Merge pull request #409 from NeilBickford-NV/nbickford/update-stb-image
Updates stb_image and stb_image_write
diff --git a/stb_image.h b/stb_image.h
index 8b71060..d0aaf2a 100644
--- a/stb_image.h
+++ b/stb_image.h
@@ -1,4 +1,4 @@
-/* stb_image - v2.21 - public domain image loader - http://nothings.org/stb
+/* stb_image - v2.28 - public domain image loader - http://nothings.org/stb
no warranty implied; use at your own risk
Do this:
@@ -48,8 +48,15 @@
RECENT REVISION HISTORY:
+ 2.28 (2023-01-29) many error fixes, security errors, just tons of stuff
+ 2.27 (2021-07-11) document stbi_info better, 16-bit PNM support, bug fixes
+ 2.26 (2020-07-13) many minor fixes
+ 2.25 (2020-02-02) fix warnings
+ 2.24 (2020-02-02) fix warnings; thread-local failure_reason and flip_vertically
+ 2.23 (2019-08-11) fix clang static analysis warning
+ 2.22 (2019-03-04) gif fixes, fix warnings
2.21 (2019-02-25) fix typo in comment
- 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs
+ 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs
2.19 (2018-02-11) fix warning
2.18 (2018-01-30) fix warnings
2.17 (2018-01-29) bugfix, 1-bit BMP, 16-bitness query, fix warnings
@@ -84,26 +91,37 @@
Jeremy Sawicki (handle all ImageNet JPGs)
Optimizations & bugfixes Mikhail Morozov (1-bit BMP)
Fabian "ryg" Giesen Anael Seghezzi (is-16-bit query)
- Arseny Kapoulkine
+ Arseny Kapoulkine Simon Breuss (16-bit PNM)
John-Mark Allen
Carmelo J Fdez-Aguera
Bug & warning fixes
- Marc LeBlanc David Woo Guillaume George Martins Mozeiko
- Christpher Lloyd Jerry Jansson Joseph Thomson Phil Jordan
- Dave Moore Roy Eltham Hayaki Saito Nathan Reed
- Won Chun Luke Graham Johan Duparc Nick Verigakis
- the Horde3D community Thomas Ruf Ronny Chevalier github:rlyeh
- Janez Zemva John Bartholomew Michal Cichon github:romigrou
- Jonathan Blow Ken Hamada Tero Hanninen github:svdijk
- Laurent Gomila Cort Stratton Sergio Gonzalez github:snagar
- Aruelien Pocheville Thibault Reuille Cass Everitt github:Zelex
- Ryamond Barbiero Paul Du Bois Engin Manap github:grim210
- Aldo Culquicondor Philipp Wiesemann Dale Weiler github:sammyhw
- Oriol Ferrer Mesia Josh Tobin Matthew Gregan github:phprus
- Julian Raschke Gregory Mullen Baldur Karlsson github:poppolopoppo
- Christian Floisand Kevin Schmidt JR Smith github:darealshinji
- Blazej Dariusz Roszkowski github:Michaelangel007
+ Marc LeBlanc David Woo Guillaume George Martins Mozeiko
+ Christpher Lloyd Jerry Jansson Joseph Thomson Blazej Dariusz Roszkowski
+ Phil Jordan Dave Moore Roy Eltham
+ Hayaki Saito Nathan Reed Won Chun
+ Luke Graham Johan Duparc Nick Verigakis the Horde3D community
+ Thomas Ruf Ronny Chevalier github:rlyeh
+ Janez Zemva John Bartholomew Michal Cichon github:romigrou
+ Jonathan Blow Ken Hamada Tero Hanninen github:svdijk
+ Eugene Golushkov Laurent Gomila Cort Stratton github:snagar
+ Aruelien Pocheville Sergio Gonzalez Thibault Reuille github:Zelex
+ Cass Everitt Ryamond Barbiero github:grim210
+ Paul Du Bois Engin Manap Aldo Culquicondor github:sammyhw
+ Philipp Wiesemann Dale Weiler Oriol Ferrer Mesia github:phprus
+ Josh Tobin Neil Bickford Matthew Gregan github:poppolopoppo
+ Julian Raschke Gregory Mullen Christian Floisand github:darealshinji
+ Baldur Karlsson Kevin Schmidt JR Smith github:Michaelangel007
+ Brad Weinberger Matvey Cherevko github:mosra
+ Luca Sas Alexander Veselov Zack Middleton [reserved]
+ Ryan C. Gordon [reserved] [reserved]
+ DO NOT ADD YOUR NAME HERE
+
+ Jacko Dirks
+
+ To add your name to the credits, pick a random blank space in the middle and fill it.
+ 80% of merge conflicts on stb PRs are due to people adding their name at the end
+ of the credits.
*/
#ifndef STBI_INCLUDE_STB_IMAGE_H
@@ -123,7 +141,7 @@
// // ... x = width, y = height, n = # 8-bit components per pixel ...
// // ... replace '0' with '1'..'4' to force that many components per pixel
// // ... but 'n' will always be the number that it would have been if you said 0
-// stbi_image_free(data)
+// stbi_image_free(data);
//
// Standard parameters:
// int *x -- outputs image width in pixels
@@ -162,6 +180,32 @@
//
// Paletted PNG, BMP, GIF, and PIC images are automatically depalettized.
//
+// To query the width, height and component count of an image without having to
+// decode the full file, you can use the stbi_info family of functions:
+//
+// int x,y,n,ok;
+// ok = stbi_info(filename, &x, &y, &n);
+// // returns ok=1 and sets x, y, n if image is a supported format,
+// // 0 otherwise.
+//
+// Note that stb_image pervasively uses ints in its public API for sizes,
+// including sizes of memory buffers. This is now part of the API and thus
+// hard to change without causing breakage. As a result, the various image
+// loaders all have certain limits on image size; these differ somewhat
+// by format but generally boil down to either just under 2GB or just under
+// 1GB. When the decoded image would be larger than this, stb_image decoding
+// will fail.
+//
+// Additionally, stb_image will reject image files that have any of their
+// dimensions set to a larger value than the configurable STBI_MAX_DIMENSIONS,
+// which defaults to 2**24 = 16777216 pixels. Due to the above memory limit,
+// the only way to have an image with such dimensions load correctly
+// is for it to have a rather extreme aspect ratio. Either way, the
+// assumption here is that such larger images are likely to be malformed
+// or malicious. If you do need to load an image with individual dimensions
+// larger than that, and it still fits in the overall size limit, you can
+// #define STBI_MAX_DIMENSIONS on your own to be something larger.
+//
// ===========================================================================
//
// UNICODE:
@@ -267,11 +311,10 @@
//
// iPhone PNG support:
//
-// By default we convert iphone-formatted PNGs back to RGB, even though
-// they are internally encoded differently. You can disable this conversion
-// by calling stbi_convert_iphone_png_to_rgb(0), in which case
-// you will always just get the native iphone "format" through (which
-// is BGR stored in RGB).
+// We optionally support converting iPhone-formatted PNGs (which store
+// premultiplied BGRA) back to RGB, even though they're internally encoded
+// differently. To enable this conversion, call
+// stbi_convert_iphone_png_to_rgb(1).
//
// Call stbi_set_unpremultiply_on_load(1) as well to force a divide per
// pixel to remove any premultiplied alpha *only* if the image file explicitly
@@ -313,7 +356,14 @@
// - If you use STBI_NO_PNG (or _ONLY_ without PNG), and you still
// want the zlib decoder to be available, #define STBI_SUPPORT_ZLIB
//
-
+// - If you define STBI_MAX_DIMENSIONS, stb_image will reject images greater
+// than that size (in either width or height) without further processing.
+// This is to let programs in the wild set an upper bound to prevent
+// denial-of-service attacks on untrusted data, as one could generate a
+// valid image of gigantic dimensions and force stb_image to allocate a
+// huge block of memory and spend disproportionate time decoding it. By
+// default this is set to (1 << 24), which is 16777216, but that's still
+// very big.
#ifndef STBI_NO_STDIO
#include <stdio.h>
@@ -339,11 +389,13 @@
extern "C" {
#endif
+#ifndef STBIDEF
#ifdef STB_IMAGE_STATIC
#define STBIDEF static
#else
#define STBIDEF extern
#endif
+#endif
//////////////////////////////////////////////////////////////////////////////
//
@@ -430,7 +482,7 @@
// get a VERY brief reason for failure
-// NOT THREADSAFE
+// on most compilers (and ALL modern mainstream compilers) this is threadsafe
STBIDEF const char *stbi_failure_reason (void);
// free the loaded image -- this is just free()
@@ -463,6 +515,13 @@
// flip the image vertically, so the first pixel in the output array is the bottom left
STBIDEF void stbi_set_flip_vertically_on_load(int flag_true_if_should_flip);
+// as above, but only applies to images loaded on the thread that calls the function
+// this function is only available if your compiler supports thread-local variables;
+// calling it will fail to link if your compiler doesn't
+STBIDEF void stbi_set_unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply);
+STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert);
+STBIDEF void stbi_set_flip_vertically_on_load_thread(int flag_true_if_should_flip);
+
// ZLIB client - used by PNG, available for other purposes
STBIDEF char *stbi_zlib_decode_malloc_guesssize(const char *buffer, int len, int initial_size, int *outlen);
@@ -559,8 +618,25 @@
#define stbi_inline __forceinline
#endif
+#ifndef STBI_NO_THREAD_LOCALS
+ #if defined(__cplusplus) && __cplusplus >= 201103L
+ #define STBI_THREAD_LOCAL thread_local
+ #elif defined(__GNUC__) && __GNUC__ < 5
+ #define STBI_THREAD_LOCAL __thread
+ #elif defined(_MSC_VER)
+ #define STBI_THREAD_LOCAL __declspec(thread)
+ #elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 201112L && !defined(__STDC_NO_THREADS__)
+ #define STBI_THREAD_LOCAL _Thread_local
+ #endif
-#ifdef _MSC_VER
+ #ifndef STBI_THREAD_LOCAL
+ #if defined(__GNUC__)
+ #define STBI_THREAD_LOCAL __thread
+ #endif
+ #endif
+#endif
+
+#if defined(_MSC_VER) || defined(__SYMBIAN32__)
typedef unsigned short stbi__uint16;
typedef signed short stbi__int16;
typedef unsigned int stbi__uint32;
@@ -589,7 +665,7 @@
#ifdef STBI_HAS_LROTL
#define stbi_lrot(x,y) _lrotl(x,y)
#else
- #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (32 - (y))))
+ #define stbi_lrot(x,y) (((x) << (y)) | ((x) >> (-(y) & 31)))
#endif
#if defined(STBI_MALLOC) && defined(STBI_FREE) && (defined(STBI_REALLOC) || defined(STBI_REALLOC_SIZED))
@@ -703,14 +779,21 @@
#ifdef STBI_NEON
#include <arm_neon.h>
-// assume GCC or Clang on ARM targets
+#ifdef _MSC_VER
+#define STBI_SIMD_ALIGN(type, name) __declspec(align(16)) type name
+#else
#define STBI_SIMD_ALIGN(type, name) type name __attribute__((aligned(16)))
#endif
+#endif
#ifndef STBI_SIMD_ALIGN
#define STBI_SIMD_ALIGN(type, name) type name
#endif
+#ifndef STBI_MAX_DIMENSIONS
+#define STBI_MAX_DIMENSIONS (1 << 24)
+#endif
+
///////////////////////////////////////////////
//
// stbi__context struct and start_xxx functions
@@ -728,6 +811,7 @@
int read_from_callbacks;
int buflen;
stbi_uc buffer_start[128];
+ int callback_already_read;
stbi_uc *img_buffer, *img_buffer_end;
stbi_uc *img_buffer_original, *img_buffer_original_end;
@@ -741,6 +825,7 @@
{
s->io.read = NULL;
s->read_from_callbacks = 0;
+ s->callback_already_read = 0;
s->img_buffer = s->img_buffer_original = (stbi_uc *) buffer;
s->img_buffer_end = s->img_buffer_original_end = (stbi_uc *) buffer+len;
}
@@ -752,7 +837,8 @@
s->io_user_data = user;
s->buflen = sizeof(s->buffer_start);
s->read_from_callbacks = 1;
- s->img_buffer_original = s->buffer_start;
+ s->callback_already_read = 0;
+ s->img_buffer = s->img_buffer_original = s->buffer_start;
stbi__refill_buffer(s);
s->img_buffer_original_end = s->img_buffer_end;
}
@@ -766,12 +852,17 @@
static void stbi__stdio_skip(void *user, int n)
{
+ int ch;
fseek((FILE*) user, n, SEEK_CUR);
+ ch = fgetc((FILE*) user); /* have to read a byte to reset feof()'s flag */
+ if (ch != EOF) {
+ ungetc(ch, (FILE *) user); /* push byte back onto stream if valid. */
+ }
}
static int stbi__stdio_eof(void *user)
{
- return feof((FILE*) user);
+ return feof((FILE*) user) || ferror((FILE *) user);
}
static stbi_io_callbacks stbi__stdio_callbacks =
@@ -867,21 +958,27 @@
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);
+static int stbi__pnm_is16(stbi__context *s);
#endif
-// this is not threadsafe
-static const char *stbi__g_failure_reason;
+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)
{
@@ -920,11 +1017,13 @@
return a <= INT_MAX/b;
}
+#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR)
// returns 1 if "a*b + add" has no negative terms/factors and doesn't overflow
static int stbi__mad2sizes_valid(int a, int b, int add)
{
return stbi__mul2sizes_valid(a, b) && stbi__addsizes_valid(a*b, add);
}
+#endif
// returns 1 if "a*b*c + add" has no negative terms/factors and doesn't overflow
static int stbi__mad3sizes_valid(int a, int b, int c, int add)
@@ -934,7 +1033,7 @@
}
// returns 1 if "a*b*c*d + add" has no negative terms/factors and doesn't overflow
-#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR)
+#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM)
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) &&
@@ -942,12 +1041,14 @@
}
#endif
+#if !defined(STBI_NO_JPEG) || !defined(STBI_NO_PNG) || !defined(STBI_NO_TGA) || !defined(STBI_NO_HDR)
// mallocs with size overflow checking
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)
{
@@ -955,7 +1056,7 @@
return stbi__malloc(a*b*c + add);
}
-#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR)
+#if !defined(STBI_NO_LINEAR) || !defined(STBI_NO_HDR) || !defined(STBI_NO_PNM)
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;
@@ -963,6 +1064,23 @@
}
#endif
+// returns 1 if the sum of two signed ints is valid (between -2^31 and 2^31-1 inclusive), 0 on overflow.
+static int stbi__addints_valid(int a, int b)
+{
+ if ((a >= 0) != (b >= 0)) return 1; // a and b have different signs, so no overflow
+ if (a < 0 && b < 0) return a >= INT_MIN - b; // same as a + b >= INT_MIN; INT_MIN - b cannot overflow since b < 0.
+ return a <= INT_MAX - b;
+}
+
+// returns 1 if the product of two signed shorts is valid, 0 on overflow.
+static int stbi__mul2shorts_valid(short a, short b)
+{
+ if (b == 0 || b == -1) return 1; // multiplication by 0 is always 0; check for -1 so SHRT_MIN/b doesn't overflow
+ if ((a >= 0) == (b >= 0)) return a <= SHRT_MAX/b; // product is positive, so similar to mul2sizes_valid
+ if (b < 0) return a <= SHRT_MIN / b; // same as a * b >= SHRT_MIN
+ return a >= SHRT_MIN / b;
+}
+
// stbi__err - error
// stbi__errpf - error returning pointer to float
// stbi__errpuc - error returning pointer to unsigned char
@@ -991,13 +1109,29 @@
static stbi_uc *stbi__hdr_to_ldr(float *data, int x, int y, int comp);
#endif
-static int stbi__vertically_flip_on_load = 0;
+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 = 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 // STBI_THREAD_LOCAL
+
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)); // make sure it's initialized if we add new fields
@@ -1005,9 +1139,8 @@
ri->channel_order = STBI_ORDER_RGB; // all current input & output are this, but this is here so we can add BGR order
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
+ // test the formats with a very explicit header first (at least a FOURCC
+ // or distinctive magic number first)
#ifndef STBI_NO_PNG
if (stbi__png_test(s)) return stbi__png_load(s,x,y,comp,req_comp, ri);
#endif
@@ -1019,10 +1152,19 @@
#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
+
+ // then the formats that can end up attempting to load with just 1 or 2
+ // bytes matching expectations; these are prone to false positives, so
+ // try them later
+ #ifndef STBI_NO_JPEG
+ if (stbi__jpeg_test(s)) return stbi__jpeg_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
@@ -1107,8 +1249,8 @@
stbi_uc *bytes = (stbi_uc *)image;
for (slice = 0; slice < z; ++slice) {
- stbi__vertical_flip(bytes, w, h, bytes_per_pixel);
- bytes += slice_size;
+ stbi__vertical_flip(bytes, w, h, bytes_per_pixel);
+ bytes += slice_size;
}
}
#endif
@@ -1121,8 +1263,10 @@
if (result == NULL)
return NULL;
+ // it is the responsibility of the loaders to make sure we get either 8 or 16 bit.
+ STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16);
+
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;
}
@@ -1145,8 +1289,10 @@
if (result == NULL)
return NULL;
+ // it is the responsibility of the loaders to make sure we get either 8 or 16 bit.
+ STBI_ASSERT(ri.bits_per_channel == 8 || ri.bits_per_channel == 16);
+
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;
}
@@ -1174,31 +1320,31 @@
#ifndef STBI_NO_STDIO
-#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8)
+#if defined(_WIN32) && 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)
+#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8)
STBIDEF int stbi_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input)
{
- return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, bufferlen, NULL, NULL);
+ return WideCharToMultiByte(65001 /* UTF8 */, 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)
+#if defined(_WIN32) && defined(STBI_WINDOWS_UTF8)
wchar_t wMode[64];
wchar_t wFilename[1024];
- if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename)))
- return 0;
-
- if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode)))
+ if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename)/sizeof(*wFilename)))
return 0;
-#if _MSC_VER >= 1400
+ if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode)/sizeof(*wMode)))
+ return 0;
+
+#if defined(_MSC_VER) && _MSC_VER >= 1400
if (0 != _wfopen_s(&f, wFilename, wMode))
f = 0;
#else
@@ -1296,15 +1442,15 @@
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);
-
+ 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 );
+ stbi__vertical_flip_slices( result, *x, *y, *z, *comp );
}
- return result;
+ return result;
}
#endif
@@ -1449,6 +1595,7 @@
static void stbi__refill_buffer(stbi__context *s)
{
int n = (s->io.read)(s->io_user_data,(char*)s->buffer_start,s->buflen);
+ s->callback_already_read += (int) (s->img_buffer - s->img_buffer_original);
if (n == 0) {
// at end of file, treat same as if from memory, but need to handle case
// where s->img_buffer isn't pointing to safe memory, e.g. 0-byte file
@@ -1473,6 +1620,9 @@
return 0;
}
+#if defined(STBI_NO_JPEG) && defined(STBI_NO_HDR) && defined(STBI_NO_PIC) && defined(STBI_NO_PNM)
+// nothing
+#else
stbi_inline static int stbi__at_eof(stbi__context *s)
{
if (s->io.read) {
@@ -1484,9 +1634,14 @@
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)
+// nothing
+#else
static void stbi__skip(stbi__context *s, int n)
{
+ if (n == 0) return; // already there!
if (n < 0) {
s->img_buffer = s->img_buffer_end;
return;
@@ -1501,7 +1656,11 @@
}
s->img_buffer += n;
}
+#endif
+#if defined(STBI_NO_PNG) && defined(STBI_NO_TGA) && defined(STBI_NO_HDR) && defined(STBI_NO_PNM)
+// nothing
+#else
static int stbi__getn(stbi__context *s, stbi_uc *buffer, int n)
{
if (s->io.read) {
@@ -1525,18 +1684,27 @@
} else
return 0;
}
+#endif
+#if defined(STBI_NO_JPEG) && defined(STBI_NO_PNG) && defined(STBI_NO_PSD) && defined(STBI_NO_PIC)
+// nothing
+#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)
+// nothing
+#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)
// nothing
@@ -1552,13 +1720,16 @@
static stbi__uint32 stbi__get32le(stbi__context *s)
{
stbi__uint32 z = stbi__get16le(s);
- return z + (stbi__get16le(s) << 16);
+ z += (stbi__uint32)stbi__get16le(s) << 16;
+ return z;
}
#endif
#define STBI__BYTECAST(x) ((stbi_uc) ((x) & 255)) // truncate int to byte without warnings
-
+#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)
+// nothing
+#else
//////////////////////////////////////////////////////////////////////////////
//
// generic converter from built-in img_n to req_comp
@@ -1574,7 +1745,11 @@
{
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)
+// nothing
+#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;
@@ -1610,7 +1785,7 @@
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);
+ default: STBI_ASSERT(0); STBI_FREE(data); STBI_FREE(good); return stbi__errpuc("unsupported", "Unsupported format conversion");
}
#undef STBI__CASE
}
@@ -1618,12 +1793,20 @@
STBI_FREE(data);
return good;
}
+#endif
+#if defined(STBI_NO_PNG) && defined(STBI_NO_PSD)
+// nothing
+#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)
+// nothing
+#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;
@@ -1659,7 +1842,7 @@
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);
+ default: STBI_ASSERT(0); STBI_FREE(data); STBI_FREE(good); return (stbi__uint16*) stbi__errpuc("unsupported", "Unsupported format conversion");
}
#undef STBI__CASE
}
@@ -1667,6 +1850,7 @@
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)
@@ -1819,9 +2003,12 @@
int i,j,k=0;
unsigned int code;
// build size list for each symbol (from JPEG spec)
- for (i=0; i < 16; ++i)
- for (j=0; j < count[i]; ++j)
+ for (i=0; i < 16; ++i) {
+ for (j=0; j < count[i]; ++j) {
h->size[k++] = (stbi_uc) (i+1);
+ if(k >= 257) return stbi__err("bad size list","Corrupt JPEG");
+ }
+ }
h->size[k] = 0;
// compute actual symbols (from jpeg spec)
@@ -1946,6 +2133,8 @@
// convert the huffman code to the symbol id
c = ((j->code_buffer >> (32 - k)) & stbi__bmask[k]) + h->delta[k];
+ if(c < 0 || c >= 256) // symbol id out of bounds!
+ return -1;
STBI_ASSERT((((j->code_buffer) >> (32 - h->size[c])) & stbi__bmask[h->size[c]]) == h->code[c]);
// convert the id to a symbol
@@ -1964,14 +2153,14 @@
unsigned int k;
int sgn;
if (j->code_bits < n) stbi__grow_buffer_unsafe(j);
+ if (j->code_bits < n) return 0; // ran out of bits from stream, return 0s intead of continuing
- sgn = (stbi__int32)j->code_buffer >> 31; // sign bit is always in MSB
+ sgn = j->code_buffer >> 31; // sign bit always in MSB; 0 if MSB clear (positive), 1 if MSB set (negative)
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);
+ return k + (stbi__jbias[n] & (sgn - 1));
}
// get some unsigned bits
@@ -1979,6 +2168,7 @@
{
unsigned int k;
if (j->code_bits < n) stbi__grow_buffer_unsafe(j);
+ if (j->code_bits < n) return 0; // ran out of bits from stream, return 0s intead of continuing
k = stbi_lrot(j->code_buffer, n);
j->code_buffer = k & ~stbi__bmask[n];
k &= stbi__bmask[n];
@@ -1990,6 +2180,7 @@
{
unsigned int k;
if (j->code_bits < 1) stbi__grow_buffer_unsafe(j);
+ if (j->code_bits < 1) return 0; // ran out of bits from stream, return 0s intead of continuing
k = j->code_buffer;
j->code_buffer <<= 1;
--j->code_bits;
@@ -2021,14 +2212,16 @@
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");
+ if (t < 0 || t > 15) return stbi__err("bad huffman code","Corrupt JPEG");
// 0 all the ac values now so we can do it 32-bits at a time
memset(data,0,64*sizeof(data[0]));
diff = t ? stbi__extend_receive(j, t) : 0;
+ if (!stbi__addints_valid(j->img_comp[b].dc_pred, diff)) return stbi__err("bad delta","Corrupt JPEG");
dc = j->img_comp[b].dc_pred + diff;
j->img_comp[b].dc_pred = dc;
+ if ((dc > SHRT_MAX) || (dequant[0] > SHRT_MAX) || !stbi__mul2shorts_valid((short) dc, (short) dequant[0])) return stbi__err("can't merge dc and ac", "Corrupt JPEG");
data[0] = (short) (dc * dequant[0]);
// decode AC components, see JPEG spec
@@ -2042,6 +2235,7 @@
if (r) { // fast-AC path
k += (r >> 4) & 15; // run
s = r & 15; // combined length
+ if (s > j->code_bits) return stbi__err("bad huffman code", "Combined length longer than code bits available");
j->code_buffer <<= s;
j->code_bits -= s;
// decode into unzigzag'd location
@@ -2078,11 +2272,14 @@
// first scan for DC coefficient, must be first
memset(data,0,64*sizeof(data[0])); // 0 all the ac values now
t = stbi__jpeg_huff_decode(j, hdc);
+ if (t < 0 || t > 15) return stbi__err("can't merge dc and ac", "Corrupt JPEG");
diff = t ? stbi__extend_receive(j, t) : 0;
+ if (!stbi__addints_valid(j->img_comp[b].dc_pred, diff)) return stbi__err("bad delta", "Corrupt JPEG");
dc = j->img_comp[b].dc_pred + diff;
j->img_comp[b].dc_pred = dc;
- data[0] = (short) (dc << j->succ_low);
+ if ((dc > SHRT_MAX) || !stbi__mul2shorts_valid((short) dc, 1 << j->succ_low)) return stbi__err("can't merge dc and ac", "Corrupt JPEG");
+ data[0] = (short) (dc * (1 << j->succ_low));
} else {
// refinement scan for DC coefficient
if (stbi__jpeg_get_bit(j))
@@ -2116,10 +2313,11 @@
if (r) { // fast-AC path
k += (r >> 4) & 15; // run
s = r & 15; // combined length
+ if (s > j->code_bits) return stbi__err("bad huffman code", "Combined length longer than code bits available");
j->code_buffer <<= s;
j->code_bits -= s;
zig = stbi__jpeg_dezigzag[k++];
- data[zig] = (short) ((r >> 8) << shift);
+ data[zig] = (short) ((r >> 8) * (1 << shift));
} else {
int rs = stbi__jpeg_huff_decode(j, hac);
if (rs < 0) return stbi__err("bad huffman code","Corrupt JPEG");
@@ -2137,7 +2335,7 @@
} else {
k += r;
zig = stbi__jpeg_dezigzag[k++];
- data[zig] = (short) (stbi__extend_receive(j,s) << shift);
+ data[zig] = (short) (stbi__extend_receive(j,s) * (1 << shift));
}
}
} while (k <= j->spec_end);
@@ -2936,6 +3134,7 @@
sizes[i] = stbi__get8(z->s);
n += sizes[i];
}
+ if(n > 256) return stbi__err("bad DHT header","Corrupt JPEG"); // Loop over i < n would write past end of values!
L -= 17;
if (tc == 0) {
if (!stbi__build_huffman(z->huff_dc+th, sizes)) return 0;
@@ -3068,6 +3267,8 @@
p = stbi__get8(s); if (p != 8) return stbi__err("only 8-bit","JPEG format not supported: 8-bit only"); // JPEG baseline
s->img_y = stbi__get16be(s); if (s->img_y == 0) return stbi__err("no header height", "JPEG format not supported: delayed height"); // Legal, but we don't handle it--but neither does IJG
s->img_x = stbi__get16be(s); if (s->img_x == 0) return stbi__err("0 width","Corrupt JPEG"); // JPEG requires
+ if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)");
+ if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)");
c = stbi__get8(s);
if (c != 3 && c != 1 && c != 4) return stbi__err("bad component count","Corrupt JPEG");
s->img_n = c;
@@ -3099,6 +3300,13 @@
if (z->img_comp[i].v > v_max) v_max = z->img_comp[i].v;
}
+ // check that plane subsampling factors are integer ratios; our resamplers can't deal with fractional ratios
+ // and I've never seen a non-corrupted JPEG file actually use them
+ for (i=0; i < s->img_n; ++i) {
+ if (h_max % z->img_comp[i].h != 0) return stbi__err("bad H","Corrupt JPEG");
+ if (v_max % z->img_comp[i].v != 0) return stbi__err("bad V","Corrupt JPEG");
+ }
+
// compute interleaved mcu info
z->img_h_max = h_max;
z->img_v_max = v_max;
@@ -3176,6 +3384,28 @@
return 1;
}
+static stbi_uc stbi__skip_jpeg_junk_at_end(stbi__jpeg *j)
+{
+ // some JPEGs have junk at end, skip over it but if we find what looks
+ // like a valid marker, resume there
+ while (!stbi__at_eof(j->s)) {
+ stbi_uc x = stbi__get8(j->s);
+ while (x == 255) { // might be a marker
+ if (stbi__at_eof(j->s)) return STBI__MARKER_none;
+ x = stbi__get8(j->s);
+ if (x != 0x00 && x != 0xff) {
+ // not a stuffed zero or lead-in to another marker, looks
+ // like an actual marker, return it
+ return x;
+ }
+ // stuffed zero has x=0 now which ends the loop, meaning we go
+ // back to regular scan loop.
+ // repeated 0xff keeps trying to read the next byte of the marker.
+ }
+ }
+ return STBI__MARKER_none;
+}
+
// decode image to YCbCr format
static int stbi__decode_jpeg_image(stbi__jpeg *j)
{
@@ -3192,25 +3422,22 @@
if (!stbi__process_scan_header(j)) return 0;
if (!stbi__parse_entropy_coded_data(j)) return 0;
if (j->marker == STBI__MARKER_none ) {
- // handle 0s at the end of image data from IP Kamera 9060
- while (!stbi__at_eof(j->s)) {
- int x = stbi__get8(j->s);
- if (x == 255) {
- j->marker = stbi__get8(j->s);
- break;
- }
- }
+ j->marker = stbi__skip_jpeg_junk_at_end(j);
// if we reach eof without hitting a marker, stbi__get_marker() below will fail and we'll eventually return 0
}
+ m = stbi__get_marker(j);
+ if (STBI__RESTART(m))
+ m = stbi__get_marker(j);
} 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");
+ m = stbi__get_marker(j);
} else {
- if (!stbi__process_marker(j, m)) return 0;
+ if (!stbi__process_marker(j, m)) return 1;
+ m = stbi__get_marker(j);
}
- m = stbi__get_marker(j);
}
if (j->progressive)
stbi__jpeg_finish(j);
@@ -3654,12 +3881,16 @@
else
decode_n = z->s->img_n;
+ // nothing to do if no components requested; check this now to avoid
+ // accessing uninitialized coutput[0] later
+ if (decode_n <= 0) { stbi__cleanup_jpeg(z); return NULL; }
+
// resample and color-convert
{
int k;
unsigned int i,j;
stbi_uc *output;
- stbi_uc *coutput[4];
+ stbi_uc *coutput[4] = { NULL, NULL, NULL, NULL };
stbi__resample res_comp[4];
@@ -3796,6 +4027,8 @@
{
unsigned char* result;
stbi__jpeg* j = (stbi__jpeg*) stbi__malloc(sizeof(stbi__jpeg));
+ if (!j) return stbi__errpuc("outofmem", "Out of memory");
+ memset(j, 0, sizeof(stbi__jpeg));
STBI_NOTUSED(ri);
j->s = s;
stbi__setup_jpeg(j);
@@ -3808,6 +4041,8 @@
{
int r;
stbi__jpeg* j = (stbi__jpeg*)stbi__malloc(sizeof(stbi__jpeg));
+ if (!j) return stbi__err("outofmem", "Out of memory");
+ memset(j, 0, sizeof(stbi__jpeg));
j->s = s;
stbi__setup_jpeg(j);
r = stbi__decode_jpeg_header(j, STBI__SCAN_type);
@@ -3832,6 +4067,8 @@
{
int result;
stbi__jpeg* j = (stbi__jpeg*) (stbi__malloc(sizeof(stbi__jpeg)));
+ if (!j) return stbi__err("outofmem", "Out of memory");
+ memset(j, 0, sizeof(stbi__jpeg));
j->s = s;
result = stbi__jpeg_info_raw(j, x, y, comp);
STBI_FREE(j);
@@ -3851,6 +4088,7 @@
// fast-way is faster to check than jpeg huffman, but slow way is slower
#define STBI__ZFAST_BITS 9 // accelerate all cases in default tables
#define STBI__ZFAST_MASK ((1 << STBI__ZFAST_BITS) - 1)
+#define STBI__ZNSYMS 288 // number of symbols in literal/length alphabet
// zlib-style huffman encoding
// (jpegs packs from left, zlib from right, so can't share code)
@@ -3860,8 +4098,8 @@
stbi__uint16 firstcode[16];
int maxcode[17];
stbi__uint16 firstsymbol[16];
- stbi_uc size[288];
- stbi__uint16 value[288];
+ stbi_uc size[STBI__ZNSYMS];
+ stbi__uint16 value[STBI__ZNSYMS];
} stbi__zhuffman;
stbi_inline static int stbi__bitreverse16(int n)
@@ -3938,6 +4176,7 @@
{
stbi_uc *zbuffer, *zbuffer_end;
int num_bits;
+ int hit_zeof_once;
stbi__uint32 code_buffer;
char *zout;
@@ -3948,16 +4187,23 @@
stbi__zhuffman z_length, z_distance;
} stbi__zbuf;
+stbi_inline static int stbi__zeof(stbi__zbuf *z)
+{
+ return (z->zbuffer >= z->zbuffer_end);
+}
+
stbi_inline static stbi_uc stbi__zget8(stbi__zbuf *z)
{
- if (z->zbuffer >= z->zbuffer_end) return 0;
- return *z->zbuffer++;
+ return stbi__zeof(z) ? 0 : *z->zbuffer++;
}
static void stbi__fill_bits(stbi__zbuf *z)
{
do {
- STBI_ASSERT(z->code_buffer < (1U << z->num_bits));
+ if (z->code_buffer >= (1U << z->num_bits)) {
+ z->zbuffer = z->zbuffer_end; /* treat this as EOF so we fail. */
+ return;
+ }
z->code_buffer |= (unsigned int) stbi__zget8(z) << z->num_bits;
z->num_bits += 8;
} while (z->num_bits <= 24);
@@ -3982,10 +4228,11 @@
for (s=STBI__ZFAST_BITS+1; ; ++s)
if (k < z->maxcode[s])
break;
- if (s == 16) return -1; // invalid code!
+ if (s >= 16) return -1; // invalid code!
// code size is s, so:
b = (k >> (16-s)) - z->firstcode[s] + z->firstsymbol[s];
- STBI_ASSERT(z->size[b] == s);
+ if (b >= STBI__ZNSYMS) return -1; // some data was corrupt somewhere!
+ if (z->size[b] != s) return -1; // was originally an assert, but report failure instead.
a->code_buffer >>= s;
a->num_bits -= s;
return z->value[b];
@@ -3994,7 +4241,23 @@
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);
+ if (a->num_bits < 16) {
+ if (stbi__zeof(a)) {
+ if (!a->hit_zeof_once) {
+ // This is the first time we hit eof, insert 16 extra padding btis
+ // to allow us to keep going; if we actually consume any of them
+ // though, that is invalid data. This is caught later.
+ a->hit_zeof_once = 1;
+ a->num_bits += 16; // add 16 implicit zero bits
+ } else {
+ // We already inserted our extra 16 padding bits and are again
+ // out, this stream is actually prematurely terminated.
+ return -1;
+ }
+ } else {
+ stbi__fill_bits(a);
+ }
+ }
b = z->fast[a->code_buffer & STBI__ZFAST_MASK];
if (b) {
s = b >> 9;
@@ -4008,13 +4271,16 @@
static int stbi__zexpand(stbi__zbuf *z, char *zout, int n) // need to make room for n bytes
{
char *q;
- int cur, limit, old_limit;
+ unsigned 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)
+ cur = (unsigned int) (z->zout - z->zout_start);
+ limit = old_limit = (unsigned) (z->zout_end - z->zout_start);
+ if (UINT_MAX - cur < (unsigned) n) return stbi__err("outofmem", "Out of memory");
+ while (cur + n > limit) {
+ if(limit > UINT_MAX / 2) return stbi__err("outofmem", "Out of memory");
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");
@@ -4055,17 +4321,25 @@
int len,dist;
if (z == 256) {
a->zout = zout;
+ if (a->hit_zeof_once && a->num_bits < 16) {
+ // The first time we hit zeof, we inserted 16 extra zero bits into our bit
+ // buffer so the decoder can just do its speculative decoding. But if we
+ // actually consumed any of those bits (which is the case when num_bits < 16),
+ // the stream actually read past the end so it is malformed.
+ return stbi__err("unexpected end","Corrupt PNG");
+ }
return 1;
}
+ if (z >= 286) return stbi__err("bad huffman code","Corrupt PNG"); // per DEFLATE, length codes 286 and 287 must not appear in compressed data
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");
+ if (z < 0 || z >= 30) return stbi__err("bad huffman code","Corrupt PNG"); // per DEFLATE, distance codes 30 and 31 must not appear in compressed data
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 (len > a->zout_end - zout) {
if (!stbi__zexpand(a, zout, len)) return 0;
zout = a->zout;
}
@@ -4112,11 +4386,12 @@
c = stbi__zreceive(a,2)+3;
if (n == 0) return stbi__err("bad codelengths", "Corrupt PNG");
fill = lencodes[n-1];
- } else if (c == 17)
+ } else if (c == 17) {
c = stbi__zreceive(a,3)+3;
- else {
- STBI_ASSERT(c == 18);
+ } else if (c == 18) {
c = stbi__zreceive(a,7)+11;
+ } else {
+ return stbi__err("bad codelengths", "Corrupt PNG");
}
if (ntot - n < c) return stbi__err("bad codelengths", "Corrupt PNG");
memset(lencodes+n, fill, c);
@@ -4142,7 +4417,7 @@
a->code_buffer >>= 8;
a->num_bits -= 8;
}
- STBI_ASSERT(a->num_bits == 0);
+ if (a->num_bits < 0) return stbi__err("zlib corrupt","Corrupt PNG");
// now fill header the normal way
while (k < 4)
header[k++] = stbi__zget8(a);
@@ -4164,6 +4439,7 @@
int cm = cmf & 15;
/* int cinfo = cmf >> 4; */
int flg = stbi__zget8(a);
+ if (stbi__zeof(a)) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec
if ((cmf*256+flg) % 31 != 0) return stbi__err("bad zlib header","Corrupt PNG"); // zlib spec
if (flg & 32) return stbi__err("no preset dict","Corrupt PNG"); // preset dictionary not allowed in png
if (cm != 8) return stbi__err("bad compression","Corrupt PNG"); // DEFLATE required for png
@@ -4171,7 +4447,7 @@
return 1;
}
-static const stbi_uc stbi__zdefault_length[288] =
+static const stbi_uc stbi__zdefault_length[STBI__ZNSYMS] =
{
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,
@@ -4207,6 +4483,7 @@
if (!stbi__parse_zlib_header(a)) return 0;
a->num_bits = 0;
a->code_buffer = 0;
+ a->hit_zeof_once = 0;
do {
final = stbi__zreceive(a,1);
type = stbi__zreceive(a,2);
@@ -4217,7 +4494,7 @@
} else {
if (type == 1) {
// use fixed code lengths
- if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , 288)) return 0;
+ if (!stbi__zbuild_huffman(&a->z_length , stbi__zdefault_length , STBI__ZNSYMS)) return 0;
if (!stbi__zbuild_huffman(&a->z_distance, stbi__zdefault_distance, 32)) return 0;
} else {
if (!stbi__compute_huffman_codes(a)) return 0;
@@ -4425,7 +4702,7 @@
return stbi__err("invalid filter","Corrupt PNG");
if (depth < 8) {
- STBI_ASSERT(img_width_bytes <= x);
+ if (img_width_bytes > x) return stbi__err("invalid width","Corrupt PNG");
cur += x*out_n - img_width_bytes; // store output to the rightmost img_len bytes, so we can decode in place
filter_bytes = 1;
width = img_width_bytes;
@@ -4613,6 +4890,7 @@
// de-interlacing
final = (stbi_uc *) stbi__malloc_mad3(a->s->img_x, a->s->img_y, out_bytes, 0);
+ if (!final) return stbi__err("outofmem", "Out of memory");
for (p=0; p < 7; ++p) {
int xorig[] = { 0,4,0,2,0,1,0 };
int yorig[] = { 0,0,4,0,2,0,1 };
@@ -4733,19 +5011,46 @@
return 1;
}
-static int stbi__unpremultiply_on_load = 0;
-static int stbi__de_iphone_flag = 0;
+static int stbi__unpremultiply_on_load_global = 0;
+static int stbi__de_iphone_flag_global = 0;
STBIDEF void stbi_set_unpremultiply_on_load(int flag_true_if_should_unpremultiply)
{
- stbi__unpremultiply_on_load = flag_true_if_should_unpremultiply;
+ stbi__unpremultiply_on_load_global = 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;
+ stbi__de_iphone_flag_global = flag_true_if_should_convert;
}
+#ifndef STBI_THREAD_LOCAL
+#define stbi__unpremultiply_on_load stbi__unpremultiply_on_load_global
+#define stbi__de_iphone_flag stbi__de_iphone_flag_global
+#else
+static STBI_THREAD_LOCAL int stbi__unpremultiply_on_load_local, stbi__unpremultiply_on_load_set;
+static STBI_THREAD_LOCAL int stbi__de_iphone_flag_local, stbi__de_iphone_flag_set;
+
+STBIDEF void stbi_set_unpremultiply_on_load_thread(int flag_true_if_should_unpremultiply)
+{
+ stbi__unpremultiply_on_load_local = flag_true_if_should_unpremultiply;
+ stbi__unpremultiply_on_load_set = 1;
+}
+
+STBIDEF void stbi_convert_iphone_png_to_rgb_thread(int flag_true_if_should_convert)
+{
+ stbi__de_iphone_flag_local = flag_true_if_should_convert;
+ stbi__de_iphone_flag_set = 1;
+}
+
+#define stbi__unpremultiply_on_load (stbi__unpremultiply_on_load_set \
+ ? stbi__unpremultiply_on_load_local \
+ : stbi__unpremultiply_on_load_global)
+#define stbi__de_iphone_flag (stbi__de_iphone_flag_set \
+ ? stbi__de_iphone_flag_local \
+ : stbi__de_iphone_flag_global)
+#endif // STBI_THREAD_LOCAL
+
static void stbi__de_iphone(stbi__png *z)
{
stbi__context *s = z->s;
@@ -4820,8 +5125,10 @@
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?)");
+ s->img_x = stbi__get32be(s);
+ s->img_y = stbi__get32be(s);
+ if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)");
+ if (s->img_x > STBI_MAX_DIMENSIONS) 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");
@@ -4833,14 +5140,13 @@
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 {
// if paletted, then pal_n is our final components, and
// img_n is # components to decompress/filter.
s->img_n = 1;
if ((1 << 30) / s->img_x / 4 < s->img_y) return stbi__err("too large","Corrupt PNG");
- // if SCAN_header, have to scan to see if we have a tRNS
}
+ // even with SCAN_header, have to scan to see if we have a tRNS
break;
}
@@ -4872,6 +5178,8 @@
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;
+ // non-paletted with tRNS = constant alpha. if header-scanning, we can stop now.
+ if (scan == STBI__SCAN_header) { ++s->img_n; return 1; }
if (z->depth == 16) {
for (k = 0; k < s->img_n; ++k) tc16[k] = (stbi__uint16)stbi__get16be(s); // copy the values as-is
} else {
@@ -4884,7 +5192,13 @@
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 (scan == STBI__SCAN_header) {
+ // header scan definitely stops at first IDAT
+ if (pal_img_n)
+ s->img_n = pal_img_n;
+ return 1;
+ }
+ if (c.length > (1u << 30)) return stbi__err("IDAT size limit", "IDAT section larger than 2^30 bytes");
if ((int)(ioff + c.length) < (int)ioff) return 0;
if (ioff + c.length > idata_limit) {
stbi__uint32 idata_limit_old = idata_limit;
@@ -4938,6 +5252,8 @@
++s->img_n;
}
STBI_FREE(z->expanded); z->expanded = NULL;
+ // end of PNG chunk, read and skip CRC
+ stbi__get32be(s);
return 1;
}
@@ -4968,10 +5284,12 @@
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)
+ if (p->depth <= 8)
ri->bits_per_channel = 8;
+ else if (p->depth == 16)
+ ri->bits_per_channel = 16;
else
- ri->bits_per_channel = p->depth;
+ return stbi__errpuc("bad bits_per_channel", "PNG not supported: unsupported color depth");
result = p->out;
p->out = NULL;
if (req_comp && req_comp != p->s->img_out_n) {
@@ -5076,7 +5394,7 @@
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; z >>= 1; }
+ if (z >= 0x00002) { n += 1;/* >>= 1;*/ }
return n;
}
@@ -5107,7 +5425,7 @@
v <<= -shift;
else
v >>= shift;
- STBI_ASSERT(v >= 0 && v < 256);
+ STBI_ASSERT(v < 256);
v >>= (8-bits);
STBI_ASSERT(bits >= 0 && bits <= 8);
return (int) ((unsigned) v * mul_table[bits]) >> shift_table[bits];
@@ -5117,8 +5435,35 @@
{
int bpp, offset, hsz;
unsigned int mr,mg,mb,ma, all_a;
+ int extra_read;
} stbi__bmp_data;
+static int stbi__bmp_set_mask_defaults(stbi__bmp_data *info, int compress)
+{
+ // BI_BITFIELDS specifies masks explicitly, don't override
+ if (compress == 3)
+ return 1;
+
+ if (compress == 0) {
+ if (info->bpp == 16) {
+ info->mr = 31u << 10;
+ info->mg = 31u << 5;
+ info->mb = 31u << 0;
+ } else if (info->bpp == 32) {
+ info->mr = 0xffu << 16;
+ info->mg = 0xffu << 8;
+ info->mb = 0xffu << 0;
+ info->ma = 0xffu << 24;
+ info->all_a = 0; // if all_a is 0 at end, then we loaded alpha channel but it was all 0
+ } else {
+ // otherwise, use defaults, which is all-0
+ info->mr = info->mg = info->mb = info->ma = 0;
+ }
+ return 1;
+ }
+ return 0; // error
+}
+
static void *stbi__bmp_parse_header(stbi__context *s, stbi__bmp_data *info)
{
int hsz;
@@ -5129,6 +5474,9 @@
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 (info->offset < 0) return stbi__errpuc("bad BMP", "bad BMP");
if (hsz != 12 && hsz != 40 && hsz != 56 && hsz != 108 && hsz != 124) return stbi__errpuc("unknown BMP", "BMP type not supported: unknown");
if (hsz == 12) {
@@ -5143,6 +5491,8 @@
if (hsz != 12) {
int compress = stbi__get32le(s);
if (compress == 1 || compress == 2) return stbi__errpuc("BMP RLE", "BMP type not supported: RLE");
+ if (compress >= 4) return stbi__errpuc("BMP JPEG/PNG", "BMP type not supported: unsupported compression"); // this includes PNG/JPEG modes
+ if (compress == 3 && info->bpp != 16 && info->bpp != 32) return stbi__errpuc("bad BMP", "bad BMP"); // bitfields requires 16 or 32 bits/pixel
stbi__get32le(s); // discard sizeof
stbi__get32le(s); // discard hres
stbi__get32le(s); // discard vres
@@ -5157,21 +5507,12 @@
}
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; // if all_a is 0 at end, then we loaded alpha channel but it was all 0
- } else {
- info->mr = 31u << 10;
- info->mg = 31u << 5;
- info->mb = 31u << 0;
- }
+ stbi__bmp_set_mask_defaults(info, compress);
} else if (compress == 3) {
info->mr = stbi__get32le(s);
info->mg = stbi__get32le(s);
info->mb = stbi__get32le(s);
+ info->extra_read += 12;
// not documented, but generated by photoshop and handled by mspaint
if (info->mr == info->mg && info->mg == info->mb) {
// ?!?!?
@@ -5181,6 +5522,7 @@
return stbi__errpuc("bad BMP", "bad BMP");
}
} else {
+ // V4/V5 header
int i;
if (hsz != 108 && hsz != 124)
return stbi__errpuc("bad BMP", "bad BMP");
@@ -5188,6 +5530,8 @@
info->mg = stbi__get32le(s);
info->mb = stbi__get32le(s);
info->ma = stbi__get32le(s);
+ if (compress != 3) // override mr/mg/mb unless in BI_BITFIELDS mode, as per docs
+ stbi__bmp_set_mask_defaults(info, compress);
stbi__get32le(s); // discard color space
for (i=0; i < 12; ++i)
stbi__get32le(s); // discard color space parameters
@@ -5220,6 +5564,9 @@
flip_vertically = ((int) s->img_y) > 0;
s->img_y = abs((int) s->img_y);
+ if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+ if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+
mr = info.mr;
mg = info.mg;
mb = info.mb;
@@ -5228,13 +5575,35 @@
if (info.hsz == 12) {
if (info.bpp < 24)
- psize = (info.offset - 14 - 24) / 3;
+ psize = (info.offset - info.extra_read - 24) / 3;
} else {
if (info.bpp < 16)
- psize = (info.offset - 14 - info.hsz) >> 2;
+ psize = (info.offset - info.extra_read - info.hsz) >> 2;
+ }
+ if (psize == 0) {
+ // accept some number of extra bytes after the header, but if the offset points either to before
+ // the header ends or implies a large amount of extra data, reject the file as malformed
+ int bytes_read_so_far = s->callback_already_read + (int)(s->img_buffer - s->img_buffer_original);
+ int header_limit = 1024; // max we actually read is below 256 bytes currently.
+ int extra_data_limit = 256*4; // what ordinarily goes here is a palette; 256 entries*4 bytes is its max size.
+ if (bytes_read_so_far <= 0 || bytes_read_so_far > header_limit) {
+ return stbi__errpuc("bad header", "Corrupt BMP");
+ }
+ // we established that bytes_read_so_far is positive and sensible.
+ // the first half of this test rejects offsets that are either too small positives, or
+ // negative, and guarantees that info.offset >= bytes_read_so_far > 0. this in turn
+ // ensures the number computed in the second half of the test can't overflow.
+ if (info.offset < bytes_read_so_far || info.offset - bytes_read_so_far > extra_data_limit) {
+ return stbi__errpuc("bad offset", "Corrupt BMP");
+ } else {
+ stbi__skip(s, info.offset - bytes_read_so_far);
+ }
}
- s->img_n = ma ? 4 : 3;
+ if (info.bpp == 24 && ma == 0xff000000)
+ s->img_n = 3;
+ else
+ s->img_n = ma ? 4 : 3;
if (req_comp && req_comp >= 3) // we can directly decode 3 or 4
target = req_comp;
else
@@ -5256,7 +5625,7 @@
if (info.hsz != 12) stbi__get8(s);
pal[i][3] = 255;
}
- stbi__skip(s, info.offset - 14 - info.hsz - psize * (info.hsz == 12 ? 3 : 4));
+ 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;
@@ -5305,7 +5674,7 @@
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 - 14 - info.hsz);
+ 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 /* bpp = 32 and pad = 0 */ width=0;
@@ -5323,6 +5692,7 @@
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);
+ if (rcount > 8 || gcount > 8 || bcount > 8 || acount > 8) { STBI_FREE(out); return stbi__errpuc("bad masks", "Corrupt BMP"); }
}
for (j=0; j < (int) s->img_y; ++j) {
if (easy) {
@@ -5544,6 +5914,11 @@
int RLE_repeating = 0;
int read_next_pixel = 1;
STBI_NOTUSED(ri);
+ STBI_NOTUSED(tga_x_origin); // @TODO
+ STBI_NOTUSED(tga_y_origin); // @TODO
+
+ if (tga_height > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+ if (tga_width > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
// do a tiny bit of precessing
if ( tga_image_type >= 8 )
@@ -5584,6 +5959,11 @@
// do I need to load a palette?
if ( tga_indexed)
{
+ if (tga_palette_len == 0) { /* you have to have at least one entry! */
+ STBI_FREE(tga_data);
+ return stbi__errpuc("bad palette", "Corrupt TGA");
+ }
+
// any data to skip? (offset usually = 0)
stbi__skip(s, tga_palette_start );
// load the palette
@@ -5707,6 +6087,7 @@
// Microsoft's C compilers happy... [8^(
tga_palette_start = tga_palette_len = tga_palette_bits =
tga_x_origin = tga_y_origin = 0;
+ STBI_NOTUSED(tga_palette_start);
// OK, done
return tga_data;
}
@@ -5791,6 +6172,9 @@
h = stbi__get32be(s);
w = stbi__get32be(s);
+ if (h > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+ if (w > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+
// Make sure the depth is 8 bits.
bitdepth = stbi__get16be(s);
if (bitdepth != 8 && bitdepth != 16)
@@ -6145,6 +6529,10 @@
x = stbi__get16be(s);
y = stbi__get16be(s);
+
+ if (y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+ if (x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+
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");
@@ -6154,11 +6542,12 @@
// intermediate buffer is RGBA
result = (stbi_uc *) stbi__malloc_mad3(x, y, 4, 0);
+ if (!result) return stbi__errpuc("outofmem", "Out of memory");
memset(result, 0xff, x*y*4);
if (!stbi__pic_load_core(s,x,y,comp, result)) {
STBI_FREE(result);
- result=0;
+ return 0;
}
*px = x;
*py = y;
@@ -6192,7 +6581,7 @@
int w,h;
stbi_uc *out; // output buffer (always 4 components)
stbi_uc *background; // The current "background" as far as a gif is concerned
- stbi_uc *history;
+ stbi_uc *history;
int flags, bgindex, ratio, transparent, eflags;
stbi_uc pal[256][4];
stbi_uc lpal[256][4];
@@ -6253,6 +6642,9 @@
g->ratio = stbi__get8(s);
g->transparent = -1;
+ if (g->w > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)");
+ if (g->h > STBI_MAX_DIMENSIONS) return stbi__err("too large","Very large image (corrupt?)");
+
if (comp != 0) *comp = 4; // can't actually tell whether it's 3 or 4 until we parse the comments
if (is_info) return 1;
@@ -6266,6 +6658,7 @@
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 (!g) return stbi__err("outofmem", "Out of memory");
if (!stbi__gif_header(s, g, comp, 1)) {
STBI_FREE(g);
stbi__rewind( s );
@@ -6280,7 +6673,7 @@
static void stbi__out_gif_code(stbi__gif *g, stbi__uint16 code)
{
stbi_uc *p, *c;
- int idx;
+ int idx;
// recurse to decode the prefixes, since the linked-list is backwards,
// and working backwards through an interleaved image would be nasty
@@ -6289,12 +6682,12 @@
if (g->cur_y >= g->max_y) return;
- idx = g->cur_x + g->cur_y;
+ idx = g->cur_x + g->cur_y;
p = &g->out[idx];
- g->history[idx / 4] = 1;
+ g->history[idx / 4] = 1;
c = &g->color_table[g->codes[code].suffix * 4];
- if (c[3] > 128) { // don't render transparent pixels;
+ if (c[3] > 128) { // don't render transparent pixels;
p[0] = c[2];
p[1] = c[1];
p[2] = c[0];
@@ -6403,32 +6796,36 @@
// two back is the image from two frames ago, used for a very specific disposal format
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;
+ int dispose;
+ int first_frame;
+ int pi;
+ int pcount;
STBI_NOTUSED(req_comp);
// on first frame, any non-written pixels get the background colour (non-transparent)
- first_frame = 0;
+ first_frame = 0;
if (g->out == 0) {
- if (!stbi__gif_header(s, g, comp,0)) return 0; // stbi__g_failure_reason set by stbi__gif_header
- g->out = (stbi_uc *) stbi__malloc(4 * g->w * g->h);
- g->background = (stbi_uc *) stbi__malloc(4 * g->w * g->h);
- g->history = (stbi_uc *) stbi__malloc(g->w * g->h);
- if (g->out == 0) return stbi__errpuc("outofmem", "Out of memory");
+ if (!stbi__gif_header(s, g, comp,0)) return 0; // stbi__g_failure_reason set by stbi__gif_header
+ 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");
- // image is treated as "transparent" at the start - ie, nothing overwrites the current background;
+ // image is treated as "transparent" at the start - ie, nothing overwrites the current background;
// background colour is only used for pixels that are not rendered first frame, after that "background"
- // color refers to the color that was there the previous frame.
- memset( g->out, 0x00, 4 * g->w * g->h );
- memset( g->background, 0x00, 4 * g->w * g->h ); // state of the background (starts transparent)
- memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame
- first_frame = 1;
+ // color refers to the color that was there the previous frame.
+ memset(g->out, 0x00, 4 * pcount);
+ memset(g->background, 0x00, 4 * pcount); // state of the background (starts transparent)
+ memset(g->history, 0x00, pcount); // pixels that were affected previous frame
+ first_frame = 1;
} else {
- // second frame - how do we dispoase of the previous one?
- dispose = (g->eflags & 0x1C) >> 2;
- pcount = g->w * g->h;
+ // second frame - how do we dispose of the previous one?
+ dispose = (g->eflags & 0x1C) >> 2;
+ pcount = g->w * g->h;
if ((dispose == 3) && (two_back == 0)) {
dispose = 2; // if I don't have an image to revert back to, default to the old background
@@ -6437,32 +6834,32 @@
if (dispose == 3) { // use previous graphic
for (pi = 0; pi < pcount; ++pi) {
if (g->history[pi]) {
- memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 );
+ memcpy( &g->out[pi * 4], &two_back[pi * 4], 4 );
}
}
- } else if (dispose == 2) {
- // restore what was changed last frame to background before that frame;
+ } else if (dispose == 2) {
+ // restore what was changed last frame to background before that frame;
for (pi = 0; pi < pcount; ++pi) {
if (g->history[pi]) {
- memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 );
+ memcpy( &g->out[pi * 4], &g->background[pi * 4], 4 );
}
}
} else {
- // This is a non-disposal case eithe way, so just
+ // This is a non-disposal case eithe way, so just
// leave the pixels as is, and they will become the new background
// 1: do not dispose
// 0: not specified.
}
- // background is what out is after the undoing of the previou frame;
- memcpy( g->background, g->out, 4 * g->w * g->h );
+ // background is what out is after the undoing of the previou frame;
+ memcpy( g->background, g->out, 4 * g->w * g->h );
}
- // clear my history;
+ // clear my history;
memset( g->history, 0x00, g->w * g->h ); // pixels that were affected previous frame
for (;;) {
- int tag = stbi__get8(s);
+ int tag = stbi__get8(s);
switch (tag) {
case 0x2C: /* Image Descriptor */
{
@@ -6484,6 +6881,13 @@
g->cur_x = g->start_x;
g->cur_y = g->start_y;
+ // if the width of the specified rectangle is 0, that means
+ // we may not see *any* pixels or the image is malformed;
+ // to make sure this is caught, move the current y down to
+ // max_y (which is what out_gif_code checks).
+ if (w == 0)
+ g->cur_y = g->max_y;
+
g->lflags = stbi__get8(s);
if (g->lflags & 0x40) {
@@ -6500,19 +6904,19 @@
} 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 == NULL) return NULL;
+ return stbi__errpuc("missing color table", "Corrupt GIF");
- // if this was the first frame,
- pcount = g->w * g->h;
+ o = stbi__process_gif_raster(s, g);
+ if (!o) return NULL;
+
+ // if this was the first frame,
+ pcount = g->w * g->h;
if (first_frame && (g->bgindex > 0)) {
// if first frame, any pixel not drawn to gets the background color
for (pi = 0; pi < pcount; ++pi) {
if (g->history[pi] == 0) {
- g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It will be reset next frame if need be;
- memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 );
+ g->pal[g->bgindex][3] = 255; // just in case it was made transparent, undo that; It will be reset next frame if need be;
+ memcpy( &g->out[pi * 4], &g->pal[g->bgindex], 4 );
}
}
}
@@ -6523,7 +6927,7 @@
case 0x21: // Comment Extension.
{
int len;
- int ext = stbi__get8(s);
+ int ext = stbi__get8(s);
if (ext == 0xF9) { // Graphic Control Extension.
len = stbi__get8(s);
if (len == 4) {
@@ -6532,23 +6936,23 @@
// unset old transparent
if (g->transparent >= 0) {
- g->pal[g->transparent][3] = 255;
- }
+ 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;
+ g->pal[g->transparent][3] = 0;
}
} else {
// don't need transparent
- stbi__skip(s, 1);
- g->transparent = -1;
+ stbi__skip(s, 1);
+ g->transparent = -1;
}
} else {
stbi__skip(s, len);
break;
}
- }
+ }
while ((len = stbi__get8(s)) != 0) {
stbi__skip(s, len);
}
@@ -6564,18 +6968,35 @@
}
}
+static void *stbi__load_gif_main_outofmem(stbi__gif *g, stbi_uc *out, int **delays)
+{
+ STBI_FREE(g->out);
+ STBI_FREE(g->history);
+ STBI_FREE(g->background);
+
+ if (out) STBI_FREE(out);
+ if (delays && *delays) STBI_FREE(*delays);
+ return stbi__errpuc("outofmem", "Out of memory");
+}
+
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;
+ int layers = 0;
stbi_uc *u = 0;
stbi_uc *out = 0;
- stbi_uc *two_back = 0;
+ stbi_uc *two_back = 0;
stbi__gif g;
- int stride;
+ int stride;
+ int out_size = 0;
+ int delays_size = 0;
+
+ STBI_NOTUSED(out_size);
+ STBI_NOTUSED(delays_size);
+
memset(&g, 0, sizeof(g));
if (delays) {
- *delays = 0;
+ *delays = 0;
}
do {
@@ -6585,44 +7006,61 @@
if (u) {
*x = g.w;
*y = g.h;
- ++layers;
- stride = g.w * g.h * 4;
-
+ ++layers;
+ stride = g.w * g.h * 4;
+
if (out) {
- out = (stbi_uc*) STBI_REALLOC( out, layers * stride );
+ void *tmp = (stbi_uc*) STBI_REALLOC_SIZED( out, out_size, layers * stride );
+ if (!tmp)
+ return stbi__load_gif_main_outofmem(&g, out, delays);
+ else {
+ out = (stbi_uc*) tmp;
+ out_size = layers * stride;
+ }
+
if (delays) {
- *delays = (int*) STBI_REALLOC( *delays, sizeof(int) * layers );
+ int *new_delays = (int*) STBI_REALLOC_SIZED( *delays, delays_size, sizeof(int) * layers );
+ if (!new_delays)
+ return stbi__load_gif_main_outofmem(&g, out, delays);
+ *delays = new_delays;
+ delays_size = layers * sizeof(int);
}
} else {
- out = (stbi_uc*)stbi__malloc( layers * stride );
+ out = (stbi_uc*)stbi__malloc( layers * stride );
+ if (!out)
+ return stbi__load_gif_main_outofmem(&g, out, delays);
+ out_size = layers * stride;
if (delays) {
- *delays = (int*) stbi__malloc( layers * sizeof(int) );
+ *delays = (int*) stbi__malloc( layers * sizeof(int) );
+ if (!*delays)
+ return stbi__load_gif_main_outofmem(&g, out, delays);
+ delays_size = layers * sizeof(int);
}
}
- memcpy( out + ((layers - 1) * stride), u, stride );
+ memcpy( out + ((layers - 1) * stride), u, stride );
if (layers >= 2) {
- two_back = out - 2 * stride;
+ two_back = out - 2 * stride;
}
if (delays) {
- (*delays)[layers - 1U] = g.delay;
+ (*delays)[layers - 1U] = g.delay;
}
}
- } while (u != 0);
+ } while (u != 0);
- // free temp buffer;
- STBI_FREE(g.out);
- STBI_FREE(g.history);
- STBI_FREE(g.background);
+ // free temp buffer;
+ STBI_FREE(g.out);
+ STBI_FREE(g.history);
+ STBI_FREE(g.background);
- // do the final conversion after loading everything;
+ // do the final conversion after loading everything;
if (req_comp && req_comp != 4)
out = stbi__convert_format(out, 4, req_comp, layers * g.w, g.h);
- *z = layers;
+ *z = layers;
return out;
} else {
- return stbi__errpuc("not GIF", "Image was not as a gif type.");
+ return stbi__errpuc("not GIF", "Image was not as a gif type.");
}
}
@@ -6640,14 +7078,17 @@
*y = g.h;
// moved conversion to after successful load so that the same
- // can be done for multiple frames.
+ // can be done for multiple frames.
if (req_comp && req_comp != 4)
u = stbi__convert_format(u, 4, req_comp, g.w, g.h);
+ } else if (g.out) {
+ // if there was an error and we allocated an image buffer, free it!
+ STBI_FREE(g.out);
}
- // free buffers needed for multiple frame loading;
+ // free buffers needed for multiple frame loading;
STBI_FREE(g.history);
- STBI_FREE(g.background);
+ STBI_FREE(g.background);
return u;
}
@@ -6772,6 +7213,9 @@
token += 3;
width = (int) strtol(token, NULL, 10);
+ if (height > STBI_MAX_DIMENSIONS) return stbi__errpf("too large","Very large image (corrupt?)");
+ if (width > STBI_MAX_DIMENSIONS) return stbi__errpf("too large","Very large image (corrupt?)");
+
*x = width;
*y = height;
@@ -6840,12 +7284,12 @@
// Run
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"); }
+ if ((count == 0) || (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 {
// Dump
- if (count > nleft) { STBI_FREE(hdr_data); STBI_FREE(scanline); return stbi__errpf("corrupt", "bad RLE data in HDR"); }
+ if ((count == 0) || (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);
}
@@ -6914,12 +7358,18 @@
info.all_a = 255;
p = stbi__bmp_parse_header(s, &info);
- stbi__rewind( s );
- if (p == NULL)
+ if (p == NULL) {
+ stbi__rewind( s );
return 0;
+ }
if (x) *x = s->img_x;
if (y) *y = s->img_y;
- if (comp) *comp = info.ma ? 4 : 3;
+ if (comp) {
+ if (info.bpp == 24 && info.ma == 0xff000000)
+ *comp = 3;
+ else
+ *comp = info.ma ? 4 : 3;
+ }
return 1;
}
#endif
@@ -6977,8 +7427,8 @@
stbi__rewind( s );
return 0;
}
- (void) stbi__get32be(s);
- (void) stbi__get32be(s);
+ STBI_NOTUSED(stbi__get32be(s));
+ STBI_NOTUSED(stbi__get32be(s));
depth = stbi__get16be(s);
if (depth != 16) {
stbi__rewind( s );
@@ -7057,7 +7507,6 @@
// Known limitations:
// Does not support comments in the header section
// Does not support ASCII image data (formats P2 and P3)
-// Does not support 16-bit-per-channel
#ifndef STBI_NO_PNM
@@ -7078,22 +7527,33 @@
stbi_uc *out;
STBI_NOTUSED(ri);
- if (!stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n))
+ ri->bits_per_channel = stbi__pnm_info(s, (int *)&s->img_x, (int *)&s->img_y, (int *)&s->img_n);
+ if (ri->bits_per_channel == 0)
return 0;
+ if (s->img_y > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+ if (s->img_x > STBI_MAX_DIMENSIONS) return stbi__errpuc("too large","Very large image (corrupt?)");
+
*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))
+ if (!stbi__mad4sizes_valid(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 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);
+ out = (stbi_uc *) stbi__malloc_mad4(s->img_n, s->img_x, s->img_y, ri->bits_per_channel / 8, 0);
if (!out) return stbi__errpuc("outofmem", "Out of memory");
- stbi__getn(s, out, s->img_n * s->img_x * s->img_y);
+ if (!stbi__getn(s, out, s->img_n * s->img_x * s->img_y * (ri->bits_per_channel / 8))) {
+ STBI_FREE(out);
+ return stbi__errpuc("bad PNM", "PNM file truncated");
+ }
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 (ri->bits_per_channel == 16) {
+ out = (stbi_uc *) stbi__convert_format16((stbi__uint16 *) out, s->img_n, req_comp, s->img_x, s->img_y);
+ } else {
+ out = stbi__convert_format(out, s->img_n, req_comp, s->img_x, s->img_y);
+ }
if (out == NULL) return out; // stbi__convert_format frees input on failure
}
return out;
@@ -7130,6 +7590,8 @@
while (!stbi__at_eof(s) && stbi__pnm_isdigit(*c)) {
value = value*10 + (*c - '0');
*c = (char) stbi__get8(s);
+ if((value > 214748364) || (value == 214748364 && *c > '7'))
+ return stbi__err("integer parse overflow", "Parsing an integer in the PPM header overflowed a 32-bit int");
}
return value;
@@ -7160,17 +7622,29 @@
stbi__pnm_skip_whitespace(s, &c);
*x = stbi__pnm_getinteger(s, &c); // read width
+ if(*x == 0)
+ return stbi__err("invalid width", "PPM image header had zero or overflowing width");
stbi__pnm_skip_whitespace(s, &c);
*y = stbi__pnm_getinteger(s, &c); // read height
+ if (*y == 0)
+ return stbi__err("invalid width", "PPM image header had zero or overflowing width");
stbi__pnm_skip_whitespace(s, &c);
maxv = stbi__pnm_getinteger(s, &c); // read max value
-
- if (maxv > 255)
- return stbi__err("max value > 255", "PPM image not 8-bit");
+ if (maxv > 65535)
+ return stbi__err("max value > 65535", "PPM image supports only 8-bit and 16-bit images");
+ else if (maxv > 255)
+ return 16;
else
- return 1;
+ return 8;
+}
+
+static int stbi__pnm_is16(stbi__context *s)
+{
+ if (stbi__pnm_info(s, NULL, NULL, NULL) == 16)
+ return 1;
+ return 0;
}
#endif
@@ -7226,6 +7700,9 @@
if (stbi__psd_is16(s)) return 1;
#endif
+ #ifndef STBI_NO_PNM
+ if (stbi__pnm_is16(s)) return 1;
+ #endif
return 0;
}
@@ -7305,7 +7782,7 @@
/*
revision history:
- 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs
+ 2.20 (2019-02-07) support utf8 filenames in Windows; fix warnings and platform ifdefs
2.19 (2018-02-11) fix warning
2.18 (2018-01-30) fix warnings
2.17 (2018-01-29) change sbti__shiftsigned to avoid clang -O2 bug
diff --git a/stb_image_write.h b/stb_image_write.h
index ebaa0c5..e4b32ed 100644
--- a/stb_image_write.h
+++ b/stb_image_write.h
@@ -1,4 +1,4 @@
-/* stb_image_write - v1.11 - public domain - http://nothings.org/stb/stb_image_write.h
+/* stb_image_write - v1.16 - public domain - http://nothings.org/stb
writes out PNG/BMP/TGA/JPEG/HDR images to C stdio - Sean Barrett 2010-2015
no warranty implied; use at your own risk
@@ -10,11 +10,6 @@
Will probably not work correctly with strict-aliasing optimizations.
- If using a modern Microsoft Compiler, non-safe versions of CRT calls may cause
- compilation warnings or even errors. To avoid this, also before #including,
-
- #define STBI_MSC_SECURE_CRT
-
ABOUT:
This header file is a library for writing images to C stdio or a callback.
@@ -110,7 +105,7 @@
TGA supports RLE or non-RLE compressed data. To use non-RLE-compressed
data, set the global variable 'stbi_write_tga_with_rle' to 0.
-
+
JPEG does ignore alpha channels in input data; quality is between 1 and 100.
Higher quality looks better but results in a bigger image.
JPEG baseline (no JPEG progressive).
@@ -118,7 +113,7 @@
CREDITS:
- Sean Barrett - PNG/BMP/TGA
+ Sean Barrett - PNG/BMP/TGA
Baldur Karlsson - HDR
Jean-Sebastien Guay - TGA monochrome
Tim Kelsey - misc enhancements
@@ -145,6 +140,7 @@
Ivan Tikhonov
github:ignotion
Adam Schackart
+ Andrew Kensler
LICENSE
@@ -171,9 +167,9 @@
#endif
#ifndef STB_IMAGE_WRITE_STATIC // C++ forbids static forward declarations
-extern int stbi_write_tga_with_rle;
-extern int stbi_write_png_compression_level;
-extern int stbi_write_force_png_filter;
+STBIWDEF int stbi_write_tga_with_rle;
+STBIWDEF int stbi_write_png_compression_level;
+STBIWDEF int stbi_write_force_png_filter;
#endif
#ifndef STBI_WRITE_NO_STDIO
@@ -183,7 +179,7 @@
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
+#ifdef STBIW_WINDOWS_UTF8
STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input);
#endif
#endif
@@ -252,17 +248,17 @@
#define STBIW_UCHAR(x) (unsigned char) ((x) & 0xff)
#ifdef STB_IMAGE_WRITE_STATIC
-static int stbi__flip_vertically_on_write=0;
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__flip_vertically_on_write=0;
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;
@@ -272,6 +268,8 @@
{
stbi_write_func *func;
void *context;
+ unsigned char buffer[64];
+ int buf_used;
} stbi__write_context;
// initialize a callback-based context
@@ -288,7 +286,7 @@
fwrite(data,1,size,(FILE*) context);
}
-#if defined(_MSC_VER) && defined(STBI_WINDOWS_UTF8)
+#if defined(_WIN32) && defined(STBIW_WINDOWS_UTF8)
#ifdef __cplusplus
#define STBIW_EXTERN extern "C"
#else
@@ -299,25 +297,25 @@
STBIWDEF int stbiw_convert_wchar_to_utf8(char *buffer, size_t bufferlen, const wchar_t* input)
{
- return WideCharToMultiByte(65001 /* UTF8 */, 0, input, -1, buffer, bufferlen, NULL, NULL);
+ return WideCharToMultiByte(65001 /* UTF8 */, 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)
+#if defined(_WIN32) && defined(STBIW_WINDOWS_UTF8)
wchar_t wMode[64];
wchar_t wFilename[1024];
- if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename)))
- return 0;
-
- if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode)))
+ if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, filename, -1, wFilename, sizeof(wFilename)/sizeof(*wFilename)))
return 0;
-#if _MSC_VER >= 1400
- if (0 != _wfopen_s(&f, wFilename, wMode))
- f = 0;
+ if (0 == MultiByteToWideChar(65001 /* UTF8 */, 0, mode, -1, wMode, sizeof(wMode)/sizeof(*wMode)))
+ return 0;
+
+#if defined(_MSC_VER) && _MSC_VER >= 1400
+ if (0 != _wfopen_s(&f, wFilename, wMode))
+ f = 0;
#else
f = _wfopen(wFilename, wMode);
#endif
@@ -385,16 +383,36 @@
va_end(v);
}
+static void stbiw__write_flush(stbi__write_context *s)
+{
+ if (s->buf_used) {
+ s->func(s->context, &s->buffer, s->buf_used);
+ s->buf_used = 0;
+ }
+}
+
static void stbiw__putc(stbi__write_context *s, unsigned char c)
{
s->func(s->context, &c, 1);
}
+static void stbiw__write1(stbi__write_context *s, unsigned char a)
+{
+ if ((size_t)s->buf_used + 1 > sizeof(s->buffer))
+ stbiw__write_flush(s);
+ s->buffer[s->buf_used++] = a;
+}
+
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);
+ int n;
+ if ((size_t)s->buf_used + 3 > sizeof(s->buffer))
+ stbiw__write_flush(s);
+ n = s->buf_used;
+ s->buf_used = n+3;
+ s->buffer[n+0] = a;
+ s->buffer[n+1] = b;
+ s->buffer[n+2] = c;
}
static void stbiw__write_pixel(stbi__write_context *s, int rgb_dir, int comp, int write_alpha, int expand_mono, unsigned char *d)
@@ -403,7 +421,7 @@
int k;
if (write_alpha < 0)
- s->func(s->context, &d[comp - 1], 1);
+ stbiw__write1(s, d[comp - 1]);
switch (comp) {
case 2: // 2 pixels = mono + alpha, alpha is written separately, so same as 1-channel case
@@ -411,7 +429,7 @@
if (expand_mono)
stbiw__write3(s, d[0], d[0], d[0]); // monochrome bmp
else
- s->func(s->context, d, 1); // monochrome TGA
+ stbiw__write1(s, d[0]); // monochrome TGA
break;
case 4:
if (!write_alpha) {
@@ -427,7 +445,7 @@
break;
}
if (write_alpha > 0)
- s->func(s->context, &d[comp - 1], 1);
+ stbiw__write1(s, d[comp - 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)
@@ -441,16 +459,18 @@
if (stbi__flip_vertically_on_write)
vdir *= -1;
- if (vdir < 0)
- j_end = -1, j = y-1;
- else
- j_end = y, j = 0;
+ 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);
}
+ stbiw__write_flush(s);
s->func(s->context, &zero, scanline_pad);
}
}
@@ -471,16 +491,27 @@
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, // file header
- 40, x,y, 1,24, 0,0,0,0,0,0); // bitmap header
+ if (comp != 4) {
+ // write RGB bitmap
+ 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, // file header
+ 40, x,y, 1,24, 0,0,0,0,0,0); // bitmap header
+ } else {
+ // RGBA bitmaps need a v4 header
+ // use BI_BITFIELDS mode with 32bpp and alpha mask
+ // (straight BI_RGB with alpha mask doesn't work in most readers)
+ return stbiw__outfile(s,-1,-1,x,y,comp,1,(void *)data,1,0,
+ "11 4 22 4" "4 44 22 444444 4444 4 444 444 444 444",
+ 'B', 'M', 14+108+x*y*4, 0, 0, 14+108, // file header
+ 108, x,y, 1,32, 3,0,0,0,0,0, 0xff0000,0xff00,0xff,0xff000000u, 0, 0,0,0, 0,0,0, 0,0,0, 0,0,0); // bitmap V4 header
+ }
}
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__write_context s = { 0 };
stbi__start_write_callbacks(&s, func, context);
return stbi_write_bmp_core(&s, x, y, comp, data);
}
@@ -488,7 +519,7 @@
#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;
+ stbi__write_context s = { 0 };
if (stbi__start_write_file(&s,filename)) {
int r = stbi_write_bmp_core(&s, x, y, comp, data);
stbi__end_write_file(&s);
@@ -561,24 +592,25 @@
if (diff) {
unsigned char header = STBIW_UCHAR(len - 1);
- s->func(s->context, &header, 1);
+ stbiw__write1(s, header);
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__write1(s, header);
stbiw__write_pixel(s, -1, comp, has_alpha, 0, begin);
}
}
}
+ stbiw__write_flush(s);
}
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__write_context s = { 0 };
stbi__start_write_callbacks(&s, func, context);
return stbi_write_tga_core(&s, x, y, comp, (void *) data);
}
@@ -586,7 +618,7 @@
#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;
+ stbi__write_context s = { 0 };
if (stbi__start_write_file(&s,filename)) {
int r = stbi_write_tga_core(&s, x, y, comp, (void *) data);
stbi__end_write_file(&s);
@@ -602,6 +634,8 @@
#define stbiw__max(a, b) ((a) > (b) ? (a) : (b))
+#ifndef STBI_WRITE_NO_STDIO
+
static void stbiw__linear_to_rgbe(unsigned char *rgbe, float *linear)
{
int exponent;
@@ -736,8 +770,8 @@
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 STBI_MSC_SECURE_CRT
- len = sprintf_s(buffer, "EXPOSURE= 1.0000000000000\n\n-Y %d +X %d\n", y, x);
+#ifdef __STDC_LIB_EXT1__
+ 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
@@ -752,15 +786,14 @@
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__write_context s = { 0 };
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;
+ stbi__write_context s = { 0 };
if (stbi__start_write_file(&s,filename)) {
int r = stbi_write_hdr_core(&s, x, y, comp, (float *) data);
stbi__end_write_file(&s);
@@ -778,7 +811,7 @@
#ifndef STBIW_ZLIB_COMPRESS
// stretchy buffer; stbiw__sbpush() == vector<>::push_back() -- stbiw__sbcount() == vector<>::size()
-#define stbiw__sbraw(a) ((int *) (a) - 2)
+#define stbiw__sbraw(a) ((int *) (void *) (a) - 2)
#define stbiw__sbm(a) stbiw__sbraw(a)[0]
#define stbiw__sbn(a) stbiw__sbraw(a)[1]
@@ -872,7 +905,7 @@
unsigned int bitbuf=0;
int i,j, bitcount=0;
unsigned char *out = NULL;
- unsigned char ***hash_table = (unsigned char***) STBIW_MALLOC(stbiw__ZHASH * sizeof(char**));
+ unsigned char ***hash_table = (unsigned char***) STBIW_MALLOC(stbiw__ZHASH * sizeof(unsigned char**));
if (hash_table == NULL)
return NULL;
if (quality < 5) quality = 5;
@@ -895,7 +928,7 @@
for (j=0; j < n; ++j) {
if (hlist[j]-data > i-32768) { // if entry lies within window
int d = stbiw__zlib_countm(hlist[j], data+i, data_len-i);
- if (d >= best) best=d,bestloc=hlist[j];
+ if (d >= best) { best=d; bestloc=hlist[j]; }
}
}
// when hash table entry is too long, delete half the entries
@@ -948,14 +981,31 @@
(void) stbiw__sbfree(hash_table[i]);
STBIW_FREE(hash_table);
+ // store uncompressed instead if compression was worse
+ if (stbiw__sbn(out) > data_len + 2 + ((data_len+32766)/32767)*5) {
+ stbiw__sbn(out) = 2; // truncate to DEFLATE 32K window and FLEVEL = 1
+ for (j = 0; j < data_len;) {
+ int blocklen = data_len - j;
+ if (blocklen > 32767) blocklen = 32767;
+ stbiw__sbpush(out, data_len - j == blocklen); // BFINAL = ?, BTYPE = 0 -- no compression
+ stbiw__sbpush(out, STBIW_UCHAR(blocklen)); // LEN
+ stbiw__sbpush(out, STBIW_UCHAR(blocklen >> 8));
+ stbiw__sbpush(out, STBIW_UCHAR(~blocklen)); // NLEN
+ stbiw__sbpush(out, STBIW_UCHAR(~blocklen >> 8));
+ memcpy(out+stbiw__sbn(out), data+j, blocklen);
+ stbiw__sbn(out) += blocklen;
+ j += blocklen;
+ }
+ }
+
{
// compute adler32 on input
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;
+ for (i=0; i < blocklen; ++i) { s1 += data[j+i]; s2 += s1; }
+ s1 %= 65521; s2 %= 65521;
j += blocklen;
blocklen = 5552;
}
@@ -1048,13 +1098,13 @@
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;
}
- // first loop isn't optimized since it's just one pixel
+ // first loop isn't optimized since it's just one pixel
for (i = 0; i < n; ++i) {
switch (type) {
case 1: line_buffer[i] = z[i]; break;
@@ -1275,26 +1325,31 @@
bits[0] = val & ((1<<bits[1])-1);
}
-static int stbiw__jpg_processDU(stbi__write_context *s, int *bitBuf, int *bitCnt, float *CDU, float *fdtbl, int DC, const unsigned short HTDC[256][2], const unsigned short HTAC[256][2]) {
+static int stbiw__jpg_processDU(stbi__write_context *s, int *bitBuf, int *bitCnt, float *CDU, int du_stride, float *fdtbl, int DC, const unsigned short HTDC[256][2], const unsigned short HTAC[256][2]) {
const unsigned short EOB[2] = { HTAC[0x00][0], HTAC[0x00][1] };
const unsigned short M16zeroes[2] = { HTAC[0xF0][0], HTAC[0xF0][1] };
- int dataOff, i, diff, end0pos;
+ int dataOff, i, j, n, diff, end0pos, x, y;
int DU[64];
// DCT rows
- for(dataOff=0; dataOff<64; dataOff+=8) {
+ for(dataOff=0, n=du_stride*8; dataOff<n; dataOff+=du_stride) {
stbiw__jpg_DCT(&CDU[dataOff], &CDU[dataOff+1], &CDU[dataOff+2], &CDU[dataOff+3], &CDU[dataOff+4], &CDU[dataOff+5], &CDU[dataOff+6], &CDU[dataOff+7]);
}
// DCT columns
for(dataOff=0; dataOff<8; ++dataOff) {
- stbiw__jpg_DCT(&CDU[dataOff], &CDU[dataOff+8], &CDU[dataOff+16], &CDU[dataOff+24], &CDU[dataOff+32], &CDU[dataOff+40], &CDU[dataOff+48], &CDU[dataOff+56]);
+ stbiw__jpg_DCT(&CDU[dataOff], &CDU[dataOff+du_stride], &CDU[dataOff+du_stride*2], &CDU[dataOff+du_stride*3], &CDU[dataOff+du_stride*4],
+ &CDU[dataOff+du_stride*5], &CDU[dataOff+du_stride*6], &CDU[dataOff+du_stride*7]);
}
// Quantize/descale/zigzag the coefficients
- for(i=0; i<64; ++i) {
- float v = CDU[i]*fdtbl[i];
- // DU[stbiw__jpg_ZigZag[i]] = (int)(v < 0 ? ceilf(v - 0.5f) : floorf(v + 0.5f));
- // ceilf() and floorf() are C99, not C89, but I /think/ they're not needed here anyway?
- DU[stbiw__jpg_ZigZag[i]] = (int)(v < 0 ? v - 0.5f : v + 0.5f);
+ for(y = 0, j=0; y < 8; ++y) {
+ for(x = 0; x < 8; ++x,++j) {
+ float v;
+ i = y*du_stride+x;
+ v = CDU[i]*fdtbl[j];
+ // DU[stbiw__jpg_ZigZag[j]] = (int)(v < 0 ? ceilf(v - 0.5f) : floorf(v + 0.5f));
+ // ceilf() and floorf() are C99, not C89, but I /think/ they're not needed here anyway?
+ DU[stbiw__jpg_ZigZag[j]] = (int)(v < 0 ? v - 0.5f : v + 0.5f);
+ }
}
// Encode DC
@@ -1409,10 +1464,10 @@
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,
+ 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;
+ int row, col, i, k, subsample;
float fdtbl_Y[64], fdtbl_UV[64];
unsigned char YTable[64], UVTable[64];
@@ -1421,6 +1476,7 @@
}
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;
@@ -1443,7 +1499,7 @@
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),
- 3,1,0x11,0,2,0x11,1,3,0x11,1,0xFF,0xC4,0x01,0xA2,0 };
+ 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);
@@ -1466,36 +1522,74 @@
// Encode 8x8 macroblocks
{
static const unsigned short fillBits[] = {0x7F, 7};
- const unsigned char *imageData = (const unsigned char *)data;
int DCY=0, DCU=0, DCV=0;
int bitBuf=0, bitCnt=0;
// comp == 2 is grey+alpha (alpha is ignored)
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;
- for(y = 0; y < height; y += 8) {
- for(x = 0; x < width; x += 8) {
- float YDU[64], UDU[64], VDU[64];
- for(row = y, pos = 0; row < y+8; ++row) {
- // row >= height => use last input 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) {
- float r, g, b;
- // if col >= width => use pixel from last input column
- int p = base_p + ((col < width) ? col : (width-1))*comp;
+ 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) {
+ // row >= height => use last input 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) {
+ // if col >= width => use pixel from last input column
+ 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);
- r = imageData[p+0];
- g = imageData[p+ofsG];
- b = imageData[p+ofsB];
- YDU[pos]=+0.29900f*r+0.58700f*g+0.11400f*b-128;
- UDU[pos]=-0.16874f*r-0.33126f*g+0.50000f*b;
- VDU[pos]=+0.50000f*r-0.41869f*g-0.08131f*b;
+ // subsample U,V
+ {
+ 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) {
+ // row >= height => use last input 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) {
+ // if col >= width => use pixel from last input column
+ 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, YDU, fdtbl_Y, DCY, YDC_HT, YAC_HT);
- DCU = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, UDU, fdtbl_UV, DCU, UVDC_HT, UVAC_HT);
- DCV = stbiw__jpg_processDU(s, &bitBuf, &bitCnt, VDU, fdtbl_UV, DCV, UVDC_HT, UVAC_HT);
+ 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);
+ }
}
}
@@ -1512,7 +1606,7 @@
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__write_context s = { 0 };
stbi__start_write_callbacks(&s, func, context);
return stbi_write_jpg_core(&s, x, y, comp, (void *) data, quality);
}
@@ -1521,7 +1615,7 @@
#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;
+ stbi__write_context s = { 0 };
if (stbi__start_write_file(&s,filename)) {
int r = stbi_write_jpg_core(&s, x, y, comp, data, quality);
stbi__end_write_file(&s);
@@ -1534,8 +1628,17 @@
#endif // STB_IMAGE_WRITE_IMPLEMENTATION
/* Revision history
+ 1.16 (2021-07-11)
+ make Deflate code emit uncompressed blocks when it would otherwise expand
+ support writing BMPs with alpha channel
+ 1.15 (2020-07-13) unknown
+ 1.14 (2020-02-02) updated JPEG writer to downsample chroma channels
+ 1.13
+ 1.12
+ 1.11 (2019-08-11)
+
1.10 (2019-02-07)
- support utf8 filenames in Windows; fix warnings and platform ifdefs
+ support utf8 filenames in Windows; fix warnings and platform ifdefs
1.09 (2018-02-11)
fix typo in zlib quality API, improve STB_I_W_STATIC in C++
1.08 (2018-01-29)
@@ -1566,7 +1669,7 @@
add HDR output
fix monochrome BMP
0.95 (2014-08-17)
- add monochrome TGA output
+ add monochrome TGA output
0.94 (2014-05-31)
rename private functions to avoid conflicts with stb_image.h
0.93 (2014-05-27)
@@ -1584,38 +1687,38 @@
------------------------------------------------------------------------------
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
+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
+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
+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,
+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
+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
+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.
------------------------------------------------------------------------------
*/