17B7BE43, 0x60B08ED5, + 0xD6D6A3E8, 0xA1D1937E, 0x38D8C2C4, 0x4FDFF252, 0xD1BB67F1, 0xA6BC5767, 0x3FB506DD, 0x48B2364B, + 0xD80D2BDA, 0xAF0A1B4C, 0x36034AF6, 0x41047A60, 0xDF60EFC3, 0xA867DF55, 0x316E8EEF, 0x4669BE79, + 0xCB61B38C, 0xBC66831A, 0x256FD2A0, 0x5268E236, 0xCC0C7795, 0xBB0B4703, 0x220216B9, 0x5505262F, + 0xC5BA3BBE, 0xB2BD0B28, 0x2BB45A92, 0x5CB36A04, 0xC2D7FFA7, 0xB5D0CF31, 0x2CD99E8B, 0x5BDEAE1D, + 0x9B64C2B0, 0xEC63F226, 0x756AA39C, 0x026D930A, 0x9C0906A9, 0xEB0E363F, 0x72076785, 0x05005713, + 0x95BF4A82, 0xE2B87A14, 0x7BB12BAE, 0x0CB61B38, 0x92D28E9B, 0xE5D5BE0D, 0x7CDCEFB7, 0x0BDBDF21, + 0x86D3D2D4, 0xF1D4E242, 0x68DDB3F8, 0x1FDA836E, 0x81BE16CD, 0xF6B9265B, 0x6FB077E1, 0x18B74777, + 0x88085AE6, 0xFF0F6A70, 0x66063BCA, 0x11010B5C, 0x8F659EFF, 0xF862AE69, 0x616BFFD3, 0x166CCF45, + 0xA00AE278, 0xD70DD2EE, 0x4E048354, 0x3903B3C2, 0xA7672661, 0xD06016F7, 0x4969474D, 0x3E6E77DB, + 0xAED16A4A, 0xD9D65ADC, 0x40DF0B66, 0x37D83BF0, 0xA9BCAE53, 0xDEBB9EC5, 0x47B2CF7F, 0x30B5FFE9, + 0xBDBDF21C, 0xCABAC28A, 0x53B39330, 0x24B4A3A6, 0xBAD03605, 0xCDD70693, 0x54DE5729, 0x23D967BF, + 0xB3667A2E, 0xC4614AB8, 0x5D681B02, 0x2A6F2B94, 0xB40BBE37, 0xC30C8EA1, 0x5A05DF1B, 0x2D02EF8D + }; + unsigned int crc = ~0u; + int i; + for (i=0; i < len; ++i) + crc = (crc >> 8) ^ crc_table[buffer[i] ^ (crc & 0xff)]; + return ~crc; +#endif +} +#define stbiw__wpng4(o,a,b,c,d) ((o)[0]=STBIW_UCHAR(a),(o)[1]=STBIW_UCHAR(b),(o)[2]=STBIW_UCHAR(c),(o)[3]=STBIW_UCHAR(d),(o)+=4) +#define stbiw__wp32(data,v) stbiw__wpng4(data, (v)>>24,(v)>>16,(v)>>8,(v)); +#define stbiw__wptag(data,s) stbiw__wpng4(data, s[0],s[1],s[2],s[3]) +static void stbiw__wpcrc(unsigned char **data, int len) +{ + unsigned int crc = stbiw__crc32(*data - len - 4, len+4); + stbiw__wp32(*data, crc); +} +static unsigned char stbiw__paeth(int a, int b, int c) +{ + int p = a + b - c, pa = abs(p-a), pb = abs(p-b), pc = abs(p-c); + if (pa <= pb && pa <= pc) return STBIW_UCHAR(a); + if (pb <= pc) return STBIW_UCHAR(b); + return STBIW_UCHAR(c); +} +static void stbiw__encode_png_line(unsigned char *pixels, int stride_bytes, int width, int height, int y, int n, int filter_type, signed char *line_buffer) +{ + static int mapping[] = { 0,1,2,3,4 }; + static int firstmap[] = { 0,1,0,5,6 }; + int *mymap = (y != 0) ? mapping : firstmap; + int i; + int type = mymap[filter_type]; + unsigned char *z = pixels + stride_bytes * (stbi__flip_vertically_on_write ? height-1-y : y); + int signed_stride = stbi__flip_vertically_on_write ? -stride_bytes : stride_bytes; + if (type==0) { + memcpy(line_buffer, z, width*n); + return; + } + + for (i = 0; i < n; ++i) { + switch (type) { + case 1: line_buffer[i] = z[i]; break; + case 2: line_buffer[i] = z[i] - z[i-signed_stride]; break; + case 3: line_buffer[i] = z[i] - (z[i-signed_stride]>>1); break; + case 4: line_buffer[i] = (signed char) (z[i] - stbiw__paeth(0,z[i-signed_stride],0)); break; + case 5: line_buffer[i] = z[i]; break; + case 6: line_buffer[i] = z[i]; break; + } + } + switch (type) { + case 1: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - z[i-n]; break; + case 2: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - z[i-signed_stride]; break; + case 3: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - ((z[i-n] + z[i-signed_stride])>>1); break; + case 4: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - stbiw__paeth(z[i-n], z[i-signed_stride], z[i-signed_stride-n]); break; + case 5: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - (z[i-n]>>1); break; + case 6: for (i=n; i < width*n; ++i) line_buffer[i] = z[i] - stbiw__paeth(z[i-n], 0,0); break; + } +} +STBIWDEF unsigned char *stbi_write_png_to_mem(const unsigned char *pixels, int stride_bytes, int x, int y, int n, int *out_len) +{ + int force_filter = stbi_write_force_png_filter; + int ctype[5] = { -1, 0, 4, 2, 6 }; + unsigned char sig[8] = { 137,80,78,71,13,10,26,10 }; + unsigned char *out,*o, *filt, *zlib; + signed char *line_buffer; + int j,zlen; + if (stride_bytes == 0) + stride_bytes = x * n; + if (force_filter >= 5) { + force_filter = -1; + } + filt = (unsigned char *) STBIW_MALLOC((x*n+1) * y); if (!filt) return 0; + line_buffer = (signed char *) STBIW_MALLOC(x * n); if (!line_buffer) { STBIW_FREE(filt); return 0; } + for (j=0; j < y; ++j) { + int filter_type; + if (force_filter > -1) { + filter_type = force_filter; + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, force_filter, line_buffer); + } else { + int best_filter = 0, best_filter_val = 0x7fffffff, est, i; + for (filter_type = 0; filter_type < 5; filter_type++) { + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, filter_type, line_buffer); + + est = 0; + for (i = 0; i < x*n; ++i) { + est += abs((signed char) line_buffer[i]); + } + if (est < best_filter_val) { + best_filter_val = est; + best_filter = filter_type; + } + } + if (filter_type != best_filter) { + stbiw__encode_png_line((unsigned char*)(pixels), stride_bytes, x, y, j, n, best_filter, line_buffer); + filter_type = best_filter; + } + } + + filt[j*(x*n+1)] = (unsigned char) filter_type; + STBIW_MEMMOVE(filt+j*(x*n+1)+1, line_buffer, x*n); + } + STBIW_FREE(line_buffer); + zlib = stbi_zlib_compress(filt, y*( x*n+1), &zlen, stbi_write_png_compression_level); + STBIW_FREE(filt); + if (!zlib) return 0; + + out = (unsigned char *) STBIW_MALLOC(8 + 12+13 + 12+zlen + 12); + if (!out) return 0; + *out_len = 8 + 12+13 + 12+zlen + 12; + o=out; + STBIW_MEMMOVE(o,sig,8); o+= 8; + stbiw__wp32(o, 13); + stbiw__wptag(o, "IHDR"); + stbiw__wp32(o, x); + stbiw__wp32(o, y); + *o++ = 8; + *o++ = STBIW_UCHAR(ctype[n]); + *o++ = 0; + *o++ = 0; + *o++ = 0; + stbiw__wpcrc(&o,13); + stbiw__wp32(o, zlen); + stbiw__wptag(o, "IDAT"); + STBIW_MEMMOVE(o, zlib, zlen); + o += zlen; + STBIW_FREE(zlib); + stbiw__wpcrc(&o, zlen); + stbiw__wp32(o,0); + stbiw__wptag(o, "IEND"); + stbiw__wpcrc(&o,0); + STBIW_ASSERT(o == out + *out_len); + return out; +} +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_png(char const *filename, int x, int y, int comp, const void *data, int stride_bytes) +{ + FILE *f; + int len; + unsigned char *png = stbi_write_png_to_mem((const unsigned char *) data, stride_bytes, x, y, comp, &len); + if (png == NULL) return 0; + f = stbiw__fopen(filename, "wb"); + if (!f) { STBIW_FREE(png); return 0; } + fwrite(png, 1, len, f); + fclose(f); + STBIW_FREE(png); + return 1; +} +#endif +STBIWDEF int stbi_write_png_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int stride_bytes) +{ + int len; + unsigned char *png = stbi_write_png_to_mem((const unsigned char *) data, stride_bytes, x, y, comp, &len); + if (png == NULL) return 0; + func(context, png, len); + STBIW_FREE(png); + return 1; +} +static const unsigned char stbiw__jpg_ZigZag[] = { 0,1,5,6,14,15,27,28,2,4,7,13,16,26,29,42,3,8,12,17,25,30,41,43,9,11,18, + 24,31,40,44,53,10,19,23,32,39,45,52,54,20,22,33,38,46,51,55,60,21,34,37,47,50,56,59,61,35,36,48,49,57,58,62,63 }; +static void stbiw__jpg_writeBits(stbi__write_context *s, int *bitBufP, int *bitCntP, const unsigned short *bs) { + int bitBuf = *bitBufP, bitCnt = *bitCntP; + bitCnt += bs[1]; + bitBuf |= bs[0] << (24 - bitCnt); + while(bitCnt >= 8) { + unsigned char c = (bitBuf >> 16) & 255; + stbiw__putc(s, c); + if(c == 255) { + stbiw__putc(s, 0); + } + bitBuf <<= 8; + bitCnt -= 8; + } + *bitBufP = bitBuf; + *bitCntP = bitCnt; +} +static void stbiw__jpg_DCT(float *d0p, float *d1p, float *d2p, float *d3p, float *d4p, float *d5p, float *d6p, float *d7p) { + float d0 = *d0p, d1 = *d1p, d2 = *d2p, d3 = *d3p, d4 = *d4p, d5 = *d5p, d6 = *d6p, d7 = *d7p; + float z1, z2, z3, z4, z5, z11, z13; + float tmp0 = d0 + d7; + float tmp7 = d0 - d7; + float tmp1 = d1 + d6; + float tmp6 = d1 - d6; + float tmp2 = d2 + d5; + float tmp5 = d2 - d5; + float tmp3 = d3 + d4; + float tmp4 = d3 - d4; + + float tmp10 = tmp0 + tmp3; + float tmp13 = tmp0 - tmp3; + float tmp11 = tmp1 + tmp2; + float tmp12 = tmp1 - tmp2; + d0 = tmp10 + tmp11; + d4 = tmp10 - tmp11; + z1 = (tmp12 + tmp13) * 0.707106781f; + d2 = tmp13 + z1; + d6 = tmp13 - z1; + + tmp10 = tmp4 + tmp5; + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + z5 = (tmp10 - tmp12) * 0.382683433f; + z2 = tmp10 * 0.541196100f + z5; + z4 = tmp12 * 1.306562965f + z5; + z3 = tmp11 * 0.707106781f; + z11 = tmp7 + z3; + z13 = tmp7 - z3; + *d5p = z13 + z2; + *d3p = z13 - z2; + *d1p = z11 + z4; + *d7p = z11 - z4; + *d0p = d0; *d2p = d2; *d4p = d4; *d6p = d6; +} +static void stbiw__jpg_calcBits(int val, unsigned short bits[2]) { + int tmp1 = val < 0 ? -val : val; + val = val < 0 ? val-1 : val; + bits[1] = 1; + while(tmp1 >>= 1) { + ++bits[1]; + } + bits[0] = val & ((1<0)&&(DU[end0pos]==0); --end0pos) { + } + + if(end0pos == 0) { + stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB); + return DU[0]; + } + for(i = 1; i <= end0pos; ++i) { + int startpos = i; + int nrzeroes; + unsigned short bits[2]; + for (; DU[i]==0 && i<=end0pos; ++i) { + } + nrzeroes = i-startpos; + if ( nrzeroes >= 16 ) { + int lng = nrzeroes>>4; + int nrmarker; + for (nrmarker=1; nrmarker <= lng; ++nrmarker) + stbiw__jpg_writeBits(s, bitBuf, bitCnt, M16zeroes); + nrzeroes &= 15; + } + stbiw__jpg_calcBits(DU[i], bits); + stbiw__jpg_writeBits(s, bitBuf, bitCnt, HTAC[(nrzeroes<<4)+bits[1]]); + stbiw__jpg_writeBits(s, bitBuf, bitCnt, bits); + } + if(end0pos != 63) { + stbiw__jpg_writeBits(s, bitBuf, bitCnt, EOB); + } + return DU[0]; +} +static int stbi_write_jpg_core(stbi__write_context *s, int width, int height, int comp, const void* data, int quality) { + + static const unsigned char std_dc_luminance_nrcodes[] = {0,0,1,5,1,1,1,1,1,1,0,0,0,0,0,0,0}; + static const unsigned char std_dc_luminance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11}; + static const unsigned char std_ac_luminance_nrcodes[] = {0,0,2,1,3,3,2,4,3,5,5,4,4,0,0,1,0x7d}; + static const unsigned char std_ac_luminance_values[] = { + 0x01,0x02,0x03,0x00,0x04,0x11,0x05,0x12,0x21,0x31,0x41,0x06,0x13,0x51,0x61,0x07,0x22,0x71,0x14,0x32,0x81,0x91,0xa1,0x08, + 0x23,0x42,0xb1,0xc1,0x15,0x52,0xd1,0xf0,0x24,0x33,0x62,0x72,0x82,0x09,0x0a,0x16,0x17,0x18,0x19,0x1a,0x25,0x26,0x27,0x28, + 0x29,0x2a,0x34,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58,0x59, + 0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x83,0x84,0x85,0x86,0x87,0x88,0x89, + 0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4,0xb5,0xb6, + 0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda,0xe1,0xe2, + 0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf1,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa + }; + static const unsigned char std_dc_chrominance_nrcodes[] = {0,0,3,1,1,1,1,1,1,1,1,1,0,0,0,0,0}; + static const unsigned char std_dc_chrominance_values[] = {0,1,2,3,4,5,6,7,8,9,10,11}; + static const unsigned char std_ac_chrominance_nrcodes[] = {0,0,2,1,2,4,4,3,4,7,5,4,4,0,1,2,0x77}; + static const unsigned char std_ac_chrominance_values[] = { + 0x00,0x01,0x02,0x03,0x11,0x04,0x05,0x21,0x31,0x06,0x12,0x41,0x51,0x07,0x61,0x71,0x13,0x22,0x32,0x81,0x08,0x14,0x42,0x91, + 0xa1,0xb1,0xc1,0x09,0x23,0x33,0x52,0xf0,0x15,0x62,0x72,0xd1,0x0a,0x16,0x24,0x34,0xe1,0x25,0xf1,0x17,0x18,0x19,0x1a,0x26, + 0x27,0x28,0x29,0x2a,0x35,0x36,0x37,0x38,0x39,0x3a,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x53,0x54,0x55,0x56,0x57,0x58, + 0x59,0x5a,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,0x82,0x83,0x84,0x85,0x86,0x87, + 0x88,0x89,0x8a,0x92,0x93,0x94,0x95,0x96,0x97,0x98,0x99,0x9a,0xa2,0xa3,0xa4,0xa5,0xa6,0xa7,0xa8,0xa9,0xaa,0xb2,0xb3,0xb4, + 0xb5,0xb6,0xb7,0xb8,0xb9,0xba,0xc2,0xc3,0xc4,0xc5,0xc6,0xc7,0xc8,0xc9,0xca,0xd2,0xd3,0xd4,0xd5,0xd6,0xd7,0xd8,0xd9,0xda, + 0xe2,0xe3,0xe4,0xe5,0xe6,0xe7,0xe8,0xe9,0xea,0xf2,0xf3,0xf4,0xf5,0xf6,0xf7,0xf8,0xf9,0xfa + }; + + static const unsigned short YDC_HT[256][2] = { {0,2},{2,3},{3,3},{4,3},{5,3},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9}}; + static const unsigned short UVDC_HT[256][2] = { {0,2},{1,2},{2,2},{6,3},{14,4},{30,5},{62,6},{126,7},{254,8},{510,9},{1022,10},{2046,11}}; + static const unsigned short YAC_HT[256][2] = { + {10,4},{0,2},{1,2},{4,3},{11,4},{26,5},{120,7},{248,8},{1014,10},{65410,16},{65411,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {12,4},{27,5},{121,7},{502,9},{2038,11},{65412,16},{65413,16},{65414,16},{65415,16},{65416,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {28,5},{249,8},{1015,10},{4084,12},{65417,16},{65418,16},{65419,16},{65420,16},{65421,16},{65422,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {58,6},{503,9},{4085,12},{65423,16},{65424,16},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {59,6},{1016,10},{65430,16},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {122,7},{2039,11},{65438,16},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {123,7},{4086,12},{65446,16},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {250,8},{4087,12},{65454,16},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {504,9},{32704,15},{65462,16},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {505,9},{65470,16},{65471,16},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {506,9},{65479,16},{65480,16},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1017,10},{65488,16},{65489,16},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1018,10},{65497,16},{65498,16},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2040,11},{65506,16},{65507,16},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {65515,16},{65516,16},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2041,11},{65525,16},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0} + }; + static const unsigned short UVAC_HT[256][2] = { + {0,2},{1,2},{4,3},{10,4},{24,5},{25,5},{56,6},{120,7},{500,9},{1014,10},{4084,12},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {11,4},{57,6},{246,8},{501,9},{2038,11},{4085,12},{65416,16},{65417,16},{65418,16},{65419,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {26,5},{247,8},{1015,10},{4086,12},{32706,15},{65420,16},{65421,16},{65422,16},{65423,16},{65424,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {27,5},{248,8},{1016,10},{4087,12},{65425,16},{65426,16},{65427,16},{65428,16},{65429,16},{65430,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {58,6},{502,9},{65431,16},{65432,16},{65433,16},{65434,16},{65435,16},{65436,16},{65437,16},{65438,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {59,6},{1017,10},{65439,16},{65440,16},{65441,16},{65442,16},{65443,16},{65444,16},{65445,16},{65446,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {121,7},{2039,11},{65447,16},{65448,16},{65449,16},{65450,16},{65451,16},{65452,16},{65453,16},{65454,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {122,7},{2040,11},{65455,16},{65456,16},{65457,16},{65458,16},{65459,16},{65460,16},{65461,16},{65462,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {249,8},{65463,16},{65464,16},{65465,16},{65466,16},{65467,16},{65468,16},{65469,16},{65470,16},{65471,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {503,9},{65472,16},{65473,16},{65474,16},{65475,16},{65476,16},{65477,16},{65478,16},{65479,16},{65480,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {504,9},{65481,16},{65482,16},{65483,16},{65484,16},{65485,16},{65486,16},{65487,16},{65488,16},{65489,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {505,9},{65490,16},{65491,16},{65492,16},{65493,16},{65494,16},{65495,16},{65496,16},{65497,16},{65498,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {506,9},{65499,16},{65500,16},{65501,16},{65502,16},{65503,16},{65504,16},{65505,16},{65506,16},{65507,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {2041,11},{65508,16},{65509,16},{65510,16},{65511,16},{65512,16},{65513,16},{65514,16},{65515,16},{65516,16},{0,0},{0,0},{0,0},{0,0},{0,0},{0,0}, + {16352,14},{65517,16},{65518,16},{65519,16},{65520,16},{65521,16},{65522,16},{65523,16},{65524,16},{65525,16},{0,0},{0,0},{0,0},{0,0},{0,0}, + {1018,10},{32707,15},{65526,16},{65527,16},{65528,16},{65529,16},{65530,16},{65531,16},{65532,16},{65533,16},{65534,16},{0,0},{0,0},{0,0},{0,0},{0,0} + }; + static const int YQT[] = {16,11,10,16,24,40,51,61,12,12,14,19,26,58,60,55,14,13,16,24,40,57,69,56,14,17,22,29,51,87,80,62,18,22, + 37,56,68,109,103,77,24,35,55,64,81,104,113,92,49,64,78,87,103,121,120,101,72,92,95,98,112,100,103,99}; + static const int UVQT[] = {17,18,24,47,99,99,99,99,18,21,26,66,99,99,99,99,24,26,56,99,99,99,99,99,47,66,99,99,99,99,99,99, + 99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99,99}; + static const float aasf[] = { 1.0f * 2.828427125f, 1.387039845f * 2.828427125f, 1.306562965f * 2.828427125f, 1.175875602f * 2.828427125f, + 1.0f * 2.828427125f, 0.785694958f * 2.828427125f, 0.541196100f * 2.828427125f, 0.275899379f * 2.828427125f }; + int row, col, i, k, subsample; + float fdtbl_Y[64], fdtbl_UV[64]; + unsigned char YTable[64], UVTable[64]; + if(!data || !width || !height || comp > 4 || comp < 1) { + return 0; + } + quality = quality ? quality : 90; + subsample = quality <= 90 ? 1 : 0; + quality = quality < 1 ? 1 : quality > 100 ? 100 : quality; + quality = quality < 50 ? 5000 / quality : 200 - quality * 2; + for(i = 0; i < 64; ++i) { + int uvti, yti = (YQT[i]*quality+50)/100; + YTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (yti < 1 ? 1 : yti > 255 ? 255 : yti); + uvti = (UVQT[i]*quality+50)/100; + UVTable[stbiw__jpg_ZigZag[i]] = (unsigned char) (uvti < 1 ? 1 : uvti > 255 ? 255 : uvti); + } + for(row = 0, k = 0; row < 8; ++row) { + for(col = 0; col < 8; ++col, ++k) { + fdtbl_Y[k] = 1 / (YTable [stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]); + fdtbl_UV[k] = 1 / (UVTable[stbiw__jpg_ZigZag[k]] * aasf[row] * aasf[col]); + } + } + + { + static const unsigned char head0[] = { 0xFF,0xD8,0xFF,0xE0,0,0x10,'J','F','I','F',0,1,1,0,0,1,0,1,0,0,0xFF,0xDB,0,0x84,0 }; + static const unsigned char head2[] = { 0xFF,0xDA,0,0xC,3,1,0,2,0x11,3,0x11,0,0x3F,0 }; + const unsigned char head1[] = { 0xFF,0xC0,0,0x11,8,(unsigned char)(height>>8),STBIW_UCHAR(height),(unsigned char)(width>>8),STBIW_UCHAR(width), + 3,1,(unsigned char)(subsample?0x22:0x11),0,2,0x11,1,3,0x11,1,0xFF,0xC4,0x01,0xA2,0 }; + s->func(s->context, (void*)head0, sizeof(head0)); + s->func(s->context, (void*)YTable, sizeof(YTable)); + stbiw__putc(s, 1); + s->func(s->context, UVTable, sizeof(UVTable)); + s->func(s->context, (void*)head1, sizeof(head1)); + s->func(s->context, (void*)(std_dc_luminance_nrcodes+1), sizeof(std_dc_luminance_nrcodes)-1); + s->func(s->context, (void*)std_dc_luminance_values, sizeof(std_dc_luminance_values)); + stbiw__putc(s, 0x10); + s->func(s->context, (void*)(std_ac_luminance_nrcodes+1), sizeof(std_ac_luminance_nrcodes)-1); + s->func(s->context, (void*)std_ac_luminance_values, sizeof(std_ac_luminance_values)); + stbiw__putc(s, 1); + s->func(s->context, (void*)(std_dc_chrominance_nrcodes+1), sizeof(std_dc_chrominance_nrcodes)-1); + s->func(s->context, (void*)std_dc_chrominance_values, sizeof(std_dc_chrominance_values)); + stbiw__putc(s, 0x11); + s->func(s->context, (void*)(std_ac_chrominance_nrcodes+1), sizeof(std_ac_chrominance_nrcodes)-1); + s->func(s->context, (void*)std_ac_chrominance_values, sizeof(std_ac_chrominance_values)); + s->func(s->context, (void*)head2, sizeof(head2)); + } + + { + static const unsigned short fillBits[] = {0x7F, 7}; + int DCY=0, DCU=0, DCV=0; + int bitBuf=0, bitCnt=0; + + int ofsG = comp > 2 ? 1 : 0, ofsB = comp > 2 ? 2 : 0; + const unsigned char *dataR = (const unsigned char *)data; + const unsigned char *dataG = dataR + ofsG; + const unsigned char *dataB = dataR + ofsB; + int x, y, pos; + if(subsample) { + for(y = 0; y < height; y += 16) { + for(x = 0; x < width; x += 16) { + float Y[256], U[256], V[256]; + for(row = y, pos = 0; row < y+16; ++row) { + + int clamped_row = (row < height) ? row : height - 1; + int base_p = (stbi__flip_vertically_on_write ? (height-1-clamped_row) : clamped_row)*width*comp; + for(col = x; col < x+16; ++col, ++pos) { + + int p = base_p + ((col < width) ? col : (width-1))*comp; + float r = dataR[p], g = dataG[p], b = dataB[p]; + Y[pos]= +0.29900f*r + 0.58700f*g + 0.11400f*b - 128; + U[pos]= -0.16874f*r - 0.33126f*g + 0.50000f*b; + V[pos]= +0.50000f*r - 0.41869f*g - 0.08131f*b; + } + } + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, Y+0, 16, fdtbl_Y, DCY, YDC_HT, YAC_HT); + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, Y+8, 16, fdtbl_Y, DCY, YDC_HT, YAC_HT); + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, Y+128, 16, fdtbl_Y, DCY, YDC_HT, YAC_HT); + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, Y+136, 16, fdtbl_Y, DCY, YDC_HT, YAC_HT); + + { + float subU[64], subV[64]; + int yy, xx; + for(yy = 0, pos = 0; yy < 8; ++yy) { + for(xx = 0; xx < 8; ++xx, ++pos) { + int j = yy*32+xx*2; + subU[pos] = (U[j+0] + U[j+1] + U[j+16] + U[j+17]) * 0.25f; + subV[pos] = (V[j+0] + V[j+1] + V[j+16] + V[j+17]) * 0.25f; + } + } + DCU = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, subU, 8, fdtbl_UV, DCU, UVDC_HT, UVAC_HT); + DCV = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, subV, 8, fdtbl_UV, DCV, UVDC_HT, UVAC_HT); + } + } + } + } else { + for(y = 0; y < height; y += 8) { + for(x = 0; x < width; x += 8) { + float Y[64], U[64], V[64]; + for(row = y, pos = 0; row < y+8; ++row) { + + int clamped_row = (row < height) ? row : height - 1; + int base_p = (stbi__flip_vertically_on_write ? (height-1-clamped_row) : clamped_row)*width*comp; + for(col = x; col < x+8; ++col, ++pos) { + + int p = base_p + ((col < width) ? col : (width-1))*comp; + float r = dataR[p], g = dataG[p], b = dataB[p]; + Y[pos]= +0.29900f*r + 0.58700f*g + 0.11400f*b - 128; + U[pos]= -0.16874f*r - 0.33126f*g + 0.50000f*b; + V[pos]= +0.50000f*r - 0.41869f*g - 0.08131f*b; + } + } + DCY = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, Y, 8, fdtbl_Y, DCY, YDC_HT, YAC_HT); + DCU = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, U, 8, fdtbl_UV, DCU, UVDC_HT, UVAC_HT); + DCV = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, V, 8, fdtbl_UV, DCV, UVDC_HT, UVAC_HT); + } + } + } + + stbiw__jpg_writeBits(s, &bitBuf, &bitCnt, fillBits); + } + + stbiw__putc(s, 0xFF); + stbiw__putc(s, 0xD9); + return 1; +} +STBIWDEF int stbi_write_jpg_to_func(stbi_write_func *func, void *context, int x, int y, int comp, const void *data, int quality) +{ + stbi__write_context s; + stbi__start_write_callbacks(&s, func, context); + return stbi_write_jpg_core(&s, x, y, comp, (void *) data, quality); +} +#ifndef STBI_WRITE_NO_STDIO +STBIWDEF int stbi_write_jpg(char const *filename, int x, int y, int comp, const void *data, int quality) +{ + stbi__write_context s; + if (stbi__start_write_file(&s,filename)) { + int r = stbi_write_jpg_core(&s, x, y, comp, data, quality); + stbi__end_write_file(&s); + return r; + } else + return 0; +} +#endif +#endif diff --git a/src/catcoon.c b/src/catcoon.c new file mode 100644 index 0000000..6a5257d --- /dev/null +++ b/src/catcoon.c @@ -0,0 +1,15 @@ +#include "catcoon.h" +#include "util_log.h" + +void cc_print_info(void) +{ + int auto_tsrmgr_flag = 0; +#ifdef AUTO_TSRMGR + auto_tsrmgr_flag = 1; +#endif + utlog_format(UTLOG_INFO, "Catcoon Version: %.1f\n", CATCOON_VER); + if (auto_tsrmgr_flag) + utlog_format(UTLOG_INFO, "Auto tensor management: On\n"); + else + utlog_format(UTLOG_INFO, "Auto tensor management: Off\n"); +} diff --git a/src/catcoon.h b/src/catcoon.h new file mode 100644 index 0000000..3690251 --- /dev/null +++ b/src/catcoon.h @@ -0,0 +1,30 @@ +#ifndef _CARCOON_H_ +#define _CARCOON_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_assert.h" +#include "cc_basic.h" +#include "cc_conv2d.h" +#include "cc_fmap2d.h" +#include "cc_fullycon.h" +#include "cc_image.h" +#include "cc_pad2d.h" +#include "cc_pool2d.h" +#include "cc_tsrmgr.h" +#include "cc_tensor.h" + +#define cc_clear() \ + cc_tsrmgr_clear(); + +#define CATCOON_VER 0.0F + +void cc_print_info(void); + +#ifdef __cplusplus + } +#endif + +#endif /* _CARCOON_H_ */ diff --git a/src/cc_array.c b/src/cc_array.c new file mode 100644 index 0000000..89b5d80 --- /dev/null +++ b/src/cc_array.c @@ -0,0 +1,377 @@ +#include +#include +#include + +#include "util_log.h" +#include "cc_dtype.h" +#include "cc_array.h" + +#define ELEM_OPS_ON_ARRAY(op, elem, arr, arrlen, dtype) \ + for (i = 0; i < arrlen; ++i) { \ + *((dtype*)arr + i) op ## = *(dtype*)elem; \ + } + +#define ARRAY_ELEM_OPS(op, elem, inp, oup, arrlen, dtype) \ + for (i = 0; i < arrlen; ++i) { \ + *((dtype*)oup + i) = *((dtype*)inp + i) \ + op *(dtype*)elem; \ + } + +#define ARRAY_SUM(arr, arrlen, dtype, sum) \ + *(dtype*)sum = 0; \ + for (i = 0; i < arrlen; ++i) { \ + *(dtype*)sum += *((dtype*)arr + i); \ + } + +#define ARRAY_OPS(op, dst, src, arrlen, dtype) \ + for (i = 0; i < arrlen; ++i) { \ + *((dtype*)dst + i) += *((dtype*)src + i); \ + } + +#define ARRAY_CAST_CASE(_DT, _srcdt, _dstdt) \ +case _DT: \ + for (i = 0; i < arrlen; ++i) \ + *((_dstdt*)dst + i) = (_dstdt)*((_srcdt*)src + i); \ + break; + +#define CC_ARRAY_CAST_IMPLEMENTATION(dtype) \ +void cc_array_cast_ ## dtype(void *dst, void *src, int arrlen, int dt) \ +{ \ + cc_int32 i; \ + switch (dt) { \ + ARRAY_CAST_CASE(CC_UINT8, cc_uint8, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_UINT16, cc_uint16, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_UINT32, cc_uint32, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_UINT64, cc_uint64, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_INT8, cc_int8, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_INT16, cc_int16, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_INT32, cc_int32, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_INT64, cc_int64, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_FLOAT32, cc_float32, cc_ ## dtype); \ + ARRAY_CAST_CASE(CC_FLOAT64, cc_float64, cc_ ## dtype); \ + default: \ + utlog_format(UTLOG_ERR, \ + "cc_array: unsupported dtype %x\n", dt); \ + break; \ + } \ +} + +CC_ARRAY_CAST_IMPLEMENTATION (uint8) +CC_ARRAY_CAST_IMPLEMENTATION (uint16) +CC_ARRAY_CAST_IMPLEMENTATION (uint32) +CC_ARRAY_CAST_IMPLEMENTATION (uint64) +CC_ARRAY_CAST_IMPLEMENTATION (int8) +CC_ARRAY_CAST_IMPLEMENTATION (int16) +CC_ARRAY_CAST_IMPLEMENTATION (int32) +CC_ARRAY_CAST_IMPLEMENTATION (int64) +CC_ARRAY_CAST_IMPLEMENTATION (float32) +CC_ARRAY_CAST_IMPLEMENTATION (float64) + +#define ARRAY_ADD_BY_CASE(_DT, _dt) \ +case _DT: \ + ELEM_OPS_ON_ARRAY(+, x, arr, arrlen, _dt); \ + break; +void cc_array_add_by(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_ADD_BY_CASE(CC_UINT8, cc_uint8); + ARRAY_ADD_BY_CASE(CC_UINT16, cc_uint16); + ARRAY_ADD_BY_CASE(CC_UINT32, cc_uint32); + ARRAY_ADD_BY_CASE(CC_UINT64, cc_uint64); + ARRAY_ADD_BY_CASE(CC_INT8, cc_int8); + ARRAY_ADD_BY_CASE(CC_INT16, cc_int16); + ARRAY_ADD_BY_CASE(CC_INT32, cc_int32); + ARRAY_ADD_BY_CASE(CC_INT64, cc_int64); + ARRAY_ADD_BY_CASE(CC_FLOAT32, cc_float32); + ARRAY_ADD_BY_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_SUB_BY_CASE(_DT, _dt) \ +case _DT: \ + ELEM_OPS_ON_ARRAY(-, x, arr, arrlen, _dt); \ + break; +void cc_array_sub_by(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_SUB_BY_CASE(CC_UINT8, cc_uint8); + ARRAY_SUB_BY_CASE(CC_UINT16, cc_uint16); + ARRAY_SUB_BY_CASE(CC_UINT32, cc_uint32); + ARRAY_SUB_BY_CASE(CC_UINT64, cc_uint64); + ARRAY_SUB_BY_CASE(CC_INT8, cc_int8); + ARRAY_SUB_BY_CASE(CC_INT16, cc_int16); + ARRAY_SUB_BY_CASE(CC_INT32, cc_int32); + ARRAY_SUB_BY_CASE(CC_INT64, cc_int64); + ARRAY_SUB_BY_CASE(CC_FLOAT32, cc_float32); + ARRAY_SUB_BY_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_MUL_BY_CASE(_DT, _dt) \ +case _DT: \ + ELEM_OPS_ON_ARRAY(*, x, arr, arrlen, _dt); \ + break; +void cc_array_mul_by(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_MUL_BY_CASE(CC_UINT8, cc_uint8); + ARRAY_MUL_BY_CASE(CC_UINT16, cc_uint16); + ARRAY_MUL_BY_CASE(CC_UINT32, cc_uint32); + ARRAY_MUL_BY_CASE(CC_UINT64, cc_uint64); + ARRAY_MUL_BY_CASE(CC_INT8, cc_int8); + ARRAY_MUL_BY_CASE(CC_INT16, cc_int16); + ARRAY_MUL_BY_CASE(CC_INT32, cc_int32); + ARRAY_MUL_BY_CASE(CC_INT64, cc_int64); + ARRAY_MUL_BY_CASE(CC_FLOAT32, cc_float32); + ARRAY_MUL_BY_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_DIV_BY_CASE(_DT, _dt) \ +case _DT: \ + ELEM_OPS_ON_ARRAY(/, x, arr, arrlen, _dt); \ + break; +void cc_array_div_by(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_DIV_BY_CASE(CC_UINT8, cc_uint8); + ARRAY_DIV_BY_CASE(CC_UINT16, cc_uint16); + ARRAY_DIV_BY_CASE(CC_UINT32, cc_uint32); + ARRAY_DIV_BY_CASE(CC_UINT64, cc_uint64); + ARRAY_DIV_BY_CASE(CC_INT8, cc_int8); + ARRAY_DIV_BY_CASE(CC_INT16, cc_int16); + ARRAY_DIV_BY_CASE(CC_INT32, cc_int32); + ARRAY_DIV_BY_CASE(CC_INT64, cc_int64); + ARRAY_DIV_BY_CASE(CC_FLOAT32, cc_float32); + ARRAY_DIV_BY_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_ADD_EW_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_ELEM_OPS(+, x, inp, oup, arrlen, _dt); \ + break; +void cc_array_add_ew(void *inp, void *oup, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_ADD_EW_CASE(CC_UINT8, cc_uint8); + ARRAY_ADD_EW_CASE(CC_UINT16, cc_uint16); + ARRAY_ADD_EW_CASE(CC_UINT32, cc_uint32); + ARRAY_ADD_EW_CASE(CC_UINT64, cc_uint64); + ARRAY_ADD_EW_CASE(CC_INT8, cc_int8); + ARRAY_ADD_EW_CASE(CC_INT16, cc_int16); + ARRAY_ADD_EW_CASE(CC_INT32, cc_int32); + ARRAY_ADD_EW_CASE(CC_INT64, cc_int64); + ARRAY_ADD_EW_CASE(CC_FLOAT32, cc_float32); + ARRAY_ADD_EW_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_MUL_EW_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_ELEM_OPS(*, x, inp, oup, arrlen, _dt); \ + break; +void cc_array_mul_ew(void *inp, void *oup, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_MUL_EW_CASE(CC_UINT8, cc_uint8); + ARRAY_MUL_EW_CASE(CC_UINT16, cc_uint16); + ARRAY_MUL_EW_CASE(CC_UINT32, cc_uint32); + ARRAY_MUL_EW_CASE(CC_UINT64, cc_uint64); + ARRAY_MUL_EW_CASE(CC_INT8, cc_int8); + ARRAY_MUL_EW_CASE(CC_INT16, cc_int16); + ARRAY_MUL_EW_CASE(CC_INT32, cc_int32); + ARRAY_MUL_EW_CASE(CC_INT64, cc_int64); + ARRAY_MUL_EW_CASE(CC_FLOAT32, cc_float32); + ARRAY_MUL_EW_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_SUM_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_SUM(arr, arrlen, _dt, x); \ + break; +void cc_array_sum(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_SUM_CASE(CC_UINT8, cc_uint8); + ARRAY_SUM_CASE(CC_UINT16, cc_uint16); + ARRAY_SUM_CASE(CC_UINT32, cc_uint32); + ARRAY_SUM_CASE(CC_UINT64, cc_uint64); + ARRAY_SUM_CASE(CC_INT8, cc_int8); + ARRAY_SUM_CASE(CC_INT16, cc_int16); + ARRAY_SUM_CASE(CC_INT32, cc_int32); + ARRAY_SUM_CASE(CC_INT64, cc_int64); + ARRAY_SUM_CASE(CC_FLOAT32, cc_float32); + ARRAY_SUM_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ARRAY_MEAN_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_SUM(arr, arrlen, _dt, x); \ + *(_dt*)x /= arrlen; \ + break; +void cc_array_mean(void *arr, int arrlen, void *x, int dt) +{ + cc_int32 i; + switch (dt) { + ARRAY_MEAN_CASE(CC_UINT8, cc_uint8); + ARRAY_MEAN_CASE(CC_UINT16, cc_uint16); + ARRAY_MEAN_CASE(CC_UINT32, cc_uint32); + ARRAY_MEAN_CASE(CC_UINT64, cc_uint64); + ARRAY_MEAN_CASE(CC_INT8, cc_int8); + ARRAY_MEAN_CASE(CC_INT16, cc_int16); + ARRAY_MEAN_CASE(CC_INT32, cc_int32); + ARRAY_MEAN_CASE(CC_INT64, cc_int64); + ARRAY_MEAN_CASE(CC_FLOAT32, cc_float32); + ARRAY_MEAN_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define ADD_TO_ARRAY_EW_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_OPS(+, dst, src, arrlen, _dt); \ + break; +void cc_add_to_array_ew(void *dst, void *src, int arrlen, int dt) +{ + cc_int32 i; + switch (dt) { + ADD_TO_ARRAY_EW_CASE(CC_UINT8, cc_uint8); + ADD_TO_ARRAY_EW_CASE(CC_UINT16, cc_uint16); + ADD_TO_ARRAY_EW_CASE(CC_UINT32, cc_uint32); + ADD_TO_ARRAY_EW_CASE(CC_UINT64, cc_uint64); + ADD_TO_ARRAY_EW_CASE(CC_INT8, cc_int8); + ADD_TO_ARRAY_EW_CASE(CC_INT16, cc_int16); + ADD_TO_ARRAY_EW_CASE(CC_INT32, cc_int32); + ADD_TO_ARRAY_EW_CASE(CC_INT64, cc_int64); + ADD_TO_ARRAY_EW_CASE(CC_FLOAT32, cc_float32); + ADD_TO_ARRAY_EW_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define MUL_TO_ARRAY_EW_CASE(_DT, _dt) \ +case _DT: \ + ARRAY_OPS(*, dst, src, arrlen, _dt); \ + break; +void cc_mul_to_array_ew(void *dst, void *src, int arrlen, int dt) +{ + cc_int32 i; + switch (dt) { + MUL_TO_ARRAY_EW_CASE(CC_UINT8, cc_uint8); + MUL_TO_ARRAY_EW_CASE(CC_UINT16, cc_uint16); + MUL_TO_ARRAY_EW_CASE(CC_UINT32, cc_uint32); + MUL_TO_ARRAY_EW_CASE(CC_UINT64, cc_uint64); + MUL_TO_ARRAY_EW_CASE(CC_INT8, cc_int8); + MUL_TO_ARRAY_EW_CASE(CC_INT16, cc_int16); + MUL_TO_ARRAY_EW_CASE(CC_INT32, cc_int32); + MUL_TO_ARRAY_EW_CASE(CC_INT64, cc_int64); + MUL_TO_ARRAY_EW_CASE(CC_FLOAT32, cc_float32); + MUL_TO_ARRAY_EW_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } +} + +#define PRINT_ARRAY_CASE(_DT, _dt) \ +case _DT: \ + fprintf(fp, pat, *((_dt*)a + i)); \ + break; +void cc_print_array(void *a, int arrlen, int dt, void *stream) +{ + cc_int32 i; + const char *pat = NULL; + FILE *fp = stream ? (FILE*)stream : stdout; + switch (dt) { + case CC_FLOAT32: + pat = "%f "; + break; + case CC_FLOAT64: + pat = "%lf "; + break; + case CC_UINT8: + case CC_UINT16: + case CC_UINT32: + pat = "%d "; + break; + case CC_UINT64: + pat = "%ld "; + break; + case CC_INT8: + case CC_INT16: + case CC_INT32: + pat = "%d "; + break; + case CC_INT64: + pat = "%ld "; + break; + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } + for (i = 0; i < arrlen; ++i) { + switch(dt) { + PRINT_ARRAY_CASE(CC_UINT8, cc_uint8); + PRINT_ARRAY_CASE(CC_UINT16, cc_uint16); + PRINT_ARRAY_CASE(CC_UINT32, cc_uint32); + PRINT_ARRAY_CASE(CC_UINT64, cc_uint64); + PRINT_ARRAY_CASE(CC_INT8, cc_int8); + PRINT_ARRAY_CASE(CC_INT16, cc_int16); + PRINT_ARRAY_CASE(CC_INT32, cc_int32); + PRINT_ARRAY_CASE(CC_INT64, cc_int64); + PRINT_ARRAY_CASE(CC_FLOAT32, cc_float32); + PRINT_ARRAY_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_array: unsupported dtype %x\n", dt); + break; + } + } + printf("\n"); +} diff --git a/src/cc_array.h b/src/cc_array.h new file mode 100644 index 0000000..e51a1d2 --- /dev/null +++ b/src/cc_array.h @@ -0,0 +1,46 @@ +#ifndef _CC_ARRAY_H_ +#define _CC_ARRAY_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +/* + * int <---> cc_int32 + */ + +#define CC_ARRAY_CAST_DEFINITION(dtype) \ +void cc_array_cast_ ## dtype(void *dst, void *src, int arrlen, int dt); + +CC_ARRAY_CAST_DEFINITION (uint8) +CC_ARRAY_CAST_DEFINITION (uint16) +CC_ARRAY_CAST_DEFINITION (uint32) +CC_ARRAY_CAST_DEFINITION (uint64) +CC_ARRAY_CAST_DEFINITION (int8) +CC_ARRAY_CAST_DEFINITION (int16) +CC_ARRAY_CAST_DEFINITION (int32) +CC_ARRAY_CAST_DEFINITION (int64) +CC_ARRAY_CAST_DEFINITION (float32) +CC_ARRAY_CAST_DEFINITION (float64) + +void cc_array_add_by(void *arr, int arrlen, void *x, int dt); +void cc_array_sub_by(void *arr, int arrlen, void *x, int dt); +void cc_array_mul_by(void *arr, int arrlen, void *x, int dt); +void cc_array_div_by(void *arr, int arrlen, void *x, int dt); + +void cc_array_add_ew(void *inp, void *oup, int arrlen, void *x, int dt); +void cc_array_mul_ew(void *inp, void *oup, int arrlen, void *x, int dt); + +void cc_array_sum(void *arr, int arrlen, void *x, int dt); +void cc_array_mean(void *arr, int arrlen, void *x, int dt); + +void cc_add_to_array_ew(void *dst, void *src, int arrlen, int dt); +void cc_mul_to_array_ew(void *dst, void *src, int arrlen, int dt); + +void cc_print_array(void *a, int arrlen, int dt, void *stream); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_ARRAY_H_ */ diff --git a/src/cc_assert.h b/src/cc_assert.h new file mode 100644 index 0000000..5a9ea64 --- /dev/null +++ b/src/cc_assert.h @@ -0,0 +1,47 @@ +#ifndef _CC_ASSERT_H_ +#define _CC_ASSERT_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "util_log.h" + +#if defined(ENABLE_CC_ASSERT) /* Assertions */ +#define cc_assert(expr) \ + if (!(expr)) \ + utlog_format(UTLOG_ERR, \ + "Assertion failed @["#expr "] in %s line: %d\n", \ + __FILE__, __LINE__); + +#define cc_assert_zero(expr) \ + if (expr) \ + utlog_format(UTLOG_ERR, \ + "Assertion failed @["#expr "] in %s line: %d\n", \ + __FILE__, __LINE__); + +#define cc_assert_ptr(ptr) \ + if (!(ptr)) \ + utlog_format(UTLOG_ERR, \ + "Assertion failed @["#ptr "] in %s line: %d\n", \ + __FILE__, __LINE__); + +#define cc_assert_alloc(ptr) \ + if (!(ptr)) \ + utlog_format(UTLOG_ERR, \ + "Memory alloc failed @["#ptr "] in %s line: %d\n", \ + __FILE__, __LINE__); +#else /* not defined(ENABLE_CC_ASSERT) */ + +#define cc_assert(expr) (expr) +#define cc_assert_zero(expr) (expr) +#define cc_assert_ptr(ptr) (ptr) +#define cc_assert_alloc(ptr) (ptr) + +#endif + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_ASSERT_H_ */ diff --git a/src/cc_basic.c b/src/cc_basic.c new file mode 100644 index 0000000..c8d6152 --- /dev/null +++ b/src/cc_basic.c @@ -0,0 +1,277 @@ +#include +#include +#include + +#include "cc_array.h" +#include "cc_assert.h" +#include "cc_dtype.h" +#include "cc_tensor.h" +#include "cc_tsrmgr.h" +#include "util_list.h" +#include "util_log.h" +#include "cc_basic.h" + +/* #include "global_fn_cfg.h" */ +extern void (*_activation_relu) + (void *inp, cc_int32 elems, cc_dtype dt); +extern void (*_activation_relu6) + (void *inp, cc_int32 elems, cc_dtype dt); +extern void (*_activation_softmax) + (void *inp, cc_int32 elems, cc_dtype dt); + +static cc_int32 _calc_elems(const cc_int32 *shape) +{ + cc_int32 elems; + elems = *shape; + while (*++shape) + elems *= *shape; + return elems; +} + +cc_int32 cc_tensor_elements(cc_tensor_t *tensor) +{ + cc_int32 elems; + if (!tensor) + return 0; + elems = _calc_elems(tensor->shape); + return elems; +} + +void cc_tensor_shape_fix(cc_int32 *shape, cc_int32 elems) +{ + cc_int32 v, i = 0, f = 0, s = 1; + const cc_int32 *sptr = shape; + while ((v = *sptr)) { + if (v == -1) { +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(f); +#endif + i = sptr - shape; + f = (v = 1); + } + s *= v; + sptr++; + } + if (s != elems) { +#ifdef ENABLE_CC_ASSERT + cc_assert(f); +#endif + shape[i] = elems / s; + } +} + +cc_int32 cc_tensor_dimension(cc_tensor_t *tensor) +{ + cc_int32 dim = 0; + const cc_int32 *sptr; + if (!tensor) + return 0; + sptr = tensor->shape; + while (*sptr++) + ++dim; + return dim; +} + +cc_tensor_t *cc_tensor_reshape(cc_tensor_t *tensor, cc_int32 *shape) +{ + cc_int32 elems; + const cc_int32 *sptr; + cc_tensor_shape_fix(shape, _calc_elems(tensor->shape)); +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(_calc_elems(tensor->shape) - _calc_elems(shape)); +#endif + sptr = shape; + elems = *shape; + while (*++sptr) + elems *= *sptr; + cc_assert_zero( + list_del_record(tensor->container, CC_TENSOR_SHAPE)); + cc_assert_zero( + list_set_record(tensor->container, CC_TENSOR_SHAPE, + shape, (sptr - shape + 1) * sizeof(cc_int32))); + cc_assert_ptr( + tensor->shape = (cc_int32*) + list_get_record(tensor->container, CC_TENSOR_SHAPE)); + return tensor; +} + +int cc_tsrcmp_by_shape(cc_tensor_t *a, cc_tensor_t *b) +{ + int ret = 0; + const cc_int32 *ptra = a->shape; + const cc_int32 *ptrb = b->shape; + while(!(ret = *(cc_int32*)ptra - *(cc_int32*)ptrb) && *ptra) { + ptra++; + ptrb++; + } + if (ret < 0) + return -1; + else if (ret > 0) + return 1; + return 0; +} + +void cc_print_tensor(cc_tensor_t *tensor) +{ + utlog_format(UTLOG_INFO, ""); + cc_print_array(tensor->data, + cc_tensor_elements(tensor), *tensor->dtype, + utlog_get_ostream()); +} + +cc_tensor_t *cc_cast_tensor(cc_tensor_t *tensor, + cc_dtype dtype, const char *name) +{ + cc_tensor_t *cast; + const cc_int32 *sptr = tensor->shape; + cc_int32 memsize, elems = *tensor->shape; + while (*++sptr) + elems *= *sptr; + memsize = cc_dtype_size(dtype) * elems; + if (!memsize) + return NULL; + cc_assert_ptr(cast = cc_create_tensor( + tensor->shape, dtype, NULL)); + switch (dtype) { + case CC_INT8: + cc_array_cast_int8(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_UINT8: + cc_array_cast_uint8(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_INT16: + cc_array_cast_int16(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_UINT16: + cc_array_cast_uint16(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_INT32: + cc_array_cast_int32(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_UINT32: + cc_array_cast_uint32(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_INT64: + cc_array_cast_int64(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_UINT64: + cc_array_cast_uint64(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_FLOAT32: + cc_array_cast_float32(cast->data, + tensor->data, elems, *tensor->dtype); + break; + case CC_FLOAT64: + cc_array_cast_float64(cast->data, + tensor->data, elems, *tensor->dtype); + break; + default: + utlog_format(UTLOG_ERR, + "cc_cast_tensor: unsupported dtype %x\n", + dtype); + cc_free_tensor(cast); + return NULL; + } + list_set_name(cast->container, name); + cast->name = cast->container->name; + if (!name || !strcmp(name, tensor->name)) { +#ifdef AUTO_TSRMGR + cc_tsrmgr_replace(cast); +#else + cc_free_tensor(tensor); +#endif + return cast; + } +#ifdef AUTO_TSRMGR + cc_tsrmgr_auto_reg(cast); +#endif + return cast; +} + +cc_tensor_t *cc_tensor_by_scalar(cc_tensor_t *tensor, + char op, void *data, const char *name) +{ + cc_tensor_t *yield; + cc_int32 elems = *tensor->shape; + const cc_int32 *sptr = tensor->shape; + while (*++sptr) + elems *= *sptr; + if (!name || !strcmp(name, tensor->name)) + yield = tensor; + else + yield = cc_copy_tensor(tensor, name); + switch (op) { + case '+': + cc_array_add_by( + yield->data, elems, data, *tensor->dtype); + break; + case '-': + cc_array_sub_by( + yield->data, elems, data, *tensor->dtype); + break; + case '*': + cc_array_mul_by( + yield->data, elems, data, *tensor->dtype); + break; + case '/': + cc_array_div_by( + yield->data, elems, data, *tensor->dtype); + break; + default: + utlog_format(UTLOG_ERR, + "cc_tensor_by_scalar: unsupported operator [%c]\n", + op); +#ifdef AUTO_TSRMGR + if (strcmp(name, tensor->name)) + cc_tsrmgr_del(name); +#else + cc_free_tensor(yield); +#endif + return NULL; + } + return yield; +} + +cc_tensor_t *cc_relu(cc_tensor_t *tensor, const char *name) +{ + cc_tensor_t *relu; + if (!name || !strcmp(name, tensor->name)) + relu = tensor; + else + relu = cc_copy_tensor(tensor, name); + _activation_relu(relu->data, + cc_tensor_elements(relu), *relu->dtype); + return relu; +} + +cc_tensor_t *cc_relu6(cc_tensor_t *tensor, const char *name) +{ + cc_tensor_t *relu; + if (!name || !strcmp(name, tensor->name)) + relu = tensor; + else + relu = cc_copy_tensor(tensor, name); + _activation_relu6(relu->data, + cc_tensor_elements(relu), *relu->dtype); + return relu; +} + +cc_tensor_t *cc_softmax(cc_tensor_t *tensor, const char *name) +{ + cc_tensor_t *softmax; + if (!name || !strcmp(name, tensor->name)) + softmax = tensor; + else + softmax = cc_copy_tensor(tensor, name); + _activation_softmax(softmax->data, + cc_tensor_elements(softmax), *softmax->dtype); + return softmax; +} diff --git a/src/cc_basic.h b/src/cc_basic.h new file mode 100644 index 0000000..f2bc6fe --- /dev/null +++ b/src/cc_basic.h @@ -0,0 +1,38 @@ +#ifndef _CC_BASIC_H_ +#define _CC_BASIC_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +cc_int32 cc_tensor_elements(cc_tensor_t *tensor); + +cc_int32 cc_tensor_dimension(cc_tensor_t *tensor); + +void cc_tensor_shape_fix(cc_int32 *shape, cc_int32 elems); + +cc_tensor_t *cc_tensor_reshape(cc_tensor_t *tensor, cc_int32 *shape); + +int cc_tsrcmp_by_shape(cc_tensor_t *a, cc_tensor_t *b); + +void cc_print_tensor(cc_tensor_t *tensor); + +cc_tensor_t *cc_cast_tensor(cc_tensor_t *tensor, + cc_dtype dtype, const char *name); + +cc_tensor_t *cc_tensor_by_scalar(cc_tensor_t *tensor, + char op, void *data, const char *name); + +cc_tensor_t *cc_relu (cc_tensor_t *tensor, const char *name); +cc_tensor_t *cc_relu6(cc_tensor_t *tensor, const char *name); + +cc_tensor_t *cc_softmax(cc_tensor_t *tensor, const char *name); + + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_BASIC_H_ */ diff --git a/src/cc_conv2d.c b/src/cc_conv2d.c new file mode 100644 index 0000000..6536685 --- /dev/null +++ b/src/cc_conv2d.c @@ -0,0 +1,138 @@ +#include +#include +#include + +#include "cc_array.h" +#include "cc_assert.h" +#include "cc_basic.h" +#include "cc_fmap2d.h" +#include "cc_pad2d.h" +#include "cc_tsrmgr.h" +#include "util_list.h" +#include "util_log.h" +#include "cc_conv2d.h" + +#ifdef ENABLE_OPENMP + #include +#endif + +/* #include "global_fn_cfg.h" */ +extern void (*_conv2d)(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 oup_x, + cc_int32 oup_y, cc_int32 sx, cc_int32 sy, + void *filter, cc_int32 fw, cc_dtype dt); + +cc_int32 cc_conv2d_shape_calc(cc_int32 i, + cc_int32 k, cc_int32 s, cc_int32 p) +{ + return (cc_int32)((i - k + 2 * p) / s) + 1; +} + +cc_tensor_t *cc_conv2d(cc_tensor_t *inp, + cc_tensor_t *kernel, cc_tensor_t *bias, cc_int32 s, + cc_int32 p, cc_int32 off, const char *name) +{ + cc_uint8 *omp_out_buf = NULL; + cc_tensor_t *inp_pad, *oup = NULL; + cc_int32 o_ch_size, p_ch_mem_size, o_ch_mem_size, + k_ch_mem_size, k_mem_size, num_omp_threads, i, j; + cc_int32 shape[CC_CNN2D_SHAPE_LEN] = {0}; + char pad_name[CC_CONV2D_PAD_NAME_LEN]; +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(cc_tensor_dimension(inp) - CC_CNN2D_DIM); + cc_assert_zero(cc_tensor_dimension(kernel) - CC_CONV2D_KERNEL_DIM); + cc_assert_zero(*inp->dtype - *kernel->dtype); + cc_assert_zero(inp->shape[CC_CNN2D_SHAPE_C] + - kernel->shape[CC_CONV2D_KERNEL_I]); +#endif + if (p) { + sprintf(pad_name, "%s%s", + inp->name, CC_CONV2D_PAD_NAME_SURFFIX); + inp_pad = cc_pad2d(inp, p, off, pad_name); + } + else + inp_pad = inp; +#ifdef AUTO_TSRMGR + oup = cc_tsrmgr_get(name); +#endif + if (!oup) { + shape[CC_CNN2D_SHAPE_C] = kernel->shape[CC_CONV2D_KERNEL_O]; + shape[CC_CNN2D_SHAPE_H] = cc_conv2d_shape_calc( + inp->shape[CC_CNN2D_SHAPE_H], + kernel->shape[CC_CONV2D_KERNEL_H], s, p); + shape[CC_CNN2D_SHAPE_W] = cc_conv2d_shape_calc( + inp->shape[CC_CNN2D_SHAPE_W], + kernel->shape[CC_CONV2D_KERNEL_W], s, p); + oup = cc_create_tensor(shape, *inp->dtype, name); + } + o_ch_size = oup->shape[CC_CNN2D_SHAPE_W] * + oup->shape[CC_CNN2D_SHAPE_H]; + o_ch_mem_size = o_ch_size * cc_dtype_size(*oup->dtype); + p_ch_mem_size = inp_pad->shape[CC_CNN2D_SHAPE_W] * + inp_pad->shape[CC_CNN2D_SHAPE_H] * + cc_dtype_size(*inp->dtype); + k_ch_mem_size = kernel->shape[CC_CONV2D_KERNEL_W] * + kernel->shape[CC_CONV2D_KERNEL_H] * + cc_dtype_size(*kernel->dtype); + k_mem_size = k_ch_mem_size * kernel->shape[CC_CONV2D_KERNEL_I]; + num_omp_threads = 1; +#ifdef ENABLE_OPENMP + num_omp_threads = omp_get_max_threads(); +#endif + cc_assert_alloc(omp_out_buf = + (cc_uint8*)malloc(o_ch_mem_size * num_omp_threads)); +#ifdef AUTO_TSRMGR + memset(oup->data, 0, + list_get_record_len(oup->container, CC_TENSOR_DATA)); +#endif +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i, j) +#endif + for (i = 0; i < kernel->shape[CC_CONV2D_KERNEL_O]; ++i) { + for (j = 0; j < kernel->shape[CC_CONV2D_KERNEL_I]; ++j) + { +#ifdef ENABLE_OPENMP + _conv2d((inp_pad->data + j * p_ch_mem_size), + omp_out_buf + omp_get_thread_num() * o_ch_mem_size, + inp_pad->shape[CC_CNN2D_SHAPE_W], + inp_pad->shape[CC_CNN2D_SHAPE_H], + oup->shape[CC_CNN2D_SHAPE_W], + oup->shape[CC_CNN2D_SHAPE_H], s, s, + (kernel->data + (k_mem_size * i) + + k_ch_mem_size * j), + kernel->shape[CC_CONV2D_KERNEL_W], + *kernel->dtype); + cc_add_to_array_ew(oup->data + o_ch_mem_size * i, + omp_out_buf + omp_get_thread_num() * o_ch_mem_size, + o_ch_size, *oup->dtype); +#else + _conv2d((inp_pad->data + p_ch_mem_size * j), + omp_out_buf, inp_pad->shape[CC_CNN2D_SHAPE_W], + inp_pad->shape[CC_CNN2D_SHAPE_H], + oup->shape[CC_CNN2D_SHAPE_W], + oup->shape[CC_CNN2D_SHAPE_H], s, s, + (kernel->data + (k_mem_size * i) + + k_ch_mem_size * j), + kernel->shape[CC_CONV2D_KERNEL_W], + *kernel->dtype); + cc_add_to_array_ew(oup->data + o_ch_mem_size * i, + omp_out_buf, o_ch_size, *oup->dtype); +#endif + } + } + free(omp_out_buf); + if (!bias){ +#ifndef AUTO_TSRMGR + if (p) + cc_free_tensor(inp_pad); +#endif + return oup; + } else { + oup = cc_fmap2d_bias(oup, bias, oup->name); + } +#ifndef AUTO_TSRMGR + if (p) + cc_free_tensor(inp_pad); +#endif + return oup; +} diff --git a/src/cc_conv2d.h b/src/cc_conv2d.h new file mode 100644 index 0000000..1a5d279 --- /dev/null +++ b/src/cc_conv2d.h @@ -0,0 +1,32 @@ +#ifndef _CC_CONV2D_H_ +#define _CC_CONV2D_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#define CC_CONV2D_PAD_NAME_LEN 128 +#define CC_CONV2D_PAD_NAME_SURFFIX "_pad" + +enum cc_conv2d_kernel { + CC_CONV2D_KERNEL_H, + CC_CONV2D_KERNEL_W, + CC_CONV2D_KERNEL_I, + CC_CONV2D_KERNEL_O, + CC_CONV2D_KERNEL_DIM +}; + +#include "cc_tensor.h" + +cc_int32 cc_conv2d_shape_calc(cc_int32 i, + cc_int32 k, cc_int32 s, cc_int32 p); + +cc_tensor_t *cc_conv2d(cc_tensor_t *inp, + cc_tensor_t *kernel, cc_tensor_t *bias, cc_int32 s, + cc_int32 p, cc_int32 off, const char *name); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_CONV2D_H_ */ diff --git a/src/cc_cpufn.c b/src/cc_cpufn.c new file mode 100644 index 0000000..a9e1fd0 --- /dev/null +++ b/src/cc_cpufn.c @@ -0,0 +1,417 @@ +#include +#include + +#include "cc_assert.h" +#include "util_log.h" +#include "cc_cpufn.h" + +#define UNSUPPORTED_DTYPE_LOG(dt) \ +utlog_format(UTLOG_ERR, \ +"cc_cpufn: unsupported dtype [%x]@%s: %d\n",\ +dt, __FILE__, __LINE__); + +#define CC_CPU_RELU_CASE_TEMP(_dt) \ +for (i = 0; i < elems; ++i) { \ + *(((_dt*)inp) + i) = \ + *(((_dt*)inp) + i) > 0 ? \ + *(((_dt*)inp) + i) : 0; \ +} + +void cc_cpu_activation_relu(void *inp, cc_int32 elems, cc_dtype dt) +{ + cc_int32 i; + switch (dt) { + case CC_UINT8: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_uint8); + break; + case CC_UINT16: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_uint16); + break; + case CC_UINT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_uint32); + break; + case CC_UINT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_uint64); + break; + case CC_INT8: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_int8); + break; + case CC_INT16: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_int16); + break; + case CC_INT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_int32); + break; + case CC_INT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_int64); + break; + case CC_FLOAT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_float32); + break; + case CC_FLOAT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU_CASE_TEMP(cc_float64); + break; + default: + utlog_format(UTLOG_ERR, + "cc_cpufn: unsupported dtype %x\n", dt); + } +} + +#define CC_CPU_RELU6_CASE_TEMP(_dt) \ +for (i = 0; i < elems; ++i) { \ + *(((_dt*)inp) + i) = \ + *(((_dt*)inp) + i) > 0 ? \ + *(((_dt*)inp) + i) : 0; \ + *(((_dt*)inp) + i) = \ + *(((_dt*)inp) + i) > 6 ? \ + 6 : *(((_dt*)inp) + i); \ +} + +void cc_cpu_activation_relu6(void *inp, cc_int32 elems, cc_dtype dt) +{ + cc_int32 i; + switch (dt) { + case CC_UINT8: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_uint8); + break; + case CC_UINT16: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_uint16); + break; + case CC_UINT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_uint32); + break; + case CC_UINT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_uint64); + break; + case CC_INT8: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_int8); + break; + case CC_INT16: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_int16); + break; + case CC_INT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_int32); + break; + case CC_INT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_int64); + break; + case CC_FLOAT32: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_float32); + break; + case CC_FLOAT64: +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + CC_CPU_RELU6_CASE_TEMP(cc_float64); + break; + default: + utlog_format(UTLOG_ERR, + "cc_cpufn: unsupported dtype %x\n", dt); + } +} + +static void cc_cpu_activation_softmax_float32( + cc_float32 *inp, cc_int32 elems) +{ + cc_int32 i; + cc_float32 v, s = 0; + for (i = 0; i < elems; ++i) { +#ifdef CONFIG_STD_C89 + v = (cc_float32)exp(inp[i]); + inp[i] = v; + s += v; +#else + v = expf(inp[i]); + inp[i] = v; + s += v; +#endif + } + for (i = 0; i < elems; ++i) { + inp[i] = inp[i] / s; + } +} + +static void cc_cpu_activation_softmax_float64( + cc_float64 *inp, cc_int32 elems) +{ + cc_int32 i; + cc_float64 v, s = 0; + for (i = 0; i < elems; ++i) { + v = exp(inp[i]); + inp[i] = v; + s += v; + } + for (i = 0; i < elems; ++i) { + inp[i] = inp[i] / s; + } +} + +void cc_cpu_activation_softmax(void *inp, cc_int32 elems, cc_dtype dt) +{ + switch (dt) { + case CC_FLOAT32: + cc_cpu_activation_softmax_float32( + (cc_float32*)inp, elems); + break; + case CC_FLOAT64: + cc_cpu_activation_softmax_float64( + (cc_float64*)inp, elems); + break; + default: + utlog_format(UTLOG_ERR, + "cc_cpufn: unsupported dtype %x\n", dt); + } +} + +#define CC_CPU_MAX_POOL2D_CASE(DT, dtype) \ +case DT: \ +for (i = 0; i < o_y; ++i) { \ + for (j = 0; j < o_x; ++j) \ + { \ + *(dtype*)v_max = *((dtype*)inp + s * i * x + s * j); \ + for (k = 0; k < s; ++k) \ + { \ + for (l = 0; l < s; ++l) \ + { \ + curr = (dtype*)inp + \ + (s * i + k) * x + j * s + l; \ + *(dtype*)v_max = *((dtype*)curr) > \ + *(dtype*)v_max ? \ + *((dtype*)curr) : *(dtype*)v_max; \ + } \ + } \ + *((dtype*)oup + i * o_x + j) = *(dtype*)v_max; \ + } \ +} \ +break; + +void cc_cpu_max_pool2d(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 s, cc_dtype dt) +{ + cc_int32 o_x = x / s; + cc_int32 o_y = y / s; + cc_int32 i, j, k, l; + void *curr, *v_max; + cc_assert_alloc(v_max = malloc(cc_dtype_size(dt))); + switch (dt) { + CC_CPU_MAX_POOL2D_CASE(CC_UINT8, cc_uint8); + CC_CPU_MAX_POOL2D_CASE(CC_UINT16, cc_uint16); + CC_CPU_MAX_POOL2D_CASE(CC_UINT32, cc_uint32); + CC_CPU_MAX_POOL2D_CASE(CC_UINT64, cc_uint64); + CC_CPU_MAX_POOL2D_CASE(CC_INT8, cc_int8); + CC_CPU_MAX_POOL2D_CASE(CC_INT16, cc_int16); + CC_CPU_MAX_POOL2D_CASE(CC_INT32, cc_int32); + CC_CPU_MAX_POOL2D_CASE(CC_INT64, cc_int64); + CC_CPU_MAX_POOL2D_CASE(CC_FLOAT32, cc_float32); + CC_CPU_MAX_POOL2D_CASE(CC_FLOAT64, cc_float64); + default: + utlog_format(UTLOG_ERR, + "cc_cpufn: unsupported dtype %x\n", dt); + } + free(v_max); +} + +#define CC_CPU_CONV2D_IMPLEMENTATION(dt) \ +static void cc_cpu_conv2d_ ## dt (cc_ ## dt *inp, cc_ ## dt *oup, \ + cc_int32 x, cc_int32 y, cc_int32 oup_x, cc_int32 oup_y, \ + cc_int32 sx, cc_int32 sy, cc_ ## dt *filter, cc_int32 fw) \ +{ \ + cc_int32 i, j, k, l, oup_i, oup_j; \ + cc_int32 half_fl = fw >> 1; \ + cc_ ## dt sum; \ + for (i = half_fl; i < y - half_fl; i += sy) { \ + for (j = half_fl; j < x - half_fl; j += sx) { \ + sum = 0; \ + for (k = -half_fl; k <= half_fl; ++k) { \ + for (l = -half_fl; l <= half_fl; ++l) { \ + sum += *(inp + (i + k) * x + (j + l)) * \ + *(filter + (k + half_fl) * fw + (l + half_fl)); \ + } \ + } \ + oup_i = ((i - half_fl) / sy); \ + oup_j = ((j - half_fl) / sx); \ + *(oup + oup_i * oup_x + oup_j) = sum; \ + } \ + } \ +} + +CC_CPU_CONV2D_IMPLEMENTATION (uint8) +CC_CPU_CONV2D_IMPLEMENTATION (uint16) +CC_CPU_CONV2D_IMPLEMENTATION (uint32) +CC_CPU_CONV2D_IMPLEMENTATION (uint64) +CC_CPU_CONV2D_IMPLEMENTATION (int8) +CC_CPU_CONV2D_IMPLEMENTATION (int16) +CC_CPU_CONV2D_IMPLEMENTATION (int32) +CC_CPU_CONV2D_IMPLEMENTATION (int64) +CC_CPU_CONV2D_IMPLEMENTATION (float32) +CC_CPU_CONV2D_IMPLEMENTATION (float64) + + +void cc_cpu_conv2d(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 oup_x, cc_int32 oup_y, + cc_int32 sx, cc_int32 sy, void *filter, cc_int32 fw, cc_dtype dt) +{ + switch (dt) { + case CC_UINT8: + cc_cpu_conv2d_uint8((cc_uint8*)inp, + (cc_uint8*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_uint8*)filter, fw); + break; + case CC_UINT16: + cc_cpu_conv2d_uint16((cc_uint16*)inp, + (cc_uint16*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_uint16*)filter, fw); + break; + case CC_UINT32: + cc_cpu_conv2d_uint32((cc_uint32*)inp, + (cc_uint32*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_uint32*)filter, fw); + break; + case CC_UINT64: + cc_cpu_conv2d_uint64((cc_uint64*)inp, + (cc_uint64*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_uint64*)filter, fw); + break; + case CC_INT8: + cc_cpu_conv2d_int8((cc_int8*)inp, + (cc_int8*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_int8*)filter, fw); + break; + case CC_INT16: + cc_cpu_conv2d_int16((cc_int16*)inp, + (cc_int16*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_int16*)filter, fw); + break; + case CC_INT32: + cc_cpu_conv2d_int32((cc_int32*)inp, + (cc_int32*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_int32*)filter, fw); + break; + case CC_INT64: + cc_cpu_conv2d_int64((cc_int64*)inp, + (cc_int64*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_int64*)filter, fw); + break; + case CC_FLOAT32: + cc_cpu_conv2d_float32((cc_float32*)inp, + (cc_float32*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_float32*)filter, fw); + break; + case CC_FLOAT64: + cc_cpu_conv2d_float64((cc_float64*)inp, + (cc_float64*)oup, x, y, oup_x, oup_y, + sx, sy, (cc_float64*)filter, fw); + break; + default: + utlog_format(UTLOG_ERR, + "cc_cpufn: unsupported dtype %x\n", dt); + } +} + +#define CC_CPU_FULLY_CONNECTED_PROCESS \ +for (i = 0; i < ow; ++i) { \ + oup[i] = 0; \ + for (j = 0; j < iw; ++j) { \ + oup[i] += (*(inp + j)) * (*(w + iw * i + j)); \ + } \ + if (b) \ + oup[i] += b[i]; \ +} + +static void cc_cpu_fully_connected_float32( + cc_float32 *inp, cc_float32 *oup, cc_float32 *w, + cc_float32 *b, cc_int32 iw, cc_int32 ow) +{ + cc_int32 i, j; +#ifdef ENABLE_OPENMP +#pragma omp parallel for private(i, j) +#endif + CC_CPU_FULLY_CONNECTED_PROCESS; +} + +static void cc_cpu_fully_connected_float64( + cc_float64 *inp, cc_float64 *oup, cc_float64 *w, + cc_float64 *b, cc_int32 iw, cc_int32 ow) +{ + cc_int32 i, j; +#ifdef ENABLE_OPENMP +#pragma omp parallel for private(i, j) +#endif + CC_CPU_FULLY_CONNECTED_PROCESS; +} + +void cc_cpu_fully_connected(void *inp, void *oup, + void *w, void *b, cc_int32 iw, cc_int32 ow, cc_dtype dt) +{ + switch (dt) { + case CC_FLOAT32: + cc_cpu_fully_connected_float32( + (cc_float32*)inp, (cc_float32*)oup, + (cc_float32*)w, (cc_float32*)b, iw, ow); + break; + case CC_FLOAT64: + cc_cpu_fully_connected_float64( + (cc_float64*)inp, (cc_float64*)oup, + (cc_float64*)w, (cc_float64*)b, iw, ow); + break; + default: + UNSUPPORTED_DTYPE_LOG(dt); + } +} diff --git a/src/cc_cpufn.h b/src/cc_cpufn.h new file mode 100644 index 0000000..66cc610 --- /dev/null +++ b/src/cc_cpufn.h @@ -0,0 +1,30 @@ +#ifndef _CC_CPUFN_H_ +#define _CC_CPUFN_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_dtype.h" + +void cc_cpu_activation_relu(void *inp, cc_int32 elems, cc_dtype dt); + +void cc_cpu_activation_relu6(void *inp, cc_int32 elems, cc_dtype dt); + +void cc_cpu_activation_softmax(void *inp, cc_int32 elems, cc_dtype dt); + +void cc_cpu_max_pool2d(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 s, cc_dtype dt); + +void cc_cpu_conv2d(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 oup_x, cc_int32 oup_y, + cc_int32 sx, cc_int32 sy, void *filter, cc_int32 fw, cc_dtype dt); + +void cc_cpu_fully_connected(void *inp, void *oup, + void *w, void *b, cc_int32 iw, cc_int32 ow, cc_dtype dt); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_CPUFN_H_ */ diff --git a/src/cc_dtype.c b/src/cc_dtype.c new file mode 100644 index 0000000..a5180d7 --- /dev/null +++ b/src/cc_dtype.c @@ -0,0 +1,65 @@ +#include "cc_dtype.h" +#include "util_log.h" + +#define COMPILER_ASSERT(condition)\ + ((void)sizeof(char[1 - 2*!!(condition)])) +/*============ compile-time test ============*/ +void __________cc_datatype_check__________() +{ + COMPILER_ASSERT(CC_1B_LEN - sizeof(cc_uint8)); + COMPILER_ASSERT(CC_2B_LEN - sizeof(cc_uint16)); + COMPILER_ASSERT(CC_4B_LEN - sizeof(cc_uint32)); + COMPILER_ASSERT(CC_8B_LEN - sizeof(cc_uint64)); + COMPILER_ASSERT(CC_4B_LEN - sizeof(cc_float32)); + COMPILER_ASSERT(CC_8B_LEN - sizeof(cc_float64)); +} + +int cc_dtype_size(cc_dtype dt) +{ + switch (dt) { + case CC_INT8: + case CC_UINT8: + return CC_1B_LEN; + case CC_INT16: + case CC_UINT16: + return CC_2B_LEN; + case CC_INT32: + case CC_UINT32: + case CC_FLOAT32: + return CC_4B_LEN; + case CC_INT64: + case CC_UINT64: + case CC_FLOAT64: + return CC_8B_LEN; + default: + return 0; + } +} + +const char *cc_dtype_to_string(cc_dtype dt) +{ + switch (dt) { + case CC_UINT8: + return "cc_uint8"; + case CC_UINT16: + return "cc_uint16"; + case CC_UINT32: + return "cc_uint32"; + case CC_UINT64: + return "cc_uint64"; + case CC_INT8: + return "cc_int8"; + case CC_INT16: + return "cc_int16"; + case CC_INT32: + return "cc_int32"; + case CC_INT64: + return "cc_int64"; + case CC_FLOAT32: + return "cc_float32"; + case CC_FLOAT64: + return "cc_float64"; + default: + return "unsupported dtype"; + } +} diff --git a/src/cc_dtype.h b/src/cc_dtype.h new file mode 100644 index 0000000..5567ab7 --- /dev/null +++ b/src/cc_dtype.h @@ -0,0 +1,96 @@ +#ifndef _CC_DTYPE_H_ +#define _CC_DTYPE_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#define _DT_UINT8 unsigned char +#define _DT_INT8 char +#define _DT_UINT16 unsigned short +#define _DT_INT16 short +#define _DT_UINT32 unsigned int +#define _DT_INT32 int +#ifdef _MSC_VER + #define _DT_UINT64 unsigned __int64 + #define _DT_INT64 __int64 +#else /* 64bit integer for gcc, clang... */ + #define _DT_UINT64 unsigned long long + #define _DT_INT64 long long +#endif +/* #define _DT_FLOAT16 */ +#define _DT_FLOAT32 float +#define _DT_FLOAT64 double + +typedef _DT_UINT8 cc_dtype; +typedef _DT_UINT8 cc_uint8; +typedef _DT_UINT16 cc_uint16; +typedef _DT_UINT32 cc_uint32; +typedef _DT_UINT64 cc_uint64; +typedef _DT_INT8 cc_int8; +typedef _DT_INT16 cc_int16; +typedef _DT_INT32 cc_int32; +typedef _DT_INT64 cc_int64; +/* typedef _DT_FLOAT32 cc_float16; */ +typedef _DT_FLOAT32 cc_float32; +typedef _DT_FLOAT64 cc_float64; + +/* + * + * 7 6 5 4 3 ~ 0 + * | | | | |_____| + * | | | | |____sub-class code + * | | | + * | | + * | |______integer or float(0/1) + * | + * |______unsigned or signed(0/1) + * + * Sub class code: + * bits = 2 ^ n + * + */ + +#define CC_FLAG_POINT 6 +#define CC_FLAG_SIGN 7 + +#define CC_UINT8 0x03 +#define CC_UINT16 0x04 +#define CC_UINT32 0x05 +#define CC_UINT64 0x06 +#define CC_INT8 0x83 +#define CC_INT16 0x84 +#define CC_INT32 0x85 +#define CC_INT64 0x86 +/* #define CC_FLOAT16 0xC4 */ +#define CC_FLOAT32 0xC5 +#define CC_FLOAT64 0xC6 + +#define CC_1B_LEN 1 +#define CC_2B_LEN 2 +#define CC_4B_LEN 4 +#define CC_8B_LEN 8 + +#define CC_INT8_LEN CC_1B_LEN +#define CC_INT16_LEN CC_2B_LEN +#define CC_INT32_LEN CC_4B_LEN +#define CC_INT64_LEN CC_8B_LEN +#define CC_UINT8_LEN CC_1B_LEN +#define CC_UINT16_LEN CC_2B_LEN +#define CC_UINT32_LEN CC_4B_LEN +#define CC_UINT64_LEN CC_8B_LEN +/* #define CC_FLOAT16_LEN CC_2B_LEN */ +#define CC_FLOAT32_LEN CC_4B_LEN +#define CC_FLOAT64_LEN CC_8B_LEN + +#define cc_dtype_check(dt1, dt2) (!(dt1 - dt2)) + +int cc_dtype_size(cc_dtype dt); + +const char *cc_dtype_to_string(cc_dtype dt); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_DTYPE_H_ */ diff --git a/src/cc_fmap2d.c b/src/cc_fmap2d.c new file mode 100644 index 0000000..844fa47 --- /dev/null +++ b/src/cc_fmap2d.c @@ -0,0 +1,77 @@ +#include +#include + +#include "cc_array.h" +#include "cc_assert.h" +#include "cc_basic.h" +#include "cc_fmap2d.h" +#include "cc_tsrmgr.h" + +#ifdef ENABLE_OPENMP + #include +#endif + +cc_tensor_t *cc_fmap2d_bias(cc_tensor_t *inp, + cc_tensor_t *bias, const char *name) +{ + cc_tensor_t *fmap; + cc_int32 i, ch_size, ch_mem_size, dt_size; +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(cc_tensor_dimension(inp) - CC_CNN2D_DIM); + cc_assert_zero(*inp->dtype - *bias->dtype); + cc_assert_zero(inp->shape[CC_CNN2D_SHAPE_C] + - bias->shape[CC_CNN2D_SHAPE_C]); + cc_assert_zero(bias->shape[CC_CNN2D_SHAPE_H]); /* [C, \0] */ +#endif + if (!name || !strcmp(name, inp->name)) + fmap = inp; + else + fmap = cc_copy_tensor(inp, name); + dt_size = cc_dtype_size(*fmap->dtype); + ch_size = fmap->shape[CC_CNN2D_SHAPE_H] * + fmap->shape[CC_CNN2D_SHAPE_W]; + ch_mem_size = ch_size * dt_size; +#ifdef ENABLE_OPENMP + #pragma omp parallel for private(i) +#endif + for (i = 0; i < bias->shape[CC_CNN2D_SHAPE_C]; ++i) { + cc_array_add_by(fmap->data + ch_mem_size * i, + ch_size, bias->data + dt_size * i, *fmap->dtype); + } + return fmap; +} + +cc_tensor_t *cc_fmap2d_flat(cc_tensor_t *inp, const char *name) +{ + cc_tensor_t *flat = NULL; + cc_uint8 *sptr, *dptr; + cc_int32 shape[CC_CNN2D_SHAPE_LEN] = {0}; + cc_int32 i, j ,ch_size, dt_size; +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(cc_tensor_dimension(inp) - CC_CNN2D_DIM); +#endif +#ifdef AUTO_TSRMGR + flat = cc_tsrmgr_get(name); +#endif + if (!flat) { + shape[CC_CNN2D_SHAPE_C] = cc_tensor_elements(inp); + shape[CC_CNN2D_SHAPE_H] = 1; + shape[CC_CNN2D_SHAPE_W] = 1; + cc_assert_ptr(flat = + cc_create_tensor(shape, *inp->dtype, name)); + } + sptr = (cc_uint8*)inp->data; + dptr = (cc_uint8*)flat->data; + dt_size = cc_dtype_size(*inp->dtype); + ch_size = inp->shape[CC_CNN2D_SHAPE_H] * + inp->shape[CC_CNN2D_SHAPE_W]; + for (i = 0; i < ch_size; ++i) { + for (j = 0; j < inp->shape[CC_CNN2D_SHAPE_C]; ++j) { + memcpy((dptr + + (i * inp->shape[CC_CNN2D_SHAPE_C] + j) * + dt_size), (sptr + (j * ch_size + i) * + dt_size), dt_size); + } + } + return flat; +} diff --git a/src/cc_fmap2d.h b/src/cc_fmap2d.h new file mode 100644 index 0000000..9557175 --- /dev/null +++ b/src/cc_fmap2d.h @@ -0,0 +1,27 @@ +#ifndef _CC_FMAP2D_H_ +#define _CC_FMAP2D_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +enum cc_cnn2d_shape { + CC_CNN2D_SHAPE_C, + CC_CNN2D_SHAPE_H, + CC_CNN2D_SHAPE_W, + CC_CNN2D_DIM, + CC_CNN2D_SHAPE_LEN +}; + +cc_tensor_t *cc_fmap2d_bias(cc_tensor_t *inp, + cc_tensor_t *bias, const char *name); + +cc_tensor_t *cc_fmap2d_flat(cc_tensor_t *inp, const char *name); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_FMAP2D_H_ */ diff --git a/src/cc_fullycon.c b/src/cc_fullycon.c new file mode 100644 index 0000000..4630f79 --- /dev/null +++ b/src/cc_fullycon.c @@ -0,0 +1,46 @@ +#include + +#include "cc_assert.h" +#include "cc_basic.h" +#include "cc_conv2d.h" +#include "cc_fmap2d.h" +#include "cc_tsrmgr.h" +#include "cc_fullycon.h" + +/* #include "global_fn_cfg.h" */ +extern void (*_fully_connected)(void *inp, + void *oup, void *w, void *b, cc_int32 iw, + cc_int32 ow, cc_dtype dt); + +cc_tensor_t *cc_fully_connected(cc_tensor_t *inp, + cc_tensor_t *w, cc_tensor_t *b, const char *name) +{ + cc_tensor_t *oup = NULL; + cc_int32 shape[CC_CNN2D_SHAPE_LEN] = {0}; +#ifdef ENABLE_CC_ASSERT + cc_assert_zero(cc_tensor_dimension(inp) - CC_CNN2D_DIM); + cc_assert_zero(cc_tensor_dimension(w) - CC_CONV2D_KERNEL_DIM); + cc_assert_zero(*inp->dtype - *w->dtype); + cc_assert_zero(*inp->dtype - *b->dtype); + cc_assert_zero(inp->shape[CC_CNN2D_SHAPE_C] + - w->shape[CC_CONV2D_KERNEL_I]); +#endif +#ifdef AUTO_TSRMGR + oup = cc_tsrmgr_get(name); +#endif + if (!oup) { + shape[CC_CNN2D_SHAPE_C] = w->shape[CC_CONV2D_KERNEL_O]; + shape[CC_CNN2D_SHAPE_H] = 1; + shape[CC_CNN2D_SHAPE_W] = 1; + oup = cc_create_tensor(shape, *inp->dtype, name); + } + if (b) + _fully_connected(inp->data, oup->data, + w->data, b->data, w->shape[CC_CONV2D_KERNEL_I], + w->shape[CC_CONV2D_KERNEL_O], *inp->dtype); + else + _fully_connected(inp->data, oup->data, + w->data, NULL, w->shape[CC_CONV2D_KERNEL_I], + w->shape[CC_CONV2D_KERNEL_O], *inp->dtype); + return oup; +} diff --git a/src/cc_fullycon.h b/src/cc_fullycon.h new file mode 100644 index 0000000..2bd7eef --- /dev/null +++ b/src/cc_fullycon.h @@ -0,0 +1,17 @@ +#ifndef _CC_FULLYCON_H_ +#define _CC_FULLYCON_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +cc_tensor_t *cc_fully_connected(cc_tensor_t *inp, + cc_tensor_t *w, cc_tensor_t *b, const char *name); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_FULLYCON_H_ */ diff --git a/src/cc_image.c b/src/cc_image.c new file mode 100644 index 0000000..5d948e6 --- /dev/null +++ b/src/cc_image.c @@ -0,0 +1,54 @@ +#include +#include + +#include "cc_assert.h" +#include "cc_array.h" +#include "cc_basic.h" +#define _CC_IMAGE_C_ +#include "cc_image.h" + +cc_tensor_t *cc_image2tensor(utim_image_t *img, const char *name) +{ + cc_uint8 *ptr; + cc_int32 i, j, ch_size; + cc_int32 shape[4] = {0}; /* [C, H, W, \0] */ + cc_tensor_t *tensor; + shape[CC_IT_SHAPE_C] = img->channels; + shape[CC_IT_SHAPE_H] = img->ysize; + shape[CC_IT_SHAPE_W] = img->xsize; + cc_assert_ptr(tensor = + cc_create_tensor(shape, CC_UINT8, name)); + ch_size = img->xsize * img->ysize; + ptr = tensor->data; + for (i = 0; i < ch_size; ++i) { + for (j = 0; j < img->channels; ++j) { + *(ptr + ch_size * j) = + *(img->pixels + i * img->channels + j); + } + ptr++; + } + return tensor; +} + +utim_image_t *cc_tensor2image(cc_tensor_t *tensor) +{ + utim_image_t *img; + cc_uint8 *pxls, *ptr; + cc_int32 i, j, ch_size, npxls; + const cc_int32 *sptr = tensor->shape; + if (cc_tensor_dimension(tensor) != 3) /* [C, H, W] */ + return NULL; + img = utim_create(sptr[CC_IT_SHAPE_W], + sptr[CC_IT_SHAPE_H], sptr[CC_IT_SHAPE_C], 0); + ch_size = img->xsize * img->ysize; + npxls = ch_size * img->channels; + cc_assert_alloc(pxls = (cc_uint8*)malloc(npxls)); + cc_array_cast_uint8(pxls, tensor->data, npxls, *tensor->dtype); + ptr = img->pixels; + for (i = 0; i < ch_size; ++i) { + for (j = 0; j < img->channels; ++j) + *ptr++ = *(pxls + j * ch_size + i); + } + free(pxls); + return img; +} diff --git a/src/cc_image.h b/src/cc_image.h new file mode 100644 index 0000000..5d99336 --- /dev/null +++ b/src/cc_image.h @@ -0,0 +1,32 @@ +#ifndef _CC_IMAGE_H_ +#define _CC_IMAGE_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" +#include "util_image.h" + +/* + * Create tensor from image, [C, H, W] format + */ +#define CC_IT_SHAPE_C 0 +#define CC_IT_SHAPE_H 1 +#define CC_IT_SHAPE_W 2 + +cc_tensor_t *cc_image2tensor(utim_image_t *img, const char *name); + +utim_image_t *cc_tensor2image(cc_tensor_t *tensor); + +#ifndef _CC_IMAGE_C_ + #undef CC_IT_SHAPE_C + #undef CC_IT_SHAPE_H + #undef CC_IT_SHAPE_W +#endif + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_IMAGE_H_ */ diff --git a/src/cc_pad2d.c b/src/cc_pad2d.c new file mode 100644 index 0000000..65b5e26 --- /dev/null +++ b/src/cc_pad2d.c @@ -0,0 +1,51 @@ +#include +#include + +#include "cc_dtype.h" +#include "cc_fmap2d.h" +#include "cc_pad2d.h" +#include "cc_tsrmgr.h" + +#define PAD_MEMCPY \ + memcpy((pad->data) + p_ch_mem_size * c + \ + (i + p - poffset) * p_row_mem_size + (p + j - poffset) * dtsize, \ + inp->data + i_ch_mem_size * c + i * i_row_mem_size + j * dtsize, \ + dtsize); + +cc_tensor_t *cc_pad2d(cc_tensor_t *inp, cc_int32 p, + cc_int32 offset, const char *name) +{ + cc_tensor_t *pad = NULL; + cc_int32 shape[CC_CNN2D_SHAPE_LEN] = {0}; + cc_int32 soffset = offset ? 1 : 0; + cc_int32 poffset = offset > 0 ? 1 : 0; + cc_int32 i, j, c, dtsize = cc_dtype_size(*inp->dtype); + cc_int32 i_ch_size, i_ch_mem_size, i_row_mem_size, + p_ch_size, p_ch_mem_size, p_row_mem_size; + i_ch_size = inp->shape[CC_CNN2D_SHAPE_W] * + inp->shape[CC_CNN2D_SHAPE_H]; + i_ch_mem_size = i_ch_size * dtsize; + i_row_mem_size = inp->shape[CC_CNN2D_SHAPE_W] * dtsize; +#ifdef AUTO_TSRMGR + pad = cc_tsrmgr_get(name); +#endif + if (!pad) { + shape[CC_CNN2D_SHAPE_C] = inp->shape[CC_CNN2D_SHAPE_C]; + shape[CC_CNN2D_SHAPE_H] = + inp->shape[CC_CNN2D_SHAPE_H] + p + p - soffset; + shape[CC_CNN2D_SHAPE_W] = + inp->shape[CC_CNN2D_SHAPE_W] + p + p - soffset; + pad = cc_create_tensor(shape, *inp->dtype, name); + } + p_ch_size = pad->shape[CC_CNN2D_SHAPE_W] * + pad->shape[CC_CNN2D_SHAPE_H]; + p_ch_mem_size = p_ch_size * dtsize; + p_row_mem_size = pad->shape[CC_CNN2D_SHAPE_W] * dtsize; + for (c = 0; c < inp->shape[CC_CNN2D_SHAPE_C]; ++c) { + for (i = 0; i < inp->shape[CC_CNN2D_SHAPE_H]; ++i) { + for (j = 0; j < inp->shape[CC_CNN2D_SHAPE_W]; ++j) + PAD_MEMCPY; + } + } + return pad; +} diff --git a/src/cc_pad2d.h b/src/cc_pad2d.h new file mode 100644 index 0000000..d5b2909 --- /dev/null +++ b/src/cc_pad2d.h @@ -0,0 +1,17 @@ +#ifndef _CC_PAD2D_H_ +#define _CC_PAD2D_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +cc_tensor_t *cc_pad2d(cc_tensor_t *inp, cc_int32 p, + cc_int32 offset, const char *name); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_PAD2D_H_ */ diff --git a/src/cc_pool2d.c b/src/cc_pool2d.c new file mode 100644 index 0000000..a8020cb --- /dev/null +++ b/src/cc_pool2d.c @@ -0,0 +1,40 @@ +#include + +#include "cc_cpufn.h" +#include "cc_fmap2d.h" +#include "cc_tsrmgr.h" +#include "cc_pool2d.h" + +/* #include "global_fn_cfg.h" */ +extern void (*_max_pool2d)(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 s, cc_dtype dt); + +cc_tensor_t *cc_max_pool2d(cc_tensor_t *inp, cc_int32 s, const char *name) +{ + cc_int32 i, i_ch_size, i_ch_mem_size, o_ch_size, o_ch_mem_size; + cc_int32 shape[CC_CNN2D_SHAPE_LEN] = {0}; + cc_tensor_t *pool = NULL; +#ifdef AUTO_TSRMGR + pool = cc_tsrmgr_get(name); +#endif + if (!pool) { + shape[CC_CNN2D_SHAPE_C] = inp->shape[CC_CNN2D_SHAPE_C]; + shape[CC_CNN2D_SHAPE_H] = inp->shape[CC_CNN2D_SHAPE_H] / s; + shape[CC_CNN2D_SHAPE_W] = inp->shape[CC_CNN2D_SHAPE_W] / s; + pool = cc_create_tensor(shape, *inp->dtype, name); + } + i_ch_size = inp->shape[CC_CNN2D_SHAPE_H] * + inp->shape[CC_CNN2D_SHAPE_W]; + i_ch_mem_size = i_ch_size * cc_dtype_size(*inp->dtype); + o_ch_size = pool->shape[CC_CNN2D_SHAPE_H] * + pool->shape[CC_CNN2D_SHAPE_W]; + o_ch_mem_size = o_ch_size * cc_dtype_size(*pool->dtype); + for (i = 0; i < inp->shape[CC_CNN2D_SHAPE_C]; ++i) { + _max_pool2d(inp->data + i_ch_mem_size * i, + pool->data + o_ch_mem_size * i, + inp->shape[CC_CNN2D_SHAPE_W], + inp->shape[CC_CNN2D_SHAPE_H], s, + *inp->dtype); + } + return pool; +} diff --git a/src/cc_pool2d.h b/src/cc_pool2d.h new file mode 100644 index 0000000..ac1686c --- /dev/null +++ b/src/cc_pool2d.h @@ -0,0 +1,16 @@ +#ifndef _CC_POOL2D_H_ +#define _CC_POOL2D_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +cc_tensor_t *cc_max_pool2d(cc_tensor_t *inp, cc_int32 s, const char *name); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_POOL2D_H_ */ diff --git a/src/cc_tensor.c b/src/cc_tensor.c new file mode 100644 index 0000000..7c0671a --- /dev/null +++ b/src/cc_tensor.c @@ -0,0 +1,153 @@ +#include +#include +#include + +#include "cc_assert.h" +#include "cc_tensor.h" +#include "cc_tsrmgr.h" + +cc_tensor_t *cc_create_tensor( const cc_int32 *shape, + cc_dtype dtype, const char *name) +{ + const cc_int32 *sptr = shape; + cc_int32 memsize, elems = *shape; + cc_tensor_t *tensor; + while (*++sptr) + elems *= *sptr; + memsize = cc_dtype_size(dtype) * elems; + if (!memsize) + return NULL; + cc_assert_alloc( + tensor = (cc_tensor_t*)malloc(sizeof(cc_tensor_t))); + cc_assert_ptr( + tensor->container = list_new_dynamic(CC_TENSOR_ITEMS)); + cc_assert_alloc( + list_alloc(tensor->container, CC_TENSOR_DATA, memsize)); + tensor->data = (unsigned char*) + list_get_record(tensor->container, CC_TENSOR_DATA); + memset(tensor->data, 0, memsize); + cc_assert_zero( + list_set_record(tensor->container, CC_TENSOR_SHAPE, + shape, (sptr - shape + 1) * sizeof(cc_int32))); + cc_assert_ptr( + tensor->shape = (cc_int32*) + list_get_record(tensor->container, CC_TENSOR_SHAPE)); + cc_assert_zero( + list_set_record(tensor->container, + CC_TENSOR_DTYPE, &dtype, sizeof(cc_dtype))); + cc_assert_ptr( + tensor->dtype = (cc_dtype*) + list_get_record(tensor->container, CC_TENSOR_DTYPE)); + cc_assert_zero(list_set_name(tensor->container, name)); + tensor->name = tensor->container->name; +#ifdef AUTO_TSRMGR + if (name) { + cc_tsrmgr_auto_reg(tensor); + } +#endif + return tensor; +} + +cc_tensor_t *cc_copy_tensor(cc_tensor_t *tensor, const char *name) +{ + cc_tensor_t *copied; + cc_assert_alloc( + copied = (cc_tensor_t*)malloc(sizeof(cc_tensor_t))); + cc_assert_ptr( + copied->container = list_clone(tensor->container)); + cc_assert_ptr( + copied->data = (unsigned char*) + list_get_record(copied->container, CC_TENSOR_DATA)); + cc_assert_ptr( + copied->shape = (cc_int32*) + list_get_record(copied->container, CC_TENSOR_SHAPE)); + cc_assert_ptr( + copied->dtype = (cc_dtype*) + list_get_record(copied->container, CC_TENSOR_DTYPE)); + cc_assert_zero(list_set_name(copied->container, name)); + copied->name = copied->container->name; +#ifdef AUTO_TSRMGR + if (name) { + cc_tsrmgr_auto_reg(copied); + } +#endif + return copied; +} + +cc_tensor_t *cc_load_tensor(const char *filename) +{ + cc_tensor_t *tensor; + cc_assert_alloc( + tensor = (cc_tensor_t*)malloc(sizeof(cc_tensor_t))); + if (!(tensor->container = list_import(filename))) { + free(tensor); + return NULL; + } + cc_assert_ptr( + tensor->data = (unsigned char*) + list_get_record(tensor->container, CC_TENSOR_DATA)); + cc_assert_ptr( + tensor->shape = (cc_int32*) + list_get_record(tensor->container, CC_TENSOR_SHAPE)); + cc_assert_ptr( + tensor->dtype = (cc_dtype*) + list_get_record(tensor->container, CC_TENSOR_DTYPE)); + tensor->name = tensor->container->name; +#ifdef AUTO_TSRMGR + cc_tsrmgr_del(tensor->name); + cc_tsrmgr_auto_reg(tensor); +#endif + return tensor; +} + +cc_tensor_t *cc_load_bin(const char *filename, + const cc_int32 *shape, cc_dtype dtype, const char *name) +{ + FILE *fp; + cc_int32 memsize; + cc_tensor_t *tensor; + cc_assert_ptr(tensor = cc_create_tensor(shape, dtype, name)); + cc_assert_ptr(fp = fopen(filename, "rb")); + memsize = list_get_record_len(tensor->container, CC_TENSOR_DATA); + cc_assert(fread(tensor->data, memsize, 1, fp)); + fclose(fp); + return tensor; +} + +void cc_save_tensor(cc_tensor_t *tensor, const char *filename) +{ + list_export(tensor->container, filename, NULL); +} + +void cc_free_tensor(cc_tensor_t *tensor) +{ + if (tensor) { + list_del(tensor->container); + free(tensor); + } +} + +#define BUFLEN 128 +#define BUFLIM 100 +void cc_print_tensor_property(cc_tensor_t *tensor) +{ + char buf[BUFLEN]; + char *bptr = buf; + const cc_int32 *sptr; + if (!tensor) { + utlog_format(UTLOG_WARN, "invalid tensor\n"); + return; + } + sptr = tensor->shape; + while (*sptr) { + bptr += sprintf(bptr, "%d, ", *sptr++); + if ((bptr - buf) > BUFLIM) { + sprintf(bptr, "..., "); + break; + } + } + buf[strlen(buf) - 2] = '\0'; /* clear ", " */ + utlog_format(UTLOG_INFO, + "tensor: \"%s\", dtype: \"%s\", shape: [%s]\n", + tensor->name, cc_dtype_to_string(*tensor->dtype), buf); +} diff --git a/src/cc_tensor.h b/src/cc_tensor.h new file mode 100644 index 0000000..e35e483 --- /dev/null +++ b/src/cc_tensor.h @@ -0,0 +1,52 @@ +#ifndef _CC_TENSOR_H_ +#define _CC_TENSOR_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "util_list.h" +#include "cc_dtype.h" + +enum cc_tensor_items { + CC_TENSOR_DATA, /* 0 */ + CC_TENSOR_SHAPE, /* 1 */ + CC_TENSOR_DTYPE, /* 2 */ + CC_TENSOR_ITEMS /* 3 */ +}; +/* + * container____ data (0) + * | |__ shape (1) + * | |__ dtype (2) + * |_____ name + */ + +typedef struct { + list_t *container; + const char *name; + unsigned char *data; + const cc_dtype *dtype; + const cc_int32 *shape; +} cc_tensor_t; + +cc_tensor_t *cc_create_tensor(const cc_int32 *shape, + cc_dtype dtype, const char *name); + +cc_tensor_t *cc_copy_tensor(cc_tensor_t *tensor, const char *name); + +cc_tensor_t *cc_load_tensor(const char *filename); + +cc_tensor_t *cc_load_bin(const char *filename, + const cc_int32 *shape, cc_dtype dtype, const char *name); + +void cc_save_tensor(cc_tensor_t *tensor, const char *filename); + +void cc_free_tensor(cc_tensor_t *tensor); + +void cc_print_tensor_property(cc_tensor_t *tensor); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_TENSOR_H_ */ diff --git a/src/cc_tsrmgr.c b/src/cc_tsrmgr.c new file mode 100644 index 0000000..4c67b1e --- /dev/null +++ b/src/cc_tsrmgr.c @@ -0,0 +1,219 @@ +#include +#include +#include + +#include "util_rbt.h" +#include "util_log.h" +#include "cc_assert.h" +#include "cc_tsrmgr.h" + +static int global_counter; +static rbt_t *global_tsrmgr_table; + +struct pair { + char *key; + void *dat; +}; + +static void *getkey(struct rbt_node *node) +{ + struct pair *pair = (struct pair*)node->data; + return pair->key; +} + +static int compare(void *a, void *b) +{ + return strcmp((char*)a, (char*)b); +} + +static void free_pair(struct pair *pair) +{ + /* if (pair) */ + cc_free_tensor((cc_tensor_t*)pair->dat); + free(pair); +} + +static struct pair *mkpair(cc_tensor_t *tensor) +{ + struct pair *pair = (struct pair*)malloc(sizeof(struct pair)); + cc_assert_ptr(pair->dat = tensor); + cc_assert_ptr(pair->key = (char*)tensor->name); + return pair; +} + +void cc_tsrmgr_init(void) +{ + if (global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: already initialized\n"); + return; + } + global_tsrmgr_table = new_rbt(getkey, compare); + return; +} + +void cc_tsrmgr_clear(void) +{ + rbt_iterator *it; + struct pair *pair; + if (!global_tsrmgr_table) return; + it = new_rbt_iterator(global_tsrmgr_table); + while (rbt_iterator_has_next(it)) { + pair = (struct pair*)rbt_iterator_next(it); + free_pair(pair); + } + free_rbt_iterator(it); + free_rbt(global_tsrmgr_table); + global_tsrmgr_table = NULL; + global_counter = 0; + return; +} + +int cc_tsrmgr_status(void) +{ + return global_tsrmgr_table ? 1 : 0; +} + +void cc_tsrmgr_reg(cc_tensor_t *tensor) +{ + struct pair *ptr; +#ifdef ENABLE_CC_ASSERT + cc_assert_ptr(tensor->name); +#endif + if (!global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: not initialized\n"); + return; + } + if (!strlen(tensor->name)) { + utlog_format(UTLOG_ERR, "cc_tsrmgr: illegal tensor name\n"); + return; + } + ptr = (struct pair*) + rbt_insert(global_tsrmgr_table, mkpair(tensor)); + if (ptr) { + utlog_format(UTLOG_ERR, + "cc_tsrmgr: tensor \"%s\" is exist\n", + tensor->name); + if (tensor == ptr->dat) + free(ptr); + else + free_pair(ptr); + } else { + global_counter++; + } +} + +void cc_tsrmgr_replace(cc_tensor_t *tensor) +{ + struct pair *ptr; +#ifdef ENABLE_CC_ASSERT + cc_assert_ptr(tensor->name); +#endif + if (!global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: not initialized\n"); + return; + } + if (!strlen(tensor->name)) { + utlog_format(UTLOG_ERR, "cc_tsrmgr: illegal tensor name\n"); + return; + } + ptr = (struct pair*) + rbt_insert(global_tsrmgr_table, mkpair(tensor)); + if (ptr) { + if (tensor == ptr->dat) + free(ptr); + else + free_pair(ptr); + } else { + global_counter++; + } +} + +void cc_tsrmgr_del(const char *name) +{ + struct pair *ptr; +#ifdef ENABLE_CC_ASSERT + cc_assert_ptr(name); +#endif + if (!global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: not initialized\n"); + return; + } + ptr = (struct pair*) + rbt_delete(global_tsrmgr_table, (void*)name); + if (ptr) { + free_pair(ptr); + global_counter--; + } +} + +cc_tensor_t *cc_tsrmgr_get(const char *name) +{ + struct pair *ptr; +#ifdef ENABLE_CC_ASSERT + cc_assert_ptr(name); +#endif + if (!global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: not initialized\n"); + return NULL; + } + ptr = (struct pair*) + rbt_search(global_tsrmgr_table, (void*)name); + if (ptr) + return (cc_tensor_t*)ptr->dat; + return NULL; +} + +#define BUFLEN 128 +#define BUFLIM 100 +static void _print_tensor_property(cc_tensor_t *tensor) +{ + char buf[BUFLEN]; + char *bptr = buf; + const cc_int32 *sptr; + if (!tensor) { + utlog_format(UTLOG_WARN, "invalid tensor\n"); + return; + } + sptr = tensor->shape; + while (*sptr) { + bptr += sprintf(bptr, "%d, ", *sptr++); + if ((bptr - buf) > BUFLIM) { + sprintf(bptr, "..., "); + break; + } + } + buf[strlen(buf) - 2] = '\0'; /* clear ", " */ + fprintf((FILE*)utlog_get_ostream(), + "dtype: \"%s\", shape: [%s]\n", + cc_dtype_to_string(*tensor->dtype), buf); +} + +#define FMTLEN 32 +static void _print_pair(struct pair *pair) +{ + char fmt[FMTLEN]; + sprintf(fmt, "<0x%%0%dlx>: \"%%s\", ", (int)(sizeof(void*) << 1)); + fprintf((FILE*)utlog_get_ostream(), fmt, pair->dat, + ((cc_tensor_t*)pair->dat)->name); + _print_tensor_property((cc_tensor_t*)pair->dat); +} + +static void _cc_tsrmgr_list_traversal(struct rbt_node *n) +{ + if (n != rbt_nil()) { + _cc_tsrmgr_list_traversal(n->left); + _print_pair((struct pair*)n->data); + _cc_tsrmgr_list_traversal(n->right); + } +} + +void cc_tsrmgr_list(void) +{ + if (!global_tsrmgr_table) { + utlog_format(UTLOG_WARN, "cc_tsrmgr: not initialized\n"); + return; + } + utlog_format(UTLOG_INFO, + "cc_tsrmgr: handling %d tensor(s)\n", global_counter); + _cc_tsrmgr_list_traversal(global_tsrmgr_table->root); +} diff --git a/src/cc_tsrmgr.h b/src/cc_tsrmgr.h new file mode 100644 index 0000000..045f65c --- /dev/null +++ b/src/cc_tsrmgr.h @@ -0,0 +1,33 @@ +#ifndef _CC_TSRMGR_H_ +#define _CC_TSRMGR_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_tensor.h" + +void cc_tsrmgr_init(void); +void cc_tsrmgr_clear(void); + +void cc_tsrmgr_reg(cc_tensor_t *tensor); +void cc_tsrmgr_del(const char *name); +void cc_tsrmgr_replace(cc_tensor_t *tensor); + +int cc_tsrmgr_status(void); + +cc_tensor_t *cc_tsrmgr_get(const char *name); + +void cc_tsrmgr_list(void); + +/* AUTO_TSRMGR */ +#define cc_tsrmgr_auto_reg(tensor) \ + if (!cc_tsrmgr_status()) \ + cc_tsrmgr_init(); \ + cc_tsrmgr_reg(tensor); + +#ifdef __cplusplus + } +#endif + +#endif /* _CC_TSRMGR_H_ */ diff --git a/src/global_fn_cfg.c b/src/global_fn_cfg.c new file mode 100644 index 0000000..0fd14bd --- /dev/null +++ b/src/global_fn_cfg.c @@ -0,0 +1,3 @@ +#include "global_fn_cfg.h" + +void __gfn_check__(void) {return;} diff --git a/src/global_fn_cfg.h b/src/global_fn_cfg.h new file mode 100644 index 0000000..5c52104 --- /dev/null +++ b/src/global_fn_cfg.h @@ -0,0 +1,40 @@ +/* + * Global function configuration + */ + +#ifndef _GLOBAL_FN_CFG_H_ +#define _GLOBAL_FN_CFG_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#include "cc_dtype.h" +#include "cc_cpufn.h" + +void (*_activation_relu)(void *inp, cc_int32 elems, cc_dtype dt) + = cc_cpu_activation_relu; + +void (*_activation_relu6)(void *inp, cc_int32 elems, cc_dtype dt) + = cc_cpu_activation_relu6; + +void (*_activation_softmax)(void *inp, cc_int32 elems, cc_dtype dt) + = cc_cpu_activation_softmax; + +void (*_max_pool2d)(void *inp, void *oup, cc_int32 x, cc_int32 y, + cc_int32 s, cc_dtype dt) = cc_cpu_max_pool2d; + +void (*_conv2d)(void *inp, void *oup, + cc_int32 x, cc_int32 y, cc_int32 oup_x, cc_int32 oup_y, + cc_int32 sx, cc_int32 sy, void *filter, cc_int32 fw, + cc_dtype dt) = cc_cpu_conv2d; + +void (*_fully_connected)(void *inp, void *oup, void *w, void *b, + cc_int32 iw, cc_int32 ow, cc_dtype dt) = cc_cpu_fully_connected; + + +#ifdef __cplusplus + } +#endif + +#endif /* _GLOBAL_FN_CFG_H_ */ diff --git a/src/util_image.c b/src/util_image.c new file mode 100644 index 0000000..50cb540 --- /dev/null +++ b/src/util_image.c @@ -0,0 +1,1401 @@ +/* + * #define CONFIG_STD_C89 + * #define RUN_TIME_ENDIAN_CHECK + * #define BYTE_ORDER_BIG_ENDIAN + * #define USE3RD_STB_IMAGE + */ + +#include +#include +#include +#include + +#include "util_image.h" + +#ifndef byte + #define byte unsigned char +#endif + +#define ASSERT_ON_BUILD(condition)\ + ((void)sizeof(char[1 - 2*!!(condition)])) +/*=========== compile-time test ===========*/ +void ____utim_datatype_test____() +{ + /* + * Ref: utim_write(), utim_write_ctrl() + * At simple bmp operations... + */ + ASSERT_ON_BUILD(sizeof(utim_int32_t) - 4); + /* + * Psf2 font file support + */ + ASSERT_ON_BUILD(sizeof(struct psf2_header) - 32); +} +#ifdef RUN_TIME_ENDIAN_CHECK + static int _is_little_endian() + { + unsigned int i = 1; + unsigned char *c = (unsigned char*)&i; + if (*c) + return 1; + return 0; + } + + static void _memrev(void *i, unsigned int s) + { + unsigned int j = s >> 1; + unsigned char *c = (unsigned char*)i; + for (s--; j > 0; j--) { + *c ^= *(c + s); + *(c + s) ^= *c; + *c ^= *(c + s); + c++, s -= 2; + } + } +#endif + +#ifndef CONFIG_STD_C89 + inline +#endif +static utim_int32_t _index_calc(int xsize, int ysize, int x, int y) +{ + utim_int32_t index; + if (x >= xsize || y >= ysize || x < 0 || y < 0) + return -1; + index = xsize * y + x; + return index; +} + +static const char *file_extension(const char *filename) +{ + const char *p = filename + strlen(filename); + while (filename != p--) + if (*p == '.') + break; + return filename < p ? p + 1 : NULL; +} + +#ifndef USE3RD_STB_IMAGE +/* 8, 24, 32 bit NON-COMPRESSED BMP supported */ +static byte *_bmp_read(const char *filename, + int *xsize, int *ysize, int *channels) +{ + byte header[54]; + byte color_dp; + utim_uint32_t pxl_begin, *x; + utim_uint32_t r, tmp, windows_byte_cnt = 0; + byte *image_tail, *pixels; + FILE *fp; + if (!(fp = fopen(filename, "rb"))) + return NULL; + fread(header, sizeof(byte), 54, fp); + x = (utim_uint32_t*)(header + 18); + *xsize = x[0]; + *ysize = x[1]; + pxl_begin = *(utim_uint32_t*)(header + 10); +#if defined(RUN_TIME_ENDIAN_CHECK) + if (!_is_little_endian) { + _memrev(xsize, sizeof(utim_uint32_t)); + _memrev(ysize, sizeof(utim_uint32_t)); + _memrev(&pxl_begin, sizeof(utim_uint32_t)); + } +#elif defined(BYTE_ORDER_BIG_ENDIAN) + _memrev(xsize, sizeof(utim_uint32_t)); + _memrev(ysize, sizeof(utim_uint32_t)); + _memrev(&pxl_begin, sizeof(utim_uint32_t)); +#endif + color_dp = *(byte*)(header + 28); + switch (color_dp) { + case 8: + case 24: + case 32: + *channels = color_dp >> 3; + break; + default: + fclose(fp); + return NULL; + } + if (!(((*xsize) * (*channels)) % 4)) { + windows_byte_cnt = 0; + } else { + tmp = (*xsize * (*channels)) / 4; + windows_byte_cnt = ((tmp + 1) * 4) - *xsize * (*channels); + } + pixels = (byte*)malloc((*xsize) * (*ysize) * (*channels)); + if (!pixels) { + fclose(fp); + return NULL; + } + image_tail = pixels + ((*ysize) * (*xsize) * (*channels)); + fseek(fp, pxl_begin, SEEK_SET); + for (r = 0; r < (*ysize); ++r) { + image_tail -= ((*xsize) * (*channels)); + fread(image_tail , sizeof(byte), (*xsize) * (*channels), fp); + fseek(fp, windows_byte_cnt, SEEK_CUR); + } + fclose(fp); + return pixels; +} + +static void _write_palette_gray(FILE *fp) +{ + int i; + byte c[4] = {0}; + for (i = 0; i < 256; ++i) { + memset(c, i, 3); + fwrite(c, sizeof(4), 1, fp); + } +} + +static int _bmp_write(const char *filename, + byte *pixels, int xsize, int ysize, int channels) +{ + byte header[54] = { + 0x42, 0x4d, 0, 0, 0, 0, 0, 0, 0, 0, + 54, 0, 0, 0, 40, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 24, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0 + }; + FILE *fp; + byte *image_tail; + utim_uint32_t windows_byte_cnt = 0; + utim_uint32_t windows_byte_set = 0; + utim_uint32_t tmp, r = (xsize * channels) % 4; + utim_uint32_t pxl_begin, *x; + switch (channels) { + case 1: + pxl_begin = 54 + 1024; + break; + case 3: + case 4: + pxl_begin = 54; + break; + default: + return -1; + } + if (!r) { + windows_byte_cnt = 0; + } else { + tmp = (xsize * channels ) / 4; + windows_byte_cnt = ((tmp + 1) * 4) - xsize * channels; + } + x = (utim_uint32_t*)(header + 2); + x[0] = (xsize * channels + windows_byte_cnt) * ysize + 54; + x = (utim_uint32_t*)(header + 18); + x[0] = xsize; + x[1] = ysize; + *(utim_uint32_t*)(header + 10) = pxl_begin; +#if defined(RUN_TIME_ENDIAN_CHECK) + if (!_is_little_endian) { + _memrev(x + 0, sizeof(utim_uint32_t)); + _memrev(x + 1, sizeof(utim_uint32_t)); + _memrev((utim_uint32_t*)(header + 10), + sizeof(utim_uint32_t)); /* pxl_begin */ + } +#elif defined(BYTE_ORDER_BIG_ENDIAN) + _memrev(x + 0, sizeof(utim_uint32_t)); + _memrev(x + 1, sizeof(utim_uint32_t)); + _memrev((utim_uint32_t*)(header + 10), + sizeof(utim_uint32_t)); /* pxl_begin */ +#endif + *(byte*)(header + 28) = channels << 3; + if (!(fp = fopen(filename, "wb+"))) + return -1; + fwrite(header, sizeof(byte), 54, fp); + image_tail = pixels + (ysize * xsize * channels); + if (channels == 1) { /* 8bit palette, support gray image only */ + _write_palette_gray(fp); + } + fseek(fp, pxl_begin, SEEK_SET); + for (r = 0; r < ysize; ++r) { + image_tail -= (xsize * channels); + fwrite(image_tail, sizeof(byte) * channels, xsize, fp); + fwrite(&windows_byte_set, 1, windows_byte_cnt, fp); + } + fclose(fp); + return 0; +} +#endif /* not defined USE3RD_STB_IMAGE */ + +/* + * Use stb image. stb_image.h, stb_image_write.h are included. + */ +#ifdef USE3RD_STB_IMAGE + #define STB_IMAGE_IMPLEMENTATION + #include "stb_image.h" + #define STB_IMAGE_WRITE_IMPLEMENTATION + #include "stb_image_write.h" + #define STB_IMAGE_RESIZE_IMPLEMENTATION + #include "stb_image_resize.h" + int jpg_default_quality = 80; + int png_default_compress = 8; +#endif + +utim_image_t *utim_read(const char*filename) +{ +#ifndef USE3RD_STB_IMAGE + const char *t; +#endif + utim_image_t *img = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!img) + return NULL; +#ifdef USE3RD_STB_IMAGE + img->pixels = stbi_load(filename, &img->xsize, &img->ysize, + &img->channels, 0); + if (!img->pixels) { + free(img); + return NULL; + } +#else /* Support 24bit bmp image only */ + t = file_extension(filename); + if (strcmp(t, "bmp")) { + free(img); + return NULL; + } + img->pixels = _bmp_read(filename, + &img->xsize, &img->ysize, &img->channels); + if (!img->pixels) { + free(img); + return NULL; + } +#endif + return img; +} + +int utim_write(const char *filename, utim_image_t *img) +{ + return utim_write_ctrl(filename, img, 0, 0); +} + +int utim_write_ctrl(const char *filename, + utim_image_t *img, int comp, int quality) +{ + const char *t; + char tbuf[8]; + char *p = tbuf; +#ifdef USE3RD_STB_IMAGE + int state; +#endif + t = file_extension(filename); + strcpy(tbuf, t); + while ((*p = tolower(*p))) ++p; + if (strlen(t) < 3) + return -1; /* Not supported file format */ +#if defined(RUN_TIME_ENDIAN_CHECK) + if (!_is_little_endian()) + _memrev(tbuf, strlen(tbuf) + 1); +#elif defined(BYTE_ORDER_BIG_ENDIAN) + _memrev(tbuf, strlen(tbuf) + 1); +#endif +#ifdef USE3RD_STB_IMAGE + switch (*(utim_int32_t*)tbuf) { + case UTIM_BMP: + if (!comp) + comp = img->channels; + state = stbi_write_bmp(filename, img->xsize, + img->ysize, comp, img->pixels); + if (state != 1) + return state; + return 0; + case UTIM_PNG: + if (!comp) + comp = img->channels; + quality = quality ? quality : png_default_compress; + stbi_write_png_compression_level = quality; + state = stbi_write_png(filename, img->xsize, + img->ysize, img->channels, img->pixels, + img->xsize * img->channels); + if (state != 1) + return state; + return 0; + case UTIM_JPG: + if (!comp) + comp = img->channels; + quality = quality ? quality : jpg_default_quality; + state = stbi_write_jpg(filename, img->xsize, + img->ysize, comp, img->pixels, quality); + if (state != 1) + return state; + return 0; + case UTIM_TGA: + if (!comp) + comp = img->channels; + state = stbi_write_tga(filename, img->xsize, + img->ysize, comp, img->pixels); + if (state != 1) + return state; + return 0; + default: + return -1; /* Not supported file format */ + } +#else /* Support 8, 24, 32 bit bmp image only */ + if (strcmp(tbuf, "bmp")) + return -1; /* Not supported file format */ + switch (img->channels) { + case 1: + case 3: + case 4: + break; + default: + return -1; + } + return _bmp_write(filename, + img->pixels, img->xsize, img->ysize, img->channels); +#endif +} + +void utim_free_image(utim_image_t *img) +{ + if (!img) + return; + if (img->pixels) + free(img->pixels); + free(img); +} + +utim_image_t *utim_clone(utim_image_t *img) +{ + utim_image_t *clone; + int size = img->xsize * img->ysize * img->channels; + if (!size) + return NULL; + clone = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!clone) + return NULL; + memcpy(clone, img, sizeof(utim_image_t)); + clone->pixels = (byte*)malloc(size); + if (!clone->pixels) { + free(clone); + return NULL; + } + memcpy(clone->pixels, img->pixels, size); + return clone; +} + +static byte *_image_resize_nearest(utim_image_t *img, int x, int y) +{ + byte *p; + int i, j, ch_size = x * y; + p = (byte*)malloc(ch_size * img->channels); + if (!p) + return NULL; + for (j = 0; j < y; ++j) { + for (i = 0; i < x; ++i) { + memcpy(p + _index_calc(x, y, i, j) * img->channels, + img->pixels + _index_calc + (img->xsize, img->ysize, + i * img->xsize / x, j * img->ysize / y) * + img->channels,img->channels); + } + } + return p; +} + +#ifdef USE3RD_STB_IMAGE +static byte *_image_resize_linear(utim_image_t *img, int x, int y) +{ + byte *p; + int ch_size = x * y; + p = (byte*)malloc(ch_size * img->channels); + if (!p) + return NULL; + stbir_resize_uint8(img->pixels, + img->xsize, img->ysize, 0, p, x, y, 0, img->channels); + return p; +} +#endif + +utim_image_t *utim_resize(utim_image_t *img, int x, int y, int mode) +{ + utim_image_t *resize; + resize = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!resize) + return NULL; + switch (mode) { + case UTIM_RESIZE_NEAREST: + resize->pixels = _image_resize_nearest(img, x, y); + if (!resize->pixels) { + free(resize); + return NULL; + } + break; +#ifdef USE3RD_STB_IMAGE + case UTIM_RESIZE_LINEAR: + resize->pixels = _image_resize_linear(img, x, y); + if (!resize->pixels) { + free(resize); + return NULL; + } + break; +#endif + default: + return NULL; + } + resize->xsize = x; + resize->ysize = y; + resize->channels = img->channels; + return resize; +} + +int utim_swap_chl(utim_image_t *img, int a, int b) +{ + byte c; + int i, ch_size; + if (a == b || a >= img->channels || b >= img->channels) + return UTIM_ERR_BAD_ARG; + ch_size = img->xsize * img->ysize; + for (i = 0; i < ch_size; ++i) { + c = *(img->pixels + img->channels * i + a); + *(img->pixels + img->channels * i + a) = + *(img->pixels + img->channels * i + b); + *(img->pixels + img->channels * i + b) = c; + } + return 0; +} + +int utim_img2rgb(utim_image_t *img) +{ + switch (img->channels) { + case 1: + return utim_gray2rgb(img); + case 3: + return 0; + case 4: + return utim_rgba2rgb(img); + default: + return UTIM_ERR_BAD_ARG; + } +} + +int utim_img2gray(utim_image_t *img) +{ + switch (img->channels) { + case 1: + return 0; + case 3: + return utim_rgb2gray(img); + case 4: + return utim_rgba2gray(img); + default: + return UTIM_ERR_BAD_ARG; + } +} + +int utim_img2rgba(utim_image_t *img) +{ + switch (img->channels) { + case 1: + return utim_gray2rgba(img); + case 3: + return utim_rgb2rgba(img); + case 4: + return 0; + default: + return UTIM_ERR_BAD_ARG; + } +} + +int utim_rgba2rgb(utim_image_t *rgba) +{ + byte *new_pixels, *p; + int i, ch_size = rgba->xsize * rgba->ysize; + if (!ch_size || rgba->channels != 4) + return UTIM_ERR_BAD_ARG; + new_pixels = (byte*)malloc(ch_size * 3); + if (!new_pixels) + return UTIM_ERR_ALLOC; + p = new_pixels; + for (i = 0; i < ch_size; ++i) { + *p++ = *(rgba->pixels + i * 4); + *p++ = *(rgba->pixels + i * 4 + 1); + *p++ = *(rgba->pixels + i * 4 + 2); + } + free(rgba->pixels); + rgba->pixels = new_pixels; + rgba->channels = 3; + return 0; +} + +utim_image_t *utim_rgb_by_rgba(utim_image_t *rgba) +{ + utim_image_t *rgb; + byte *p; + int i, ch_size = rgba->xsize * rgba->ysize; + if (!ch_size || rgba->channels != 4) + return NULL; + rgb = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!rgb) + return NULL; + rgb->pixels = (byte*)malloc(ch_size * 3); + if (!rgb->pixels) { + free(rgb); + return NULL; + } + p = rgb->pixels; + for (i = 0; i < ch_size; ++i) { + *p++ = *(rgba->pixels + i * 4); + *p++ = *(rgba->pixels + i * 4 + 1); + *p++ = *(rgba->pixels + i * 4 + 2); + } + rgb->xsize = rgba->xsize; + rgb->ysize = rgba->ysize; + rgb->channels = 3; + return rgb; +} + +int utim_rgb2bgr(utim_image_t *rgb) +{ + int n; + byte *p1, *p2; + if (rgb->channels != 3) + return UTIM_ERR_BAD_ARG; + n = rgb->xsize * rgb->ysize; + p1 = rgb->pixels; + p2 = rgb->pixels + 2; + while (n--) { + *(p1) ^= *(p2); + *(p2) ^= *(p1); + *(p1) ^= *(p2); + p1 += 3; + p2 += 3; + } + return 0; +} + +utim_image_t *utim_bgr_by_rgb(utim_image_t *rgb) +{ + utim_image_t *bgr = utim_clone(rgb); + if (!bgr) + return NULL; + if (utim_rgb2bgr(bgr)) { + utim_free_image(bgr); + return NULL; + } + return bgr; +} + +int utim_rgb2gray(utim_image_t *rgb) +{ + byte *new_pixels, *p; + int i, sum, ch_size; + if (rgb->channels != 3) + return UTIM_ERR_BAD_ARG; + ch_size = rgb->xsize * rgb->ysize; + new_pixels = (byte*)malloc(ch_size); + if (!new_pixels) + return UTIM_ERR_ALLOC; + p = new_pixels; + for (i = 0; i < ch_size; ++i) { + sum = 0; + sum += *(rgb->pixels + i * 3); + sum += *(rgb->pixels + i * 3 + 1); + sum += *(rgb->pixels + i * 3 + 2); + *p++ = sum / 3; + } + free(rgb->pixels); + rgb->pixels = new_pixels; + rgb->channels = 1; + return 0; +} + +int utim_rgba2gray(utim_image_t *rgba) +{ + byte *new_pixels, *p; + int i, sum, ch_size; + if (rgba->channels != 4) + return UTIM_ERR_BAD_ARG; + ch_size = rgba->xsize * rgba->ysize; + new_pixels = (byte*)malloc(ch_size); + if (!new_pixels) + return UTIM_ERR_ALLOC; + p = new_pixels; + for (i = 0; i < ch_size; ++i) { + sum = 0; + sum += *(rgba->pixels + i * 4); + sum += *(rgba->pixels + i * 4 + 1); + sum += *(rgba->pixels + i * 4 + 2); + *p++ = sum / 3; + } + free(rgba->pixels); + rgba->pixels = new_pixels; + rgba->channels = 1; + return 0; +} + +utim_image_t *utim_gray_by_rgb(utim_image_t *rgb) +{ + byte *p; + utim_image_t *gray; + int i, sum, ch_size; + if (rgb->channels != 3) + return NULL; + gray = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!gray) + return NULL; + ch_size = rgb->xsize * rgb->ysize; + gray->pixels = (byte*)malloc(ch_size); + if (!gray->pixels) { + free(gray); + return NULL; + } + p = gray->pixels; + for (i = 0; i < ch_size; ++i) { + sum = 0; + sum += *(rgb->pixels + i * 3); + sum += *(rgb->pixels + i * 3 + 1); + sum += *(rgb->pixels + i * 3 + 2); + *p++ = sum / 3; + } + gray->xsize = rgb->xsize; + gray->ysize = rgb->ysize; + gray->channels = 1; + return gray; +} + +int utim_gray2rgb(utim_image_t *gray) +{ + byte *new_pixels; + int i, off, ch_size; + if (gray->channels != 1) + return UTIM_ERR_BAD_ARG; + ch_size = gray->xsize * gray->ysize; + new_pixels = (byte*)malloc(ch_size * 3); + if (!new_pixels) + return UTIM_ERR_ALLOC; + for (i = 0; i < ch_size; ++i) { + off = i * 3; + *(new_pixels + off) = *(gray->pixels + i); + *(new_pixels + off + 1) = *(gray->pixels + i); + *(new_pixels + off + 2) = *(gray->pixels + i); + } + free(gray->pixels); + gray->pixels = new_pixels; + gray->channels = 3; + return 0; +} + +int utim_gray2rgba(utim_image_t *gray) +{ + byte *new_pixels; + int i, off, ch_size; + if (gray->channels != 1) + return UTIM_ERR_BAD_ARG; + ch_size = gray->xsize * gray->ysize; + new_pixels = (byte*)malloc(ch_size * 4); + if (!new_pixels) + return UTIM_ERR_ALLOC; + for (i = 0; i < ch_size; ++i) { + off = i * 4; + *(new_pixels + off) = *(gray->pixels + i); + *(new_pixels + off + 1) = *(gray->pixels + i); + *(new_pixels + off + 2) = *(gray->pixels + i); + *(new_pixels + off + 3) = 255; + } + free(gray->pixels); + gray->pixels = new_pixels; + gray->channels = 4; + return 0; +} + +int utim_rgb2rgba(utim_image_t *rgb) +{ + byte *new_pixels; + int i, off_rgba, off_rgb, ch_size; + if (rgb->channels != 3) + return UTIM_ERR_BAD_ARG; + ch_size = rgb->xsize * rgb->ysize; + new_pixels = (byte*)malloc(ch_size * 4); + if (!new_pixels) + return UTIM_ERR_ALLOC; + for (i = 0; i < ch_size; ++i) { + off_rgb = i * 3; + off_rgba = i * 4; + *(new_pixels + off_rgba) = *(rgb->pixels + off_rgb); + *(new_pixels + off_rgba + 1) = *(rgb->pixels + off_rgb + 1); + *(new_pixels + off_rgba + 2) = *(rgb->pixels + off_rgb + 2); + *(new_pixels + off_rgba + 3) = 255; + } + free(rgb->pixels); + rgb->pixels = new_pixels; + rgb->channels = 4; + return 0; +} + +utim_image_t *utim_rgb_by_gray(utim_image_t *gray) +{ + utim_image_t *rgb; + int i, off, ch_size; + if (gray->channels != 1) + return NULL; + rgb = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!rgb) + return NULL; + ch_size = gray->xsize * gray->ysize; + rgb->pixels = (byte*)malloc(ch_size * 3); + if (!rgb->pixels) { + free(rgb); + return NULL; + } + for (i = 0; i < ch_size; ++i) { + off = i * 3; + *(rgb->pixels + off + 0) = *(gray->pixels + i); + *(rgb->pixels + off + 1) = *(gray->pixels + i); + *(rgb->pixels + off + 2) = *(gray->pixels + i); + } + rgb->xsize = gray->xsize; + rgb->ysize = gray->ysize; + rgb->channels = 3; + return rgb; +} + +utim_image_t *utim_create(int x, int y, int nch, int c) +{ + utim_image_t *img; + int size = x * y * nch; + if (!size || nch > UTIM_MAX_CHANNELS) + return NULL; + img = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!img) + return NULL; + img->pixels = (byte*)malloc(size); + if (!img->pixels) { + free(img); + return NULL; + } + memset(img->pixels, c, size); + img->xsize = x; + img->ysize = y; + img->channels = nch; + return img; +} + +utim_image_t *utim_stack(utim_image_t **chx, int nch) +{ + byte *p; + utim_image_t *img; + int i, j, ch_size = (*chx)->xsize * (*chx)->ysize; + if (nch > UTIM_MAX_CHANNELS || nch <= 0) + return NULL; + for (i = 1; i < nch; ++i) { /* size check */ + if (chx[0]->xsize != chx[i]->xsize || + chx[0]->ysize != chx[i]->ysize) + return NULL; + } + img = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!img) + return NULL; + img->pixels = (byte*)malloc(ch_size * nch); + if (!img->pixels) { + free(img); + return NULL; + } + p = img->pixels; + for (i = 0; i < ch_size; ++i) { + for (j = 0; j < nch; ++j) + *(p + i * nch + j) = chx[j]->pixels[i]; + } + img->xsize = chx[0]->xsize; + img->ysize = chx[0]->ysize; + img->channels = nch; + return img; +} + +utim_image_t *utim_pick_chl(utim_image_t *img, int ich) +{ + byte *p; + utim_image_t *ch; + int i, ch_size; + if (!(ich < img->channels)) + return NULL; + ch = (utim_image_t*)malloc(sizeof(utim_image_t)); + if (!ch) + return NULL; + ch_size = img->xsize * img->ysize; + ch->pixels = (byte*)malloc(ch_size); + if (!ch->pixels) { + free(ch); + return NULL; + } + p = ch->pixels; + for (i = 0; i < ch_size; ++i) + *p++ = *(img->pixels + i * img->channels + ich); + ch->xsize = img->xsize; + ch->ysize = img->ysize; + ch->channels = 1; + return ch; +} + +int utim_set_opacity(utim_image_t *img, int opacity) +{ + byte *c; + int i, sum, ch_size; + switch (img->channels) { + case 1: + if (!utim_gray2rgba(img)) + return UTIM_ERR_BAD_ARG; + case 3: + if (!utim_rgb2rgba(img)) + return UTIM_ERR_BAD_ARG; + case 4: /* RGBA */ + break; + default: + return UTIM_ERR_BAD_ARG; + } + ch_size = img->xsize * img->ysize; + c = img->pixels; + for (i = 0; i < ch_size; ++i) { + sum = 0; + sum += *c++; + sum += *c++; + sum += *c++; + if (sum) + *c = opacity; + c++; + } + return 0; +} + +utim_image_t *utim_set_chl(utim_image_t *img, int ich, int color) +{ + byte *c; + int i, ch_size; + if (ich >= img->channels) + return NULL; + ch_size = img->xsize * img->ysize; + c = img->pixels; + for (i = 0; i < ch_size; ++i) { + *(c + ich) = color; + c += img->channels; + } + return img; +} + +void utim_negative_color(utim_image_t *img) +{ + int i, j, ch_size = img->xsize * img->ysize; + for (i = 0; i < ch_size; ++i) { + for (j = 0; j < img->channels && j < UTIM_COLOR_A; ++j) { + img->pixels[i * img->channels + j] = + 255 - img->pixels[i * img->channels + j]; + } + } +} + +/* _compose_pixel: used for implementing utim_superpose only */ +#ifndef CONFIG_STD_C89 + inline +#endif +static void _compose_pixel(utim_image_t *bg, + utim_image_t *im, int i, int j, utim_point_t p) +{ + utim_color_t c_bg, c_im; + int k, index_bg, index_im; + byte *rb = &c_bg[UTIM_COLOR_R]; + byte *gb = &c_bg[UTIM_COLOR_G]; + byte *bb = &c_bg[UTIM_COLOR_B]; + byte *ab = &c_bg[UTIM_COLOR_A]; + byte *ri = &c_im[UTIM_COLOR_R]; + byte *gi = &c_im[UTIM_COLOR_G]; + byte *bi = &c_im[UTIM_COLOR_B]; + byte *ai = &c_im[UTIM_COLOR_A]; + index_bg = _index_calc(bg->xsize, bg->ysize, i, j); + index_im = _index_calc(im->xsize, im->ysize, + i - p[UTIM_POINT_X], j - p[UTIM_POINT_Y]); + if (index_im <0) + return; + if (bg->channels == 4) { /* RGBA */ + for (k = 0; k < bg->channels; ++k) + c_bg[k] = bg->pixels[index_bg * bg->channels + k]; + memset(c_im, UTIM_BACK_COLOR, sizeof(c_im)); + for (k = 0; k < im->channels; ++k) + c_im[k] = im->pixels[index_im * bg->channels + k]; + *rb = ((255 - *ai) * (*rb) + (*ai) * (*ri)) / 255; + *gb = ((255 - *ai) * (*gb) + (*ai) * (*gi)) / 255; + *bb = ((255 - *ai) * (*bb) + (*ai) * (*bi)) / 255; + *ab = *ab + ((255 - *ab) * (*ai) / 255); + for (k = 0; k < bg->channels; ++k) + bg->pixels[index_bg * bg->channels + k] = c_bg[k]; + + } else { /* Direct copy */ + for (k = 0; k < bg->channels; ++k) + bg->pixels[index_bg + k] = im->pixels[index_im + k]; + } +} + +int utim_superpose(utim_image_t *bg, + utim_image_t *img, utim_point_t p) +{ + int i, j; + if (p[UTIM_POINT_X] >= bg->xsize || + p[UTIM_POINT_Y] >= bg->ysize) + return 0; + for (j = p[UTIM_POINT_Y]; j < bg->ysize; ++j) { + for (i = p[UTIM_POINT_X]; i < bg->xsize; ++i) + _compose_pixel(bg, img, i, j, p); + } + return 0; +} + +/* + * Simple drawing functions + */ + +static int (*utim_draw_point_fn) + (utim_image_t*, utim_point_t, utim_color_t) = utim_draw_point; + +int utim_set_pixel(utim_image_t *img, utim_point_t p, utim_color_t c) +{ + int i, index = _index_calc(img->xsize, + img->ysize, p[UTIM_POINT_X], p[UTIM_POINT_Y]); + if (index < 0) + return -1; + for (i = 0; i < img->channels; ++i) + img->pixels[index * img->channels + i] = c[i]; + return 0; +} + +int utim_draw_point(utim_image_t *img, utim_point_t p, utim_color_t c) +{ + utim_color_t c_img; + byte *ri = &c[UTIM_COLOR_R]; + byte *gi = &c[UTIM_COLOR_G]; + byte *bi = &c[UTIM_COLOR_B]; + byte *ai = &c[UTIM_COLOR_A]; + byte *rb = &c_img[UTIM_COLOR_R]; + byte *gb = &c_img[UTIM_COLOR_G]; + byte *bb = &c_img[UTIM_COLOR_B]; + byte *ab = &c_img[UTIM_COLOR_A]; + int k, index = _index_calc(img->xsize, + img->ysize, p[UTIM_POINT_X], p[UTIM_POINT_Y]); + if (index < 0) + return -1; + if (img->channels == 4) { /* RGBA */ + for (k = 0; k < img->channels; ++k) + c_img[k] = img->pixels[index * img->channels + k]; + *rb = ((255 - *ai) * (*rb) + (*ai) * (*ri)) / 255; + *gb = ((255 - *ai) * (*gb) + (*ai) * (*gi)) / 255; + *bb = ((255 - *ai) * (*bb) + (*ai) * (*bi)) / 255; + *ab = *ab + ((255 - *ab) * (*ai) / 255); + for (k = 0; k < img->channels; ++k) + img->pixels[index * img->channels + k] = c_img[k]; + } else { /* Direct copy */ + for (k = 0; k < img->channels; ++k) + img->pixels[index + k] = c[k]; + } + return 0; +} + +void utim_set_draw_point_fn(int (*fn) + (utim_image_t*, utim_point_t, utim_color_t)) +{ + utim_draw_point_fn = fn; + return; +} + +static void _image_draw_rect_w1(utim_image_t *img, + utim_point_t a, int w, int h, utim_color_t c) +{ + int i; + utim_point_t p; + if (!w || !h) + return; + p[UTIM_POINT_X] = a[UTIM_POINT_X]; + p[UTIM_POINT_Y] = a[UTIM_POINT_Y]; + for (i = 0; i < w; ++i) { + utim_draw_point_fn(img, p, c); + p[UTIM_POINT_X]++; + } + for (i = 0; i < h; ++i) { + utim_draw_point_fn(img, p, c); + p[UTIM_POINT_Y]++; + } + for (i = 0; i < w; ++i) { + utim_draw_point_fn(img, p, c); + p[UTIM_POINT_X]--; + } + for (i = 0; i < h; ++i) { + utim_draw_point_fn(img, p, c); + p[UTIM_POINT_Y]--; + } + return; +} + +void utim_draw_rect(utim_image_t *img, + utim_point_t a, int w, int h, utim_color_t c, int width) +{ + int i, half; + utim_point_t p; + if (!w || !h || !width) + return; + half = width >> 1; + w -= width; h -= width; + p[UTIM_POINT_X] = a[UTIM_POINT_X] + half; + p[UTIM_POINT_Y] = a[UTIM_POINT_Y] + half; + for (i = 0; i < width; ++i) { + _image_draw_rect_w1(img, p, w, h, c); + w += 2; h += 2; + p[UTIM_POINT_X]--; + p[UTIM_POINT_Y]--; + } + return; +} + +void utim_draw_line(utim_image_t *img, + utim_point_t a, utim_point_t b, utim_color_t c, int width) +{ + int dx, dy, orientation, xa, xb, ya, yb, t; + utim_point_t p; + if (!width) + return; + width = width > 0 ? width : -width; + dx = b[UTIM_POINT_X] - a[UTIM_POINT_X]; + dx = dx < 0 ? -dx : dx; + dy = b[UTIM_POINT_Y] - a[UTIM_POINT_Y]; + dy = dy < 0 ? -dy : dy; + if ((dx == 0) && (dy == 0)) { + utim_draw_point_fn(img, a, c); + return; + } + orientation = dx < dy ? 0 : 1; + xa = a[UTIM_POINT_X]; ya = a[UTIM_POINT_Y]; + xb = b[UTIM_POINT_X]; yb = b[UTIM_POINT_Y]; + if (orientation) { + if (a[UTIM_POINT_X] > b[UTIM_POINT_X]) { + t = xa; xa = xb; xb = t; + t = ya; ya = yb; yb = t; + } + for (p[UTIM_POINT_X] = xa; + p[UTIM_POINT_X] <= xb; ++p[UTIM_POINT_X]) { + p[UTIM_POINT_Y] = + ya + (yb - ya) * (p[UTIM_POINT_X] - xa) / dx; + if (width == 1) + utim_draw_point_fn(img, p, c); + else + utim_draw_filled_circle(img, p, width, c); + } + } else { + if (a[UTIM_POINT_Y] > b[UTIM_POINT_Y]) { + t = xa; xa = xb; xb = t; + t = ya; ya = yb; yb = t; + } + for (p[UTIM_POINT_Y] = ya; + p[UTIM_POINT_Y] <= yb; ++p[UTIM_POINT_Y]) { + p[UTIM_POINT_X] = + xa + (xb - xa) * (p[UTIM_POINT_Y] - ya) / dy; + if (width == 1) + utim_draw_point_fn(img, p, c); + else + utim_draw_filled_circle(img, p, width, c); + } + } +} + +#ifndef CONFIG_STD_C89 + inline +#endif +static void _draw_circle_point(utim_image_t *img, utim_point_t p, + utim_point_t center, int x, int y, utim_color_t color, int width) +{ + p[UTIM_POINT_X] += x; + p[UTIM_POINT_Y] += y; + if (width == 1) + utim_draw_point_fn(img, p, color); + else + utim_draw_filled_circle(img, p, width, color); + p[UTIM_POINT_X] = center[UTIM_POINT_X]; + p[UTIM_POINT_Y] = center[UTIM_POINT_Y]; +} +void utim_draw_circle(utim_image_t *img, + utim_point_t center, int radius, utim_color_t color, int width) +{ + utim_point_t p; + int x, y, dx, dy, err; + if (!width) + return; + x = radius - 1; y = 0; + dx = 1; dy = 1; err = dx - (radius << 1); + p[UTIM_POINT_X] = center[UTIM_POINT_X]; + p[UTIM_POINT_Y] = center[UTIM_POINT_Y]; + width = width > 0 ? width : -width; + while(x >= y) { + _draw_circle_point(img, p, center, x, y, color, width); + _draw_circle_point(img, p, center, y, x, color, width); + _draw_circle_point(img, p, center, -y, x, color, width); + _draw_circle_point(img, p, center, -x, y, color, width); + _draw_circle_point(img, p, center, -x, -y, color, width); + _draw_circle_point(img, p, center, -y, -x, color, width); + _draw_circle_point(img, p, center, y, -x, color, width); + _draw_circle_point(img, p, center, x, -y, color, width); + if (err <= 0) { + y++; + err += dy; + dy += 2; + } + if (err > 0) { + x--; + dx += 2; + err += dx - (radius << 1); + } + } +} + +void utim_draw_filled_circle(utim_image_t *img, + utim_point_t center, int radius, utim_color_t color) +{ + int x, y, r2; + utim_point_t p, bounds[2]; + if (radius <= 0) return; + bounds[0][UTIM_POINT_X] = center[UTIM_POINT_X] - radius; + bounds[1][UTIM_POINT_X] = center[UTIM_POINT_X] + radius; + bounds[0][UTIM_POINT_Y] = center[UTIM_POINT_Y] - radius; + bounds[1][UTIM_POINT_Y] = center[UTIM_POINT_Y] + radius; + + for (x = bounds[0][UTIM_POINT_X]; + x <= bounds[1][UTIM_POINT_X]; ++x) { + for (y = bounds[0][UTIM_POINT_Y]; + y <= bounds[1][UTIM_POINT_Y]; ++y) { + p[UTIM_POINT_X] = x; + p[UTIM_POINT_Y] = y; + r2 = radius * radius; + if(((p[UTIM_POINT_X] - center[UTIM_POINT_X]) * + (p[UTIM_POINT_X] - center[UTIM_POINT_X]) + + (p[UTIM_POINT_Y] - center[UTIM_POINT_Y]) * + (p[UTIM_POINT_Y] - center[UTIM_POINT_Y])) < r2) + utim_draw_point_fn(img, p, color); + } + } +} + +/* + * PSF2 fonts + * View https://www.win.tue.nl/~aeb/linux/kbd/font-formats-1.html + */ +#define PSF2_MAGIC0 0x72 +#define PSF2_MAGIC1 0xb5 +#define PSF2_MAGIC2 0x4a +#define PSF2_MAGIC3 0x86 + +/* bits used in flags */ +#define PSF2_HAS_UNICODE_TABLE 0x01 + +/* max version recognized so far */ +#define PSF2_MAXVERSION 0 + +/* UTF8 separators */ +#define PSF2_SEPARATOR 0xFF +#define PSF2_STARTSEQ 0xFE + +static int _chk_psf2_header(struct psf2_header header) +{ + if (header.magic[0]!=PSF2_MAGIC0) return 0; + if (header.magic[1]!=PSF2_MAGIC1) return 0; + if (header.magic[2]!=PSF2_MAGIC2) return 0; + if (header.magic[3]!=PSF2_MAGIC3) return 0; + if (header.version > PSF2_MAXVERSION) return 0; + return 1; +} + +static unsigned int _read_utf8_value(byte *character, int *nb) +{ + int i; + unsigned int first, scalar; + *nb = 1; + first = character[0]; + scalar = first; + if (first <= 127) + return first; + if (first >> 5 == 6) + *nb = 2; + else if (first >> 4 == 14) + *nb = 3; + else if (first >> 3 == 30) + *nb = 4; + for (i = 1; i < *nb; ++i) + scalar = scalar * 256 + character[i]; + return scalar; +} + +static void _init_utf8_table(unsigned int *utf8index, + int nb_glyphs, unsigned char *data, int data_len) +{ + unsigned int v; + int b, i = 0, pos = 0; + while (pos < data_len && i < nb_glyphs) { + v = _read_utf8_value(data + pos, &b); + utf8index[i] = v; + if (v != 255) + pos += b; + while (pos < data_len && data[pos] != PSF2_SEPARATOR) + pos++; + pos++; i++; + } +} + +utim_font_t *utim_load_font(const char *filename) +{ + FILE *fp; + byte *utf_data; + utim_font_t *font; + int dsize, state, pos, end; + fp = fopen(filename, "rb"); + if (!fp) + return NULL; + font = (utim_font_t*)malloc(sizeof(utim_font_t)); + if (!font) { + fclose(fp); + return NULL; + } + font->utf8index = NULL; + /* _ARRT_ */ + state = fread(&(font->header), sizeof(struct psf2_header), 1, fp); + /* CHK */ + if ((state < 1) || (!_chk_psf2_header(font->header))) { + free(font); + fclose(fp); + return NULL; + } + dsize = font->header.length * font->header.charsize; + font->glyphs = (byte*)malloc(dsize); + if (!font->glyphs) { + free(font); + fclose(fp); + return NULL; + } + state = fread(font->glyphs, dsize, 1, fp); + if (state < 1) { + free(font->glyphs); + free(font); + fclose(fp); + return NULL; + } + if (font->header.flags & PSF2_HAS_UNICODE_TABLE) { + font->utf8index = (unsigned int*)malloc( + font->header.length * sizeof(unsigned int)); + if (!font->utf8index) { + free(font->glyphs); + free(font); + fclose(fp); + return NULL; + } + pos = ftell(fp); + fseek(fp, 0, SEEK_END); + end = ftell(fp); + fseek(fp, pos, SEEK_SET); + utf_data = (byte*)malloc(end - pos); + if (!utf_data) { + free(font->utf8index); + free(font->glyphs); + free(font); + fclose(fp); + return NULL; + } + state = fread(utf_data, end - pos, 1, fp); + if (!state) { + font->header.flags = font->header.flags & + (0xffffffff - PSF2_HAS_UNICODE_TABLE); + free(font->utf8index); + font->utf8index = NULL; + free(utf_data); + fclose(fp); + return font; + } + _init_utf8_table(font->utf8index, + font->header.length, utf_data, end - pos); + free(utf_data); + } + fclose(fp); + return font; +} + +void utim_free_font(utim_font_t *font) +{ + if (!font) + return; + if (font->glyphs) + free(font->glyphs); + if (font->utf8index) + free(font->utf8index); + free(font); +} + +byte *get_character(utim_font_t *font, unsigned int number) +{ + if (number >= font->header.length) + return NULL; + if (!font->glyphs) + return NULL; + if (number > 0) + return(font->glyphs + (font->header.charsize) * number); + /* If not found, use the default glyph */ + return(font->glyphs); +} + +static int _glyph_index(utim_font_t *font, unsigned int v) +{ + unsigned int i; + if (!(font->header.flags & PSF2_HAS_UNICODE_TABLE)) { + /* Only 7-bits ASCII is supported */ + if (v <= 32 || v > 127) { + /* use a default glyph */ + return -1; + } + return v; + } + for (i = 0; i < font->header.length; ++i) { + if (font->utf8index[i] == v) { + return i; + } + } + return -1; +} + +byte *get_glyph(utim_font_t *font, char *character, int *nb){ + return get_character(font, + _glyph_index(font, _read_utf8_value((byte*)character, nb))); +} + +#define UTIM_TEXT_DRAW_GLYPH_PROC \ +for (j = 0; j < charlen; ++j) { \ + for (k = 7; k > 0; --k) { \ + t = ((1 << k) &charp[j]); \ + if (t) { \ + point[UTIM_POINT_X] = \ + (pos * font->header.width) + (7 - k); \ + point[UTIM_POINT_Y] = j; \ + utim_draw_point_fn(img, point, color); \ + } \ + } \ +} + +#define UTIM_TEXT_DEFAULT_GLYPH_PROC \ +for (j = 0; j < charlen; ++j) { \ + for (k = 7; k > 0; --k) { \ + point[UTIM_POINT_X] = \ + (pos * font->header.width) + (7 - k); \ + point[UTIM_POINT_Y] = j; \ + utim_draw_point_fn(img, point, color); \ + } \ +} + +utim_image_t *utim_text(utim_font_t *font, char *text, utim_color_t color) +{ + byte *charp; + utim_image_t *img; + utim_point_t point; + int i, j, k, t, l, charlen, nb, pos = 0; + if (!font || !text) + return NULL; + l = strlen(text); + img = utim_create(font->header.width * l, font->header.height, 4, 0); + for (i = 0; i < l; i = i + nb) { + charlen = font->header.charsize; + charp = get_glyph(font, text + i, &nb); + if (charp) { + UTIM_TEXT_DRAW_GLYPH_PROC; + } else { /* failled getting the character */ + UTIM_TEXT_DEFAULT_GLYPH_PROC; + } + pos++; + } + return img; +} diff --git a/src/util_image.h b/src/util_image.h new file mode 100644 index 0000000..eb253a1 --- /dev/null +++ b/src/util_image.h @@ -0,0 +1,172 @@ +#ifndef _UTIL_IMAGE_H_ +#define _UTIL_IMAGE_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#define UTIM_INT32 int + +typedef UTIM_INT32 utim_int32_t; +typedef unsigned UTIM_INT32 utim_uint32_t; + +#ifndef byte + #define byte unsigned char +#endif + +/* + * Defined as little-endian + */ +#define UTIM_BMP 0x00706d62 /* b m p \0 */ +#define UTIM_PNG 0x00676e70 /* p n g \0 */ +#define UTIM_JPG 0x0067706a /* j p g \0 */ +#define UTIM_TGA 0x00616774 /* t g a \0 */ + +#define UTIM_MAX_CHANNELS 4 + +enum { + UTIM_ERR_ALLOC, + UTIM_ERR_FILE_OPS, + UTIM_ERR_BAD_ARG +}; + +typedef struct { + byte *pixels; + int xsize; + int ysize; + int channels; +} utim_image_t; /* For 8bit color-depth images */ + +typedef int utim_point_t[2]; /* X, Y */ +typedef byte utim_color_t[4]; /* R, G, B, A */ + +/* + * PSF2 fonts + * View https://www.win.tue.nl/~aeb/linux/kbd/font-formats-1.html + */ +struct psf2_header { + unsigned char magic[4]; + unsigned int version; + unsigned int headersize; /* offset of bitmaps in file */ + unsigned int flags; + unsigned int length; /* number of glyphs */ + unsigned int charsize; /* number of bytes for each character */ + unsigned int height, width; /* max dimensions of glyphs */ + /* charsize = height * ((width + 7) / 8) */ +}; + +typedef struct { + struct psf2_header header; + byte *glyphs; + unsigned int *utf8index; +} utim_font_t; + +#define UTIM_POINT_X 0 +#define UTIM_POINT_Y 1 + +#define UTIM_COLOR_R 0 +#define UTIM_COLOR_G 1 +#define UTIM_COLOR_B 2 +#define UTIM_COLOR_A 3 + +#define UTIM_BACK_COLOR 255 + +#define UTIM_RESIZE_NEAREST 0 +#define UTIM_RESIZE_LINEAR 1 + +utim_image_t *utim_read(const char *filename); + +int utim_write(const char *filename, utim_image_t *img); +int utim_write_ctrl(const char *filename, + utim_image_t *img, int comp, int quality); + +utim_image_t *utim_clone(utim_image_t *img); +utim_image_t *utim_create(int x, int y, int nch, int c); + +void utim_free_image(utim_image_t *img); + +utim_image_t *utim_resize(utim_image_t *img, int x, int y, int mode); + +/* Swap 2 channels */ +int utim_swap_chl(utim_image_t *img, int a, int b); + +int utim_img2rgb (utim_image_t *img); +int utim_img2gray(utim_image_t *img); +int utim_img2rgba(utim_image_t *img); + +/* gray2rgb will copy gray channel 2 times */ +int utim_gray2rgb (utim_image_t *gray); +int utim_gray2rgba(utim_image_t *gray); +int utim_rgb2bgr (utim_image_t *rgb); +int utim_rgb2gray (utim_image_t *rgb); +int utim_rgb2rgba (utim_image_t *rgb); +int utim_rgba2rgb (utim_image_t *rgba); +int utim_rgba2gray(utim_image_t *rgba); + +utim_image_t *utim_rgb_by_rgba(utim_image_t *rgba); +utim_image_t *utim_bgr_by_rgb (utim_image_t *rgb); +utim_image_t *utim_gray_by_rgb(utim_image_t *rgb); +utim_image_t *utim_rgb_by_gray(utim_image_t *gray); + +utim_image_t *utim_stack (utim_image_t **chx, int nch); +utim_image_t *utim_pick_chl(utim_image_t *img, int ich); +utim_image_t *utim_set_chl (utim_image_t *img, int ich, int color); + +void utim_negative_color(utim_image_t *img); + +#define utim_set_point(point, x, y) \ + point[UTIM_POINT_X] = x; \ + point[UTIM_POINT_Y] = y; + +#define utim_set_color(color, r, g, b, a) \ + color[UTIM_COLOR_R] = r; \ + color[UTIM_COLOR_G] = g; \ + color[UTIM_COLOR_B] = b; \ + color[UTIM_COLOR_A] = a; + +/* + * Support Gray, RGB, RGBA images. + * If not RGBA image, will convert to RGBA image + */ +int utim_set_opacity(utim_image_t *img, int opacity); + +int utim_superpose(utim_image_t *bg, utim_image_t *img, utim_point_t p); + +/* + * Simple Drawing functions + */ +int utim_set_pixel(utim_image_t *img, utim_point_t p, utim_color_t c); + +int utim_draw_point(utim_image_t *img, utim_point_t p, utim_color_t c); + +void utim_set_draw_point_fn(int (*fn) + (utim_image_t*, utim_point_t, utim_color_t)); + +void utim_draw_line(utim_image_t *img, + utim_point_t a, utim_point_t b, utim_color_t c, int width); + +void utim_draw_rect(utim_image_t *img, + utim_point_t a, int w, int h, utim_color_t c, int width); + +void utim_draw_circle(utim_image_t *img, + utim_point_t center, int radius, utim_color_t color, int width); + +void utim_draw_filled_circle(utim_image_t *img, + utim_point_t center, int radius, utim_color_t color); + +/* + * Simple text + */ +utim_font_t *utim_load_font(const char *filename); + +void utim_free_font(utim_font_t *font); + +utim_image_t *utim_text(utim_font_t *font, char *text, utim_color_t color); + +#undef byte + +#ifdef __cplusplus + } +#endif + +#endif /* _UTIL_IMAGE_H_ */ diff --git a/src/util_list.c b/src/util_list.c new file mode 100644 index 0000000..6a956cf --- /dev/null +++ b/src/util_list.c @@ -0,0 +1,1058 @@ +#include +#include +#include + +#include "util_list.h" + +#ifdef ENABLE_SSHM + #include "sshm.h" +#endif +#ifdef ENABLE_ZLIB + /* #include */ + #include +#endif + +#ifndef byte + #define byte unsigned char +#endif +#ifndef uint + #define uint unsigned int +#endif + +#define ASSERT_ON_BUILD(condition)\ + ((void)sizeof(char[1 - 2*!!(condition)])) +/*=========== compile-time test ===========*/ +#ifdef ENABLE_FOPS +void __________compile_time_test___________() +{ + /* LIST_INFO_LEN must be 32 */ + ASSERT_ON_BUILD(24 - LIST_INFO_LEN); +} +#endif + +static uint _djb_hash(void *s, uint len, uint seed) +{ + uint hash = seed; + uint i; + for (i = 0; i < len; i++) { + hash = (hash << 5) + hash + *((byte*)s + i); + } + return hash; +} + +static int _is_little_endian() +{ + uint i = 1; + unsigned char *c = (unsigned char*)&i; + if (*c) + return 1; + return 0; +} + +static int _is_gzfile(const char *filename) +{ + const byte gzhead[] = {0x1F, 0x8B, 0x08}; + byte buf[sizeof(gzhead)] = {0}; + int r; + FILE *fp = fopen(filename, "rb"); + if (!fp) + return 0; + fread(buf, sizeof(gzhead), 1, fp); + r = memcmp(gzhead, buf, sizeof(gzhead)); + fclose(fp); + return !r; +} + +static void _memswap(void *a, void *b, int len) +{ + unsigned char *p1 = (unsigned char*)a; + unsigned char *p2 = (unsigned char*)b; + while (len--) { + *p1 ^= *p2; + *p2 ^= *p1; + *p1++ ^= *p2++; + } +} + +static void _memrev(void *i, int len) +{ + int j = len >> 1; + unsigned char *c = (unsigned char*)i; + for (len--; j > 0; j--) { + *c ^= *(c + len); + *(c + len) ^= *c; + *c ^= *(c + len); + c++, len -= 2; + } +} + +static void _switch_byte_order(list_t *list) { + _memrev(&list->length , sizeof(uint)); + _memrev(&list->key , sizeof(uint)); + _memrev(&list->counter, sizeof(uint)); + _memrev(&list->blen , sizeof(uint)); + _memrev(&list->scale , sizeof(uint)); +} + +list_t *list_new_static(uint scale, uint blen) +{ + list_t *list = (list_t*)malloc(sizeof(list_t)); + if (!scale || !blen || !list) + return NULL; + memset(list, 0, sizeof(list_t)); + SET_FLAG(list->flag, LIST_STATIC_MODE); + list->length = scale * blen; + list->mem = (byte*)malloc(list->length); + if (!list->mem) { + free(list); + return NULL; + } + memset(list->mem, 0, list->length); + list->blen = blen; + list->scale = scale; + list->counter = scale; + return list; +} + +list_t *list_new_dynamic(uint scale) +{ + list_t *list = (list_t*)malloc(sizeof(list_t)); + if (!scale || !list) + return NULL; + memset(list, 0, sizeof(list_t)); + SET_FLAG(list->flag, LIST_DYNAMIC_MODE); + list->index = (byte**)malloc(scale * sizeof(byte*)); + if (!list->index) { + free(list); + return NULL; + } + memset(list->index, 0, scale * sizeof(byte*)); + list->scale = scale; + return list; +} + +list_t *list_clone(list_t *list) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_clone_dynamic(list); + else /* LIST_STATIC_MODE */ + return list_clone_static(list); +} + +list_t *list_clone_static(list_t *list) +{ + list_t *clone; + if (!list) + return NULL; + clone = (list_t*)malloc(sizeof(list_t)); + if (!clone) + return NULL; + memcpy(clone, list, sizeof(list_t)); + if (list->name) { + clone->name = (char*)malloc(LIST_NAME_LEN); + if (!clone->name) { + free(clone); + return NULL; + } + strcpy(clone->name, list->name); + } else { + clone->name = NULL; + } + clone->mem = (byte*)malloc(list->length); + if (!clone->mem) { + if (clone->name) + free(clone->name); + free(clone); + return NULL; + } + memcpy(clone->mem, list->mem, list->length); + clone->index = NULL; + return clone; +} + +#define PROCESS_CLONE_DYN_INDEX(i) \ +if (list->index[i]) { \ + s = list_set_record(clone, i, \ + list->index[i] + sizeof(DynLenFlag), \ + *(DynLenFlag*)list->index[i]); \ + if (s) { \ + clone->status |= s; \ + return clone; \ + } \ +} + +list_t *list_clone_dynamic(list_t *list) +{ + uint i; + list_t *clone; + ls_status_t s; + if (!list) + return NULL; + clone = (list_t*)malloc(sizeof(list_t)); + if (!clone) + return NULL; + memcpy(clone, list, sizeof(list_t)); + if (list->name) { + clone->name = (char*)malloc(LIST_NAME_LEN); + if (!clone->name) { + free(clone); + return NULL; + } + strcpy(clone->name, list->name); + } else { + clone->name = NULL; + } + clone->index = (byte**)malloc(list->scale * sizeof(byte*)); + if (!clone->index) { + if (clone->name) + free(clone->name); + free(clone); + return NULL; + } + memset(clone->index, 0, list->scale * sizeof(byte*)); + clone->length = 0; + clone->counter = 0; + for (i = 0; i < list->scale; ++i) { + PROCESS_CLONE_DYN_INDEX(i); + } + return clone; +} + +ls_status_t list_del(list_t *list) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_del_dynamic(list); + else /* LIST_STATIC_MODE */ + return list_del_static(list); +} + +ls_status_t list_del_static(list_t *list) +{ + if (!list) + return LSS_SUCCESS; + if (list->name) + free(list->name); + if (list->mem) + free(list->mem); + free(list); + return LSS_SUCCESS; +} + +ls_status_t list_del_dynamic(list_t *list) +{ + if (!list) + return LSS_SUCCESS; + if (list->name) + free(list->name); + while (list->scale--) { + if (list->index[list->scale]) + free(list->index[list->scale]); + } + if (list->index) + free(list->index); + free(list); + return LSS_SUCCESS; +} + +ls_status_t list_set_name(list_t *list, const char *name) +{ + if (TEST_FLAG(list->flag, LIST_NOT_SHARED)) { + if (!name) { + if (list->name) + free(list->name); + list->name = NULL; + return LSS_SUCCESS; + } + if (strlen(name) >= LIST_NAME_LEN) + return LSS_ARG_ILL; + if (list->name) + free(list->name); + list->name = (char*)malloc(LIST_NAME_LEN); + if (!list->name) + return LSS_MALLOC_ERR; + strcpy(list->name, name); + } else { + memset((char*)list + sizeof(list_t), 0, LIST_NAME_LEN); + strcpy((char*)list + sizeof(list_t), name); + } + return LSS_SUCCESS; +} + + +ls_status_t list_resize(list_t *list, uint newScale) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_resize_dynamic(list, newScale); + else /* LIST_STATIC_MODE */ + return list_resize_static(list, newScale); +} + +ls_status_t list_resize_static(list_t *list, uint newScale) +{ + byte *newMem; + uint newLen = newScale * list->blen; + if (TEST_FLAG(list->flag, LIST_UNRESIZABLE)) + return LSS_BAD_OBJ; + newMem = (byte*)realloc(list->mem, newLen); + if (!newMem) + return LSS_MALLOC_ERR; + if (newScale > list->scale) { + memset(list->mem + list->scale * list->blen, 0, + (newScale - list->scale) * list->blen); + } + list->mem = newMem; + list->length = newLen; + list->scale = newScale; + list->counter = newScale; + return LSS_SUCCESS; +} + +ls_status_t list_resize_dynamic(list_t *list, uint newScale) +{ + byte **newIndex; + uint i = 0; + if (TEST_FLAG(list->flag, LIST_UNRESIZABLE)) + return LSS_BAD_OBJ; + if (newScale < list->scale) { + i = newScale; + for (; i < list->scale; ++i) + list_del_dynamic_record(list, i); + } + newIndex = (byte**)realloc(list->index, + newScale * sizeof(byte*)); + if (!newIndex) + return LSS_MALLOC_ERR; + list->index = newIndex; + if (newScale > list->scale) { + memset((byte*)newIndex + list->scale * sizeof(byte*), 0, + (newScale - list->scale) * sizeof(byte*)); + } + list->scale = newScale; + return LSS_SUCCESS; +} + +ls_status_t list_set_record(list_t *list, uint id, + const void *record, uint len) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_set_dynamic_record(list, id, record, len); + else /* LIST_STATIC_MODE */ + return list_set_static_record(list, id, record, len); +} + +ls_status_t list_set_static_record(list_t *list, uint id, + const void *record, uint len) +{ + if (id >= list->scale) + return LSS_BAD_ID; + len = list->blen < len ? list->blen : len; + memcpy(list->mem + id * list->blen, record, len); + return LSS_SUCCESS; +} + +ls_status_t list_set_dynamic_record(list_t *list, uint id, + const void *record, uint len) +{ + DynLenFlag *l; + uint mlen = 0; + if (id >= list->scale) + return LSS_BAD_ID; + if (list->index[id]) + return LSS_DYN_ID_EXIST; + mlen = len + sizeof(DynLenFlag); + list->index[id] = (byte*)malloc(mlen); + if (!list->index[id]) + return LSS_MALLOC_ERR; + memcpy(list->index[id] + sizeof(DynLenFlag), record, len); + l = (DynLenFlag*)list->index[id]; + *l = len; + list->counter++; + list->length += mlen; + return LSS_SUCCESS; +} + +int list_test_record(list_t *list, uint id, uint offset, + uint nbyte) +{ + byte *rec; + uint c = 0; + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + if (list->index[id]) + return 1; + else + return 0; + } else { /* LIST_STATIC_MODE */ + rec = (byte*)list_get_static_record(list, id); + while (nbyte--) { + c += rec[offset + nbyte]; + } + if (c) + return 1; + else + return 0; + } +} + +void *list_alloc(list_t *list, uint id, uint nbyte) +{ + if (id >= list->scale) + return NULL; + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + if (list->index[id]) + list_del_dynamic_record(list, id); + list->index[id] = (byte*)malloc(nbyte + sizeof(DynLenFlag)); + if (!list->index[id]) + return NULL; + *(DynLenFlag*)list->index[id] = nbyte; + list->length += (nbyte + sizeof(DynLenFlag)); + list->counter++; + return list->index[id] + sizeof(DynLenFlag); + } else { /* LIST_STATIC_MODE */ + if (nbyte > list->blen) + return NULL; + else + return list_get_static_record(list, id); + } +} + +void *list_get_record(list_t *list, uint id) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_get_dynamic_record(list, id); + else /* LIST_STATIC_MODE */ + return list_get_static_record(list, id); +} + +void *list_get_static_record(list_t *list, uint id) +{ + if (id >= list->scale) + return NULL; + else + return list->mem + id * list->blen; +} + +void *list_get_dynamic_record(list_t *list, uint id) +{ + if (id >= list->scale) + return NULL; + else { + if (list->index[id]) + return (list->index[id] + sizeof(DynLenFlag)); + else + return NULL; + } +} + +uint list_get_record_len(list_t *list, uint id) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + if (list->index[id]) + return *(DynLenFlag*)list->index[id]; + else + return 0; + } + else /* LIST_STATIC_MODE */ + return list->blen; +} + +ls_status_t list_del_record(list_t *list, uint id) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_del_dynamic_record(list, id); + else /* LIST_STATIC_MODE */ + return list_del_static_record(list, id); +} + +ls_status_t list_del_static_record(list_t *list, uint id) +{ + if (id >= list->scale) + return LSS_BAD_ID; + memset(list->mem + id * list->blen, 0, list->blen); + return LSS_SUCCESS; +} + +ls_status_t list_del_dynamic_record(list_t *list, uint id) +{ + if (id >= list->scale) + return LSS_BAD_ID; + if (list->index[id]) { + list->length -= sizeof(DynLenFlag); + list->length -= *(DynLenFlag*)list->index[id]; + list->counter--; + free(list->index[id]); + list->index[id] = NULL; + } + return LSS_SUCCESS; +} + +ls_status_t list_swap_record(list_t *list, + uint id1, uint id2) +{ + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_swap_dynamic_record(list, id1, id2); + else /* LIST_STATIC_MODE */ + return list_swap_static_record(list, id1, id2); +} + +ls_status_t list_swap_static_record(list_t *list, + uint id1, uint id2) +{ + if (id1 >= list->scale || id2 >= list->scale) + return LSS_BAD_ID; + _memswap(list->mem + id1 * list->blen, + list->mem + id2 * list->blen, list->blen); + return LSS_SUCCESS; +} + +ls_status_t list_swap_dynamic_record(list_t *list, + uint id1, uint id2) +{ + byte *p; + if (id1 >= list->scale || id2 >= list->scale) + return LSS_BAD_ID; + p = list->index[id1]; + list->index[id1] = list->index[id2]; + list->index[id2] = p; + return LSS_SUCCESS; +} + +#ifdef ENABLE_FOPS + ls_status_t list_export(list_t *list, const char *path, + const char *mode) + { + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) + return list_export_dynamic(list, path, mode); + else /* LIST_STATIC_MODE */ + return list_export_static(list, path, mode); + } + + #define PROCESS_EXPORT_STATIC(PREFIX) \ + uint name_len = 0; \ + fp = PREFIX##fopen(path, mode); \ + if (!fp) \ + return LSS_FILE_ERR; \ + if (list->name) \ + name_len = strlen(list->name); \ + if (!_is_little_endian()) { \ + _memrev(&name_len, sizeof(uint)); \ + _switch_byte_order(list); \ + } \ + PREFIX##fwrite(&name_len, sizeof(uint), 1, fp); \ + PREFIX##fwrite(list->name, name_len, 1, fp); \ + PREFIX##fwrite((byte*)list + LIST_INFO_OFFSET, \ + LIST_INFO_LEN, 1, fp); \ + PREFIX##fwrite(list->mem, list->length, 1, fp); \ + if (!_is_little_endian()) { \ + _switch_byte_order(list); \ + } + +#ifdef ENABLE_ZLIB + static ls_status_t _list_export_static_zlib(list_t *list, + const char *path, const char *mode) + { + ZLIB_FILE fp; + PROCESS_EXPORT_STATIC(ZLIB_); + ZLIB_fclose(fp); + return LSS_SUCCESS; + } +#endif + + static ls_status_t _list_export_static_ansic(list_t *list, + const char *path, const char *mode) + { + IO_FILE fp; + PROCESS_EXPORT_STATIC(IO_); + IO_fclose(fp); + return LSS_SUCCESS; + } + + ls_status_t list_export_static(list_t *list, const char *path, + const char *mode) { + if (!mode) + mode = LIST_EXPORT_MODE_DEF; +#ifdef ENABLE_ZLIB + if (strcmp(mode, LIST_EXPORT_MODE_RAW)) + return _list_export_static_zlib(list, path, mode); + else + return _list_export_static_ansic(list, path, + LIST_EXPORT_MODE_DEF); +#else + return _list_export_static_ansic(list, path, + LIST_EXPORT_MODE_DEF); +#endif + } + + #define PROCESS_EXPORT_DYN_RECORD(PREFIX) \ + if (list->index[sc]) { \ + if (!_is_little_endian()) \ + _memrev(list->index[sc], sizeof(DynLenFlag)); \ + PREFIX##fwrite(list->index[sc], sizeof(DynLenFlag) + \ + (DynLenFlag)*list->index[sc], 1, fp); \ + if (!_is_little_endian()) { \ + _memrev(list->index[sc], sizeof(DynLenFlag)); \ + _memrev(&sc, sizeof(uint)); \ + } \ + PREFIX##fwrite(&sc, sizeof(uint), 1, fp); \ + if (!_is_little_endian()) \ + _memrev(&sc, sizeof(uint)); \ + } + + #define PROCESS_EXPORT_DYNAMIC(PREFIX) \ + uint name_len = 0; \ + uint sc = list->scale; \ + fp = PREFIX##fopen(path, mode); \ + if (!fp) \ + return LSS_FILE_ERR; \ + if (list->name) \ + name_len = strlen(list->name); \ + if (!_is_little_endian()) { \ + _memrev(&name_len, sizeof(uint)); \ + _switch_byte_order(list); \ + } \ + PREFIX##fwrite(&name_len, sizeof(uint), 1, fp); \ + PREFIX##fwrite(list->name, name_len, 1, fp); \ + PREFIX##fwrite((byte*)list + LIST_INFO_OFFSET, \ + LIST_INFO_LEN, 1, fp); \ + while (sc--) { \ + PROCESS_EXPORT_DYN_RECORD(PREFIX); \ + } \ + if (!_is_little_endian()) { \ + _switch_byte_order(list); \ + } + +#ifdef ENABLE_ZLIB + static ls_status_t _list_export_dynamic_zlib(list_t *list, + const char *path, const char *mode) + { + ZLIB_FILE fp; + PROCESS_EXPORT_DYNAMIC(ZLIB_); + ZLIB_fclose(fp); + return LSS_SUCCESS; + } +#endif + + static ls_status_t _list_export_dynamic_ansic(list_t *list, + const char *path, const char *mode) + { + IO_FILE fp; + PROCESS_EXPORT_DYNAMIC(IO_); + IO_fclose(fp); + return LSS_SUCCESS; + } + + ls_status_t list_export_dynamic(list_t *list, const char *path, + const char *mode) + { + if (!mode) + mode = LIST_EXPORT_MODE_DEF; +#ifdef ENABLE_ZLIB + if (strcmp(mode, LIST_EXPORT_MODE_RAW)) + return _list_export_dynamic_zlib(list, path, mode); + else + return _list_export_dynamic_ansic(list, path, + LIST_EXPORT_MODE_DEF); +#else + return _list_export_dynamic_ansic(list, path, + LIST_EXPORT_MODE_DEF); +#endif + } + + #define PROCESS_IMPORT_DYN_RECORD(PREFIX) \ + list->index = (byte**) \ + malloc(list->scale * sizeof(byte*)); \ + if (!list->index) { \ + free(list); \ + free(name); \ + PREFIX##fclose(fp); \ + return NULL; \ + } \ + memset(list->index, 0, list->scale * sizeof(byte*)); \ + while (PREFIX##fread(&len, sizeof(DynLenFlag), 1, fp)) { \ + if (!_is_little_endian()) \ + _memrev(&len, sizeof(DynLenFlag)); \ + tmp = (byte*)malloc(len + sizeof(DynLenFlag)); \ + if (!tmp) { \ + list->status |= LIST_IMPORT_ERROR; \ + list->status |= LIST_MALLOC_ERROR; \ + PREFIX##fclose(fp); \ + return list; \ + } \ + *(DynLenFlag*)tmp = len; \ + PREFIX##fread(tmp + \ + sizeof(DynLenFlag), len, 1, fp); \ + PREFIX##fread(&id, sizeof(uint), 1, fp); \ + if (!_is_little_endian()) \ + _memrev(&id, sizeof(uint)); \ + list->index[id] = tmp; \ + } + + #define PROCESS_IMPORT_STATIC_RECORD(PREFIX) \ + list->mem = (byte*)malloc(list->length); \ + if (!list->mem) { \ + free(list); \ + free(name); \ + PREFIX##fclose(fp); \ + return NULL; \ + } \ + PREFIX##fread(list->mem, list->length, 1, fp); + + #define PROCESS_IMPORT_OPEN_FILE(PREFIX) \ + char *name; \ + byte *tmp; \ + uint name_len = 0; \ + uint id = 0; \ + DynLenFlag len = 0; \ + fp = PREFIX##fopen(path, "rb"); \ + if (!fp) \ + return NULL; \ + list = (list_t*)malloc(sizeof(list_t)); \ + if (!list) { \ + PREFIX##fclose(fp); \ + return NULL; \ + } \ + memset(list, 0, sizeof(list_t)); \ + PREFIX##fread(&name_len, sizeof(uint), 1, fp); \ + if (!_is_little_endian()) \ + _memrev(&name_len, sizeof(uint)); \ + name_len++; \ + name = (char*)malloc(name_len); \ + if (!name) { \ + free(list); \ + PREFIX##fclose(fp); \ + return NULL; \ + } \ + memset(name, 0, name_len); \ + PREFIX##fread(name, name_len - 1, 1, fp); \ + PREFIX##fread((byte*)list + LIST_INFO_OFFSET, \ + LIST_INFO_LEN, 1, fp); \ + if (!_is_little_endian()) \ + _switch_byte_order(list); \ + list->name = name; + +#ifdef ENABLE_ZLIB + static list_t *_list_import_zlib(const char *path) + { + list_t *list; + ZLIB_FILE fp; + PROCESS_IMPORT_OPEN_FILE(ZLIB_); + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + PROCESS_IMPORT_DYN_RECORD(ZLIB_); + } else { /* LIST_STATIC_MODE */ + PROCESS_IMPORT_STATIC_RECORD(ZLIB_); + } + ZLIB_fclose(fp); + return list; + } +#endif + + static list_t *_list_import_ansic(const char *path) + { + list_t *list; + FILE *fp; + PROCESS_IMPORT_OPEN_FILE(IO_); + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + PROCESS_IMPORT_DYN_RECORD(IO_); + } else { /* LIST_STATIC_MODE */ + PROCESS_IMPORT_STATIC_RECORD(IO_); + } + fclose(fp); + return list; + } + + list_t *list_import(const char *path) + { + if (_is_gzfile(path)) { +#ifdef ENABLE_ZLIB + return _list_import_zlib(path); +#else + return NULL; +#endif + } else { + return _list_import_ansic(path); + } + return NULL; + } + +#endif /* ENABLE_FOPS */ + +#ifdef ENABLE_SSHM + list_t *list_new_shared(uint scale, uint blen, uint key) + { + byte *shm; + uint shmlen = 0; + list_t *list = (list_t*)malloc(sizeof(list_t)); + if (!list) + return NULL; + memset(list, 0, sizeof(list_t)); + shmlen += LIST_NAME_LEN; + shmlen += sizeof(list_t); + SET_FLAG(list->flag, LIST_STATIC_MODE); + SET_FLAG(list->flag, LIST_UNRESIZABLE); + SET_FLAG(list->flag, LIST_SHARED_MEM); + SET_FLAG(list->flag, LIST_SHARED_OWNER); + list->length = scale * blen; + shmlen += list->length; + shm = (byte*)create_shm(shmlen, key); + if (!shm) { + free(list); + return NULL; + } + memset(shm, 0, shmlen); + memcpy(shm, list, sizeof(list_t)); + free(list); + list = (list_t*)shm; + list->name = (char*)shm + sizeof(list_t); + list->mem = shm + sizeof(list_t) + LIST_NAME_LEN; + list->key = key; + list->blen = blen; + list->scale = scale; + list->counter = scale; + return list; + } + + list_t *list_link_shared(uint len, uint key) + { + byte *shm; + list_t *list = (list_t*)malloc(sizeof(list_t)); + if (!list) + return NULL; + shm = (byte*)create_shm(len + + sizeof(list_t) + LIST_NAME_LEN, key); + if (!shm) { + free(list); + return NULL; + } + memcpy(list, shm, sizeof(list_t)); + list->name = (char*)shm + sizeof(list_t); + list->mem = shm + sizeof(list_t) + LIST_NAME_LEN; + SET_FLAG(list->flag, LIST_UNRESIZABLE); + SET_FLAG(list->flag, LIST_SHARED_MEM); + SET_FLAG(list->flag, LIST_SHARED_USER); + return list; + } + + ls_status_t list_del_shared(list_t *list) + { + uint shmlen; + if (TEST_FLAG(list->flag, LIST_SHARED_USER)) + return LSS_ARG_ILL; + shmlen = (sizeof(list_t) + + LIST_NAME_LEN + list->length); + if (free_shm(shmlen, list->key)) + return LSS_SHM_ERR; + return LSS_SUCCESS; + } + + #define PROCESS_IMPORT_SHARED(PREFIX) \ + byte *shm; \ + char *name; \ + uint name_len = 0; \ + uint shmlen = 0; \ + fp = PREFIX##fopen(path, "rb"); \ + if (!fp) \ + return NULL; \ + list = (list_t*)malloc(sizeof(list_t)); \ + if (!list) \ + return NULL; \ + PREFIX##fread(&name_len, sizeof(uint), 1, fp); \ + if (!_is_little_endian()) \ + _memrev(&name_len, sizeof(uint)); \ + name_len++; \ + name = (char*)malloc(name_len); \ + if (!name) { \ + free(list); \ + return NULL; \ + } \ + memset(name, 0, name_len); \ + PREFIX##fread(name, name_len - 1, 1, fp); \ + PREFIX##fread(list, (sizeof(list_t)), 1, fp); \ + if (!_is_little_endian()) \ + _switch_byte_order(list); \ + shmlen = sizeof(list_t) + LIST_NAME_LEN; \ + shmlen += list->length; \ + shm = (byte*)create_shm(shmlen, key); \ + if (!shm) { \ + free(list); \ + free(name); \ + PREFIX##fclose(fp); \ + return NULL; \ + } \ + PREFIX##fread(shm + sizeof(list_t) + LIST_NAME_LEN, \ + list->length, 1, fp); \ + memcpy(shm, list, sizeof(list_t)); \ + strcpy((char*)shm + sizeof(list_t), name); \ + free(name); \ + list->name = (char*)shm + sizeof(list_t); \ + list->mem = shm + sizeof(list_t) +LIST_NAME_LEN; + +#ifdef ENABLE_ZLIB + static list_t *_list_import_shm_zlib(const char *path, uint key) + { + list_t *list; + ZLIB_FILE fp; + PROCESS_IMPORT_SHARED(ZLIB_); + ZLIB_fclose(fp); + return list; + } +#endif + + static list_t *_list_import_shm_ansic(const char *path, uint key) + { + list_t *list; + FILE *fp; + PROCESS_IMPORT_SHARED(IO_); + fclose(fp); + return list; + } + + list_t *list_import_shared(const char *path, uint key) + { + if (_is_gzfile(path)) { +#ifdef ENABLE_ZLIB + return _list_import_shm_zlib(path, key); +#else + return NULL; +#endif + } else { + return _list_import_shm_ansic(path, key); + } + return NULL; + } +#endif /* ENABLE_SSHM */ + +ls_status_t list_calc_hash_id(list_t *list, + void *data, uint len, uint arg, uint *id, __hashFx) +{ + uint nhash, cid; + if (hfx == NULL) + hfx = _djb_hash; + nhash = hfx(data, len, arg) % list->scale; + cid = nhash; + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + while (list->index[cid]) { + cid++; + if (cid == list->scale) + cid = 0; + if (cid == nhash) + return LSS_ERR_LISTFULL; + } + } else { /* LIST_STATIC_MODE */ + while (list_test_record(list, cid, 0, list->blen)) { + cid++; + if (cid == list->scale) + cid = 0; + if (cid == nhash) + return LSS_ERR_LISTFULL; + } + } + *id = cid; + return LSS_SUCCESS; +} + +ls_status_t list_search_by_hash(list_t *list, + void *data, uint len, uint arg, __hashFx, + __compFx, void *pattern, uint plen, uint *fid) +{ + uint nhash, id; + if (hfx == NULL) + hfx = _djb_hash; + nhash = hfx(data, len, arg) % list->scale; + id = nhash; + if (TEST_FLAG(list->flag, LIST_DYNAMIC_MODE)) { + while (list->index[id]) { + if (!cfx(list->index[id] + + sizeof(DynLenFlag), pattern, plen)) { + *fid = id; + return LSS_SUCCESS; + } + id++; + if (id == list->scale) + id = 0; + if (id == nhash) + return LSS_OBJ_NOFOUND; + } + } else { /* LIST_STATIC_MODE */ + while (list_test_record(list, id, 0, list->blen)) { + if (!cfx(list->index[id], pattern, plen)) { + *fid = id; + return LSS_SUCCESS; + } + id++; + if (id == list->scale) + id = 0; + if (id == nhash) + return LSS_OBJ_NOFOUND; + } + } + return LSS_OBJ_NOFOUND; +} + +void list_print_info(list_t *list, void *stream) +{ + FILE *fp = stream ? (FILE*)stream : stdout; + if (!list) { + fprintf(fp, "[ERROR] : Bad object, stopped!\n"); + return; + } + if (!list->status) + fprintf(fp, "[Status] : List status: OK!\n"); + else{ + fprintf(fp, "[Status] : Something wrong, stopped!\n"); + return; + } + if (list->name) + fprintf(fp, "[%s]\n", list->name); + else + fprintf(fp, "Not named list.\n"); + if (TEST_FLAG(list->flag, LIST_STATIC_MODE)) + fprintf(fp, "[MODE] = STATIC\n"); + else + fprintf(fp, "[MODE] = DYNAMIC\n"); + fprintf(fp, "[length] = %d\n", list->length); + fprintf(fp, "[scale] = %d\n", list->scale); + fprintf(fp, "[record] = %d\n", list->counter); + fprintf(fp, "[block length] = %d\n", list->blen); +} + +void operation_status(ls_status_t ops_stat) +{ + FILE *fp = stderr; + switch (ops_stat) { + case LSS_SUCCESS: + fprintf(fp, "Operation Succeeded!\n"); + break; + case LSS_BAD_ID: + fprintf(fp, "Failed: Bad ID\n"); + break; + case LSS_MALLOC_ERR: + fprintf(fp, "Failed: Mem Alloc Failed\n"); + break; + case LSS_DYN_ID_EXIST: + fprintf(fp, "Failed: Record Exist\n"); + break; + case LSS_BAD_OBJ: + fprintf(fp, "Failed: Bad Object\n"); + break; + case LSS_ARG_ILL: + fprintf(fp, "Failed: Arg Illegal\n"); + break; + case LSS_LNAME_ERR: + fprintf(fp, "Failed: Bad LIST name\n"); + break; + case LSS_FILE_ERR: + fprintf(fp, "Failed: File Access Err\n"); + break; + case LSS_SHM_ERR: + fprintf(fp, "Failed: Shared Memory Err\n"); + break; + case LSS_ERR_LISTFULL: + fprintf(fp, "Failed: Full list\n"); + break; + case LSS_OBJ_NOFOUND: + fprintf(fp, "Failed: Object Not Found\n"); + break; + default : + fprintf(fp, "Failed: Unknown Error\n"); + break; + } +} diff --git a/src/util_list.h b/src/util_list.h new file mode 100644 index 0000000..4297855 --- /dev/null +++ b/src/util_list.h @@ -0,0 +1,268 @@ +#ifndef _LIST_H_ +#define _LIST_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#ifndef byte + #define byte unsigned char +#endif +#ifndef uint + #define uint unsigned int +#endif + +typedef unsigned char ls_status_t; + +/*========== operation status ===========*/ +#define LSS_STATUS_ +#define LSS_SUCCESS 0x00 +#define LSS_BAD_ID 0x01 +#define LSS_MALLOC_ERR 0x02 +#define LSS_DYN_ID_EXIST 0x03 +#define LSS_BAD_OBJ 0x04 +#define LSS_ARG_ILL 0x05 +#define LSS_LNAME_ERR 0x06 +#define LSS_FILE_ERR 0x07 +#define LSS_SHM_ERR 0x08 +#define LSS_ERR_LISTFULL 0x09 +#define LSS_OBJ_NOFOUND 0x0A + +#define SET_BIT(val, bitn) (val |=(1<<(bitn))) +#define CLR_BIT(val, bitn) (val&=~(1<<(bitn))) +#define GET_BIT(val, bitn) (val & (1<<(bitn))) + +#define SET_FLAG(flag, bft) \ + if(bft & 0x0F)SET_BIT(flag, bft >> 4); else CLR_BIT(flag, bft >> 4); +#define TEST_FLAG(flag, bft) \ + !(((bft & 0x0F) << (bft >> 4)) ^ (flag & (1 << (bft >> 4)))) + +#define DYN_LENGTH_TYPE uint +typedef DYN_LENGTH_TYPE DynLenFlag; + +/* Do not edit LIST_NAME_LEN unless necessary */ +#ifndef LIST_NAME_LEN + #define LIST_NAME_LEN 128 +#endif + +#ifndef LIST_PATH_LEN + #define LIST_PATH_LEN 256 +#endif + +/* + * Record on dynamic mode : + * uint length | record |, when export : + * uint length | record | id + * Notice: length of a dynamic record will not count DynLenFlag + * + * index record(Each record will allocate memory via "malloc") + * index0----->rec0 + * index1----->rec1 + * ... + * indexX----->recX + * X = list.scale - 1 + * ============================================================== + * Record on static mode : + * list.mem : + * record 0 | record 1 | ... | record X + * _________ + * | + * |---->list.blen + */ + +typedef struct { + char *name; /* name of LIST */ + byte *mem; /* memory pool */ + byte **index; /* index table */ + uint length; /* length of memory pool */ + uint key; /* key of shared mem */ + uint counter; /* record counter */ + uint blen; /* length of a record */ + uint scale; /* scale */ + byte flag; /* list's mode flag */ + byte status; /* work status flag */ + byte placeholder[2]; +} list_t; + +struct list_info { + uint length; /* length of memory pool */ + uint key; /* key of shared mem */ + uint counter; /* record counter */ + uint blen; /* length of a record */ + uint scale; /* scale */ + byte flag; /* list's mode flag */ + byte status; /* work status flag */ + byte placeholder[2]; +}; + +#define LIST_INFO_LEN sizeof(struct list_info) +#define LIST_INFO_OFFSET sizeof(list_t) - LIST_INFO_LEN + +/*============== mode flag ==============*/ +#define LIST_FLAG_DEFAULT 0x00 +#define LIST_STATIC_MODE 0x00 +#define LIST_DYNAMIC_MODE 0x01 +#define LIST_RESIZABLE 0x50 +#define LIST_UNRESIZABLE 0x51 +#define LIST_SHARED_USER 0x60 +#define LIST_SHARED_OWNER 0x61 +#define LIST_NOT_SHARED 0x70 +#define LIST_SHARED_MEM 0x71 + +/*============= status flag =============*/ +/* LIST_STATUS_ALRIGHT MUST BE 0x00 */ +#define LIST_STATUS_ALRIGHT 0x00 + +#define LIST_MALLOC_ERROR 0x01 +#define LIST_IMPORT_ERROR 0x02 + +/* Tips: + * Overload "list_new" under cpp + * list_t *list_new(uint scale); + * list_t *list_new(uint scale, uint blen); + */ +list_t *list_new_static(uint scale, uint blen); +list_t *list_new_dynamic(uint scale); + +list_t *list_clone(list_t *list); +list_t *list_clone_static(list_t *list); +list_t *list_clone_dynamic(list_t *list); + +ls_status_t list_del(list_t *list); +ls_status_t list_del_static(list_t *list); +ls_status_t list_del_dynamic(list_t *list); + +ls_status_t list_resize(list_t *list, uint newScale); +ls_status_t list_resize_static(list_t *list, uint newScale); +ls_status_t list_resize_dynamic(list_t *list, uint newScale); + +ls_status_t list_set_name(list_t *list, const char *name); + +ls_status_t list_set_record(list_t *list, uint id, + const void *record, uint len); +ls_status_t list_set_static_record(list_t *list, uint id, + const void *record, uint len); +ls_status_t list_set_dynamic_record(list_t *list, uint id, + const void *record, uint len); + +int list_test_record(list_t *list, uint id, uint offset, + uint nbyte); + +void *list_alloc(list_t *list, uint id, uint nbyte); + +void *list_get_record(list_t *list, uint id); +void *list_get_static_record(list_t *list, uint id); +void *list_get_dynamic_record(list_t *list, uint id); + +uint list_get_record_len(list_t *list, uint id); +/* #define list_get_static_record_len */ +#define list_get_dynamic_record_len(record) \ + (DynLenFlag)*(record - sizeof(DynLenFlag)) + +ls_status_t list_del_record(list_t *list, uint id); +ls_status_t list_del_static_record(list_t *list, uint id); +ls_status_t list_del_dynamic_record(list_t *list, uint id); + +ls_status_t list_swap_record(list_t *list, + uint id1, uint id2); +ls_status_t list_swap_static_record(list_t *list, + uint id1, uint id2); +ls_status_t list_swap_dynamic_record(list_t *list, + uint id1, uint id2); + +#define IO_FILE FILE* +#define IO_fopen fopen +#define IO_fclose fclose +#define IO_fread fread +#define IO_fwrite fwrite + +/* #define ENABLE_ZLIB */ + +#ifdef ENABLE_ZLIB + #define ZLIB_FILE gzFile + #define ZLIB_fopen gzopen + #define ZLIB_fclose gzclose + #define ZLIB_fread(buf, size, count, fp) \ + gzread(fp, buf, size) + #define ZLIB_fwrite(buf, size, count, fp) \ + gzwrite(fp, buf, size) +#endif /* ENABLE_ZLIB */ + +#define LIST_EXPORT_MODE_RAW "raw" + +#ifdef ENABLE_ZLIB + #define LIST_EXPORT_MODE_DEF "wb0" +#else + #define LIST_EXPORT_MODE_DEF "wb" +#endif + +ls_status_t list_export(list_t *list, const char *path, + const char *mode); + +/* + * Exported file structure of Static LIST + * name_len | list->name | list | list->mem + */ +ls_status_t list_export_static(list_t *list, const char *path, + const char *mode); + +/* + * Exported file structure of Dynamic LIST + * name_len | list->name | list | *index[0:N] + */ +ls_status_t list_export_dynamic(list_t *list, const char *path, + const char *mode); + +list_t *list_import(const char *path); + +/* + * Sharedlist is tlist's extension to create list_t at shared memory space. + * Notice that sharedlist not support tlist which was created as Dynamic + * mode, cuz that will create too many shm handles. + */ + +/* #define ENABLE_SSHM */ + +/* list_t | name | mem */ + +#ifdef ENABLE_SSHM + list_t *list_new_shared(uint scale, uint blen, uint key); + list_t *list_link_shared(uint len, uint key); + #define list_shared_shm_len(list) \ + (sizeof(LIST) + LIST_NAME_LEN + list->length) + ls_status_t list_del_shared(list_t *list); + #define list_export_shared list_export_static + list_t *list_import_shared(const char *path, uint key); +#endif /* ENABLE_SSHM */ + +/* + * simple hash function + */ +#define __hashFx uint hfx(void*, uint, uint) +#define __compFx int cfx(void*, void*, uint) + +#define list_set_unresizable(list) \ + SET_FLAG(list->flag, LIST_UNRESIZABLE) +#define list_set_resizable(list) \ + SET_FLAG(list->flag, LIST_RESIZABLE) + +ls_status_t list_calc_hash_id(list_t *list, + void *data, uint len, uint arg, uint *id, __hashFx); + +ls_status_t list_search_by_hash(list_t *list, + void *data, uint len, uint arg, __hashFx, + __compFx, void *pattern, uint plen, uint *fid); + +/*================= Debug =================*/ +#define list_ops operation_status +void list_print_info(list_t *list, void *stream); +void operation_status(ls_status_t ops_stat); + +#undef byte +#undef uint + +#ifdef __cplusplus + } +#endif + +#endif /* _LIST_H_ */ diff --git a/src/util_log.c b/src/util_log.c new file mode 100644 index 0000000..a8367c8 --- /dev/null +++ b/src/util_log.c @@ -0,0 +1,142 @@ +#include +#include +#include +#include +#include + +#include "util_log.h" + +static int highlight_flag = UTLOG_HIGHLIGHT_ON; +static FILE *user_log_ostream = NULL; +static int error_action_flag = UTLOG_ERR_ACT_ABORT; + +void utlog_highlight_on(void) +{ + highlight_flag = UTLOG_HIGHLIGHT_ON; +} + +void utlog_highlight_off(void) +{ + highlight_flag = UTLOG_HIGHLIGHT_OFF; +} + +void utlog_set_ostream(void *stream) +{ + user_log_ostream = (FILE*)stream; +} + +void *utlog_get_ostream(void) +{ + if (!user_log_ostream) + return UTLOG_DEFAULT_OSTREAM; + else + return user_log_ostream; +} + +void utlog_set_error_action(int act) +{ + error_action_flag = act; +} + +const char *_log_clock_style_str(int logtype) +{ + switch (logtype) { + case UTLOG_ERR: + return "\033[;31m"; + case UTLOG_WARN: + return "\033[;33m"; + case UTLOG_INFO: + return "\033[0;36m"; + default: + return NULL; + } +} + +static int _account_symbol(char *s, char ch) +{ + int c = 0; + while(*s) + if(*s++ == ch) + c++; + return c; +} + +static char *_find_symbol(char *s, char ch) +{ + if(*s == ch) + return s; + while(*s) + if(*++s == ch) + return s; + return NULL; +} + +#define UTLOG_FORMAT_RET \ +if (logtype == UTLOG_ERR) { \ + switch (error_action_flag) { \ + case UTLOG_ERR_ACT_ABORT: \ + exit(-1); /* abort() */ \ + break; \ + case UTLOG_ERR_ACT_WARNING: \ + /* print warning info only */ \ + break; \ + default: \ + break; \ + } \ +} \ +return; + +void utlog_format(int logtype, const char *fmt, ...) +{ + int n; + char *p, *f, *pf; + /* time_t t; + * time(&t); + */ + FILE *log_ostream; + clock_t clk = clock(); + va_list ap; + if (user_log_ostream) + log_ostream = user_log_ostream; + else + log_ostream = UTLOG_DEFAULT_OSTREAM; + f = pf = (char*)malloc(strlen(fmt) + 1); + strcpy(f, fmt); + /* fprintf(log_ostream, "[%ld:%ld] : ", t, clk); */ + if (getenv("TERM") && + highlight_flag == UTLOG_HIGHLIGHT_ON) { + fprintf(log_ostream, "[%s%08ld\033[0m]: ", + _log_clock_style_str(logtype), + clk / UTLOG_CLK_FRAC); + } else { + fprintf(log_ostream, "[%08ld]: ", + clk / UTLOG_CLK_FRAC); + } + va_start(ap, fmt); + n = _account_symbol(f, '%'); + p = _find_symbol(f, '%'); + if(p == NULL){ + fprintf(log_ostream, "%s", f); + free(pf); + UTLOG_FORMAT_RET; + } + *p = '\0'; + fprintf(log_ostream, "%s", f); + f = p; + *f = '%'; + while(n-- > 1){ + p = _find_symbol(f + 1, '%'); + if(p == NULL){ + free(pf); + UTLOG_FORMAT_RET; + } + *p = '\0'; + fprintf(log_ostream, f, va_arg(ap, void*)); + f = p; + *f = '%'; + } + fprintf(log_ostream, f, va_arg(ap, void*)); + free(pf); + va_end(ap); + UTLOG_FORMAT_RET; +} diff --git a/src/util_log.h b/src/util_log.h new file mode 100644 index 0000000..5893d62 --- /dev/null +++ b/src/util_log.h @@ -0,0 +1,43 @@ +#ifndef _UTIL_LOG_H_ +#define _UTIL_LOG_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#define UTLOG_CLK_FRAC 1000 + +#define UTLOG_DEFAULT_OSTREAM stderr + +enum utlog_type { + UTLOG_ERR, + UTLOG_WARN, + UTLOG_INFO +}; + +enum utlog_err_act { + UTLOG_ERR_ACT_ABORT, + UTLOG_ERR_ACT_WARNING +}; + +enum utlog_highlight { + UTLOG_HIGHLIGHT_ON, + UTLOG_HIGHLIGHT_OFF +}; + + +void utlog_set_ostream(void *stream); +void *utlog_get_ostream(void); + +void utlog_set_error_action(int act); + +void utlog_highlight_on(void); +void utlog_highlight_off(void); + +void utlog_format(int logtype, const char *fmt, ...); + +#ifdef __cplusplus + } +#endif + +#endif /* _UTIL_LOG_H_ */ diff --git a/src/util_rbt.c b/src/util_rbt.c new file mode 100644 index 0000000..c8990a9 --- /dev/null +++ b/src/util_rbt.c @@ -0,0 +1,357 @@ +#include +#include + +#include "util_rbt.h" + +static struct rbt_node node_nil = { + NULL, /* .data = NULL, */ + &node_nil, /* .parent = &node_nil, */ + &node_nil, /* .left = &node_nil, */ + &node_nil, /* .right = &node_nil, */ + RBT_COLOR_BLACK /* .color = RBT_COLOR_BLACK */ +}; + +static void* __pointer(struct rbt_node *node) { + return node->data; +} + +static int __compare(void *a, void *b) { + return (unsigned char*)a - (unsigned char*)b; +} + +struct rbt_node *rbt_nil(void) +{ + return &node_nil; +} + +rbt_t *new_rbt(void*(*get_key)(struct rbt_node *), + int (*compare)(void*, void*)) +{ + rbt_t *t = (rbt_t*)malloc(sizeof(rbt_t)); + t->root = &node_nil; + t->get_key = get_key; + t->compare = compare; + t->get_key = t->get_key ? t->get_key : __pointer; + t->compare = t->compare ? t->compare : __compare; + return t; +} + +static struct rbt_node *new_rbt_node(void *d) +{ + struct rbt_node *z = (struct rbt_node*) + malloc(sizeof(struct rbt_node)); + z->data = d; + z->parent = &node_nil; + z->left = &node_nil; + z->right = &node_nil; + z->color = RBT_COLOR_RED; + return z; +} + +static void left_rotate(rbt_t *t, struct rbt_node *node) +{ + struct rbt_node *y = node->right; + node->right = y->left; + if (y->left != &node_nil) + y->left->parent = node; + y->parent = node->parent; + if(node->parent == &node_nil) + t->root = y; + else if(node == node->parent->left) + node->parent->left = y; + else + node->parent->right = y; + y->left = node; + node->parent = y; +} + +static void right_rotate(rbt_t *t, struct rbt_node *node) +{ + struct rbt_node *y = node->left; + node->left = y->right; + if(y->right != &node_nil) + y->right->parent = node; + y->parent = node->parent; + if(node->parent == &node_nil) + t->root = y; + else if(node == node->parent->right) + node->parent->right = y; + else + node->parent->left = y; + y->right = node; + node->parent = y; +} + +static void rbt_insert_fixup(rbt_t *t, struct rbt_node *z) +{ + struct rbt_node *y; + while (z->parent->color == RBT_COLOR_RED) { + if (z->parent == z->parent->parent->left) { + y = z->parent->parent->right; + if (y->color == RBT_COLOR_RED) { + z->parent->color = RBT_COLOR_BLACK; + y->color = RBT_COLOR_BLACK; + z->parent->parent->color = RBT_COLOR_RED; + z = z->parent->parent; + } else { + if (z == z->parent->right) { + z = z->parent; + left_rotate(t, z); + } + z->parent->color = RBT_COLOR_BLACK; + z->parent->parent->color = RBT_COLOR_RED; + right_rotate(t, z->parent->parent); + } + } else { + y = z->parent->parent->left; + if (y->color == RBT_COLOR_RED) { + z->parent->color = RBT_COLOR_BLACK; + y->color = RBT_COLOR_BLACK; + z->parent->parent->color = RBT_COLOR_RED; + z = z->parent->parent; + } else { + if (z == z->parent->left) { + z = z->parent; + right_rotate(t, z); + } + z->parent->color = RBT_COLOR_BLACK; + z->parent->parent->color = RBT_COLOR_RED; + left_rotate(t, z->parent->parent); + } + } + } + t->root->color = RBT_COLOR_BLACK; +} + +void *rbt_insert(rbt_t *t, void *d) +{ + void *holder; + struct rbt_node *y = &node_nil, *x = t->root; + struct rbt_node *z = new_rbt_node(d); + while (x != &node_nil) { + y = x; + if (!t->compare(t->get_key(z), t->get_key(x))) { + holder = x->data; + free(z); + x->data = d; + return holder; + } + if (t->compare(t->get_key(z), t->get_key(x)) < 0) + x = x->left; + else + x = x->right; + } + z->parent = y; + if (y == &node_nil) + t->root = z; + else { + if (t->compare(t->get_key(z), t->get_key(y)) < 0) + y->left = z; + else + y->right = z; + } + rbt_insert_fixup(t, z); + return NULL; +} + +static struct rbt_node *__search_rbt_node(rbt_t *t, void *key) +{ + struct rbt_node *z = t->root; + while (z != &node_nil) { + if (t->compare(t->get_key(z), key) == 0) + return z; + if (t->compare(t->get_key(z), key) < 0) + z = z->right; + else + z = z->left; + } + return NULL; +} + +void *rbt_search(rbt_t *t, void *key) +{ + struct rbt_node *z = __search_rbt_node(t, key); + if (z) + return z->data; + return NULL; +} + +static void __free_rbt(struct rbt_node *root) +{ + struct rbt_node *l, *r; + if (root == &node_nil) + return; + l = root->left; + r = root->right; + free(root); + __free_rbt(l); + __free_rbt(r); +} + +void free_rbt(rbt_t *t) +{ + __free_rbt(t->root); + free(t); +} + +static void __rb_transplant(rbt_t *t, + struct rbt_node *u, struct rbt_node *v) +{ + if (u->parent == &node_nil) + t->root = v; + else if (u == u->parent->left) + u->parent->left = v; + else + u->parent->right = v; + v->parent = u->parent; +} + +static struct rbt_node *__rb_tree_minimum(struct rbt_node *z) +{ + for(; z->left != &node_nil; z = z->left); + return z; +} + +static void __rb_tree_delete_fixup(rbt_t *t, struct rbt_node *x) +{ + struct rbt_node *w; + while (x != t->root && x->color == RBT_COLOR_BLACK) { + if (x == x->parent->left) { + w = x->parent->right; + if (w->color == RBT_COLOR_RED) { + w->color = RBT_COLOR_BLACK; + x->parent->color = RBT_COLOR_RED; + left_rotate(t, x->parent); + w = x->parent->right; + } + if (w->left->color == RBT_COLOR_BLACK && + w->right->color == RBT_COLOR_BLACK){ + w->color = RBT_COLOR_RED; + x = x->parent; + } else { + if (w->right->color == RBT_COLOR_BLACK) { + w->left->color = RBT_COLOR_BLACK; + w->color = RBT_COLOR_RED; + right_rotate(t, w); + w = w->parent->right; + } + w->color = x->parent->color; + x->parent->color = RBT_COLOR_BLACK; + w->right->color = RBT_COLOR_BLACK; + left_rotate(t, x->parent); + x = t->root; + } + } else { + w = x->parent->left; + if (w->color == RBT_COLOR_RED) { + w->color = RBT_COLOR_BLACK; + x->parent->color = RBT_COLOR_RED; + right_rotate(t, x->parent); + w = x->parent->left; + } + if (w->right->color == RBT_COLOR_BLACK && + w->left->color == RBT_COLOR_BLACK) { + w->color = RBT_COLOR_RED; + x = x->parent; + } else { + if (w->left->color == RBT_COLOR_BLACK) { + w->right->color = RBT_COLOR_BLACK; + w->color = RBT_COLOR_RED; + left_rotate(t, w); + w = w->parent->left; + } + w->color = x->parent->color; + x->parent->color = RBT_COLOR_BLACK; + w->left->color = RBT_COLOR_BLACK; + right_rotate(t, x->parent); + x = t->root; + } + } + } + x->color = RBT_COLOR_BLACK; +} + +void *rbt_delete(rbt_t *t, void *key) +{ + struct rbt_node *y, *z, *x, *hold_node_to_delete; + unsigned char y_original_color; + void *node_to_return; + hold_node_to_delete = y = z = __search_rbt_node(t, key); + if (y == NULL) /* Node not exist */ + return NULL; + node_to_return = y->data; + y_original_color = y->color; + if (z->left == &node_nil) { + x = z->right; + __rb_transplant(t, z, z->right); + } else if (z->right == &node_nil) { + x = z->left; + __rb_transplant(t, z, z->left); + } else { + y = __rb_tree_minimum(z->right); + y_original_color = y->color; + x = y->right; + if (y->parent == z) + x->parent = y; + else { + __rb_transplant(t, y, y->right); + y->right = z->right; + y->right->parent = y; + } + __rb_transplant(t, z, y); + y->left = z->left; + y->left->parent = y; + y->color = z->color; + } + if (y_original_color == RBT_COLOR_BLACK) + __rb_tree_delete_fixup(t, x); + free(hold_node_to_delete); + return node_to_return; +} + +rbt_iterator *new_rbt_iterator(rbt_t *t) +{ + struct rbt_node *aux = t->root; + rbt_iterator *it = (rbt_iterator*)malloc(sizeof(rbt_iterator)); + while(aux->left != &node_nil || aux->right != &node_nil) { + while(aux->left != &node_nil) + aux = aux->left; + if(aux->right != &node_nil) + aux = aux->right; + } + it->current = aux; + return it; +} + +int rbt_iterator_has_next(rbt_iterator *it) +{ + if(it->current != &node_nil) + return 1; + return 0; +} + +void *rbt_iterator_next(rbt_iterator *it) +{ + struct rbt_node *aux; + struct rbt_node *tn = it->current; + if (tn->parent != &node_nil && + tn->parent->right != &node_nil && + tn->parent->right != tn) { + aux = tn->parent->right; + while (aux->left != &node_nil || aux->right != &node_nil) { + while(aux->left != &node_nil) + aux = aux->left; + if (aux->right != &node_nil) + aux = aux->right; + } + it->current = aux; + } + else + it->current = it->current->parent; + return tn->data; +} + +void free_rbt_iterator(rbt_iterator *it) +{ + free(it); +} diff --git a/src/util_rbt.h b/src/util_rbt.h new file mode 100644 index 0000000..36a9c60 --- /dev/null +++ b/src/util_rbt.h @@ -0,0 +1,54 @@ +#ifndef _RBT_H_ +#define _RBT_H_ + +#ifdef __cplusplus + extern "C" { +#endif + +#define RBT_COLOR_RED 1 +#define RBT_COLOR_BLACK 2 + +struct rbt_node { + void *data; + struct rbt_node *parent; + struct rbt_node *left; + struct rbt_node *right; + unsigned char color; +}; + +typedef struct { + struct rbt_node *root; + void *(*get_key)(struct rbt_node*); + int (*compare)(void*, void*); +} rbt_t; + +typedef struct { + struct rbt_node *current; +} rbt_iterator; + +rbt_t *new_rbt(void*(*get_key)(struct rbt_node *), + int (*compare)(void*, void*)); + +void free_rbt(rbt_t *t); + +struct rbt_node *rbt_nil(void); + +void *rbt_insert(rbt_t *t, void *d); + +void *rbt_delete(rbt_t *t, void *key); + +void *rbt_search(rbt_t *t, void *key); + +rbt_iterator *new_rbt_iterator(rbt_t *t); + +int rbt_iterator_has_next(rbt_iterator *it); + +void *rbt_iterator_next(rbt_iterator *it); + +void free_rbt_iterator(rbt_iterator* it); + +#ifdef __cplusplus + } +#endif + +#endif /* _RBT_H_ */