1 #if 0
   2 CHANGES - changes for libpng
   4 version 0.1 [March 29, 1995]
   5   initial work-in-progress release
   7 version 0.2 [April 1, 1995]
   8   added reader into png.h
   9   fixed small problems in stub file
  11 version 0.3 [April 8, 1995]
  12   added pull reader
  13   split up pngwrite.c to several files
  14   added pnglib.txt
  15   added example.c
  16   cleaned up writer, adding a few new transformations
  17   fixed some bugs in writer
  18   interfaced with zlib 0.5
  19   added K&R support
  20   added check for 64 KB blocks for 16 bit machines
  22 version 0.4 [April 26, 1995]
  23   cleaned up code and commented code
  24   simplified time handling into png_time
  25   created png_color_16 and png_color_8 to handle color needs
  26   cleaned up color type defines
  27   fixed various bugs
  28   made various names more consistent
  29   interfaced with zlib 0.71
  30   cleaned up zTXt reader and writer (using zlib's Reset functions)
  31   split transformations into pngrtran.c and pngwtran.c
  33 version 0.5 [April 30, 1995]
  34   interfaced with zlib 0.8
  35   fixed many reading and writing bugs
  36   saved using 3 spaces instead of tabs
  38 version 0.6 [May 1, 1995]
  39   first beta release
  40   added png_large_malloc() and png_large_free()
  41   added png_size_t
  42   cleaned up some compiler warnings
  43   added png_start_read_image()
  45 version 0.7 [June 24, 1995]
  46   cleaned up lots of bugs
  47   finished dithering and other stuff
  48   added test program
  49   changed name from pnglib to libpng
  51 version 0.71 [June 26, 1995]
  52   changed pngtest.png for zlib 0.93
  53   fixed error in libpng.txt and example.c
  55 version 0.8 [August 20, 1995]
  56   cleaned up some bugs
  57   added png_set_filler()
  58   split up pngstub.c into pngmem.c, pngio.c, and pngerror.c
  59   added #define's to remove unwanted code
  60   moved png_info_init() to png.c
  61   added old_size into png_realloc()
  62   added functions to manually set filtering and compression info
  63   changed compression parameters based on image type
  64   optimized filter selection code
  65   added version info
  66   changed external functions passing floats to doubles (k&r problems?)
  67   put all the configurable stuff in pngconf.h
  68   enabled png_set_shift to work with paletted images on read
  69   added png_read_update_info() - updates info structure with transformations
  71 Version 0.81 [August, 1995]
  72   incorporated Tim Wegner's medium model code (thanks, Tim)
  74 Version 0.82 [September, 1995]
  75   [unspecified changes]
  77 Version 0.85 [December, 1995]
  78   added more medium model code (almost everything's a far)
  79   added i/o, error, and memory callback functions
  80   fixed some bugs (16-bit, 4-bit interlaced, etc.)
  81   added first run progressive reader (barely tested)
  83 Version 0.86 [January, 1996]
  84   fixed bugs
  85   improved documentation
  87 Version 0.87 [January, 1996]
  88   fixed medium model bugs
  89   fixed other bugs introduced in 0.85 and 0.86
  90   added some minor documentation
  92 Version 0.88 [January, 1996]
  93   fixed progressive bugs
  94   replaced tabs with spaces
  95   cleaned up documentation
  96   added callbacks for read/write and warning/error functions
  98 Version 0.89 [June 5, 1996]
  99   Added new initialization API to make libpng work better with shared libs
 100     we now have png_create_read_struct(), png_create_write_struct(),
 101     png_create_info_struct(), png_destroy_read_struct(), and
 102     png_destroy_write_struct() instead of the separate calls to
 103     malloc and png_read_init(), png_info_init(), and png_write_init()
 104   Changed warning/error callback functions to fix bug - this means you
 105     should use the new initialization API if you were using the old
 106     png_set_message_fn() calls, and that the old API no longer exists
 107     so that people are aware that they need to change their code
 108   Changed filter selection API to allow selection of multiple filters
 109     since it didn't work in previous versions of libpng anyways
 110   Optimized filter selection code
 111   Fixed png_set_background() to allow using an arbitrary RGB color for
 112     paletted images
 113   Fixed gamma and background correction for paletted images, so
 114     png_correct_palette is not needed unless you are correcting an
 115     external palette (you will need to #define PNG_CORRECT_PALETTE_SUPPORTED
 116     in pngconf.h) - if nobody uses this, it may disappear in the future.
 117   Fixed bug with Borland 64K memory allocation (Alexander Lehmann)
 118   Fixed bug in interlace handling (Smarasderagd, I think)
 119   Added more error checking for writing and image to reduce invalid files
 120   Separated read and write functions so that they won't both be linked
 121     into a binary when only reading or writing functionality is used
 122   New pngtest image also has interlacing and zTXt
 123   Updated documentation to reflect new API
 125 Version 0.89c [June 17, 1996]
 126   Bug fixes.
 128 Version 0.90 [January, 1997]
 129   Made CRC errors/warnings on critical and ancillary chunks configurable
 130   libpng will use the zlib CRC routines by (compile-time) default
 131   Changed DOS small/medium model memory support - needs zlib 1.04 (Tim Wegner)
 132   Added external C++ wrapper statements to png.h (Gilles Dauphin)
 133   Allow PNG file to be read when some or all of file signature has already
 134     been read from the beginning of the stream.  ****This affects the size
 135     of info_struct and invalidates all programs that use a shared libpng****
 136   Fixed png_filler() declarations
 137   Fixed? background color conversions
 138   Fixed order of error function pointers to match documentation
 139   Current chunk name is now available in png_struct to reduce the number
 140     of nearly identical error messages (will simplify multi-lingual
 141     support when available)
 142   Try to get ready for unknown-chunk callback functions:
 143     - previously read critical chunks are flagged, so the chunk handling
 144       routines can determine if the chunk is in the right place
 145     - all chunk handling routines have the same prototypes, so we will
 146       be able to handle all chunks via a callback mechanism
 147   Try to fix Linux "setjmp" buffer size problems
 148   Removed png_large_malloc, png_large_free, and png_realloc functions.
 150 Version 0.95 [March, 1997]
 151   Fixed bug in pngwutil.c allocating "up_row" twice and "avg_row" never
 152   Fixed bug in PNG file signature compares when start != 0
 153   Changed parameter type of png_set_filler(...filler...) from png_byte
 154     to png_uint_32
 155   Added test for MACOS to ensure that both math.h and fp.h are not #included
 156   Added macros for libpng to be compiled as a Windows DLL (Andreas Kupries)
 157   Added "packswap" transformation, which changes the endianness of
 158     packed-pixel bytes (Kevin Bracey)
 159   Added "strip_alpha" transformation, which removes the alpha channel of
 160     input images without using it (not necessarily a good idea)
 161   Added "swap_alpha" transformation, which puts the alpha channel in front
 162     of the color bytes instead of after
 163   Removed all implicit variable tests which assume NULL == 0 (I think)
 164   Changed several variables to "png_size_t" to show 16/32-bit limitations
 165   Added new pCAL chunk read/write support
 166   Added experimental filter selection weighting (Greg Roelofs)
 167   Removed old png_set_rgbx() and png_set_xrgb() functions that have been
 168     obsolete for about 2 years now (use png_set_filler() instead)
 169   Added macros to read 16- and 32-bit ints directly from buffer, to be
 170     used only on those systems that support it (namely PowerPC and 680x0)
 171     With some testing, this may become the default for MACOS/PPC systems.
 172   Only calculate CRC on data if we are going to use it
 173   Added macros for zTXt compression type PNG_zTXt_COMPRESSION_???
 174   Added macros for simple libpng debugging output selectable at compile time
 175   Removed PNG_READ_END_MODE in progressive reader (Smarasderagd)
 176   More description of info_struct in libpng.txt and png.h
 177   More instructions in example.c
 178   More chunk types tested in pngtest.c
 179   Renamed pngrcb.c to pngset.c, and all png_read_<chunk> functions to be
 180     png_set_<chunk>.  We now have corresponding png_get_<chunk>
 181     functions in pngget.c to get information in info_ptr.  This isolates
 182     the application from the internal organization of png_info_struct
 183     (good for shared library implementations).
 185 Version 0.96 [May, 1997]
 186   Fixed serious bug with < 8bpp images introduced in 0.95
 187   Fixed 256-color transparency bug (Greg Roelofs)
 188   Fixed up documentation (Greg Roelofs, Laszlo Nyul)
 189   Fixed "error" in pngconf.h for Linux setjmp() behavior
 190   Fixed DOS medium model support (Tim Wegner)
 191   Fixed png_check_keyword() for case with error in static string text
 192   Added read of CRC after IEND chunk for embedded PNGs (Laszlo Nyul)
 193   Added typecasts to quiet compiler errors
 194   Added more debugging info
 196 Version 0.97 [January, 1998]
 197   Removed PNG_USE_OWN_CRC capability
 198   Relocated png_set_crc_action from pngrutil.c to pngrtran.c
 199   Fixed typecasts of "new_key", etc. (Andreas Dilger)
 200   Added RFC 1152 [sic] date support
 201   Fixed bug in gamma handling of 4-bit grayscale
 202   Added 2-bit grayscale gamma handling (Glenn R-P)
 203   Added more typecasts. 65536L becomes (png_uint_32)65536L, etc. (Glenn R-P)
 204   Minor corrections in libpng.txt
 205   Added simple sRGB support (Glenn R-P)
 206   Easier conditional compiling, e.g.,
 208     all configurable options can be selected from command-line instead
 209     of having to edit pngconf.h (Glenn R-P)
 210   Fixed memory leak in pngwrite.c (free info_ptr->text) (Glenn R-P)
 211   Added more conditions for png_do_background, to avoid changing
 212     black pixels to background when a background is supplied and
 213     no pixels are transparent
 214   Repaired PNG_NO_STDIO behavior
 215   Tested NODIV support and made it default behavior (Greg Roelofs)
 216   Added "-m" option and PNGTEST_DEBUG_MEMORY to pngtest (John Bowler)
 217   Regularized version numbering scheme and bumped shared-library major
 218     version number to 2 to avoid problems with libpng 0.89 apps
 219     (Greg Roelofs)
 221 Version 0.98 [January, 1998]
 222   Cleaned up some typos in libpng.txt and in code documentation
 223   Fixed memory leaks in pCAL chunk processing (Glenn R-P and John Bowler)
 224   Cosmetic change "display_gamma" to "screen_gamma" in pngrtran.c
 225   Changed recommendation about file_gamma for PC images to .51 from .45,
 226     in example.c and libpng.txt, added comments to distinguish between
 227     screen_gamma, viewing_gamma, and display_gamma.
 228   Changed all references to RFC1152 to read RFC1123 and changed the
 230   Added png_invert_alpha capability (Glenn R-P -- suggestion by Jon Vincent)
 231   Changed srgb_intent from png_byte to int to avoid compiler bugs
 233 Version 0.99 [January 30, 1998]
 234   Free info_ptr->text instead of end_info_ptr->text in pngread.c (John Bowler)
 235   Fixed a longstanding "packswap" bug in pngtrans.c
 236   Fixed some inconsistencies in pngconf.h that prevented compiling with
 238   Fixed some typos and made other minor rearrangement of libpng.txt (Andreas)
 239   Changed recommendation about file_gamma for PC images to .50 from .51 in
 240     example.c and libpng.txt, and changed file_gamma for sRGB images to .45
 241   Added a number of functions to access information from the png structure
 242     png_get_image_height(), etc. (Glenn R-P, suggestion by Brad Pettit)
 243   Added TARGET_MACOS similar to zlib-1.0.8
 244   Define PNG_ALWAYS_EXTERN when __MWERKS__ && WIN32 are defined
 245   Added type casting to all png_malloc() function calls
 247 Version 0.99a [January 31, 1998]
 248   Added type casts and parentheses to all returns that return a value.(Tim W.)
 250 Version 0.99b [February 4, 1998]
 251   Added type cast png_uint_32 on malloc function calls where needed.
 252   Changed type of num_hist from png_uint_32 to int (same as num_palette).
 253   Added checks for rowbytes overflow, in case png_size_t is less than 32 bits.
 254   Renamed makefile.elf to makefile.lnx.
 256 Version 0.99c [February 7, 1998]
 257   More type casting.  Removed erroneous overflow test in pngmem.c.
 258   Added png_buffered_memcpy() and png_buffered_memset(), apply them to rowbytes.
 259   Added UNIX manual pages libpng.3 (incorporating libpng.txt) and  png.5.
 261 Version 0.99d [February 11, 1998]
 262   Renamed "far_to_near()" "png_far_to_near()"
 263   Revised libpng.3
 264   Version 99c "buffered" operations didn't work as intended.  Replaced them
 265     with png_memcpy_check() and png_memset_check().
 266   Added many "if (png_ptr == NULL) return" to quell compiler warnings about
 267     unused png_ptr, mostly in pngget.c and pngset.c.
 268   Check for overlength tRNS chunk present when indexed-color PLTE is read.
 269   Cleaned up spelling errors in libpng.3/libpng.txt
 270   Corrected a problem with png_get_tRNS() which returned undefined trans array
 272 Version 0.99e [February 28, 1998]
 273   Corrected png_get_tRNS() again.
 274   Add parentheses for easier reading of pngget.c, fixed "||" should be "&&".
 275   Touched up example.c to make more of it compileable, although the entire
 276     file still can't be compiled (Willem van Schaik)
 277   Fixed a bug in png_do_shift() (Bryan Tsai)
 278   Added a space in png.h prototype for png_write_chunk_start()
 279   Replaced pngtest.png with one created with zlib 1.1.1
 280   Changed pngtest to report PASS even when file size is different (Jean-loup G.)
 281   Corrected some logic errors in png_do_invert_alpha() (Chris Patterson)
 283 Version 0.99f [March 5, 1998]
 284   Corrected a bug in pngpread() introduced in version 99c (Kevin Bracey)
 285   Moved makefiles into a "scripts" directory, and added INSTALL instruction file
 286   Added makefile.os2 and pngos2.def (A. Zabolotny) and makefile.s2x (W. Sebok)
 287   Added pointers to "note on libpng versions" in makefile.lnx and README
 288   Added row callback feature when reading and writing nonprogressive rows
 289     and added a test of this feature in pngtest.c
 290   Added user transform callbacks, with test of the feature in pngtest.c
 292 Version 0.99g [March 6, 1998, morning]
 293   Minor changes to pngtest.c to suppress compiler warnings.
 294   Removed "beta" language from documentation.
 296 Version 0.99h [March 6, 1998, evening]
 297   Minor changes to previous minor changes to pngtest.c
 300   Added user transform capability
 302 Version 1.00 [March 7, 1998]
 303   Changed several typedefs in pngrutil.c
 304   Added makefile.wat (Pawel Mrochen), updated makefile.tc3 (Willem van Schaik)
 305   Replaced "while(1)" with "for(;;)"
 306   Added PNGARG() to prototypes in pngtest.c and removed some prototypes
 307   Updated some of the makefiles (Tom Lane)
 308   Changed some typedefs (s_start, etc.) in pngrutil.c
 309   Fixed dimensions of "short_months" array in pngwrite.c
 310   Replaced ansi2knr.c with the one from jpeg-v6
 312 Version 1.0.0 [March 8, 1998]
 313   Changed name from 1.00 to 1.0.0 (Adam Costello)
 314   Added smakefile.ppc (with SCOPTIONS.ppc) for Amiga PPC (Andreas Kleinert)
 316 Version 1.0.0a [March 9, 1998]
 317   Fixed three bugs in pngrtran.c to make gamma+background handling consistent
 318     (Greg Roelofs)
 319   Changed format of the PNG_LIBPNG_VER integer to xyyzz instead of xyz
 320     for major, minor, and bugfix releases.  This is 10001. (Adam Costello,
 321     Tom Lane)
 322   Make months range from 1-12 in png_convert_to_rfc1123
 324 Version 1.0.0b [March 13, 1998]
 325   Quieted compiler complaints about two empty "for" loops in pngrutil.c
 326   Minor changes to makefile.s2x
 327   Removed #ifdef/#endif around a png_free() in pngread.c
 329 Version 1.0.1 [March 14, 1998]
 330   Changed makefile.s2x to reduce security risk of using a relative pathname
 331   Fixed some typos in the documentation (Greg).
 332   Fixed a problem with value of "channels" returned by png_read_update_info()
 334 Version 1.0.1a [April 21, 1998]
 335   Optimized Paeth calculations by replacing abs() function calls with intrinsics
 336   plus other loop optimizations. Improves avg decoding speed by about 20%.
 337   Commented out i386istic "align" compiler flags in makefile.lnx.
 338   Reduced the default warning level in some makefiles, to make them consistent.
 339   Removed references to IJG and JPEG in the ansi2knr.c copyright statement.
 340   Fixed a bug in png_do_strip_filler with XXRRGGBB => RRGGBB transformation.
 341   Added grayscale and 16-bit capability to png_do_read_filler().
 342   Fixed a bug in pngset.c, introduced in version 0.99c, that sets rowbytes
 343     too large when writing an image with bit_depth < 8 (Bob Dellaca).
 344   Corrected some bugs in the experimental weighted filtering heuristics.
 345   Moved a misplaced pngrutil code block that truncates tRNS if it has more
 346     than num_palette entries -- test was done before num_palette was defined.
 347   Fixed a png_convert_to_rfc1123() bug that converts day 31 to 0 (Steve Eddins).
 348   Changed compiler flags in makefile.wat for better optimization
 349     (Pawel Mrochen).
 351 Version 1.0.1b [May 2, 1998]
 352   Relocated png_do_gray_to_rgb() within png_do_read_transformations() (Greg).
 353   Relocated the png_composite macros from pngrtran.c to png.h (Greg).
 354   Added makefile.sco (contributed by Mike Hopkirk).
 355   Fixed two bugs (missing definitions of "istop") introduced in libpng-1.0.1a.
 356   Fixed a bug in pngrtran.c that would set channels=5 under some circumstances.
 357   More work on the Paeth-filtering, achieving imperceptible speedup
 358     (A Kleinert).
 359   More work on loop optimization which may help when compiled with C++
 360     compilers.
 361   Added warnings when people try to use transforms they've defined out.
 362   Collapsed 4 "i" and "c" loops into single "i" loops in pngrtran and pngwtran.
 363   Revised paragraph about png_set_expand() in libpng.txt and libpng.3 (Greg)
 365 Version 1.0.1c [May 11, 1998]
 366   Fixed a bug in pngrtran.c (introduced in libpng-1.0.1a) where the masks for
 367     filler bytes should have been 0xff instead of 0xf.
 368   Added max_pixel_depth=32 in pngrutil.c when using FILLER with palette images.
 370     out of the PNG_WRITE_TRANSFORMS_NOT_SUPPORTED block of pngconf.h
 371   Added "PNG_NO_WRITE_TRANSFORMS" etc., as alternatives for *_NOT_SUPPORTED,
 372     for consistency, in pngconf.h
 373   Added individual "ifndef PNG_NO_[CAPABILITY]" in pngconf.h to make it easier
 374     to remove unwanted capabilities via the compile line
 375   Made some corrections to grammar (which, it's) in documentation (Greg).
 376   Corrected example.c, use of row_pointers in png_write_image().
 378 Version 1.0.1d [May 24, 1998]
 379   Corrected several statements that used side effects illegally in pngrutil.c
 380     and pngtrans.c, that were introduced in version 1.0.1b
 381   Revised png_read_rows() to avoid repeated if-testing for NULL (A Kleinert)
 382   More corrections to example.c, use of row_pointers in png_write_image()
 383     and png_read_rows().
 384   Added pngdll.mak and pngdef.pas to scripts directory, contributed by
 385     Bob Dellaca, to make a png32bd.dll with Borland C++ 4.5
 386   Fixed error in example.c with png_set_text: num_text is 3, not 2 (Guido V.)
 387   Changed several loops from count-down to count-up, for consistency.
 389 Version 1.0.1e [June 6, 1998]
 390   Revised libpng.txt and libpng.3 description of png_set_read|write_fn(), and
 391     added warnings when people try to set png_read_fn and png_write_fn in
 392     the same structure.
 393   Added a test such that png_do_gamma will be done when num_trans==0
 394     for truecolor images that have defined a background.  This corrects an
 395     error that was introduced in libpng-0.90 that can cause gamma processing
 396     to be skipped.
 397   Added tests in png.h to include "trans" and "trans_values" in structures
 399   Add png_free(png_ptr->time_buffer) in png_destroy_read_struct()
 400   Moved png_convert_to_rfc_1123() from pngwrite.c to png.c
 401   Added capability for user-provided malloc_fn() and free_fn() functions,
 402     and revised pngtest.c to demonstrate their use, replacing the
 403     PNGTEST_DEBUG_MEM feature.
 404   Added makefile.w32, for Microsoft C++ 4.0 and later (Tim Wegner).
 406 Version 1.0.2 [June 14, 1998]
 407   Fixed two bugs in makefile.bor .
 409 Version 1.0.2a [December 30, 1998]
 410   Replaced and extended code that was removed from png_set_filler() in 1.0.1a.
 411   Fixed a bug in png_do_filler() that made it fail to write filler bytes in
 412     the left-most pixel of each row (Kevin Bracey).
 413   Changed "static pngcharp tIME_string" to "static char tIME_string[30]"
 414     in pngtest.c (Duncan Simpson).
 415   Fixed a bug in pngtest.c that caused pngtest to try to write a tIME chunk
 416     even when no tIME chunk was present in the source file.
 417   Fixed a problem in pngrutil.c: gray_to_rgb didn't always work with 16-bit.
 418   Fixed a problem in png_read_push_finish_row(), which would not skip some
 419     passes that it should skip, for images that are less than 3 pixels high.
 420   Interchanged the order of calls to png_do_swap() and png_do_shift()
 421     in pngwtran.c (John Cromer).
 422   Added #ifdef PNG_DEBUG/#endif surrounding use of PNG_DEBUG in png.h .
 423   Changed "bad adaptive filter type" from error to warning in pngrutil.c .
 424   Fixed a documentation error about default filtering with 8-bit indexed-color.
 425   Separated the PNG_NO_STDIO macro into PNG_NO_STDIO and PNG_NO_CONSOLE_IO
 426     (L. Peter Deutsch).
 427   Added png_set_rgb_to_gray() and png_get_rgb_to_gray_status() functions.
 428   Added png_get_copyright() and png_get_header_version() functions.
 429   Revised comments on png_set_progressive_read_fn() in libpng.txt and example.c
 430   Added information about debugging in libpng.txt and libpng.3 .
 431   Changed "ln -sf" to "ln -s -f" in makefile.s2x, makefile.lnx, and
 432     makefile.sco.
 433   Removed lines after Dynamic Dependencies" in makefile.aco .
 434   Revised makefile.dec to make a shared library (Jeremie Petit).
 435   Removed trailing blanks from all files.
 437 Version 1.0.2a [January 6, 1999]
 438   Removed misplaced #endif and #ifdef PNG_NO_EXTERN near the end of png.h
 439   Added "if" tests to silence complaints about unused png_ptr in png.h and png.c
 440   Changed "check_if_png" function in example.c to return true (nonzero) if PNG.
 441   Changed libpng.txt to demonstrate png_sig_cmp() instead of png_check_sig()
 442     which is obsolete.
 444 Version 1.0.3 [January 14, 1999]
 445   Added makefile.hux, for Hewlett Packard HPUX 10.20 and 11.00 (Jim Rice)
 446   Added a statement of Y2K compliance in png.h, libpng.3, and Y2KINFO.
 448 Version 1.0.3a [August 12, 1999]
 449   Added check for PNG_READ_INTERLACE_SUPPORTED in pngread.c; issue a warning
 450     if an attempt is made to read an interlaced image when it's not supported.
 451   Added check if png_ptr->trans is defined before freeing it in pngread.c
 452   Modified the Y2K statement to include versions back to version 0.71
 453   Fixed a bug in the check for valid IHDR bit_depth/color_types in pngrutil.c
 454   Modified makefile.wat (added -zp8 flag, ".symbolic", changed some comments)
 455   Replaced leading blanks with tab characters in makefile.hux
 456   Changed "dworkin.wustl.edu" to "ccrc.wustl.edu" in various documents.
 457   Changed (float)red and (float)green to (double)red, (double)green
 458     in png_set_rgb_to_gray() to avoid "promotion" problems in AIX.
 459   Fixed a bug in pngconf.h that omitted <stdio.h> when PNG_DEBUG==0 (K Bracey).
 460   Reformatted libpng.3 and libpngpf.3 with proper fonts (script by J. vanZandt).
 461   Updated documentation to refer to the PNG-1.2 specification.
 462   Removed ansi2knr.c and left pointers to the latest source for ansi2knr.c
 463     in makefile.knr, INSTALL, and README (L. Peter Deutsch)
 464   Fixed bugs in calculation of the length of rowbytes when adding alpha
 465     channels to 16-bit images, in pngrtran.c (Chris Nokleberg)
 466   Added function png_set_user_transform_info() to store user_transform_ptr,
 467     user_depth, and user_channels into the png_struct, and a function
 468     png_get_user_transform_ptr() to retrieve the pointer (Chris Nokleberg)
 469   Added function png_set_empty_plte_permitted() to make libpng useable
 470     in MNG applications.
 471   Corrected the typedef for png_free_ptr in png.h (Jesse Jones).
 472   Correct gamma with srgb is 45455 instead of 45000 in pngrutil.c, to be
 473     consistent with PNG-1.2, and allow variance of 500 before complaining.
 474   Added assembler code contributed by Intel in file pngvcrd.c and modified
 475     makefile.w32 to use it (Nirav Chhatrapati, INTEL Corporation,
 476     Gilles Vollant)
 477   Changed "ln -s -f" to "ln -f -s" in the makefiles to make Solaris happy.
 478   Added some aliases for png_set_expand() in pngrtran.c, namely
 479     png_set_expand_PLTE(), png_set_expand_depth(), and png_set_expand_tRNS()
 480     (Greg Roelofs, in "PNG: The Definitive Guide").
 481   Added makefile.beo for BEOS on X86, contributed by Sander Stok.
 483 Version 1.0.3b [August 26, 1999]
 484   Replaced 2147483647L several places with PNG_MAX_UINT macro, defined in png.h
 485   Changed leading blanks to tabs in all makefiles.
 486   Define PNG_USE_PNGVCRD in makefile.w32, to get MMX assembler code.
 487   Made alternate versions of  png_set_expand() in pngrtran.c, namely
 488     png_set_gray_1_2_4_to_8, png_set_palette_to_rgb, and png_set_tRNS_to_alpha
 489     (Greg Roelofs, in "PNG: The Definitive Guide").  Deleted the 1.0.3a aliases.
 490   Relocated start of 'extern "C"' block in png.h so it doesn't include pngconf.h
 491   Revised calculation of num_blocks in pngmem.c to avoid a potentially
 492     negative shift distance, whose results are undefined in the C language.
 493   Added a check in pngset.c to prevent writing multiple tIME chunks.
 494   Added a check in pngwrite.c to detect invalid small window_bits sizes.
 496 Version 1.0.3d [September 4, 1999]
 497   Fixed type casting of igamma in pngrutil.c
 498   Added new png_expand functions to scripts/pngdef.pas and pngos2.def
 499   Added a demo read_user_transform_fn that examines the row filters in pngtest.c
 501 Version 1.0.4 [September 24, 1999, not distributed publicly]
 502   Define PNG_ALWAYS_EXTERN in pngconf.h if __STDC__ is defined
 503   Delete #define PNG_INTERNAL and include "png.h" from pngasmrd.h
 504   Made several minor corrections to pngtest.c
 505   Renamed the makefiles with longer but more user friendly extensions.
 506   Copied the PNG copyright and license to a separate LICENSE file.
 507   Revised documentation, png.h, and example.c to remove reference to
 508     "viewing_gamma" which no longer appears in the PNG specification.
 509   Revised pngvcrd.c to use MMX code for interlacing only on the final pass.
 510   Updated pngvcrd.c to use the faster C filter algorithms from libpng-1.0.1a
 511   Split makefile.win32vc into two versions, makefile.vcawin32 (uses MMX
 512     assembler code) and makefile.vcwin32 (doesn't).
 513   Added a CPU timing report to pngtest.c (enabled by defining PNGTEST_TIMING)
 514   Added a copy of pngnow.png to the distribution.
 516 Version 1.0.4a [September 25, 1999]
 517   Increase max_pixel_depth in pngrutil.c if a user transform needs it.
 518   Changed several division operations to right-shifts in pngvcrd.c
 520 Version 1.0.4b [September 30, 1999]
 521   Added parentheses in line 3732 of pngvcrd.c
 522   Added a comment in makefile.linux warning about buggy -O3 in pgcc 2.95.1
 524 Version 1.0.4c [October 1, 1999]
 525   Added a "png_check_version" function in png.c and pngtest.c that will generate
 526     a helpful compiler error if an old png.h is found in the search path.
 527   Changed type of png_user_transform_depth|channels from int to png_byte.
 528   Added "Libpng is OSI Certified Open Source Software" statement to png.h
 530 Version 1.0.4d [October 6, 1999]
 531   Changed 0.45 to 0.45455 in png_set_sRGB()
 532   Removed unused PLTE entries from pngnow.png
 533   Re-enabled some parts of pngvcrd.c (png_combine_row) that work properly.
 535 Version 1.0.4e [October 10, 1999]
 536   Fixed sign error in pngvcrd.c (Greg Roelofs)
 537   Replaced some instances of memcpy with simple assignments in pngvcrd (GR-P)
 539 Version 1.0.4f [October 15, 1999]
 540   Surrounded example.c code with #if 0 .. #endif to prevent people from
 541     inadvertently trying to compile it.
 542   Changed png_get_header_version() from a function to a macro in png.h
 543   Added type casting mostly in pngrtran.c and pngwtran.c
 544   Removed some pointless "ptr = NULL" in pngmem.c
 545   Added a "contrib" directory containing the source code from Greg's book.
 547 Version 1.0.5 [October 15, 1999]
 548   Minor editing of the INSTALL and README files.
 550 Version 1.0.5a [October 23, 1999]
 551   Added contrib/pngsuite and contrib/pngminus (Willem van Schaik)
 552   Fixed a typo in the png_set_sRGB() function call in example.c (Jan Nijtmans)
 553   Further optimization and bugfix of pngvcrd.c
 554   Revised pngset.c so that it does not allocate or free memory in the user's
 555     text_ptr structure.  Instead, it makes its own copy.
 556   Created separate write_end_info_struct in pngtest.c for a more severe test.
 557   Added code in pngwrite.c to free info_ptr->text[i].key to stop a memory leak.
 559 Version 1.0.5b [November 23, 1999]
 561     PNG_FLAG_WROTE_tIME from flags to mode.
 562   Added png_write_info_before_PLTE() function.
 563   Fixed some typecasting in contrib/gregbook/*.c
 564   Updated scripts/makevms.com and added makevms.com to contrib/gregbook
 565     and contrib/pngminus (Martin Zinser)
 567 Version 1.0.5c [November 26, 1999]
 568   Moved png_get_header_version from png.h to png.c, to accommodate ansi2knr.
 569   Removed all global arrays (according to PNG_NO_GLOBAL_ARRAYS macro), to
 570     accommodate making DLL's: Moved usr_png_ver from global variable to function
 571     png_get_header_ver() in png.c.  Moved png_sig to png_sig_bytes in png.c and
 572     eliminated use of png_sig in pngwutil.c.  Moved the various png_CHNK arrays
 573     into pngtypes.h.  Eliminated use of global png_pass arrays.  Declared the
 574     png_CHNK and png_pass arrays to be "const".  Made the global arrays
 575     available to applications (although none are used in libpng itself) when
 576     PNG_NO_GLOBAL_ARRAYS is not defined or when PNG_GLOBAL_ARRAYS is defined.
 577   Removed some extraneous "-I" from contrib/pngminus/makefile.std
 578   Changed the PNG_sRGB_INTENT macros in png.h to be consistent with PNG-1.2.
 579   Change PNG_SRGB_INTENT to PNG_sRGB_INTENT in libpng.txt and libpng.3
 581 Version 1.0.5d [November 29, 1999]
 582   Add type cast (png_const_charp) two places in png.c
 583   Eliminated pngtypes.h; use macros instead to declare PNG_CHNK arrays.
 584   Renamed "PNG_GLOBAL_ARRAYS" to "PNG_USE_GLOBAL_ARRAYS" and made available
 585     to applications a macro "PNG_USE_LOCAL_ARRAYS".
 586   comment out (with #ifdef) all the new declarations when
 587     PNG_USE_GLOBAL_ARRAYS is defined.
 588   Added PNG_EXPORT_VAR macro to accommodate making DLL's.
 590 Version 1.0.5e [November 30, 1999]
 591   Added iCCP, iTXt, and sPLT support; added "lang" member to the png_text
 592     structure; refactored the inflate/deflate support to make adding new chunks
 593     with trailing compressed parts easier in the future, and added new functions
 594     png_free_iCCP, png_free_pCAL, png_free_sPLT, png_free_text, png_get_iCCP,
 595     png_get_spalettes, png_set_iCCP, png_set_spalettes (Eric S. Raymond).
 596     NOTE: Applications that write text chunks MUST define png_text->lang
 597     before calling png_set_text(). It must be set to NULL if you want to
 598     write tEXt or zTXt chunks.  If you want your application to be able to
 599     run with older versions of libpng, use
 601       #ifdef PNG_iTXt_SUPPORTED
 602          png_text[i].lang = NULL;
 603       #endif
 605   Changed png_get_oFFs() and png_set_oFFs() to use signed rather than unsigned
 606     offsets (Eric S. Raymond).
 607   Combined PNG_READ_cHNK_SUPPORTED and PNG_WRITE_cHNK_SUPPORTED macros into
 608     PNG_cHNK_SUPPORTED and combined the three types of PNG_text_SUPPORTED
 609     macros, leaving the separate macros also available.
 610   Removed comments on #endifs at the end of many short, non-nested #if-blocks.
 612 Version 1.0.5f [December 6, 1999]
 613   Changed makefile.solaris to issue a warning about potential problems when
 614     the ucb "ld" is in the path ahead of the ccs "ld".
 615   Removed "- [date]" from the "synopsis" line in libpng.3 and libpngpf.3.
 616   Added sCAL chunk support (Eric S. Raymond).
 618 Version 1.0.5g [December 7, 1999]
 619   Fixed "png_free_spallettes" typo in png.h
 620   Added code to handle new chunks in pngpread.c
 621   Moved PNG_CHNK string macro definitions outside of PNG_NO_EXTERN block
 622   Added "translated_key" to png_text structure and png_write_iTXt().
 623   Added code in pngwrite.c to work around a newly discovered zlib bug.
 625 Version 1.0.5h [December 10, 1999]
 626   NOTE: regarding the note for version 1.0.5e, the following must also
 627     be included in your code:
 628         png_text[i].translated_key = NULL;
 629   Unknown chunk handling is now supported.
 630   Option to eliminate all floating point support was added.  Some new
 631     fixed-point functions such as png_set_gAMA_fixed() were added.
 632   Expanded tabs and removed trailing blanks in source files.
 634 Version 1.0.5i [December 13, 1999]
 635   Added some type casts to silence compiler warnings.
 636   Renamed "png_free_spalette" to "png_free_spalettes" for consistency.
 637   Removed leading blanks from a #define in pngvcrd.c
 638   Added some parameters to the new png_set_keep_unknown_chunks() function.
 639   Added a test for up->location != 0 in the first instance of writing
 640     unknown chunks in pngwrite.c
 641   Changed "num" to "i" in png_free_spalettes() and png_free_unknowns() to
 642     prevent recursion.
 643   Added png_free_hIST() function.
 644   Various patches to fix bugs in the sCAL and integer cHRM processing,
 645     and to add some convenience macros for use with sCAL.
 647 Version 1.0.5j [December 21, 1999]
 648   Changed "unit" parameter of png_write_sCAL from png_byte to int, to work
 649     around buggy compilers.
 650   Added new type "png_fixed_point" for integers that hold float*100000 values
 651   Restored backward compatibility of tEXt/zTXt chunk processing:
 652     Restored the first four members of png_text to the same order as v.1.0.5d.
 653     Added members "lang_key" and "itxt_length" to png_text struct.  Set
 654     text_length=0 when "text" contains iTXt data.  Use the "compression"
 655     member to distinguish among tEXt/zTXt/iTXt types.  Added
 657     The "Note" above, about backward incompatibility of libpng-1.0.5e, no
 658     longer applies.
 659   Fixed png_read|write_iTXt() to read|write parameters in the right order,
 660     and to write the iTXt chunk after IDAT if it appears in the end_ptr.
 661   Added pnggccrd.c, version of pngvcrd.c Intel assembler for gcc (Greg Roelofs)
 662   Reversed the order of trying to write floating-point and fixed-point gAMA.
 664 Version 1.0.5k [December 27, 1999]
 665   Added many parentheses, e.g., "if (a && b & c)" becomes "if (a && (b & c))"
 666   Added png_handle_as_unknown() function (Glenn)
 667   Added png_free_chunk_list() function and chunk_list and num_chunk_list members
 668     of png_ptr.
 669   Eliminated erroneous warnings about multiple sPLT chunks and sPLT-after-PLTE.
 670   Fixed a libpng-1.0.5h bug in pngrutil.c that was issuing erroneous warnings
 671     about ignoring incorrect gAMA with sRGB (gAMA was in fact not ignored)
 672   Added png_free_tRNS(); png_set_tRNS() now malloc's its own trans array (ESR).
 673   Define png_get_int_32 when oFFs chunk is supported as well as when pCAL is.
 674   Changed type of proflen from png_int_32 to png_uint_32 in png_get_iCCP().
 676 Version 1.0.5l [January 1, 2000]
 677   Added functions png_set_read_user_chunk_fn() and png_get_user_chunk_ptr()
 678     for setting a callback function to handle unknown chunks and for
 679     retrieving the associated user pointer (Glenn).
 681 Version 1.0.5m [January 7, 2000]
 682   Added high-level functions png_read_png(), png_write_png(), png_free_pixels().
 684 Version 1.0.5n [January 9, 2000]
 685   Added png_free_PLTE() function, and modified png_set_PLTE() to malloc its
 686     own memory for info_ptr->palette.  This makes it safe for the calling
 687     application to free its copy of the palette any time after it calls
 688     png_set_PLTE().
 690 Version 1.0.5o [January 20, 2000]
 691   Cosmetic changes only (removed some trailing blanks and TABs)
 693 Version 1.0.5p [January 31, 2000]
 694   Renamed pngdll.mak to makefile.bd32
 695   Cosmetic changes in pngtest.c
 697 Version 1.0.5q [February 5, 2000]
 698   Relocated the makefile.solaris warning about PATH problems.
 699   Fixed pngvcrd.c bug by pushing/popping registers in mmxsupport (Bruce Oberg)
 700   Revised makefile.gcmmx
 703 Version 1.0.5r [February 7, 2000]
 704   Removed superfluous prototype for png_get_itxt from png.h
 705   Fixed a bug in pngrtran.c that improperly expanded the background color.
 706   Return *num_text=0 from png_get_text() when appropriate, and fix documentation
 707     of png_get_text() in libpng.txt/libpng.3.
 709 Version 1.0.5s [February 18, 2000]
 710   Added "png_jmp_env()" macro to pngconf.h, to help people migrate to the
 711     new error handler that's planned for the next libpng release, and changed
 712     example.c, pngtest.c, and contrib programs to use this macro.
 713   Revised some of the DLL-export macros in pngconf.h (Greg Roelofs)
 714   Fixed a bug in png_read_png() that caused it to fail to expand some images
 715     that it should have expanded.
 716   Fixed some mistakes in the unused and undocumented INCH_CONVERSIONS functions
 717     in pngget.c
 718   Changed the allocation of palette, history, and trans arrays back to
 719     the version 1.0.5 method (linking instead of copying) which restores
 720     backward compatibility with version 1.0.5.  Added some remarks about
 721     that in example.c.  Added "free_me" member to info_ptr and png_ptr
 722     and added png_free_data() function.
 723   Updated makefile.linux and makefile.gccmmx to make directories conditionally.
 724   Made cosmetic changes to pngasmrd.h
 725   Added png_set_rows() and png_get_rows(), for use with png_read|write_png().
 726   Modified png_read_png() to allocate info_ptr->row_pointers only if it
 727     hasn't already been allocated.
 729 Version 1.0.5t [March 4, 2000]
 730   Changed png_jmp_env() migration aiding macro to png_jmpbuf().
 731   Fixed "interlace" typo (should be "interlaced") in contrib/gregbook/read2-x.c
 732   Fixed bug with use of PNG_BEFORE_IHDR bit in png_ptr->mode, introduced when
 733     PNG_FLAG_HAVE_CHUNK_HEADER was moved into png_ptr->mode in version 1.0.5b
 734   Files in contrib/gregbook were revised to use png_jmpbuf() and to select
 735     a 24-bit visual if one is available, and to allow abbreviated options.
 736   Files in contrib/pngminus were revised to use the png_jmpbuf() macro.
 737   Removed spaces in makefile.linux and makefile.gcmmx, introduced in 1.0.5s
 739 Version 1.0.5u [March 5, 2000]
 740   Simplified the code that detects old png.h in png.c and pngtest.c
 741   Renamed png_spalette (_p, _pp) to png_sPLT_t (_tp, _tpp)
 742   Increased precision of rgb_to_gray calculations from 8 to 15 bits and
 743     added png_set_rgb_to_gray_fixed() function.
 744   Added makefile.bc32 (32-bit Borland C++, C mode)
 746 Version 1.0.5v [March 11, 2000]
 747   Added some parentheses to the png_jmpbuf macro definition.
 748   Updated references to the zlib home page, which has moved to freesoftware.com.
 749   Corrected bugs in documentation regarding png_read_row() and png_write_row().
 750   Updated documentation of png_rgb_to_gray calculations in libpng.3/libpng.txt.
 751   Renamed makefile.borland,turboc3 back to makefile.bor,tc3 as in version 1.0.3,
 752     revised borland makefiles; added makefile.ibmvac3 and makefile.gcc (Cosmin)
 754 Version 1.0.6 [March 20, 2000]
 755   Minor revisions of makefile.bor, libpng.txt, and gregbook/rpng2-win.c
 756   Added makefile.sggcc (SGI IRIX with gcc)
 758 Version 1.0.6d [April 7, 2000]
 759   Changed sprintf() to strcpy() in png_write_sCAL_s() to work without STDIO
 760   Added data_length parameter to png_decompress_chunk() function
 761   Revised documentation to remove reference to abandoned png_free_chnk functions
 762   Fixed an error in png_rgb_to_gray_fixed()
 763   Revised example.c, usage of png_destroy_write_struct().
 764   Renamed makefile.ibmvac3 to makefile.ibmc, added libpng.icc IBM project file
 765   Added a check for info_ptr->free_me&PNG_FREE_TEXT when freeing text in png.c
 766   Simplify png_sig_bytes() function to remove use of non-ISO-C strdup().
 768 Version 1.0.6e [April 9, 2000]
 769   Added png_data_freer() function.
 770   In the code that checks for over-length tRNS chunks, added check of
 771     info_ptr->num_trans as well as png_ptr->num_trans (Matthias Benckmann)
 772   Minor revisions of libpng.txt/libpng.3.
 773   Check for existing data and free it if the free_me flag is set, in png_set_*()
 774     and png_handle_*().
 776     is defined.
 777   Changed several instances of PNG_NO_CONSOLE_ID to PNG_NO_STDIO in pngrutil.c
 778     and mentioned the purposes of the two macros in libpng.txt/libpng.3.
 780 Version 1.0.6f [April 14, 2000]
 781   Revised png_set_iCCP() and png_set_rows() to avoid prematurely freeing data.
 782   Add checks in png_set_text() for NULL members of the input text structure.
 783   Revised libpng.txt/libpng.3.
 784   Removed superfluous prototype for png_set_iTXt from png.h
 785   Removed "else" from pngread.c, after png_error(), and changed "0" to "length".
 786   Changed several png_errors about malformed ancillary chunks to png_warnings.
 788 Version 1.0.6g [April 24, 2000]
 789   Added png_pass-* arrays to pnggccrd.c when PNG_USE_LOCAL_ARRAYS is defined.
 790   Relocated paragraph about png_set_background() in libpng.3/libpng.txt
 791     and other revisions (Matthias Benckmann)
 792   Relocated info_ptr->free_me, png_ptr->free_me, and other info_ptr and
 793     png_ptr members to restore binary compatibility with libpng-1.0.5
 794     (breaks compatibility with libpng-1.0.6).
 796 Version 1.0.6h [April 24, 2000]
 797   Changed shared library so-number pattern from 2.x.y.z to xy.z (this builds
 798     libpng.so.10 & libpng.so.10.6h instead of libpng.so.2 & libpng.so.
 799     This is a temporary change for test purposes.
 801 Version 1.0.6i [May 2, 2000]
 802   Rearranged some members at the end of png_info and png_struct, to put
 803     unknown_chunks_num and free_me within the original size of the png_structs
 804     and free_me, png_read_user_fn, and png_free_fn within the original png_info,
 805     because some old applications allocate the structs directly instead of
 806     using png_create_*().
 807   Added documentation of user memory functions in libpng.txt/libpng.3
 808   Modified png_read_png so that it will use user_allocated row_pointers
 809     if present, unless free_me directs that it be freed, and added description
 810     of the use of png_set_rows() and png_get_rows() in libpng.txt/libpng.3.
 811   Added PNG_LEGACY_SUPPORTED macro, and #ifdef out all new (since version
 812     1.00) members of png_struct and png_info, to regain binary compatibility
 813     when you define this macro.  Capabilities lost in this event
 814     are user transforms (new in version 1.0.0),the user transform pointer
 815     (new in version 1.0.2), rgb_to_gray (new in 1.0.5), iCCP, sCAL, sPLT,
 816     the high-level interface, and unknown chunks support (all new in 1.0.6).
 817     This was necessary because of old applications that allocate the structs
 818     directly as authors were instructed to do in libpng-0.88 and earlier,
 819     instead of using png_create_*().
 821     can be used to detect codes that directly allocate the structs, and
 822     code to check these modes in png_read_init() and png_write_init() and
 823     generate a libpng error if the modes aren't set and PNG_LEGACY_SUPPORTED
 824     was not defined.
 825   Added makefile.intel and updated makefile.watcom (Pawel Mrochen)
 827 Version 1.0.6j [May 3, 2000]
 828   Overloaded png_read_init() and png_write_init() with macros that convert
 829     calls to png_read_init_2() or png_write_init_2() that check the version
 830     and structure sizes.
 832 Version 1.0.7beta11 [May 7, 2000]
 834     which are no longer used.
 835   Eliminated the three new members of png_text when PNG_LEGACY_SUPPORTED is
 836     defined or when neither PNG_READ_iTXt_SUPPORTED nor PNG_WRITE_iTXT_SUPPORTED
 837     is defined.
 838   Made PNG_NO_READ|WRITE_iTXt the default setting, to avoid memory
 839     overrun when old applications fill the info_ptr->text structure directly.
 840   Added PNGAPI macro, and added it to the definitions of all exported functions.
 841   Relocated version macro definitions ahead of the includes of zlib.h and
 842     pngconf.h in png.h.
 844 Version 1.0.7beta12 [May 12, 2000]
 845   Revised pngset.c to avoid a problem with expanding the png_debug macro.
 846   Deleted some extraneous defines from pngconf.h
 847   Made PNG_NO_CONSOLE_IO the default condition when PNG_BUILD_DLL is defined.
 848   Use MSC _RPTn debugging instead of fprintf if _MSC_VER is defined.
 849   Added png_access_version_number() function.
 850   Check for mask&PNG_FREE_CHNK (for TEXT, SCAL, PCAL) in png_free_data().
 851   Expanded libpng.3/libpng.txt information about png_data_freer().
 853 Version 1.0.7beta14 [May 17, 2000] (beta13 was not published)
 854   Changed pnggccrd.c and pngvcrd.c to handle bad adaptive filter types as
 855     warnings instead of errors, as pngrutil.c does.
 856   Set the PNG_INFO_IDAT valid flag in png_set_rows() so png_write_png()
 857     will actually write IDATs.
 858   Made the default PNG_USE_LOCAL_ARRAYS depend on PNG_DLL instead of WIN32.
 859   Make png_free_data() ignore its final parameter except when freeing data
 860     that can have multiple instances (text, sPLT, unknowns).
 861   Fixed a new bug in png_set_rows().
 862   Removed info_ptr->valid tests from png_free_data(), as in version 1.0.5.
 863   Added png_set_invalid() function.
 864   Fixed incorrect illustrations of png_destroy_write_struct() in example.c.
 866 Version 1.0.7beta15 [May 30, 2000]
 867   Revised the deliberately erroneous Linux setjmp code in pngconf.h to produce
 868     fewer error messages.
 869   Rearranged checks for Z_OK to check the most likely path first in pngpread.c
 870     and pngwutil.c.
 871   Added checks in pngtest.c for png_create_*() returning NULL, and mentioned
 872     in libpng.txt/libpng.3 the need for applications to check this.
 873   Changed names of png_default_*() functions in pngtest to pngtest_*().
 874   Changed return type of png_get_x|y_offset_*() from png_uint_32 to png_int_32.
 875   Fixed some bugs in the unused PNG_INCH_CONVERSIONS functions in pngget.c
 876   Set each pointer to NULL after freeing it in png_free_data().
 877   Worked around a problem in pngconf.h; AIX's strings.h defines an "index"
 878     macro that conflicts with libpng's png_color_16.index. (Dimitri
 879     Papadapoulos)
 880   Added "msvc" directory with MSVC++ project files (Simon-Pierre Cadieux).
 882 Version 1.0.7beta16 [June 4, 2000]
 883   Revised the workaround of AIX string.h "index" bug.
 884   Added a check for overlength PLTE chunk in pngrutil.c.
 885   Added PNG_NO_POINTER_INDEXING macro to use array-indexing instead of pointer
 886     indexing in pngrutil.c and pngwutil.c to accommodate a buggy compiler.
 887   Added a warning in png_decompress_chunk() when it runs out of data, e.g.
 888     when it tries to read an erroneous PhotoShop iCCP chunk.
 889   Added PNG_USE_DLL macro.
 890   Revised the copyright/disclaimer/license notice.
 891   Added contrib/msvctest directory
 893 Version 1.0.7rc1 [June 9, 2000]
 894   Corrected the definition of PNG_TRANSFORM_INVERT_ALPHA  (0x0400 not 0x0200)
 895   Added contrib/visupng directory (Willem van Schaik)
 897 Version 1.0.7beta18 [June 23, 2000]
 898   Revised PNGAPI definition, and pngvcrd.c to work with __GCC__
 899     and do not redefine PNGAPI if it is passed in via a compiler directive.
 900   Revised visupng/PngFile.c to remove returns from within the Try block.
 901   Removed leading underscores from "_PNG_H" and "_PNG_SAVE_BSD_SOURCE" macros.
 902   Updated contrib/visupng/cexcept.h to version 1.0.0.
 903   Fixed bugs in pngwrite.c and pngwutil.c that prevented writing iCCP chunks.
 905 Version 1.0.7rc2 [June 28, 2000]
 906   Updated license to include disclaimers required by UCITA.
 907   Fixed "DJBPP" typo in pnggccrd.c introduced in beta18.
 909 Version 1.0.7 [July 1, 2000]
 910   Revised the definition of "trans_values" in libpng.3/libpng.txt
 912 Version 1.0.8beta1 [July 8, 2000]
 913   Added png_free(png_ptr, key) two places in pngpread.c to stop memory leaks.
 914   Changed PNG_NO_STDIO to PNG_NO_CONSOLE_IO, several places in pngrutil.c and
 915     pngwutil.c.
 916   Changed PNG_EXPORT_VAR to use PNG_IMPEXP, in pngconf.h.
 917   Removed unused "#include <assert.h>" from png.c
 918   Added WindowsCE support.
 919   Revised pnggccrd.c to work with gcc-2.95.2 and in the Cygwin environment.
 921 Version 1.0.8beta2 [July 10, 2000]
 922   Added project files to the wince directory and made further revisions
 923     of pngtest.c, pngrio.c, and pngwio.c in support of WindowsCE.
 925 Version 1.0.8beta3 [July 11, 2000]
 926   Only set the PNG_FLAG_FREE_TRNS or PNG_FREE_TRNS flag in png_handle_tRNS()
 927     for indexed-color input files to avoid potential double-freeing trans array
 928     under some unusual conditions; problem was introduced in version 1.0.6f.
 929   Further revisions to pngtest.c and files in the wince subdirectory.
 931 Version 1.0.8beta4 [July 14, 2000]
 932   Added the files pngbar.png and pngbar.jpg to the distribution.
 933   Added makefile.cygwin, and cygwin support in pngconf.h
 934   Added PNG_NO_ZALLOC_ZERO macro (makes png_zalloc skip zeroing memory)
 936 Version 1.0.8rc1 [July 16, 2000]
 937   Revised png_debug() macros and statements to eliminate compiler warnings.
 939 Version 1.0.8 [July 24, 2000]
 940   Added png_flush() in pngwrite.c, after png_write_IEND().
 941   Updated makefile.hpux to build a shared library.
 943 Version 1.0.9beta1 [November 10, 2000]
 944   Fixed typo in scripts/makefile.hpux
 945   Updated makevms.com in scripts and contrib/* and contrib/* (Martin Zinser)
 946   Fixed seqence-point bug in contrib/pngminus/png2pnm (Martin Zinser)
 947   Changed "cdrom.com" in documentation to "libpng.org"
 948   Revised pnggccrd.c to get it all working, and updated makefile.gcmmx (Greg).
 949   Changed type of "params" from voidp to png_voidp in png_read|write_png().
 950   Make sure PNGAPI and PNG_IMPEXP are defined in pngconf.h.
 951   Revised the 3 instances of WRITEFILE in pngtest.c.
 952   Relocated "msvc" and "wince" project subdirectories into "dll" subdirectory.
 953   Updated png.rc in dll/msvc project
 954   Revised makefile.dec to define and use LIBPATH and INCPATH
 955   Increased size of global png_libpng_ver[] array from 12 to 18 chars.
 956   Made global png_libpng_ver[], png_sig[] and png_pass_*[] arrays const.
 957   Removed duplicate png_crc_finish() from png_handle_bKGD() function.
 958   Added a warning when application calls png_read_update_info() multiple times.
 959   Revised makefile.cygwin
 960   Fixed bugs in iCCP support in pngrutil.c and pngwutil.c.
 961   Replaced png_set_empty_plte_permitted() with png_permit_mng_features().
 963 Version 1.0.9beta2 [November 19, 2000]
 964   Renamed the "dll" subdirectory "projects".
 965   Added borland project files to "projects" subdirectory.
 966   Set VS_FF_PRERELEASE and VS_FF_PATCHED flags in msvc/png.rc when appropriate.
 967   Add error message in png_set_compression_buffer_size() when malloc fails.
 969 Version 1.0.9beta3 [November 23, 2000]
 970   Revised PNG_LIBPNG_BUILD_TYPE macro in png.h, used in the msvc project.
 971   Removed the png_flush() in pngwrite.c that crashes some applications
 972     that don't set png_output_flush_fn.
 973   Added makefile.macosx and makefile.aix to scripts directory.
 975 Version 1.0.9beta4 [December 1, 2000]
 976   Change png_chunk_warning to png_warning in png_check_keyword().
 977   Increased the first part of msg buffer from 16 to 18 in png_chunk_error().
 979 Version 1.0.9beta5 [December 15, 2000]
 980   Added support for filter method 64 (for PNG datastreams embedded in MNG).
 982 Version 1.0.9beta6 [December 18, 2000]
 983   Revised png_set_filter() to accept filter method 64 when appropriate.
 984   Added new PNG_HAVE_PNG_SIGNATURE bit to png_ptr->mode and use it to
 985     help prevent applications from using MNG features in PNG datastreams.
 986   Added png_permit_mng_features() function.
 987   Revised libpng.3/libpng.txt.  Changed "filter type" to "filter method".
 989 Version 1.0.9rc1 [December 23, 2000]
 990   Revised test for PNG_HAVE_PNG_SIGNATURE in pngrutil.c
 991   Fixed error handling of unknown compression type in png_decompress_chunk().
 992   In pngconf.h, define __cdecl when _MSC_VER is defined.
 994 Version 1.0.9beta7 [December 28, 2000]
 996   Revised memory management in png_set_hIST and png_handle_hIST in a backward
 997     compatible manner.  PLTE and tRNS were revised similarly.
 998   Revised the iCCP chunk reader to ignore trailing garbage.
1000 Version 1.0.9beta8 [January 12, 2001]
1001   Moved pngasmrd.h into pngconf.h.
1002   Improved handling of out-of-spec garbage iCCP chunks generated by PhotoShop.
1004 Version 1.0.9beta9 [January 15, 2001]
1005   Added png_set_invalid, png_permit_mng_features, and png_mmx_supported to
1006     wince and msvc project module definition files.
1007   Minor revision of makefile.cygwin.
1008   Fixed bug with progressive reading of narrow interlaced images in pngpread.c
1010 Version 1.0.9beta10 [January 16, 2001]
1011   Do not typedef png_FILE_p in pngconf.h when PNG_NO_STDIO is defined.
1012   Fixed "png_mmx_supported" typo in project definition files.
1014 Version 1.0.9beta11 [January 19, 2001]
1015   Updated makefile.sgi to make shared library.
1016   Removed png_mmx_support() function and disabled PNG_MNG_FEATURES_SUPPORTED
1017     by default, for the benefit of DLL forward compatibility.  These will
1018     be re-enabled in version 1.2.0.
1020 Version 1.0.9rc2 [January 22, 2001]
1021   Revised cygwin support.
1023 Version 1.0.9 [January 31, 2001]
1024   Added check of cygwin's ALL_STATIC in pngconf.h
1025   Added "-nommx" parameter to contrib/gregbook/rpng2-win and rpng2-x demos.
1027 Version 1.0.10beta1 [March 14, 2001]
1028   Revised makefile.dec, makefile.sgi, and makefile.sggcc; added makefile.hpgcc.
1029   Reformatted libpng.3 to eliminate bad line breaks.
1030   Added checks for _mmx_supported in the read_filter_row function of pnggccrd.c
1031   Added prototype for png_mmx_support() near the top of pnggccrd.c
1032   Moved some error checking from png_handle_IHDR to png_set_IHDR.
1034   Revised png_mmx_support() function in pnggccrd.c
1035   Restored version 1.0.8 PNG_WRITE_EMPTY_PLTE_SUPPORTED behavior in pngwutil.c
1036   Fixed memory leak in contrib/visupng/PngFile.c
1037   Fixed bugs in png_combine_row() in pnggccrd.c and pngvcrd.c (C version)
1038   Added warnings when retrieving or setting gamma=0.
1039   Increased the first part of msg buffer from 16 to 18 in png_chunk_warning().
1041 Version 1.0.10rc1 [March 23, 2001]
1042   Changed all instances of memcpy, strcpy, and strlen to png_memcpy, png_strcpy,
1043     and png_strlen.
1044   Revised png_mmx_supported() function in pnggccrd.c to return proper value.
1045   Fixed bug in progressive reading (pngpread.c) with small images (height < 8).
1047 Version 1.0.10 [March 30, 2001]
1048   Deleted extraneous space (introduced in 1.0.9) from line 42 of makefile.cygwin
1049   Added beos project files (Chris Herborth)
1051 Version 1.0.11beta1 [April 3, 2001]
1052   Added type casts on several png_malloc() calls (Dimitri Papadapoulos).
1053   Removed a no-longer needed AIX work-around from pngconf.h
1054   Changed several "//" single-line comments to C-style in pnggccrd.c
1056 Version 1.0.11beta2 [April 11, 2001]
1057   Removed PNGAPI from several functions whose prototypes did not have PNGAPI.
1058   Updated scripts/pngos2.def
1060 Version 1.0.11beta3 [April 14, 2001]
1061   Added checking the results of many instances of png_malloc() for NULL
1063 Version 1.0.11beta4 [April 20, 2001]
1064   Undid the changes from version 1.0.11beta3.  Added a check for NULL return
1065     from user's malloc_fn().
1066   Removed some useless type casts of the NULL pointer.
1067   Added makefile.netbsd
1069 Version 1.0.11 [April 27, 2001]
1070   Revised makefile.netbsd
1072 Version 1.0.12beta1 [May 14, 2001]
1073   Test for Windows platform in pngconf.h when including malloc.h (Emmanuel Blot)
1074   Updated makefile.cygwin and handling of Cygwin's ALL_STATIC in pngconf.h
1075   Added some never-to-be-executed code in pnggccrd.c to quiet compiler warnings.
1076   Eliminated the png_error about apps using png_read|write_init().  Instead,
1077     libpng will reallocate the png_struct and info_struct if they are too small.
1078     This retains future binary compatibility for old applications written for
1079     libpng-0.88 and earlier.
1081 Version 1.2.0beta1 [May 6, 2001]
1082   Bumped DLLNUM to 2.
1084     by default.
1085   Added runtime selection of MMX features.
1086   Added png_set_strip_error_numbers function and related macros.
1088 Version 1.2.0beta2 [May 7, 2001]
1089   Finished merging 1.2.0beta1 with version 1.0.11
1090   Added a check for attempts to read or write PLTE in grayscale PNG datastreams.
1092 Version 1.2.0beta3 [May 17, 2001]
1093   Enabled user memory function by default.
1094   Modified png_create_struct so it passes user mem_ptr to user memory allocator.
1095   Increased png_mng_features flag from png_byte to png_uint_32.
1096   Bumped shared-library (so-number) and dll-number to 3.
1098 Version 1.2.0beta4 [June 23, 2001]
1099   Check for missing profile length field in iCCP chunk and free chunk_data
1100     in case of truncated iCCP chunk.
1101   Bumped shared-library number to 3 in makefile.sgi and makefile.sggcc
1102   Bumped dll-number from 2 to 3 in makefile.cygwin
1103   Revised contrib/gregbook/rpng*-x.c to avoid a memory leak and to exit cleanly
1104     if user attempts to run it on an 8-bit display.
1105   Updated contrib/gregbook
1106   Use png_malloc instead of png_zalloc to allocate palette in pngset.c
1107   Updated makefile.ibmc
1108   Added some typecasts to eliminate gcc 3.0 warnings.  Changed prototypes
1109     of png_write_oFFS width and height from png_uint_32 to png_int_32.
1110   Updated example.c
1111   Revised prototypes for png_debug_malloc and png_debug_free in pngtest.c
1113 Version 1.2.0beta5 [August 8, 2001]
1114   Revised contrib/gregbook
1115   Revised makefile.gcmmx
1116   Revised pnggccrd.c to conditionally compile some thread-unsafe code only
1117     when PNG_THREAD_UNSAFE_OK is defined.
1118   Added tests to prevent pngwutil.c from writing a bKGD or tRNS chunk with
1119     value exceeding 2^bit_depth-1
1120   Revised makefile.sgi and makefile.sggcc
1121   Replaced calls to fprintf(stderr,...) with png_warning() in pnggccrd.c
1122   Removed restriction that do_invert_mono only operate on 1-bit opaque files
1124 Version 1.2.0 [September 1, 2001]
1125   Changed a png_warning() to png_debug() in pnggccrd.c
1126   Fixed contrib/gregbook/rpng-x.c, rpng2-x.c to avoid crash with XFreeGC().
1128 Version 1.2.1beta1 [October 19, 2001]
1129   Revised makefile.std in contrib/pngminus
1130   Include background_1 in png_struct regardless of gamma support.
1131   Revised makefile.netbsd and makefile.macosx, added makefile.darwin.
1132   Revised example.c to provide more details about using row_callback().
1134 Version 1.2.1beta2 [October 25, 2001]
1135   Added type cast to each NULL appearing in a function call, except for
1136     WINCE functions.
1137   Added makefile.so9.
1139 Version 1.2.1beta3 [October 27, 2001]
1140   Removed type casts from all NULLs.
1141   Simplified png_create_struct_2().
1143 Version 1.2.1beta4 [November 7, 2001]
1144   Revised png_create_info_struct() and png_creat_struct_2().
1145   Added error message if png_write_info() was omitted.
1146   Type cast NULLs appearing in function calls when _NO_PROTO or
1147     PNG_TYPECAST_NULL is defined.
1149 Version 1.2.1rc1 [November 24, 2001]
1150   Type cast NULLs appearing in function calls except when PNG_NO_TYPECAST_NULL
1151     is defined.
1152   Changed typecast of "size" argument to png_size_t in pngmem.c calls to
1153     the user malloc_fn, to agree with the prototype in png.h
1154   Added a pop/push operation to pnggccrd.c, to preserve Eflag (Maxim Sobolev)
1155   Updated makefile.sgi to recognize LIBPATH and INCPATH.
1156   Updated various makefiles so "make clean" does not remove previous major
1157     version of the shared library.
1159 Version 1.2.1rc2 [December 4, 2001]
1160   Always allocate 256-entry internal palette, hist, and trans arrays, to
1161     avoid out-of-bounds memory reference caused by invalid PNG datastreams.
1162   Added a check for prefix_length > data_length in iCCP chunk handler.
1164 Version 1.2.1 [December 7, 2001]
1165   None.
1167 Version 1.2.2beta1 [February 22, 2002]
1168   Fixed a bug with reading the length of iCCP profiles (Larry Reeves).
1169   Revised makefile.linux, makefile.gcmmx, and makefile.sgi to generate
1170     libpng.a, libpng12.so (not libpng.so.3), and libpng12/png.h
1171   Revised makefile.darwin to remove "-undefined suppress" option.
1172   Added checks for gamma and chromaticity values over 21474.83, which exceed
1173     the limit for PNG unsigned 32-bit integers when encoded.
1174   Revised calls to png_create_read_struct() and png_create_write_struct()
1175     for simpler debugging.
1176   Revised png_zalloc() so zlib handles errors (uses PNG_FLAG_MALLOC_NULL_MEM_OK)
1178 Version 1.2.2beta2 [February 23, 2002]
1179   Check chunk_length and idat_size for invalid (over PNG_MAX_UINT) lengths.
1180   Check for invalid image dimensions in png_get_IHDR.
1181   Added missing "fi;" in the install target of the SGI makefiles.
1182   Added install-static to all makefiles that make shared libraries.
1183   Always do gamma compensation when image is partially transparent.
1185 Version 1.2.2beta3 [March 7, 2002]
1186   Compute background.gray and background_1.gray even when color_type is RGB
1187     in case image gets reduced to gray later.
1188   Modified shared-library makefiles to install pkgconfig/libpngNN.pc.
1189   Export (with PNGAPI) png_zalloc, png_zfree, and png_handle_as_unknown
1190   Removed unused png_write_destroy_info prototype from png.h
1191   Eliminated incorrect use of width_mmx from pnggccrd.c in pixel_bytes == 8 case
1192   Added install-shared target to all makefiles that make shared libraries.
1193   Stopped a double free of palette, hist, and trans when not using free_me.
1194   Added makefile.32sunu for Sun Ultra 32 and makefile.64sunu for Sun Ultra 64.
1196 Version 1.2.2beta4 [March 8, 2002]
1197   Compute background.gray and background_1.gray even when color_type is RGB
1198     in case image gets reduced to gray later (Jason Summers).
1199   Relocated a misplaced /bin/rm in the "install-shared" makefile targets
1200   Added PNG_1_0_X macro which can be used to build a 1.0.x-compatible library.
1202 Version 1.2.2beta5 [March 26, 2002]
1203   Added missing PNGAPI to several function definitions.
1204   Check for invalid bit_depth or color_type in png_get_IHDR(), and
1205     check for missing PLTE or IHDR in png_push_read_chunk() (Matthias Clasen).
1206   Revised iTXt support to accept NULL for lang and lang_key.
1207   Compute gamma for color components of background even when color_type is gray.
1208   Changed "()" to "{}" in scripts/libpng.pc.in.
1209   Revised makefiles to put png.h and pngconf.h only in $prefix/include/libpngNN
1210   Revised makefiles to make symlink to libpng.so.NN in addition to libpngNN.so
1212 Version 1.2.2beta6 [March 31, 2002]
1214 Version 1.0.13beta1 [March 31, 2002]
1215   Prevent png_zalloc() from trying to memset memory that it failed to acquire.
1216   Add typecasts of PNG_MAX_UINT in pngset_cHRM_fixed() (Matt Holgate).
1217   Ensure that the right function (user or default) is used to free the
1218     png_struct after an error in png_create_read_struct_2().
1220 Version 1.2.2rc1 [April 7, 2002]
1222 Version 1.0.13rc1 [April 7, 2002]
1223   Save the ebx register in pnggccrd.c (Sami Farin)
1224   Add "mem_ptr = png_ptr->mem_ptr" in png_destroy_write_struct() (Paul Gardner).
1225   Updated makefiles to put headers in include/libpng and remove old include/*.h.
1227 Version 1.2.2 [April 15, 2002]
1229 Version 1.0.13 [April 15, 2002]
1230   Revised description of png_set_filter() in libpng.3/libpng.txt.
1231   Revised makefile.netbsd and added makefile.neNNbsd and makefile.freebsd
1233 Version 1.0.13patch01 [April 17, 2002]
1235 Version 1.2.2patch01 [April 17, 2002]
1236   Changed ${PNGMAJ}.${PNGVER} bug to ${PNGVER} in makefile.sgi and
1237     makefile.sggcc
1238   Fixed VER -> PNGVER typo in makefile.macosx and added install-static to
1239     install
1240   Added install: target to makefile.32sunu and makefile.64sunu
1242 Version 1.0.13patch03 [April 18, 2002]
1244 Version 1.2.2patch03 [April 18, 2002]
1245   Revised 15 makefiles to link libpng.a to libpngNN.a and the include libpng
1246   subdirectory to libpngNN subdirectory without the full pathname.
1247   Moved generation of libpng.pc from "install" to "all" in 15 makefiles.
1249 Version 1.2.3rc1 [April 28, 2002]
1250   Added install-man target to 15 makefiles (Dimitri Papadopolous-Orfanos).
1251   Added $(DESTDIR) feature to 24 makefiles (Tim Mooney)
1252   Fixed bug with $prefix, should be $(prefix) in makefile.hpux.
1253   Updated cygwin-specific portion of pngconf.h and revised makefile.cygwin
1254   Added a link from libpngNN.pc to libpng.pc in 15 makefiles.
1255   Added links from include/libpngNN/*.h to include/*.h in 24 makefiles.
1256   Revised makefile.darwin to make relative links without full pathname.
1257   Added setjmp() at the end of png_create_*_struct_2() in case user forgets
1258     to put one in their application.
1259   Restored png_zalloc() and png_zfree() prototypes to version 1.2.1 and
1260     removed them from module definition files.
1262 Version 1.2.3rc2 [May 1, 2002]
1263   Fixed bug in reporting number of channels in pngget.c and pngset.c,
1264     that was introduced in version 1.2.2beta5.
1265   Exported png_zalloc(), png_zfree(), png_default_read(), png_default_write(),
1266     png_default_flush(), and png_push_fill_buffer() and included them in
1267     module definition files.
1268   Added "libpng.pc" dependency to the "install-shared" target in 15 makefiles.
1270 Version 1.2.3rc3 [May 1, 2002]
1271   Revised prototype for png_default_flush()
1272   Remove old libpng.pc and libpngNN.pc before installing new ones.
1274 Version 1.2.3rc4 [May 2, 2002]
1275   Typos in *.def files (png_default_read|write -> png_default_read|write_data)
1276   In makefiles, changed rm libpng.NN.pc to rm libpngNN.pc
1277   Added libpng-config and libpngNN-config and modified makefiles to install
1278     them.
1279   Changed $(MANPATH) to $(DESTDIR)$(MANPATH) in makefiles
1280   Added "Win32 DLL VB" configuration to projects/msvc/libpng.dsp
1282 Version 1.2.3rc5 [May 11, 2002]
1283   Changed "error" and "message" in prototypes to "error_message" and
1284     "warning_message" to avoid namespace conflict.
1285   Revised 15 makefiles to build libpng-config from libpng-config-*.in
1286   Once more restored png_zalloc and png_zfree to regular nonexported form.
1287   Restored png_default_read|write_data, png_default_flush, png_read_fill_buffer
1288     to nonexported form, but with PNGAPI, and removed them from module def
1289     files.
1291 Version 1.2.3rc6 [May 14, 2002]
1292   Removed "PNGAPI" from png_zalloc() and png_zfree() in png.c
1293   Changed "Gz" to "Gd" in projects/msvc/libpng.dsp and zlib.dsp.
1294   Removed leftover libpng-config "sed" script from four makefiles.
1295   Revised libpng-config creating script in 16 makefiles.
1297 Version 1.2.3 [May 22, 2002]
1298   Revised libpng-config target in makefile.cygwin.
1299   Removed description of png_set_mem_fn() from documentation.
1300   Revised makefile.freebsd.
1301   Minor cosmetic changes to 15 makefiles, e.g., $(DI) = $(DESTDIR)/$(INCDIR).
1302   Revised projects/msvc/README.txt
1303   Changed -lpng to -lpngNN in LDFLAGS in several makefiles.
1305 Version 1.2.4beta1 [May 24, 2002]
1306   Added libpng.pc and libpng-config to "all:" target in 16 makefiles.
1307   Fixed bug in 16 makefiles: $(DESTDIR)/$(LIBPATH) to $(DESTDIR)$(LIBPATH)
1308   Added missing "\" before closing double quote in makefile.gcmmx.
1309   Plugged various memory leaks; added png_malloc_warn() and png_set_text_2()
1310     functions.
1312 Version 1.2.4beta2 [June 25, 2002]
1313   Plugged memory leak of png_ptr->current_text (Matt Holgate).
1314   Check for buffer overflow before reading CRC in pngpread.c (Warwick Allison)
1315   Added -soname to the loader flags in makefile.dec, makefile.sgi, and
1316     makefile.sggcc.
1317   Added "test-installed" target to makefile.linux, makefile.gcmmx,
1318     makefile.sgi, and makefile.sggcc.
1320 Version 1.2.4beta3 [June 28, 2002]
1321   Plugged memory leak of row_buf in pngtest.c when there is a png_error().
1322   Detect buffer overflow in pngpread.c when IDAT is corrupted with extra data.
1323   Added "test-installed" target to makefile.32sunu, makefile.64sunu,
1324     makefile.beos, makefile.darwin, makefile.dec, makefile.macosx,
1325     makefile.solaris, makefile.hpux, makefile.hpgcc, and makefile.so9.
1327 Version 1.2.4rc1 and 1.0.14rc1 [July 2, 2002]
1328   Added "test-installed" target to makefile.cygwin and makefile.sco.
1329   Revised pnggccrd.c to be able to back out version 1.0.x via PNG_1_0_X macro.
1331 Version 1.2.4 and 1.0.14 [July 8, 2002]
1332   Changed png_warning() to png_error() when width is too large to process.
1334 Version 1.2.4patch01 [July 20, 2002]
1335   Revised makefile.cygwin to use DLL number 12 instead of 13.
1337 Version 1.2.5beta1 [August 6, 2002]
1338   Added code to contrib/gregbook/readpng2.c to ignore unused chunks.
1339   Replaced toucan.png in contrib/gregbook (it has been corrupt since 1.0.11)
1340   Removed some stray *.o files from contrib/gregbook.
1341   Changed png_error() to png_warning() about "Too much data" in pngpread.c
1342     and about "Extra compressed data" in pngrutil.c.
1343   Prevent png_ptr->pass from exceeding 7 in png_push_finish_row().
1344   Updated makefile.hpgcc
1345   Updated png.c and pnggccrd.c handling of return from png_mmx_support()
1347 Version 1.2.5beta2 [August 15, 2002]
1348   Only issue png_warning() about "Too much data" in pngpread.c when avail_in
1349     is nonzero.
1350   Updated makefiles to install a separate libpng.so.3 with its own rpath.
1352 Version 1.2.5rc1 and 1.0.15rc1 [August 24, 2002]
1353   Revised makefiles to not remove previous minor versions of shared libraries.
1355 Version 1.2.5rc2 and 1.0.15rc2 [September 16, 2002]
1356   Revised 13 makefiles to remove "-lz" and "-L$(ZLIBLIB)", etc., from shared
1357     library loader directive.
1358   Added missing "$OBJSDLL" line to makefile.gcmmx.
1359   Added missing "; fi" to makefile.32sunu.
1361 Version 1.2.5rc3 and 1.0.15rc3 [September 18, 2002]
1362   Revised libpng-config script.
1364 Version 1.2.5 and 1.0.15 [October 3, 2002]
1365   Revised makefile.macosx, makefile.darwin, makefile.hpgcc, and makefile.hpux,
1366     and makefile.aix.
1367   Relocated two misplaced PNGAPI lines in pngtest.c
1369 Version 1.2.6beta1 [October 22, 2002]
1370   Commented out warning about uninitialized mmx_support in pnggccrd.c.
1371   Changed "IBMCPP__" flag to "__IBMCPP__" in pngconf.h.
1372   Relocated two more misplaced PNGAPI lines in pngtest.c
1373   Fixed memory overrun bug in png_do_read_filler() with 16-bit datastreams,
1374     introduced in version 1.0.2.
1375   Revised makefile.macosx, makefile.dec, makefile.aix, and makefile.32sunu.
1377 Version 1.2.6beta2 [November 1, 2002]
1378   Added libpng-config "--ldopts" output.
1379   Added "AR=ar" and "ARFLAGS=rc" and changed "ar rc" to "$(AR) $(ARFLAGS)"
1380     in makefiles.
1382 Version 1.2.6beta3 [July 18, 2004]
1383   Reverted makefile changes from version 1.2.6beta2 and some of the changes
1384     from version 1.2.6beta1; these will be postponed until version 1.2.7.
1385     Version 1.2.6 is going to be a simple bugfix release.
1386   Changed the one instance of "ln -sf" to "ln -f -s" in each Sun makefile.
1387   Fixed potential overrun in pngerror.c by using strncpy instead of memcpy.
1388   Added "#!/bin/sh" at the top of configure, for recognition of the
1389     'x' flag under Cygwin (Cosmin).
1390   Optimized vacuous tests that silence compiler warnings, in png.c (Cosmin).
1391   Added support for PNG_USER_CONFIG, in pngconf.h (Cosmin).
1392   Fixed the special memory handler for Borland C under DOS, in pngmem.c
1393     (Cosmin).
1394   Removed some spurious assignments in pngrutil.c (Cosmin).
1395   Replaced 65536 with 65536L, and 0xffff with 0xffffL, to silence warnings
1396     on 16-bit platforms (Cosmin).
1397   Enclosed shift op expressions in parentheses, to silence warnings (Cosmin).
1398   Used proper type png_fixed_point, to avoid problems on 16-bit platforms,
1399     in png_handle_sRGB() (Cosmin).
1400   Added compression_type to png_struct, and optimized the window size
1401     inside the deflate stream (Cosmin).
1402   Fixed definition of isnonalpha(), in pngerror.c and pngrutil.c (Cosmin).
1403   Fixed handling of unknown chunks that come after IDAT (Cosmin).
1404   Allowed png_error() and png_warning() to work even if png_ptr == NULL
1405     (Cosmin).
1406   Replaced row_info->rowbytes with row_bytes in png_write_find_filter()
1407     (Cosmin).
1408   Fixed definition of PNG_LIBPNG_VER_DLLNUM (Simon-Pierre).
1409   Used PNG_LIBPNG_VER and PNG_LIBPNG_VER_STRING instead of the hardcoded
1410     values in png.c (Simon-Pierre, Cosmin).
1411   Initialized png_libpng_ver[] with PNG_LIBPNG_VER_STRING (Simon-Pierre).
1412   Replaced PNG_LIBPNG_VER_MAJOR with PNG_LIBPNG_VER_DLLNUM in png.rc
1413     (Simon-Pierre).
1414   Moved the definition of PNG_HEADER_VERSION_STRING near the definitions
1415     of the other PNG_LIBPNG_VER_... symbols in png.h (Cosmin).
1416   Relocated #ifndef PNGAPI guards in pngconf.h (Simon-Pierre, Cosmin).
1417   Updated scripts/makefile.vc(a)win32 (Cosmin).
1418   Updated the MSVC project (Simon-Pierre, Cosmin).
1419   Updated the Borland C++ Builder project (Cosmin).
1420   Avoided access to asm_flags in pngvcrd.c, if PNG_1_0_X is defined (Cosmin).
1421   Commented out warning about uninitialized mmx_support in pngvcrd.c (Cosmin).
1422   Removed scripts/makefile.bd32 and scripts/pngdef.pas (Cosmin).
1423   Added extra guard around inclusion of Turbo C memory headers, in pngconf.h
1424     (Cosmin).
1425   Renamed projects/msvc/ to projects/visualc6/, and projects/borland/ to
1426     projects/cbuilder5/ (Cosmin).
1427   Moved projects/visualc6/png32ms.def to scripts/pngw32.def,
1428     and projects/visualc6/png.rc to scripts/pngw32.rc (Cosmin).
1429   Added projects/visualc6/pngtest.dsp; removed contrib/msvctest/ (Cosmin).
1430   Changed line endings to DOS style in cbuilder5 and visualc6 files, even
1431     in the tar.* distributions (Cosmin).
1432   Updated contrib/visupng/VisualPng.dsp (Cosmin).
1433   Updated contrib/visupng/cexcept.h to version 2.0.0 (Cosmin).
1434   Added a separate distribution with "configure" and supporting files (Junichi).
1436 Version 1.2.6beta4 [July 28, 2004]
1437   Added user ability to change png_size_t via a PNG_SIZE_T macro.
1438   Added png_sizeof() and png_convert_size() functions.
1439   Added PNG_SIZE_MAX (maximum value of a png_size_t variable.
1440   Added check in png_malloc_default() for (size_t)size != (png_uint_32)size
1441     which would indicate an overflow.
1442   Changed sPLT failure action from png_error to png_warning and abandon chunk.
1443   Changed sCAL and iCCP failures from png_error to png_warning and abandon.
1444   Added png_get_uint_31(png_ptr, buf) function.
1445   Added PNG_UINT_32_MAX macro.
1446   Renamed PNG_MAX_UINT to PNG_UINT_31_MAX.
1447   Made png_zalloc() issue a png_warning and return NULL on potential
1448     overflow.
1449   Turn on PNG_NO_ZALLOC_ZERO by default in version 1.2.x
1450   Revised "clobber list" in pnggccrd.c so it will compile under gcc-3.4.
1451   Revised Borland portion of png_malloc() to return NULL or issue
1452     png_error() according to setting of PNG_FLAG_MALLOC_NULL_MEM_OK.
1453   Added PNG_NO_SEQUENTIAL_READ_SUPPORTED macro to conditionally remove
1454     sequential read support.
1455   Added some "#if PNG_WRITE_SUPPORTED" blocks.
1456   Added #ifdef to remove some redundancy in png_malloc_default().
1457   Use png_malloc instead of png_zalloc to allocate the pallete.
1459 Version 1.0.16rc1 and 1.2.6rc1 [August 4, 2004]
1460   Fixed buffer overflow vulnerability (CVE-2004-0597) in png_handle_tRNS().
1461   Fixed NULL dereference vulnerability (CVE-2004-0598) in png_handle_iCCP().
1462   Fixed integer overflow vulnerability (CVE-2004-0599) in png_read_png().
1463   Fixed some harmless bugs in png_handle_sBIT, etc, that would cause
1464     duplicate chunk types to go undetected.
1465   Fixed some timestamps in the -config version
1466   Rearranged order of processing of color types in png_handle_tRNS().
1467   Added ROWBYTES macro to calculate rowbytes without integer overflow.
1468   Updated makefile.darwin and removed makefile.macosx from scripts directory.
1469   Imposed default one million column, one-million row limits on the image
1470     dimensions, and added png_set_user_limits() function to override them.
1471   Revised use of PNG_SET_USER_LIMITS_SUPPORTED macro.
1472   Fixed wrong cast of returns from png_get_user_width|height_max().
1473   Changed some "keep the compiler happy" from empty statements to returns,
1474   Revised libpng.txt to remove 1.2.x stuff from the 1.0.x distribution
1476 Version 1.0.16rc2 and 1.2.6rc2 [August 7, 2004]
1477   Revised makefile.darwin and makefile.solaris.  Removed makefile.macosx.
1478   Revised pngtest's png_debug_malloc() to use png_malloc() instead of
1479     png_malloc_default() which is not supposed to be exported.
1480   Fixed off-by-one error in one of the conversions to PNG_ROWBYTES() in
1481     pngpread.c.  Bug was introduced in 1.2.6rc1.
1482   Fixed bug in RGB to RGBX transformation introduced in 1.2.6rc1.
1483   Fixed old bug in RGB to Gray transformation.
1484   Fixed problem with 64-bit compilers by casting arguments to abs()
1485     to png_int_32.
1486   Changed "ln -sf" to "ln -f -s" in three makefiles (solaris, sco, so9).
1487   Changed "HANDLE_CHUNK_*" to "PNG_HANDLE_CHUNK_*" (Cosmin)
1488   Added "-@/bin/rm -f $(DL)/$(LIBNAME).so.$(PNGMAJ)" to 15 *NIX makefiles.
1489   Added code to update the row_info->colortype in png_do_read_filler() (MSB).
1491 Version 1.0.16rc3 and 1.2.6rc3 [August 9, 2004]
1492   Eliminated use of "abs()" in testing cHRM and gAMA values, to avoid
1493     trouble with some 64-bit compilers.  Created PNG_OUT_OF_RANGE() macro.
1494   Revised documentation of png_set_keep_unknown_chunks().
1495   Check handle_as_unknown status in pngpread.c, as in pngread.c previously.
1496   Moved  "PNG_HANDLE_CHUNK_*" macros out of PNG_INTERNAL section of png.h
1497   Added "rim" definitions for CONST4 and CONST6 in pnggccrd.c
1499 Version 1.0.16rc4 and 1.2.6rc4 [August 10, 2004]
1500   Fixed mistake in pngtest.c introduced in 1.2.6rc2 (declaration of
1501     "pinfo" was out of place).
1503 Version 1.0.16rc5 and 1.2.6rc5 [August 10, 2004]
1505     section of png.h where they were inadvertently placed in version rc3.
1507 Version 1.2.6 and 1.0.16 [August 15, 2004]
1508   Revised pngtest so memory allocation testing is only done when PNG_DEBUG==1.
1510 Version 1.2.7beta1 [August 26, 2004]
1511   Removed unused pngasmrd.h file.
1512   Removed references to uu.net for archived files.  Added references to
1513     PNG Spec (second edition) and the PNG ISO/IEC Standard.
1514   Added "test-dd" target in 15 makefiles, to run pngtest in DESTDIR.
1515   Fixed bug with "optimized window size" in the IDAT datastream, that
1516     causes libpng to write PNG files with incorrect zlib header bytes.
1518 Version 1.2.7beta2 [August 28, 2004]
1519   Fixed bug with sCAL chunk and big-endian machines (David Munro).
1520   Undid new code added in 1.2.6rc2 to update the color_type in
1521     png_set_filler().
1522   Added png_set_add_alpha() that updates color type.
1524 Version 1.0.17rc1 and 1.2.7rc1 [September 4, 2004]
1525   Revised png_set_strip_filler() to not remove alpha if color_type has alpha.
1527 Version 1.2.7 and 1.0.17 [September 12, 2004]
1528   Added makefile.hp64
1529   Changed projects/msvc/png32ms.def to scripts/png32ms.def in makefile.cygwin
1531 Version 1.2.8beta1 [November 1, 2004]
1532   Fixed bug in png_text_compress() that would fail to complete a large block.
1533   Fixed bug, introduced in libpng-1.2.7, that overruns a buffer during
1534     strip alpha operation in png_do_strip_filler().
1535   Added PNG_1_2_X definition in pngconf.h
1536   Use #ifdef to comment out png_info_init in png.c and png_read_init in
1537     pngread.c (as of 1.3.0)
1539 Version 1.2.8beta2 [November 2, 2004]
1540   Reduce color_type to a nonalpha type after strip alpha operation in
1541     png_do_strip_filler().
1543 Version 1.2.8beta3 [November 3, 2004]
1544   Revised definitions of PNG_MAX_UINT_32, PNG_MAX_SIZE, and PNG_MAXSUM
1546 Version 1.2.8beta4 [November 12, 2004]
1547   Fixed (again) definition of PNG_LIBPNG_VER_DLLNUM in png.h (Cosmin).
1548   Added PNG_LIBPNG_BUILD_PRIVATE in png.h (Cosmin).
1549   Set png_ptr->zstream.data_type to Z_BINARY, to avoid unnecessary detection
1550     of data type in deflate (Cosmin).
1551   Deprecated but continue to support SPECIALBUILD and PRIVATEBUILD in favor of
1554 Version 1.2.8beta5 [November 20, 2004]
1555   Use png_ptr->flags instead of png_ptr->transformations to pass
1556     PNG_STRIP_ALPHA info to png_do_strip_filler(), to preserve ABI
1557     compatibility.
1558   Revised handling of SPECIALBUILD, PRIVATEBUILD,
1561 Version 1.2.8rc1 [November 24, 2004]
1562   Moved handling of BUILD macros from pngconf.h to png.h
1563   Added definition of PNG_LIBPNG_BASE_TYPE in png.h, inadvertently
1564     omitted from beta5.
1565   Revised scripts/pngw32.rc
1566   Despammed mailing addresses by masking "@" with "at".
1567   Inadvertently installed a supposedly faster test version of pngrutil.c
1569 Version 1.2.8rc2 [November 26, 2004]
1570   Added two missing "\" in png.h
1571   Change tests in pngread.c and pngpread.c to
1572     if (png_ptr->transformations || (png_ptr->flags&PNG_FLAG_STRIP_ALPHA))
1573        png_do_read_transformations(png_ptr);
1575 Version 1.2.8rc3 [November 28, 2004]
1576   Reverted pngrutil.c to version libpng-1.2.8beta5.
1577   Added scripts/makefile.elf with supporting code in pngconf.h for symbol
1578     versioning (John Bowler).
1580 Version 1.2.8rc4 [November 29, 2004]
1581   Added projects/visualc7 (Simon-pierre).
1583 Version 1.2.8rc5 [November 29, 2004]
1584   Fixed new typo in scripts/pngw32.rc
1586 Version 1.2.8 [December 3, 2004]
1587   Removed projects/visualc7, added projects/visualc71.
1589 Version 1.2.9beta1 [February 21, 2006]
1590   Initialized some structure members in pngwutil.c to avoid gcc-4.0.0 complaints
1591   Revised man page and libpng.txt to make it clear that one should not call
1592     png_read_end or png_write_end after png_read_png or png_write_png.
1593   Updated references to png-mng-implement mailing list.
1594   Fixed an incorrect typecast in pngrutil.c
1595   Added PNG_NO_READ_SUPPORTED conditional for making a write-only library.
1597   Optimized alpha-inversion loops in pngwtran.c
1598   Moved test for nonzero gamma outside of png_build_gamma_table() in pngrtran.c
1599   Make sure num_trans is <= 256 before copying data in png_set_tRNS().
1600   Make sure num_palette is <= 256 before copying data in png_set_PLTE().
1601   Interchanged order of write_swap_alpha and write_invert_alpha transforms.
1602   Added parentheses in the definition of PNG_LIBPNG_BUILD_TYPE (Cosmin).
1603   Optimized zlib window flag (CINFO) in contrib/pngsuite/*.png (Cosmin).
1604   Updated scripts/makefile.bc32 for Borland C++ 5.6 (Cosmin).
1605   Exported png_get_uint_32, png_save_uint_32, png_get_uint_16, png_save_uint_16,
1606     png_get_int_32, png_save_int_32, png_get_uint_31 (Cosmin).
1607   Added type cast (png_byte) in png_write_sCAL() (Cosmin).
1608   Fixed scripts/makefile.cygwin (Christian Biesinger, Cosmin).
1609   Default iTXt support was inadvertently enabled.
1611 Version 1.2.9beta2 [February 21, 2006]
1612   Check for png_rgb_to_gray and png_gray_to_rgb read transformations before
1613     checking for png_read_dither in pngrtran.c
1614   Revised checking of chromaticity limits to accommodate extended RGB
1615     colorspace (John Denker).
1616   Changed line endings in some of the project files to CRLF, even in the
1617     "Unix" tar distributions (Cosmin).
1618   Made png_get_int_32 and png_save_int_32 always available (Cosmin).
1619   Updated scripts/pngos2.def, scripts/pngw32.def and projects/wince/png32ce.def
1620     with the newly exported functions.
1621   Eliminated distributions without the "configure" script.
1622   Updated INSTALL instructions.
1624 Version 1.2.9beta3 [February 24, 2006]
1625   Fixed CRCRLF line endings in contrib/visupng/VisualPng.dsp
1626   Made libpng.pc respect EXEC_PREFIX (D. P. Kreil, J. Bowler)
1627   Removed reference to pngasmrd.h from Makefile.am
1628   Renamed CHANGES to ChangeLog.
1629   Renamed LICENSE to COPYING.
1630   Renamed ANNOUNCE to NEWS.
1631   Created AUTHORS file.
1633 Version 1.2.9beta4 [March 3, 2006]
1634   Changed definition of PKGCONFIG from $prefix/lib to $libdir in configure.ac
1635   Reverted to filenames LICENSE and ANNOUNCE; removed AUTHORS and COPYING.
1636   Removed newline from the end of some error and warning messages.
1637   Removed test for sqrt() from configure.ac and configure.
1638   Made swap tables in pngtrans.c PNG_CONST (Carlo Bramix).
1639   Disabled default iTXt support that was inadvertently enabled in
1640     libpng-1.2.9beta1.
1641   Added "OS2" to list of systems that don't need underscores, in pnggccrd.c
1642   Removed libpng version and date from *.c files.
1644 Version 1.2.9beta5 [March 4, 2006]
1645   Removed trailing blanks from source files.
1646   Put version and date of latest change in each source file, and changed
1647     copyright year accordingly.
1648   More cleanup of configure.ac, Makefile.am, and associated scripts.
1649   Restored scripts/makefile.elf which was inadvertently deleted.
1651 Version 1.2.9beta6 [March 6, 2006]
1652   Fixed typo (RELEASE) in configuration files.
1654 Version 1.2.9beta7 [March 7, 2006]
1655   Removed libpng.vers and libpng.sym from libpng12_la_SOURCES in Makefile.am
1656   Fixed inconsistent #ifdef's around png_sig_bytes() and png_set_sCAL_s()
1657     in png.h.
1658   Updated makefile.elf as suggested by debian.
1659   Made cosmetic changes to some makefiles, adding LN_SF and other macros.
1660   Made some makefiles accept "exec_prefix".
1662 Version 1.2.9beta8 [March 9, 2006]
1663   Fixed some "#if defined (..." which should be "#if defined(..."
1664     Bug introduced in libpng-1.2.8.
1665   Fixed inconsistency in definition of png_default_read_data()
1666   Restored blank that was lost from makefile.sggcc "clean" target in beta7.
1667   Revised calculation of "current" and "major" for irix in ltmain.sh
1668   Changed "mkdir" to "MKDIR_P" in some makefiles.
1669   Separated PNG_EXPAND and PNG_EXPAND_tRNS.
1670   Added png_set_expand_gray_1_2_4_to_8() and deprecated
1671     png_set_gray_1_2_4_to_8() which also expands tRNS to alpha.
1673 Version 1.2.9beta9 [March 10, 2006]
1674   Include "config.h" in pngconf.h when available.
1675   Added some checks for NULL png_ptr or NULL info_ptr (timeless)
1677 Version 1.2.9beta10 [March 20, 2006]
1678   Removed extra CR from contrib/visualpng/VisualPng.dsw (Cosmin)
1679   Made pnggccrd.c PIC-compliant (Christian Aichinger).
1680   Added makefile.mingw (Wolfgang Glas).
1681   Revised pngconf.h MMX checking.
1683 Version 1.2.9beta11 [March 22, 2006]
1684   Fixed out-of-order declaration in pngwrite.c that was introduced in beta9
1685   Simplified some makefiles by using LIBSO, LIBSOMAJ, and LIBSOVER macros.
1687 Version 1.2.9rc1 [March 31, 2006]
1688   Defined PNG_USER_PRIVATEBUILD when including "pngusr.h" (Cosmin).
1689   Removed nonsensical assertion check from pngtest.c (Cosmin).
1691 Version 1.2.9 [April 14, 2006]
1692   Revised makefile.beos and added "none" selector in ltmain.sh
1694 Version 1.2.10beta1 [April 15, 2006]
1695   Renamed "config.h" to "png_conf.h" and revised Makefile.am to add
1696     -DPNG_BUILDING_LIBPNG to compile directive, and modified pngconf.h
1697     to include png_conf.h only when PNG_BUILDING_LIBPNG is defined.
1699 Version 1.2.10beta2 [April 15, 2006]
1700   Manually updated Makefile.in and configure.  Changed png_conf.h.in
1701     back to config.h.
1703 Version 1.2.10beta3 [April 15, 2006]
1704   Change png_conf.h back to config.h in pngconf.h.
1706 Version 1.2.10beta4 [April 16, 2006]
1707   Change PNG_BUILDING_LIBPNG to PNG_CONFIGURE_LIBPNG in config/Makefile*.
1709 Version 1.2.10beta5 [April 16, 2006]
1710   Added a configure check for compiling assembler code in pnggccrd.c
1712 Version 1.2.10beta6 [April 17, 2006]
1713   Revised the configure check for pnggccrd.c
1715   Added @LIBPNG_DEFINES@ to arguments when building libpng.sym
1717 Version 1.2.10beta7 [April 18, 2006]
1718   Change "exec_prefix=$prefix" to "exec_prefix=$(prefix)" in makefiles.
1720 Version 1.2.10rc1 [April 19, 2006]
1721   Ensure pngconf.h doesn't define both PNG_USE_PNGGCCRD and PNG_USE_PNGVCRD
1722   Fixed "LN_FS" typo in makefile.sco and makefile.solaris.
1724 Version 1.2.10rc2 [April 20, 2006]
1725   Added a backslash between -DPNG_CONFIGURE_LIBPNG and -DPNG_NO_ASSEMBLER_CODE
1726    in configure.ac and configure
1727   Made the configure warning about versioned symbols less arrogant.
1729 Version 1.2.10rc3 [April 21, 2006]
1730   Added a note in libpng.txt that png_set_sig_bytes(8) can be used when
1731     writing an embedded PNG without the 8-byte signature.
1732   Revised makefiles and configure to avoid making links to libpng.so.*
1734 Version 1.2.10 [April 23, 2006]
1735   Reverted configure to "rc2" state.
1737 Version 1.2.11beta1 [May 31, 2006]
1738   scripts/libpng.pc.in contained "configure" style version info and would
1739     not work with makefiles.
1740   The shared-library makefiles were linking to libpng.so.0 instead of
1741     libpng.so.3 compatibility as the library.
1743 Version 1.2.11beta2 [June 2, 2006]
1744   Increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
1745     buffer overflow.
1746   Fixed bug in example.c (png_set_palette_rgb -> png_set_palette_to_rgb)
1748 Version 1.2.11beta3 [June 5, 2006]
1749   Prepended "#! /bin/sh" to ltmail.sh and contrib/pngminus/*.sh (Cosmin).
1750   Removed the accidental leftover Makefile.in~ (Cosmin).
1751   Avoided potential buffer overflow and optimized buffer in
1752     png_write_sCAL(), png_write_sCAL_s() (Cosmin).
1753   Removed the include directories and libraries from CFLAGS and LDFLAGS
1754     in scripts/makefile.gcc (Nelson A. de Oliveira, Cosmin).
1756 Version 1.2.11beta4 [June 6, 2006]
1757   Allow zero-length IDAT chunks after the entire zlib datastream, but not
1758     after another intervening chunk type.
1760 Version 1.0.19rc1, 1.2.11rc1 [June 13, 2006]
1761   Deleted extraneous square brackets from [config.h] in configure.ac
1763 Version 1.0.19rc2, 1.2.11rc2 [June 14, 2006]
1764   Added prototypes for PNG_INCH_CONVERSIONS functions to png.h
1765   Revised INSTALL and autogen.sh
1766   Fixed typo in several makefiles (-W1 should be -Wl)
1767   Added typedef for png_int_32 and png_uint_32 on 64-bit systems.
1769 Version 1.0.19rc3, 1.2.11rc3 [June 15, 2006]
1770   Removed the new typedefs for 64-bit systems (delay until version 1.4.0)
1771   Added one zero element to png_gamma_shift[] array in pngrtran.c to avoid
1772     reading out of bounds.
1774 Version 1.0.19rc4, 1.2.11rc4 [June 15, 2006]
1775   Really removed the new typedefs for 64-bit systems.
1777 Version 1.0.19rc5, 1.2.11rc5 [June 22, 2006]
1778   Removed png_sig_bytes entry from scripts/pngw32.def
1780 Version 1.0.19, 1.2.11 [June 26, 2006]
1781   None.
1783 Version 1.0.20, 1.2.12 [June 27, 2006]
1784   Really increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
1785     buffer overflow.
1787 Version 1.2.13beta1 [October 2, 2006]
1788   Removed AC_FUNC_MALLOC from configure.ac
1789   Work around Intel-Mac compiler bug by setting PNG_NO_MMX_CODE in pngconf.h
1790   Change "logical" to "bitwise" throughout documentation.
1791   Detect and fix attempt to write wrong iCCP profile length (CVE-2006-7244)
1793 Version 1.0.21, 1.2.13 [November 14, 2006]
1794   Fix potential buffer overflow in sPLT chunk handler.
1795   Fix Makefile.am to not try to link to noexistent files.
1796   Check all exported functions for NULL png_ptr.
1798 Version 1.2.14beta1 [November 17, 2006]
1799   Relocated three misplaced tests for NULL png_ptr.
1800   Built Makefile.in with automake-1.9.6 instead of 1.9.2.
1801   Build configure with autoconf-2.60 instead of 2.59
1803 Version 1.2.14beta2 [November 17, 2006]
1804   Added some typecasts in png_zalloc().
1806 Version 1.2.14rc1 [November 20, 2006]
1807   Changed "strtod" to "png_strtod" in pngrutil.c
1809 Version 1.0.22, 1.2.14    [November 27, 2006]
1810   Added missing "$(srcdir)" in Makefile.am and Makefile.in
1812 Version 1.2.15beta1 [December 3, 2006]
1813   Generated configure with autoconf-2.61 instead of 2.60
1814   Revised configure.ac to update libpng.pc and libpng-config.
1816 Version 1.2.15beta2 [December 3, 2006]
1817   Always export MMX asm functions, just stubs if not building pnggccrd.c
1819 Version 1.2.15beta3 [December 4, 2006]
1820   Add "png_bytep" typecast to profile while calculating length in pngwutil.c
1822 Version 1.2.15beta4 [December 7, 2006]
1823   Added scripts/CMakeLists.txt
1824   Changed PNG_NO_ASSEMBLER_CODE to PNG_NO_MMX_CODE in scripts, like 1.4.0beta
1826 Version 1.2.15beta5 [December 7, 2006]
1827   Changed some instances of PNG_ASSEMBLER_* to PNG_MMX_* in pnggccrd.c
1828   Revised scripts/CMakeLists.txt
1830 Version 1.2.15beta6 [December 13, 2006]
1831   Revised scripts/CMakeLists.txt and configure.ac
1833 Version 1.2.15rc1 [December 18, 2006]
1834   Revised scripts/CMakeLists.txt
1836 Version 1.2.15rc2 [December 21, 2006]
1837   Added conditional #undef jmpbuf in pngtest.c to undo #define in AIX headers.
1838   Added scripts/makefile.nommx
1840 Version 1.2.15rc3 [December 25, 2006]
1841   Fixed shared library numbering error that was introduced in 1.2.15beta6.
1843 Version 1.2.15rc4 [December 27, 2006]
1844   Fixed handling of rgb_to_gray when png_ptr->color.gray isn't set.
1846 Version 1.2.15rc5 [December 31, 2006]
1847   Revised handling of rgb_to_gray.
1849 Version 1.2.15 [January 5, 2007]
1850   Added some (unsigned long) typecasts in pngtest.c to avoid printing errors.
1852 Version 1.2.16beta1 [January 6, 2007]
1853   Fix bugs in makefile.nommx
1855 Version 1.2.16beta2 [January 16, 2007]
1856   Revised scripts/CMakeLists.txt
1858 Version 1.2.16 [January 31, 2007]
1859   No changes.
1861 Version 1.2.17beta1 [March 6, 2007]
1862   Revised scripts/CMakeLists.txt to install both shared and static libraries.
1863   Deleted a redundant line from pngset.c.
1865 Version 1.2.17beta2 [April 26, 2007]
1866   Relocated misplaced test for png_ptr == NULL in pngpread.c
1867   Change "==" to "&" for testing PNG_RGB_TO_GRAY_ERR & PNG_RGB_TO_GRAY_WARN
1868     flags.
1869   Changed remaining instances of PNG_ASSEMBLER_* to PNG_MMX_*
1870   Added pngerror() when write_IHDR fails in deflateInit2().
1871   Added "const" to some array declarations.
1872   Mention examples of libpng usage in the libpng*.txt and libpng.3 documents.
1874 Version 1.2.17rc1 [May 4, 2007]
1875   No changes.
1877 Version 1.2.17rc2 [May 8, 2007]
1878   Moved several PNG_HAVE_* macros out of PNG_INTERNAL because applications
1879     calling set_unknown_chunk_location() need them.
1880   Changed transformation flag from PNG_EXPAND_tRNS to PNG_EXPAND in
1881     png_set_expand_gray_1_2_4_to_8().
1882   Added png_ptr->unknown_chunk to hold working unknown chunk data, so it
1883     can be free'ed in case of error.  Revised unknown chunk handling in
1884     pngrutil.c and pngpread.c to use this structure.
1886 Version 1.2.17rc3 [May 8, 2007]
1887   Revised symbol-handling in configure script.
1889 Version 1.2.17rc4 [May 10, 2007]
1890   Revised unknown chunk handling to avoid storing unknown critical chunks.
1892 Version 1.0.25 [May 15, 2007]
1893 Version 1.2.17 [May 15, 2007]
1894   Added "png_ptr->num_trans=0" before error return in png_handle_tRNS,
1895     to eliminate a vulnerability (CVE-2007-2445, CERT VU#684664)
1897 Version 1.0.26 [May 15, 2007]
1898 Version 1.2.18 [May 15, 2007]
1899   Reverted the libpng-1.2.17rc3 change to symbol-handling in configure script
1901 Version 1.2.19beta1 [May 18, 2007]
1902   Changed "const static" to "static PNG_CONST" everywhere, mostly undoing
1903     change of libpng-1.2.17beta2.  Changed other "const" to "PNG_CONST"
1904   Changed some handling of unused parameters, to avoid compiler warnings.
1905     "if (unused == NULL) return;" becomes "unused = unused".
1907 Version 1.2.19beta2 [May 18, 2007]
1908   Only use the valid bits of tRNS value in png_do_expand() (Brian Cartier)
1910 Version 1.2.19beta3 [May 19, 2007]
1911   Add some "png_byte" typecasts in png_check_keyword() and write new_key
1912   instead of key in zTXt chunk (Kevin Ryde).
1914 Version 1.2.19beta4 [May 21, 2007]
1915   Add png_snprintf() function and use it in place of sprint() for improved
1916     defense against buffer overflows.
1918 Version 1.2.19beta5 [May 21, 2007]
1919   Fixed png_handle_tRNS() to only use the valid bits of tRNS value.
1920   Changed handling of more unused parameters, to avoid compiler warnings.
1921   Removed some PNG_CONST in pngwutil.c to avoid compiler warnings.
1923 Version 1.2.19beta6 [May 22, 2007]
1924   Added some #ifdef PNG_MMX_CODE_SUPPORTED where needed in pngvcrd.c
1925   Added a special "_MSC_VER" case that defines png_snprintf to _snprintf
1927 Version 1.2.19beta7 [May 22, 2007]
1928   Squelched png_squelch_warnings() in pnggccrd.c and added
1929     an #ifdef PNG_MMX_CODE_SUPPORTED block around the declarations that caused
1930     the warnings that png_squelch_warnings was squelching.
1932 Version 1.2.19beta8 [May 22, 2007]
1933   Removed __MMX__ from test in pngconf.h.
1935 Version 1.2.19beta9 [May 23, 2007]
1936   Made png_squelch_warnings() available via PNG_SQUELCH_WARNINGS macro.
1937   Revised png_squelch_warnings() so it might work.
1938   Updated makefile.sgcc and makefile.solaris; added makefile.solaris-x86.
1940 Version 1.2.19beta10 [May 24, 2007]
1941   Resquelched png_squelch_warnings(), use "__attribute__((used))" instead.
1943 Version 1.4.0beta1 [April 20, 2006]
1944   Enabled iTXt support (changes png_struct, thus requires so-number change).
1946   Eliminated PNG_1_0_X and PNG_1_2_X macros.
1947   Removed deprecated functions png_read_init, png_write_init, png_info_init,
1948     png_permit_empty_plte, png_set_gray_1_2_4_to_8, png_check_sig, and
1949     removed the deprecated macro PNG_MAX_UINT.
1950   Moved "PNG_INTERNAL" parts of png.h and pngconf.h into pngintrn.h
1951   Removed many WIN32_WCE #ifdefs (Cosmin).
1952   Reduced dependency on C-runtime library when on Windows (Simon-Pierre)
1953   Replaced sprintf() with png_sprintf() (Simon-Pierre)
1955 Version 1.4.0beta2 [April 20, 2006]
1956   Revised makefiles and configure to avoid making links to libpng.so.*
1957   Moved some leftover MMX-related defines from pngconf.h to pngintrn.h
1958   Updated scripts/pngos2.def, pngw32.def, and projects/wince/png32ce.def
1960 Version 1.4.0beta3 [May 10, 2006]
1961   Updated scripts/pngw32.def to comment out MMX functions.
1962   Added PNG_NO_GET_INT_32 and PNG_NO_SAVE_INT_32 macros.
1963   Scripts/libpng.pc.in contained "configure" style version info and would
1964     not work with makefiles.
1965   Revised pngconf.h and added pngconf.h.in, so makefiles and configure can
1966     pass defines to libpng and applications.
1968 Version 1.4.0beta4 [May 11, 2006]
1969   Revised configure.ac, Makefile.am, and many of the makefiles to write
1970     their defines in pngconf.h.
1972 Version 1.4.0beta5 [May 15, 2006]
1973   Added a missing semicolon in Makefile.am and Makefile.in
1974   Deleted extraneous square brackets from configure.ac
1976 Version 1.4.0beta6 [June 2, 2006]
1977   Increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
1978     buffer overflow.
1979   Changed sonum from 0 to 1.
1980   Removed unused prototype for png_check_sig() from png.h
1982 Version 1.4.0beta7 [June 16, 2006]
1983   Exported png_write_sig (Cosmin).
1984   Optimized buffer in png_handle_cHRM() (Cosmin).
1985   Set pHYs = 2835 x 2835 pixels per meter, and added
1986     sCAL = 0.352778e-3 x 0.352778e-3 meters, in pngtest.png (Cosmin).
1987   Added png_set_benign_errors(), png_benign_error(), png_chunk_benign_error().
1988   Added typedef for png_int_32 and png_uint_32 on 64-bit systems.
1989   Added "(unsigned long)" typecast on png_uint_32 variables in printf lists.
1991 Version 1.4.0beta8 [June 22, 2006]
1992   Added demonstration of user chunk support in pngtest.c, to support the
1993     public sTER chunk and a private vpAg chunk.
1995 Version 1.4.0beta9 [July 3, 2006]
1996   Removed ordinals from scripts/pngw32.def and removed png_info_int and
1997     png_set_gray_1_2_4_to_8 entries.
1998   Inline call of png_get_uint_32() in png_get_uint_31().
1999   Use png_get_uint_31() to get vpAg width and height in pngtest.c
2000   Removed WINCE and Netware projects.
2001   Removed standalone Y2KINFO file.
2003 Version 1.4.0beta10 [July 12, 2006]
2004   Eliminated automatic copy of pngconf.h to pngconf.h.in from configure and
2005     some makefiles, because it was not working reliably.  Instead, distribute
2006     pngconf.h.in along with pngconf.h and cause configure and some of the
2007     makefiles to update pngconf.h from pngconf.h.in.
2008   Added pngconf.h to DEPENDENCIES in Makefile.am
2010 Version 1.4.0beta11 [August 19, 2006]
2011   Removed AC_FUNC_MALLOC from configure.ac.
2012   Added a warning when writing iCCP profile with mismatched profile length.
2013   Patched pnggccrd.c to assemble on x86_64 platforms.
2014   Moved chunk header reading into a separate function png_read_chunk_header()
2015     in pngrutil.c.  The chunk header (len+sig) is now serialized in a single
2016     operation (Cosmin).
2017   Implemented support for I/O states. Added png_ptr member io_state, and
2018     functions png_get_io_chunk_name() and png_get_io_state() in pngget.c
2019     (Cosmin).
2020   Added png_get_io_chunk_name and png_get_io_state to scripts/*.def (Cosmin).
2021   Renamed scripts/pngw32.* to scripts/pngwin.* (Cosmin).
2022   Removed the include directories and libraries from CFLAGS and LDFLAGS
2023     in scripts/makefile.gcc (Cosmin).
2024   Used png_save_uint_32() to set vpAg width and height in pngtest.c (Cosmin).
2025   Cast to proper type when getting/setting vpAg units in pngtest.c (Cosmin).
2026   Added pngintrn.h to the Visual C++ projects (Cosmin).
2027   Removed scripts/list (Cosmin).
2028   Updated copyright year in scripts/pngwin.def (Cosmin).
2029   Removed PNG_TYPECAST_NULL and used standard NULL consistently (Cosmin).
2030   Disallowed the user to redefine png_size_t, and enforced a consistent use
2031     of png_size_t across libpng (Cosmin).
2032   Changed the type of png_ptr->rowbytes, PNG_ROWBYTES() and friends
2033     to png_size_t (Cosmin).
2034   Removed png_convert_size() and replaced png_sizeof with sizeof (Cosmin).
2035   Removed some unnecessary type casts (Cosmin).
2036   Changed prototype of png_get_compression_buffer_size() and
2037     png_set_compression_buffer_size() to work with png_size_t instead of
2038     png_uint_32 (Cosmin).
2039   Removed png_memcpy_check() and png_memset_check() (Cosmin).
2040   Fixed a typo (png_byte --> png_bytep) in libpng.3 and libpng.txt (Cosmin).
2041   Clarified that png_zalloc() does not clear the allocated memory,
2042     and png_zalloc() and png_zfree() cannot be PNGAPI (Cosmin).
2043   Renamed png_mem_size_t to png_alloc_size_t, fixed its definition in
2044     pngconf.h, and used it in all memory allocation functions (Cosmin).
2045   Renamed pngintrn.h to pngpriv.h, added a comment at the top of the file
2046     mentioning that the symbols declared in that file are private, and
2047     updated the scripts and the Visual C++ projects accordingly (Cosmin).
2048   Removed circular references between pngconf.h and pngconf.h.in in
2049     scripts/makefile.vc*win32 (Cosmin).
2050   Removing trailing '.' from the warning and error messages (Cosmin).
2051   Added pngdefs.h that is built by makefile or configure, instead of
2052     pngconf.h.in (Glenn).
2053   Detect and fix attempt to write wrong iCCP profile length.
2055 Version 1.4.0beta12 [October 19, 2006]
2056   Changed "logical" to "bitwise" in the documentation.
2057   Work around Intel-Mac compiler bug by setting PNG_NO_MMX_CODE in pngconf.h
2058   Add a typecast to stifle compiler warning in pngrutil.c
2060 Version 1.4.0beta13 [November 10, 2006]
2061   Fix potential buffer overflow in sPLT chunk handler.
2062   Fix Makefile.am to not try to link to noexistent files.
2064 Version 1.4.0beta14 [November 15, 2006]
2065   Check all exported functions for NULL png_ptr.
2067 Version 1.4.0beta15 [November 17, 2006]
2068   Relocated two misplaced tests for NULL png_ptr.
2069   Built Makefile.in with automake-1.9.6 instead of 1.9.2.
2070   Build configure with autoconf-2.60 instead of 2.59
2071   Add "install: all" in Makefile.am so "configure; make install" will work.
2073 Version 1.4.0beta16 [November 17, 2006]
2074   Added a typecast in png_zalloc().
2076 Version 1.4.0beta17 [December 4, 2006]
2077   Changed "new_key[79] = '\0';" to "(*new_key)[79] = '\0';" in pngwutil.c
2078   Add "png_bytep" typecast to profile while calculating length in pngwutil.c
2080 Version 1.4.0beta18 [December 7, 2006]
2081   Added scripts/CMakeLists.txt
2083 Version 1.4.0beta19 [May 16, 2007]
2084   Revised scripts/CMakeLists.txt
2085   Rebuilt configure and Makefile.in with newer tools.
2086   Added conditional #undef jmpbuf in pngtest.c to undo #define in AIX headers.
2087   Added scripts/makefile.nommx
2089 Version 1.4.0beta20 [July 9, 2008]
2090   Moved several PNG_HAVE_* macros from pngpriv.h to png.h because applications
2091     calling set_unknown_chunk_location() need them.
2092   Moved several macro definitions from pngpriv.h to pngconf.h
2093   Merge with changes to the 1.2.X branch, as of 1.2.30beta04.
2094   Deleted all use of the MMX assembler code and Intel-licensed optimizations.
2095   Revised makefile.mingw
2097 Version 1.4.0beta21 [July 21, 2008]
2098   Moved local array "chunkdata" from pngrutil.c to the png_struct, so
2099     it will be freed by png_read_destroy() in case of a read error (Kurt
2100     Christensen).
2102 Version 1.4.0beta22 [July 21, 2008]
2103   Change "purpose" and "buffer" to png_ptr->chunkdata to avoid memory leaking.
2105 Version 1.4.0beta23 [July 22, 2008]
2106   Change "chunkdata = NULL" to "png_ptr->chunkdata = NULL" several places in
2107     png_decompress_chunk().
2109 Version 1.4.0beta24 [July 25, 2008]
2110   Change all remaining "chunkdata" to "png_ptr->chunkdata" in
2111     png_decompress_chunk(), and remove "chunkdata" from parameter list.
2112   Put a call to png_check_chunk_name() in png_read_chunk_header().
2113   Revised png_check_chunk_name() to reject a name with a lowercase 3rd byte.
2114   Removed two calls to png_check_chunk_name() occurring later in the process.
2115   Define PNG_NO_ERROR_NUMBERS by default in pngconf.h
2117 Version 1.4.0beta25 [July 30, 2008]
2118   Added a call to png_check_chunk_name() in pngpread.c
2119   Reverted png_check_chunk_name() to accept a name with a lowercase 3rd byte.
2120   Added png_push_have_buffer() function to pngpread.c
2121   Eliminated PNG_BIG_ENDIAN_SUPPORTED and associated png_get_* macros.
2122   Made inline expansion of png_get_*() optional with PNG_USE_READ_MACROS.
2124   Synced contrib directory and configure files with libpng-1.2.30beta06.
2125   Eliminated no-longer-used pngdefs.h (but it's still built in the makefiles)
2126   Relocated a misplaced "#endif /* PNG_NO_WRITE_FILTER */" in pngwutil.c
2128 Version 1.4.0beta26 [August 4, 2008]
2129   Removed png_push_have_buffer() function in pngpread.c.  It increased the
2130     compiled library size slightly.
2131   Changed "-Wall" to "-W -Wall" in the CFLAGS in all makefiles (Cosmin Truta)
2132   Declared png_ptr "volatile" in pngread.c and pngwrite.c to avoid warnings.
2133   Updated contrib/visupng/cexcept.h to version 2.0.1
2134   Added PNG_LITERAL_CHARACTER macros for #, [, and ].
2136 Version 1.4.0beta27 [August 5, 2008]
2137   Revised usage of PNG_LITERAL_SHARP in pngerror.c.
2138   Moved newline character from individual png_debug messages into the
2139     png_debug macros.
2140   Allow user to #define their own png_debug, png_debug1, and png_debug2.
2142 Version 1.4.0beta28 [August 5, 2008]
2143   Revised usage of PNG_LITERAL_SHARP in pngerror.c.
2144   Added PNG_STRING_NEWLINE macro
2146 Version 1.4.0beta29 [August 9, 2008]
2147   Revised usage of PNG_STRING_NEWLINE to work on non-ISO compilers.
2148   Added PNG_STRING_COPYRIGHT macro.
2149   Added non-ISO versions of png_debug macros.
2151 Version 1.4.0beta30 [August 14, 2008]
2152   Added premultiplied alpha feature (Volker Wiendl).
2154 Version 1.4.0beta31 [August 18, 2008]
2155   Moved png_set_premultiply_alpha from pngtrans.c to pngrtran.c
2156   Removed extra crc check at the end of png_handle_cHRM().  Bug introduced
2157     in libpng-1.4.0beta20.
2159 Version 1.4.0beta32 [August 19, 2008]
2160   Added PNG_WRITE_FLUSH_SUPPORTED block around new png_flush() call.
2161   Revised PNG_NO_STDIO version of png_write_flush()
2163 Version 1.4.0beta33 [August 20, 2008]
2164   Added png_get|set_chunk_cache_max() to limit the total number of sPLT,
2165     text, and unknown chunks that can be stored.
2167 Version 1.4.0beta34 [September 6, 2008]
2168   Shortened tIME_string to 29 bytes in pngtest.c
2169   Fixed off-by-one error introduced in png_push_read_zTXt() function in
2170     libpng-1.2.30beta04/pngpread.c (Harald van Dijk)
2172 Version 1.4.0beta35 [October 6, 2008]
2173   Changed "trans_values" to "trans_color".
2174   Changed so-number from 0 to 14.  Some OS do not like 0.
2175   Revised makefile.darwin to fix shared library numbering.
2176   Change png_set_gray_1_2_4_to_8() to png_set_expand_gray_1_2_4_to_8()
2177     in example.c (debian bug report)
2179 Version 1.4.0beta36 [October 25, 2008]
2180   Sync with tEXt vulnerability fix in libpng-1.2.33rc02.
2182 Version 1.4.0beta37 [November 13, 2008]
2183   Added png_check_cHRM in png.c and moved checking from pngget.c, pngrutil.c,
2184     and pngwrite.c
2186 Version 1.4.0beta38 [November 22, 2008]
2187   Added check for zero-area RGB cHRM triangle in png_check_cHRM() and
2188     png_check_cHRM_fixed().
2190 Version 1.4.0beta39 [November 23, 2008]
2191   Revised png_warning() to write its message on standard output by default
2192     when warning_fn is NULL.
2194 Version 1.4.0beta40 [November 24, 2008]
2195   Eliminated png_check_cHRM().  Instead, always use png_check_cHRM_fixed().
2196   In png_check_cHRM_fixed(), ensure white_y is > 0, and removed redundant
2197     check for all-zero coordinates that is detected by the triangle check.
2199 Version 1.4.0beta41 [November 26, 2008]
2200   Fixed string vs pointer-to-string error in png_check_keyword().
2201   Rearranged test expressions in png_check_cHRM_fixed() to avoid internal
2202     overflows.
2203   Added PNG_NO_CHECK_cHRM conditional.
2205 Version 1.4.0beta42, 43 [December 1, 2008]
2206   Merge png_debug with version 1.2.34beta04.
2208 Version 1.4.0beta44 [December 6, 2008]
2209   Removed redundant check for key==NULL before calling png_check_keyword()
2210     to ensure that new_key gets initialized and removed extra warning
2211     (Merge with version 1.2.34beta05 -- Arvan Pritchard).
2213 Version 1.4.0beta45 [December 9, 2008]
2214   In png_write_png(), respect the placement of the filler bytes in an earlier
2215     call to png_set_filler() (Jim Barry).
2217 Version 1.4.0beta46 [December 10, 2008]
2218   Undid previous change and added PNG_TRANSFORM_STRIP_FILLER_BEFORE and
2219     PNG_TRANSFORM_STRIP_FILLER_AFTER conditionals and deprecated
2222 Version 1.4.0beta47 [December 15, 2008]
2223   Support for dithering was disabled by default, because it has never
2224     been well tested and doesn't work very well.  The code has not
2225     been removed, however, and can be enabled by building libpng with
2228 Version 1.4.0beta48 [February 14, 2009]
2229   Added new exported function png_calloc().
2230   Combined several instances of png_malloc(); png_memset() into png_calloc().
2231   Removed prototype for png_freeptr() that was added in libpng-1.4.0beta24
2232     but was never defined.
2234 Version 1.4.0beta49 [February 28, 2009]
2235   Added png_fileno() macro to pngconf.h, used in pngwio.c
2236   Corrected order of #ifdef's in png_debug definition in png.h
2237   Fixed bug introduced in libpng-1.4.0beta48 with the memset arguments
2238     for pcal_params.
2239   Fixed order of #ifdef directives in the png_debug defines in png.h
2240     (bug introduced in libpng-1.2.34/1.4.0beta29).
2241   Revised comments in png_set_read_fn() and png_set_write_fn().
2243 Version 1.4.0beta50 [March 18, 2009]
2244   Use png_calloc() instead of png_malloc() to allocate big_row_buf when
2245     reading an interlaced file, to avoid a possible UMR.
2246   Undid revision of PNG_NO_STDIO version of png_write_flush().  Users
2247     having trouble with fflush() can build with PNG_NO_WRITE_FLUSH defined
2248     or supply their own flush_fn() replacement.
2249   Revised libpng*.txt and png.h documentation about use of png_write_flush()
2250     and png_set_write_fn().
2251   Removed fflush() from pngtest.c.
2252   Added "#define PNG_NO_WRITE_FLUSH" to contrib/pngminim/encoder/pngusr.h
2254 Version 1.4.0beta51 [March 21, 2009]
2255   Removed new png_fileno() macro from pngconf.h .
2257 Version 1.4.0beta52 [March 27, 2009]
2258   Relocated png_do_chop() ahead of building gamma tables in pngrtran.c
2259     This avoids building 16-bit gamma tables unnecessarily.
2260   Removed fflush() from pngtest.c.
2261   Added "#define PNG_NO_WRITE_FLUSH" to contrib/pngminim/encoder/pngusr.h
2262   Added a section on differences between 1.0.x and 1.2.x to libpng.3/libpng.txt
2264 Version 1.4.0beta53 [April 1, 2009]
2265   Removed some remaining MMX macros from pngpriv.h
2266   Fixed potential memory leak of "new_name" in png_write_iCCP() (Ralph Giles)
2268 Version 1.4.0beta54 [April 13, 2009]
2269   Added "ifndef PNG_SKIP_SETJMP_CHECK" block in pngconf.h to allow
2270     application code writers to bypass the check for multiple inclusion
2271     of setjmp.h when they know that it is safe to ignore the situation.
2272   Eliminated internal use of setjmp() in pngread.c and pngwrite.c
2273   Reordered ancillary chunks in pngtest.png to be the same as what
2274     pngtest now produces, and made some cosmetic changes to pngtest output.
2275   Eliminated deprecated png_read_init_3() and png_write_init_3() functions.
2277 Version 1.4.0beta55 [April 15, 2009]
2278   Simplified error handling in pngread.c and pngwrite.c by putting
2279     the new png_read_cleanup() and png_write_cleanup() functions inline.
2281 Version 1.4.0beta56 [April 25, 2009]
2282   Renamed "user_chunk_data" to "my_user_chunk_data" in pngtest.c to suppress
2283     "shadowed declaration" warning from gcc-4.3.3.
2284   Renamed "gamma" to "png_gamma" in pngset.c to avoid "shadowed declaration"
2285     warning about a global "gamma" variable in math.h on some platforms.
2287 Version 1.4.0beta57 [May 2, 2009]
2288   Removed prototype for png_freeptr() that was added in libpng-1.4.0beta24
2289     but was never defined (again).
2290   Rebuilt configure scripts with autoconf-2.63 instead of 2.62
2291   Removed pngprefs.h and MMX from makefiles
2293 Version 1.4.0beta58 [May 14, 2009]
2294   Changed pngw32.def to pngwin.def in makefile.mingw (typo was introduced
2295     in beta57).
2296   Clarified usage of sig_bit versus sig_bit_p in example.c (Vincent Torri)
2298 Version 1.4.0beta59 [May 15, 2009]
2299   Reformated sources in libpng style (3-space intentation, comment format)
2300   Fixed typo in libpng docs (PNG_FILTER_AVE should be PNG_FILTER_AVG)
2301   Added sections about the git repository and our coding style to the
2302     documentation
2303   Relocated misplaced #endif in pngwrite.c, sCAL chunk handler.
2305 Version 1.4.0beta60 [May 19, 2009]
2306   Conditionally compile png_read_finish_row() which is not used by
2307     progressive readers.
2308   Added contrib/pngminim/preader to demonstrate building minimal progressive
2309     decoder, based on contrib/gregbook with embedded libpng and zlib.
2311 Version 1.4.0beta61 [May 20, 2009]
2312   In contrib/pngminim/*, renamed "makefile.std" to "makefile", since there
2313     is only one makefile in those directories, and revised the README files
2314     accordingly.
2315   More reformatting of comments, mostly to capitalize sentences.
2317 Version 1.4.0beta62 [June 2, 2009]
2318   Added "#define PNG_NO_WRITE_SWAP" to contrib/pngminim/encoder/pngusr.h
2319     and "define PNG_NO_READ_SWAP" to decoder/pngusr.h and preader/pngusr.h
2320   Reformatted several remaining "else statement" into two lines.
2321   Added a section to the libpng documentation about using png_get_io_ptr()
2322     in configure scripts to detect the presence of libpng.
2324 Version 1.4.0beta63 [June 15, 2009]
2325   Revised libpng*.txt and libpng.3 to mention calling png_set_IHDR()
2326     multiple times and to specify the sample order in the tRNS chunk,
2327     because the ISO PNG specification has a typo in the tRNS table.
2328   Changed several PNG_UNKNOWN_CHUNK_SUPPORTED to
2329     PNG_HANDLE_AS_UNKNOWN_SUPPORTED, to make the png_set_keep mechanism
2330     available for ignoring known chunks even when not saving unknown chunks.
2331   Adopted preference for consistent use of "#ifdef" and "#ifndef" versus
2332     "#if defined()" and "if !defined()" where possible.
2334 Version 1.4.0beta64 [June 24, 2009]
2335   Eliminated PNG_LEGACY_SUPPORTED code.
2336   Moved the various unknown chunk macro definitions outside of the
2339 Version 1.4.0beta65 [June 26, 2009]
2340   Added a reference to the libpng license in each file.
2342 Version 1.4.0beta66 [June 27, 2009]
2343   Refer to the libpng license instead of the libpng license in each file.
2345 Version 1.4.0beta67 [July 6, 2009]
2346   Relocated INVERT_ALPHA within png_read_png() and png_write_png().
2347   Added high-level API transform PNG_TRANSFORM_GRAY_TO_RGB.
2348   Added an "xcode" project to the projects directory (Alam Arias).
2350 Version 1.4.0beta68 [July 19, 2009]
2351   Avoid some tests in filter selection in pngwutil.c
2353 Version 1.4.0beta69 [July 25, 2009]
2354   Simplified the new filter-selection test.  This runs faster in the
2355     common "PNG_ALL_FILTERS" and PNG_FILTER_NONE cases.
2356   Removed extraneous declaration from the new call to png_read_gray_to_rgb()
2357     (bug introduced in libpng-1.4.0beta67).
2358   Fixed up xcode project (Alam Arias)
2359   Added a prototype for png_64bit_product() in png.c
2361 Version 1.4.0beta70 [July 27, 2009]
2362   Avoid a possible NULL dereference in debug build, in png_set_text_2().
2363     (bug introduced in libpng-0.95, discovered by Evan Rouault)
2365 Version 1.4.0beta71 [July 29, 2009]
2366   Rebuilt configure scripts with autoconf-2.64.
2368 Version 1.4.0beta72 [August 1, 2009]
2369   Replaced *.tar.lzma with *.tar.xz in distribution.  Get the xz codec
2370     from <http://tukaani.org/xz>.
2372 Version 1.4.0beta73 [August 1, 2009]
2373   Reject attempt to write iCCP chunk with negative embedded profile length
2374     (JD Chen) (CVE-2009-5063).
2376 Version 1.4.0beta74 [August 8, 2009]
2377   Changed png_ptr and info_ptr member "trans" to "trans_alpha".
2379 Version 1.4.0beta75 [August 21, 2009]
2380   Removed an extra png_debug() recently added to png_write_find_filter().
2381   Fixed incorrect #ifdef in pngset.c regarding unknown chunk support.
2383 Version 1.4.0beta76 [August 22, 2009]
2384   Moved an incorrectly located test in png_read_row() in pngread.c
2386 Version 1.4.0beta77 [August 27, 2009]
2387   Removed lpXYZ.tar.bz2 (with CRLF), KNOWNBUG, libpng-x.y.z-KNOWNBUG.txt,
2388     and the "noconfig" files from the distribution.
2389   Moved CMakeLists.txt from scripts into the main libpng directory.
2390   Various bugfixes and improvements to CMakeLists.txt (Philip Lowman)
2392 Version 1.4.0beta78 [August 31, 2009]
2393   Converted all PNG_NO_* tests to PNG_*_SUPPORTED everywhere except pngconf.h
2394   Eliminated PNG_NO_FREE_ME and PNG_FREE_ME_SUPPORTED macros.
2395   Use png_malloc plus a loop instead of png_calloc() to initialize
2396     row_pointers in png_read_png().
2398 Version 1.4.0beta79 [September 1, 2009]
2399   Eliminated PNG_GLOBAL_ARRAYS and PNG_LOCAL_ARRAYS; always use local arrays.
2400   Eliminated PNG_CALLOC_SUPPORTED macro and always provide png_calloc().
2402 Version 1.4.0beta80 [September 17, 2009]
2403   Removed scripts/libpng.icc
2404   Changed typecast of filler from png_byte to png_uint_16 in png_set_filler().
2405     (Dennis Gustafsson)
2406   Fixed typo introduced in beta78 in pngtest.c ("#if def " should be "#ifdef ")
2408 Version 1.4.0beta81 [September 23, 2009]
2409   Eliminated unused PNG_FLAG_FREE_* defines from pngpriv.h
2410   Expanded TAB characters in pngrtran.c
2411   Removed PNG_CONST from all "PNG_CONST PNG_CHNK" declarations to avoid
2412     compiler complaints about doubly declaring things "const".
2413   Changed all "#if [!]defined(X)" to "if[n]def X" where possible.
2414   Eliminated unused png_ptr->row_buf_size
2416 Version 1.4.0beta82 [September 25, 2009]
2417   Moved redundant IHDR checking into new png_check_IHDR() in png.c
2418     and report all errors found in the IHDR data.
2419   Eliminated useless call to png_check_cHRM() from pngset.c
2421 Version 1.4.0beta83 [September 25, 2009]
2422   Revised png_check_IHDR() to eliminate bogus complaint about filter_type.
2424 Version 1.4.0beta84 [September 30, 2009]
2425   Fixed some inconsistent indentation in pngconf.h
2426   Revised png_check_IHDR() to add a test for width variable less than 32-bit.
2428 Version 1.4.0beta85 [October 1, 2009]
2429   Revised png_check_IHDR() again, to check info_ptr members instead of
2430     the contents of the returned parameters.
2432 Version 1.4.0beta86 [October 9, 2009]
2433   Updated the "xcode" project (Alam Arias).
2434   Eliminated a shadowed declaration of "pp" in png_handle_sPLT().
2436 Version 1.4.0rc01 [October 19, 2009]
2437   Trivial cosmetic changes.
2439 Version 1.4.0beta87 [October 30, 2009]
2440   Moved version 1.4.0 back into beta.
2442 Version 1.4.0beta88 [October 30, 2009]
2443   Revised libpng*.txt section about differences between 1.2.x and 1.4.0
2444     because most of the new features have now been ported back to 1.2.41
2446 Version 1.4.0beta89 [November 1, 2009]
2447   More bugfixes and improvements to CMakeLists.txt (Philip Lowman)
2448   Removed a harmless extra png_set_invert_alpha() from pngwrite.c
2449   Apply png_user_chunk_cache_max within png_decompress_chunk().
2450   Merged libpng-1.2.41.txt with libpng-1.4.0.txt where appropriate.
2452 Version 1.4.0beta90 [November 2, 2009]
2453   Removed all remaining WIN32_WCE #ifdefs except those involving the
2454     time.h "tm" structure
2456 Version 1.4.0beta91 [November 3, 2009]
2457   Updated scripts/pngw32.def and projects/wince/png32ce.def
2458   Copied projects/wince/png32ce.def to the scripts directory.
2459   Added scripts/makefile.wce
2460   Patched ltmain.sh for wince support.
2461   Added PNG_CONVERT_tIME_SUPPORTED macro.
2463 Version 1.4.0beta92 [November 4, 2009]
2464   Make inclusion of time.h in pngconf.h depend on PNG_CONVERT_tIME_SUPPORTED
2466   Revised libpng*.txt to describe differences from 1.2.40 to 1.4.0 (instead
2467     of differences from 1.2.41 to 1.4.0)
2469 Version 1.4.0beta93 [November 7, 2009]
2471     PNG_ALLOCATED macros to detect deprecated direct access to the
2472     png_struct or info_struct members and other deprecated usage in
2473     applications (John Bowler).
2474   Updated scripts/makefile* to add "-DPNG_CONFIGURE_LIBPNG" to CFLAGS,
2475     to prevent warnings about direct access to png structs by libpng
2476     functions while building libpng.  They need to be tested, especially
2477     those using compilers other than gcc.
2478   Updated projects/visualc6 and visualc71 with "/d PNG_CONFIGURE_LIBPNG".
2479     They should work but still need to be updated to remove
2480     references to pnggccrd.c or pngvcrd.c and ASM building.
2481   Added README.txt to the beos, cbuilder5, netware, and xcode projects warning
2482     that they need to be updated, to remove references to pnggccrd.c and
2483     pngvcrd.c and to depend on pngpriv.h
2484   Removed three direct references to read_info_ptr members in pngtest.c
2485     that were detected by the new PNG_DEPSTRUCT macro.
2486   Moved the png_debug macro definitions and the png_read_destroy(),
2487     png_write_destroy() and png_far_to_near() prototypes from png.h
2488     to pngpriv.h (John Bowler)
2489   Moved the synopsis lines for png_read_destroy(), png_write_destroy()
2490     png_debug(), png_debug1(), and png_debug2() from libpng.3 to libpngpf.3.
2492 Version 1.4.0beta94 [November 9, 2009]
2493   Removed the obsolete, unused pnggccrd.c and pngvcrd.c files.
2494   Updated CMakeLists.txt to add "-DPNG_CONFIGURE_LIBPNG" to the definitions.
2495   Removed dependency of pngtest.o on pngpriv.h in the makefiles.
2496   Only #define PNG_DEPSTRUCT, etc. in pngconf.h if not already defined.
2498 Version 1.4.0beta95 [November 10, 2009]
2499   Changed png_check_sig() to !png_sig_cmp() in contrib programs.
2500   Added -DPNG_CONFIGURE_LIBPNG to contrib/pngminm/*/makefile
2501   Changed png_check_sig() to !png_sig_cmp() in contrib programs.
2502   Corrected the png_get_IHDR() call in contrib/gregbook/readpng2.c
2503   Changed pngminim/*/gather.sh to stop trying to remove pnggccrd.c and pngvcrd.c
2504   Added dependency on pngpriv.h in contrib/pngminim/*/makefile
2506 Version 1.4.0beta96 [November 12, 2009]
2507   Renamed scripts/makefile.wce to scripts/makefile.cegcc
2508   Revised Makefile.am to use libpng.sys while building libpng.so
2509     so that only PNG_EXPORT functions are exported.
2510   Removed the deprecated png_check_sig() function/macro.
2511   Removed recently removed function names from scripts/*.def
2512   Revised pngtest.png to put chunks in the same order written by pngtest
2513     (evidently the same change made in libpng-1.0beta54 was lost).
2514   Added PNG_PRIVATE macro definition in pngconf.h for possible future use.
2516 Version 1.4.0beta97 [November 13, 2009]
2517   Restored pngtest.png to the libpng-1.4.0beta7 version.
2518   Removed projects/beos and netware.txt; no one seems to be supporting them.
2519   Revised Makefile.in
2521 Version 1.4.0beta98 [November 13, 2009]
2522   Added the "xcode" project to zip distributions,
2523   Fixed a typo in scripts/pngwin.def introduced in beta97.
2525 Version 1.4.0beta99 [November 14, 2009]
2526   Moved libpng-config.in and libpng.pc-configure.in out of the scripts
2527     directory, to libpng-config.in and libpng-pc.in, respectively, and
2528     modified Makefile.am and configure.ac accordingly.  Now "configure"
2529     needs nothing from the "scripts" directory.
2530   Avoid redefining PNG_CONST in pngconf.h
2532 Version 1.4.0beta100 [November 14, 2009]
2533   Removed ASM builds from projects/visualc6 and projects/visualc71
2534   Removed scripts/makefile.nommx and makefile.vcawin32
2535   Revised CMakeLists.txt to account for new location of libpng-config.in
2536     and libpng-pc.in
2537   Updated INSTALL to reflect removal and relocation of files.
2539 Version 1.4.0beta101 [November 14, 2009]
2540   Restored the binary files (*.jpg, *.png, some project files) that were
2541     accidentally deleted from the zip and 7z distributions when the xcode
2542     project was added.
2544 Version 1.4.0beta102 [November 18, 2009]
2545   Added libpng-config.in and libpng-pc.in to the zip and 7z distributions.
2546   Fixed a typo in projects/visualc6/pngtest.dsp, introduced in beta100.
2547   Moved descriptions of makefiles and other scripts out of INSTALL into
2548     scripts/README.txt
2549   Updated the copyright year in scripts/pngwin.rc from 2006 to 2009.
2551 Version 1.4.0beta103 [November 21, 2009]
2552   Removed obsolete comments about ASM from projects/visualc71/README_zlib.txt
2553   Align row_buf on 16-byte boundary in memory.
2554   Restored the PNG_WRITE_FLUSH_AFTER_IEND_SUPPORTED guard around the call
2555     to png_flush() after png_write_IEND().  See 1.4.0beta32, 1.4.0beta50
2556     changes above and 1.2.30, 1.2.30rc01 and rc03 in 1.2.41 CHANGES.  Someone
2557     needs this feature.
2558   Make the 'png_jmpbuf' macro expand to a call that records the correct
2559     longjmp function as well as returning a pointer to the setjmp
2560     jmp_buf buffer, and marked direct access to jmpbuf 'deprecated'.
2561     (John Bowler)
2563 Version 1.4.0beta104 [November 22, 2009]
2564   Removed png_longjmp_ptr from scripts/*.def and libpng.3
2565   Rebuilt configure scripts with autoconf-2.65
2567 Version 1.4.0beta105 [November 25, 2009]
2568   Use fast integer PNG_DIVIDE_BY_255() or PNG_DIVIDE_BY_65535()
2569     to accomplish alpha premultiplication when
2571   Changed "/255" to "/255.0" in background calculations to make it clear
2572     that the 255 is used as a double.
2574 Version 1.4.0beta106 [November 27, 2009]
2575   Removed premultiplied alpha feature.
2577 Version 1.4.0beta107 [December 4, 2009]
2578   Updated README
2579   Added "#define PNG_NO_PEDANTIC_WARNINGS" in the libpng source files.
2580   Removed "-DPNG_CONFIGURE_LIBPNG" from the makefiles and projects.
2581   Revised scripts/makefile.netbsd, makefile.openbsd, and makefile.sco
2582     to put png.h and pngconf.h in $prefix/include, like the other scripts,
2583     instead of in $prefix/include/libpng.  Also revised makefile.sco
2584     to put them in $prefix/include/libpng15 instead of in
2585     $prefix/include/libpng/libpng15.
2587 Version 1.4.0beta108 [December 11, 2009]
2588   Removed leftover "-DPNG_CONFIGURE_LIBPNG" from contrib/pngminim/*/makefile
2589   Relocated png_do_chop() to its original position in pngrtran.c; the
2590     change in version 1.2.41beta08 caused transparency to be handled wrong
2591     in some 16-bit datastreams (Yusaku Sugai).
2593 Version 1.4.0beta109 [December 13, 2009]
2594   Added "bit_depth" parameter to the private png_build_gamma_table() function.
2595   Pass bit_depth=8 to png_build_gamma_table() when bit_depth is 16 but the
2596     PNG_16_TO_8 transform has been set, to avoid unnecessary build of 16-bit
2597     tables.
2599 Version 1.4.0rc02 [December 20, 2009]
2600   Declared png_cleanup_needed "volatile" in pngread.c and pngwrite.c
2602 Version 1.4.0rc03 [December 22, 2009]
2603   Renamed libpng-pc.in back to libpng.pc.in and revised CMakeLists.txt
2604     (revising the change in 1.4.0beta99)
2606 Version 1.4.0rc04 [December 25, 2009]
2608     in pngset.c to be consistent with other changes in version 1.2.38.
2610 Version 1.4.0rc05 [December 25, 2009]
2611   Changed "libpng-pc.in" to "libpng.pc.in" in configure.ac, configure, and
2612     Makefile.in to be consistent with changes in libpng-1.4.0rc03
2614 Version 1.4.0rc06 [December 29, 2009]
2615   Reverted the gamma_table changes from libpng-1.4.0beta109.
2616   Fixed some indentation errors.
2618 Version 1.4.0rc07 [January 1, 2010]
2619   Revised libpng*.txt and libpng.3 about 1.2.x->1.4.x differences.
2620   Use png_calloc() instead of png_malloc(); png_memset() in pngrutil.c
2621   Update copyright year to 2010.
2623 Version 1.4.0rc08 [January 2, 2010]
2624   Avoid deprecated references to png_ptr-io_ptr and png_ptr->error_ptr
2625     in pngtest.c
2627 Version 1.4.0 [January 3, 2010]
2628   No changes.
2630 Version 1.4.1beta01 [January 8, 2010]
2631   Updated CMakeLists.txt for consistent indentation and to avoid an
2632     unclosed if-statement warning (Philip Lowman).
2633   Revised Makefile.am and Makefile.in to remove references to Y2KINFO,
2634     KNOWNBUG, and libpng.la (Robert Schwebel).
2635   Revised the makefiles to install the same files and symbolic
2636     links as configure, except for libpng.la and libpng14.la.
2637   Make png_set|get_compression_buffer_size() available even when
2638     PNG_WRITE_SUPPORTED is not enabled.
2639   Revised Makefile.am and Makefile.in to simplify their maintenance.
2640   Revised scripts/makefile.linux to install a link to libpng14.so.14.1
2642 Version 1.4.1beta02 [January 9, 2010]
2643   Revised the rest of the makefiles to install a link to libpng14.so.14.1
2645 Version 1.4.1beta03 [January 10, 2010]
2646   Removed png_set_premultiply_alpha() from scripts/*.def
2648 Version 1.4.1rc01 [January 16, 2010]
2649   No changes.
2651 Version 1.4.1beta04 [January 23, 2010]
2652   Revised png_decompress_chunk() to improve speed and memory usage when
2653     decoding large chunks.
2654   Added png_set|get_chunk_malloc_max() functions.
2656 Version 1.4.1beta05 [January 26, 2010]
2657   Relocated "int k" declaration in pngtest.c to minimize its scope.
2659 Version 1.4.1beta06 [January 28, 2010]
2660   Revised png_decompress_chunk() to use a two-pass method suggested by
2661     John Bowler.
2663 Version 1.4.1beta07 [February 6, 2010]
2664   Folded some long lines in the source files.
2666     and a PNG_USER_LIMITS_SUPPORTED flag.
2667   Eliminated use of png_ptr->irowbytes and reused the slot in png_ptr as
2668     png_ptr->png_user_chunk_malloc_max.
2669   Revised png_push_save_buffer() to do fewer but larger png_malloc() calls.
2671 Version 1.4.1beta08 [February 6, 2010]
2672   Minor cleanup and updating of dates and copyright year.
2674 Version 1.5.0beta01 [February 7, 2010]
2675   Moved declaration of png_struct into private pngstruct.h and png_info
2676     into pnginfo.h
2678 Version 1.4.1beta09 and 1.5.0beta02 [February 7, 2010]
2679   Reverted to original png_push_save_buffer() code.
2681 Version 1.4.1beta10 and 1.5.0beta03 [February 8, 2010]
2682   Return allocated "old_buffer" in png_push_save_buffer() before
2683     calling png_error(), to avoid a potential memory leak.
2684   Updated configure script to use SO number 15.
2686 Version 1.5.0beta04 [February 9, 2010]
2687   Removed malformed "incomplete struct declaration" of png_info from png.h
2689 Version 1.5.0beta05 [February 12, 2010]
2690   Removed PNG_DEPSTRUCT markup in pngstruct.h and pnginfo.h, and undid the
2691     linewrapping that it entailed.
2692   Revised comments in pngstruct.h and pnginfo.h and added pointers to
2693     the libpng license.
2695   Removed the cbuilder5 project, which has not been updated to 1.4.0.
2697 Version 1.4.1beta12 and 1.5.0beta06 [February 14, 2010]
2698   Fixed type declaration of png_get_chunk_malloc_max() in pngget.c (Daisuke
2699     Nishikawa)
2701 Version 1.5.0beta07 [omitted]
2703 Version 1.5.0beta08 [February 19, 2010]
2705     wherever png_snprintf() is used to construct error and warning messages.
2706   Noted in scripts/makefile.mingw that it expects to be run under MSYS.
2707   Removed obsolete unused MMX-querying support from contrib/gregbook
2708   Added exported png_longjmp() function.
2709   Removed the AIX redefinition of jmpbuf in png.h
2710   Added -D_ALLSOURCE in configure.ac, makefile.aix, and CMakeLists.txt
2711     when building on AIX.
2713 Version 1.5.0beta09 [February 19, 2010]
2714   Removed -D_ALLSOURCE from configure.ac, makefile.aix, and CMakeLists.txt.
2715   Changed the name of png_ptr->jmpbuf to png_ptr->png_jmpbuf in pngstruct.h
2717 Version 1.5.0beta10 [February 25, 2010]
2718   Removed unused gzio.c from contrib/pngminim gather and makefile scripts
2719   Removed replacement error handlers from contrib/gregbook.  Because of
2720     the new png_longjmp() function they are no longer needed.
2722 Version 1.5.0beta11 [March 6, 2010]
2723   Removed checking for already-included setjmp.h from pngconf.h
2724   Fixed inconsistent indentations and made numerous cosmetic changes.
2725   Revised the "SEE ALSO" style of libpng.3, libpngpf.3, and png.5
2727 Version 1.5.0beta12 [March 9, 2010]
2728   Moved "#include png.h" inside pngpriv.h and removed "#include png.h" from
2729     the source files, along with "#define PNG_EXPOSE_INTERNAL_STRUCTURES"
2730     and "#define PNG_NO_PEDANTIC_WARNINGS" (John Bowler).
2731   Created new pngdebug.h and moved debug definitions there.
2733 Version 1.5.0beta13 [March 10, 2010]
2734   Protect pngstruct.h, pnginfo.h, and pngdebug.h from being included twice.
2735   Revise the "#ifdef" blocks in png_inflate() so it will compile when neither
2737     is defined.
2738   Removed unused png_measure_compressed_chunk() from pngpriv.h and libpngpf.3
2739   Moved the 'config.h' support from pngconf.h to pngpriv.h
2740   Removed PNGAPI from the png_longjmp_ptr typedef.
2741   Eliminated dependence of pngtest.c on the private pngdebug.h file.
2742   Make all png_debug macros into *unterminated* statements or
2743     expressions (i.e. a trailing ';' must always be added) and correct
2744     the format statements in various png_debug messages.
2746 Version 1.5.0beta14 [March 14, 2010]
2747   Removed direct access to png_ptr->io_ptr from the Windows code in pngtest.c
2748   Revised Makefile.am to account for recent additions and replacements.
2749   Corrected CE and OS/2 DEF files (scripts/png*def) for symbols removed and
2750     added ordinal numbers to the Windows DEF file and corrected the duplicated
2751     ordinal numbers on CE symbols that are commented out.
2752   Added back in export symbols that can be present in the Windows build but
2753     are disabled by default.
2754   PNG_EXPORT changed to include an 'ordinal' field for DEF file generation.
2755     PNG_CALLBACK added to make callback definitions uniform.  PNGAPI split
2756     into PNGCAPI (base C form), PNGAPI (exports) and PNGCBAPI (callbacks),
2757     and appropriate changes made to all files.  Cygwin builds re-hinged to
2758     allow procedure call standard changes and to remove the need for the DEF
2759     file (fixes build on Cygwin).
2760   Enabled 'attribute' warnings that are relevant to library APIs and callbacks.
2761   Changed rules for generation of the various symbol files and added a new
2762     rule for a DEF file (which is also added to the distribution).
2763   Updated the symbol file generation to stop it adding spurious spaces
2764     to EOL (coming from preprocessor macro expansion).  Added a facility
2765     to join tokens in the output and rewrite *.dfn to use this.
2766   Eliminated scripts/*.def in favor of libpng.def; updated projects/visualc71
2767     and removed scripts/makefile.cygwin.
2768   Made PNG_BUILD_DLL safe: it can be set whenever a DLL is being built.
2769   Removed the include of sys/types.h - apparently unnecessary now on the
2770     platforms on which it happened (all but Mac OS and RISC OS).
2771   Moved the Mac OS test into pngpriv.h (the only place it is used.)
2773 Version 1.5.0beta15 [March 17, 2010]
2774   Added symbols.chk target to Makefile.am to validate the symbols in png.h
2775     against the new DEF file scripts/symbols.def.
2776   Changed the default DEF file back to pngwin.def.
2777   Removed makefile.mingw.
2778   Eliminated PNG_NO_EXTERN and PNG_ALL_EXTERN
2780 Version 1.5.0beta16 [April 1, 2010]
2781   Make png_text_struct independent of PNG_iTXt_SUPPORTED, so that
2782     fields are initialized in all configurations.  The READ/WRITE
2783     macros (PNG_(READ|WRITE)_iTXt_SUPPORTED) still function as
2784     before to disable code to actually read or write iTXt chunks
2785     and iTXt_SUPPORTED can be used to detect presence of either
2786     read or write support (but it is probably better to check for
2787     the one actually required - read or write.)
2788   Combined multiple png_warning() calls for a single error.
2789   Restored the macro definition of png_check_sig().
2791 Version 1.5.0beta17 [April 17, 2010]
2792   Added some "(long)" typecasts to printf calls in png_handle_cHRM().
2793   Documented the fact that png_set_dither() was disabled since libpng-1.4.0.
2794   Reenabled png_set_dither() but renamed it to png_set_quantize() to reflect
2795     more accurately what it actually does.  At the same time, renamed
2796     the PNG_DITHER_[RED,GREEN_BLUE]_BITS macros to
2798   Added some "(long)" typecasts to printf calls in png_handle_cHRM().
2799   Freeze build-time only configuration in the build.
2800     In all prior versions of libpng most configuration options
2801     controlled by compiler #defines had to be repeated by the
2802     application code that used libpng.  This patch changes this
2803     so that compilation options that can only be changed at build
2804     time are frozen in the build.  Options that are compiler
2805     dependent (and those that are system dependent) are evaluated
2806     each time - pngconf.h holds these.  Options that can be changed
2807     per-file in the application are in png.h.  Frozen options are
2808     in the new installed header file pnglibconf.h (John Bowler)
2809   Removed the xcode project because it has not been updated to work
2810     with libpng-1.5.0.
2811   Removed the ability to include optional pngusr.h
2813 Version 1.5.0beta18 [April 17, 2010]
2814   Restored the ability to include optional pngusr.h
2815   Moved replacements for png_error() and png_warning() from the
2816     contrib/pngminim project to pngerror.c, for use when warnings or
2817     errors are disabled via PNG_NO_WARN or PNG_NO_ERROR_TEXT, to avoid
2818     storing unneeded error/warning text.
2819   Updated contrib/pngminim project to work with the new pnglibconf.h
2820   Added some PNG_NO_* defines to contrib/pngminim/*/pngusr.h to save space.
2822 Version 1.5.0beta19 [April 24, 2010]
2823   Added PNG_{READ,WRITE}_INT_FUNCTIONS_SUPPORTED.  This allows the functions
2824     to read and write ints to be disabled independently of PNG_USE_READ_MACROS,
2825     which allows libpng to be built with the functions even though the default
2826     is to use the macros - this allows applications to choose at app build
2827     time whether or not to use macros (previously impossible because the
2828     functions weren't in the default build.)
2829   Changed Windows calling convention back to __cdecl for API functions.
2830     For Windows/x86 platforms only:
2831       __stdcall is no longer needed for Visual Basic, so libpng-1.5.0 uses
2832       __cdecl throughout (both API functions and callbacks) on Windows/x86
2833       platforms.
2834   Replaced visualc6 and visualc71 projects with new vstudio project
2835   Relaxed the overly-restrictive permissions of some files.
2837 Version 1.5.0beta20 [April 24, 2010]
2838   Relaxed more overly-restrictive permissions of some files.
2840 Version 1.5.0beta21 [April 27, 2010]
2841   Removed some unwanted binary bytes and changed CRLF to NEWLINE in the new
2842     vstudio project files, and some trivial editing of some files in the
2843     scripts directory.
2845     contrib/pngminim/decoder/pngusr.h to make a smaller decoder application.
2847 Version 1.5.0beta22 [April 28, 2010]
2848   Fixed dependencies of GET_INT_32 - it does not require READ_INT_FUNCTIONS
2849     because it has a macro equivalent.
2850   Improved the options.awk script; added an "everything off" option.
2851   Revised contrib/pngminim to use the "everything off" option in pngusr.dfa.
2853 Version 1.5.0beta23 [April 29, 2010]
2854   Corrected PNG_REMOVED macro to take five arguments.
2855     The macro was documented with two arguments (name,ordinal), however
2856     the symbol checking .dfn files assumed five arguments.  The five
2857     argument form seems more useful so it is changed to that.
2859     in gregbook/readpng2.c
2860   Corrected protection of png_get_user_transform_ptr. The API declaration in
2861     png.h is removed if both READ and WRITE USER_TRANSFORM are turned off
2862     but was left defined in pngtrans.c
2863   Added logunsupported=1 to cause pnglibconf.h to document disabled options.
2864     This makes the installed pnglibconf.h more readable but causes no
2865     other change.  The intention is that users of libpng will find it
2866     easier to understand if an API they need is missing.
2867   Include png_reset_zstream() in png.c only when PNG_READ_SUPPORTED is defined.
2868   Removed dummy_inflate.c from contrib/pngminim/encoder
2869   Removed contrib/pngminim/*/gather.sh; gathering is now done in the makefile.
2871 Version 1.5.0beta24 [May 7, 2010]
2872   Use bitwise "&" instead of arithmetic mod in pngrutil.c calculation of the
2873     offset of the png_ptr->rowbuf pointer into png_ptr->big_row_buf.
2874   Added more blank lines for readability.
2876 Version 1.5.0beta25 [June 18, 2010]
2877   In pngpread.c: png_push_have_row() add check for new_row > height
2878   Removed the now-redundant check for out-of-bounds new_row from example.c
2880 Version 1.5.0beta26 [June 18, 2010]
2881   In pngpread.c: png_push_process_row() add check for too many rows.
2883 Version 1.5.0beta27 [June 18, 2010]
2884   Removed the check added in beta25 as it is now redundant.
2886 Version 1.5.0beta28 [June 20, 2010]
2887   Rewrote png_process_IDAT_data to consistently treat extra data as warnings
2888     and handle end conditions more cleanly.
2889   Removed the new (beta26) check in png_push_process_row().
2891 Version 1.5.0beta29 [June 21, 2010]
2892   Revised scripts/options.awk to work on Sunos (but still doesn't work)
2893   Added comment to options.awk and contrib/pngminim/*/makefile to try nawk.
2895 Version 1.5.0beta30 [June 22, 2010]
2896   Stop memory leak when reading a malformed sCAL chunk.
2898 Version 1.5.0beta31 [June 26, 2010]
2899   Revised pngpread.c patch of beta28 to avoid an endless loop.
2900   Removed some trailing blanks.
2902 Version 1.5.0beta32 [June 26, 2010]
2903   Removed leftover scripts/options.patch and scripts/options.rej
2905 Version 1.5.0beta33 [July 6, 3010]
2906   Made FIXED and FLOATING options consistent in the APIs they enable and
2907     disable.  Corrected scripts/options.awk to handle both command line
2908     options and options specified in the .dfa files.
2909   Changed char *msg to PNG_CONST char *msg in pngrutil.c
2910   Make png_set_sRGB_gAMA_and_cHRM set values using either the fixed or
2911     floating point APIs, but not both.
2912   Reversed patch to remove error handler when the jmp_buf is stored in the
2913     main program structure, not the png_struct.
2914     The error handler is needed because the default handler in libpng will
2915     always use the jmp_buf in the library control structure; this is never
2916     set.  The gregbook code is a useful example because, even though it
2917     uses setjmp/longjmp, it shows how error handling can be implemented
2918     using control mechanisms not directly supported by libpng.  The
2919     technique will work correctly with mechanisms such as Microsoft
2920     Structure Exceptions or C++ exceptions (compiler willing - note that gcc
2921     does not by default support interworking of C and C++ error handling.)
2922   Reverted changes to call png_longjmp in contrib/gregbook where it is not
2923     appropriate.  If mainprog->jmpbuf is used by setjmp, then png_longjmp
2924     cannot be used.
2925   Changed "extern PNG_EXPORT" to "PNG_EXPORT" in png.h (Jan Nijtmans)
2926   Changed "extern" to "PNG_EXTERN" in pngpriv.h (except for the 'extern "C" {')
2928 Version 1.5.0beta34 [July 12, 2010]
2929   Put #ifndef PNG_EXTERN, #endif around the define PNG_EXTERN in pngpriv.h
2931 Version 1.5.0beta35 [July 24, 2010]
2932   Removed some newly-added TAB characters.
2933   Added -DNO_PNG_SNPRINTF to CFLAGS in scripts/makefile.dj2
2934   Moved the definition of png_snprintf() outside of the enclosing
2935     #ifdef blocks in pngconf.h
2937 Version 1.5.0beta36 [July 29, 2010]
2938   Patches by John Bowler:
2939   Fixed point APIs are now supported throughout (no missing APIs).
2940   Internal fixed point arithmetic support exists for all internal floating
2941     point operations.
2942   sCAL validates the floating point strings it is passed.
2943   Safe, albeit rudimentary, Watcom support is provided by PNG_API_RULE==2
2944   Two new APIs exist to get the number of passes without turning on the
2945     PNG_INTERLACE transform and to get the number of rows in the current
2946     pass.
2947   A new test program, pngvalid.c, validates the gamma code.
2948   Errors in the 16-bit gamma correction (overflows) have been corrected.
2949   cHRM chunk testing is done consistently (previously the floating point
2950     API bypassed it, because the test really didn't work on FP, now the test
2951     is performed on the actual values to be stored in the PNG file so it
2952     works in the FP case too.)
2953   Most floating point APIs now simply call the fixed point APIs after
2954     converting the values to the fixed point form used in the PNG file.
2955   The standard headers no longer include zlib.h, which is currently only
2956     required for pngstruct.h and can therefore be internal.
2957   Revised png_get_int_32 to undo the PNG two's complement representation of
2958     negative numbers.
2960 Version 1.5.0beta37 [July 30, 2010]
2961   Added a typecast in png_get_int_32() in png.h and pngrutil.h to avoid
2962     a compiler warning.
2963   Replaced oFFs 0,0 with oFFs -10,20 in pngtest.png
2965 Version 1.5.0beta38 [July 31, 2010]
2966   Implemented remaining "_fixed" functions.
2967   Corrected a number of recently introduced warnings mostly resulting from
2968     safe but uncast assignments to shorter integers.  Also added a zlib
2969     VStudio release library project because the latest zlib Official Windows
2970     build does not include such a thing.
2971   Revised png_get_int_16() to be similar to png_get_int_32().
2972   Restored projects/visualc71.
2974 Version 1.5.0beta39 [August 2, 2010]
2975   VisualC/GCC warning fixes, VisualC build fixes
2976   The changes include support for function attributes in VC in addition to
2977     those already present in GCC - necessary because without these some
2978     warnings are unavoidable.  Fixes include signed/unsigned fixes in
2979     pngvalid and checks with gcc -Wall -Wextra -Wunused.
2980   VC requires function attributes on function definitions as well as
2981     declarations, PNG_FUNCTION has been added to enable this and the
2982     relevant function definitions changed.
2984 Version 1.5.0beta40 [August 6, 2010]
2985   Correct use of _WINDOWS_ in pngconf.h
2986   Removed png_mem_ #defines; they are no longer used.
2987   Added the sRGB chunk to pngtest.png
2989 Version 1.5.0beta41 [August 11, 2010]
2990   Added the cHRM chunk to pngtest.png
2991   Don't try to use version-script with cygwin/mingw.
2992   Revised contrib/gregbook to work under cygwin/mingw.
2994 Version 1.5.0beta42 [August 18, 2010]
2995   Add .dll.a to the list of extensions to be symlinked by Makefile.am (Yaakov)
2996   Made all API functions that have const arguments and constant string
2997     literal pointers declare them (John Bowler).
2999 Version 1.5.0beta43 [August 20, 2010]
3000   Removed spurious tabs, shorten long lines (no source change)
3001     Also added scripts/chkfmt to validate the format of all the files that can
3002     reasonably be validated (it is suggested to run "make distclean" before
3003     checking, because some machine generated files have long lines.)
3004   Reformatted the CHANGES file to be more consistent throughout.
3005   Made changes to address various issues identified by GCC, mostly
3006     signed/unsigned and shortening problems on assignment but also a few
3007     difficult to optimize (for GCC) loops.
3008   Fixed non-GCC fixed point builds.  In png.c a declaration was misplaced
3009     in an earlier update.  Fixed to declare the auto variables at the head.
3010   Use cexcept.h in pngvalid.c.
3012 Version 1.5.0beta44 [August 24, 2010]
3013   Updated CMakeLists.txt to use CMAKE_INSTALL_LIBDIR variable; useful for
3014     installing libpng in /usr/lib64 (Funda Wang).
3015   Revised CMakeLists.txt to put the man pages in share/man/man* not man/man*
3016   Revised CMakeLists.txt to make symlinks instead of copies when installing.
3017   Changed PNG_LIB_NAME from pngNN to libpngNN in CMakeLists.txt (Philip Lowman)
3018   Implemented memory checks within pngvalid
3019   Reformatted/rearranged pngvalid.c to assist use of progressive reader.
3020   Check interlaced images in pngvalid
3021   Clarified pngusr.h comments in pnglibconf.dfa
3022   Simplified the pngvalid error-handling code now that cexcept.h is in place.
3023   Implemented progressive reader in pngvalid.c for standard tests
3024   Implemented progressive read in pngvalid.c gamma tests
3025   Turn on progressive reader in pngvalid.c by default and tidy code.
3027 Version 1.5.0beta45 [August 26, 2010]
3028   Added an explicit make step to projects/vstudio for pnglibconf.h
3029     Also corrected zlib.vcxproj into which Visual Studio had introduced
3030     what it calls an "authoring error".  The change to make pnglibconf.h
3031     simply copies the file; in the future it may actually generate the
3032     file from scripts/pnglibconf.dfa as the other build systems do.
3033   Changed pngvalid to work when floating point APIs are disabled
3034   Renamed the prebuilt scripts/pnglibconf.h to scripts/pnglibconf.h.prebuilt
3035   Supply default values for PNG_USER_PRIVATEBUILD and PNG_USER_DLLFNAME_POSTFIX
3036     in pngpriv.h in case the user neglected to define them in their pngusr.h
3038 Version 1.5.0beta46 [August 28, 2010]
3039   Added new private header files to libpng_sources in CMakeLists.txt
3040   Added PNG_READ_16BIT, PNG_WRITE_16BIT, and PNG_16BIT options.
3041   Added reference to scripts/pnglibconf.h.prebuilt in the visualc71 project.
3043 Version 1.5.0beta47 [September 11, 2010]
3044   Fixed a number of problems with 64-bit compilation reported by Visual
3045     Studio 2010 (John Bowler).
3047 Version 1.5.0beta48 [October 4, 2010]
3048   Updated CMakeLists.txt (Philip Lowman).
3049   Revised autogen.sh to recognize and use $AUTOCONF, $AUTOMAKE, $AUTOHEADER,
3051   Fixed problem with symbols creation in Makefile.am which was assuming that
3052     all versions of ccp write to standard output by default (Martin Banky). The
3053     bug was introduced in libpng-1.2.9beta5.
3054   Removed unused mkinstalldirs.
3056 Version 1.5.0beta49 [October 8, 2010]
3057   Undid Makefile.am revision of 1.5.0beta48.
3059 Version 1.5.0beta50 [October 14, 2010]
3060   Revised Makefile.in to account for mkinstalldirs being removed.
3061   Added some "(unsigned long)" typecasts in printf statements in pngvalid.c.
3062   Suppressed a compiler warning in png_handle_sPLT().
3063   Check for out-of-range text compression mode in png_set_text().
3065 Version 1.5.0beta51 [October 15, 2010]
3066   Changed embedded dates to "(PENDING RELEASE) in beta releases (and future
3067     rc releases) to minimize the difference between releases.
3069 Version 1.5.0beta52 [October 16, 2010]
3070   Restored some of the embedded dates (in png.h, png.c, documentation, etc.)
3072 Version 1.5.0beta53 [October 18, 2010]
3073   Updated INSTALL to mention using "make maintainer-clean" and to remove
3074     obsolete statement about a custom ltmain.sh
3075   Disabled "color-tests" by default in Makefile.am so it will work with
3076     automake versions earlier than 1.11.1
3077   Use document name "libpng-manual.txt" instead of "libpng-<version>.txt"
3078     to simplify version differences.
3079   Removed obsolete remarks about setjmp handling from INSTALL.
3080   Revised and renamed the typedef in png.h and png.c that was designed
3081     to catch library and header mismatch.
3083 Version 1.5.0beta54 [November 10, 2010]
3084   Require 48 bytes, not 64 bytes, for big_row_buf in overflow checks.
3085   Used a consistent structure for the pngget.c functions.
3087 Version 1.5.0beta55 [November 21, 2010]
3088   Revised png_get_uint_32, png_get_int_32, png_get_uint_16 (Cosmin)
3089   Moved reading of file signature into png_read_sig (Cosmin)
3090   Fixed atomicity of chunk header serialization (Cosmin)
3091   Added test for io_state in pngtest.c (Cosmin)
3092   Added "#!/bin/sh" at the top of contrib/pngminim/*/gather.sh scripts.
3093   Changes to remove gcc warnings (John Bowler)
3094     Certain optional gcc warning flags resulted in warnings in libpng code.
3095     With these changes only -Wconversion and -Wcast-qual cannot be turned on.
3096     Changes are trivial rearrangements of code.  -Wconversion is not possible
3097     for pngrutil.c (because of the widespread use of += et al on variables
3098     smaller than (int) or (unsigned int)) and -Wcast-qual is not possible
3099     with pngwio.c and pngwutil.c because the 'write' callback and zlib
3100     compression both fail to declare their input buffers with 'const'.
3102 Version 1.5.0beta56 [December 7, 2010]
3103   Added the private PNG_UNUSED() macro definition in pngpriv.h.
3104   Added some commentary about PNG_EXPORT in png.h and pngconf.h
3105   Revised PNG_EXPORT() macro and added PNG_EXPORTA() macro, with the
3106     objective of simplifying and improving the cosmetic appearance of png.h.
3107   Fixed some incorrect "=" macro names in pnglibconf.dfa
3108   Included documentation of changes in 1.5.0 from 1.4.x in libpng-manual.txt
3110 Version 1.5.0beta57 [December 9, 2010]
3111   Documented the pngvalid gamma error summary with additional comments and
3112     print statements.
3113   Improved missing symbol handling in checksym.awk; symbols missing in both
3114     the old and new files can now be optionally ignored, treated as errors
3115     or warnings.
3116   Removed references to pngvcrd.c and pnggccrd.c from the vstudio project.
3117   Updated "libpng14" to "libpng15" in the visualc71 project.
3118   Enabled the strip16 tests in pngvalid.`
3119   Don't display test results (except PASS/FAIL) when running "make test".
3120     Instead put them in pngtest-log.txt
3121   Added "--with-zprefix=<string>" to configure.ac
3122   Updated the prebuilt configuration files to autoconf version 2.68
3124 Version 1.5.0beta58 [December 19, 2010]
3125   Fixed interlace image handling and add test cases (John Bowler)
3126   Fixed the clean rule in Makefile.am to remove pngtest-log.txt
3127   Made minor changes to work around warnings in gcc 3.4
3129 Version 1.5.0rc01 [December 27, 2010]
3130   No changes.
3132 Version 1.5.0rc02 [December 27, 2010]
3133   Eliminated references to the scripts/*.def files in project/visualc71.
3135 Version 1.5.0rc03 [December 28, 2010]
3136   Eliminated scripts/*.def and revised Makefile.am accordingly
3138 Version 1.5.0rc04 [December 29, 2010]
3139   Fixed bug in background transformation handling in pngrtran.c (it was
3140     looking for the flag in png_ptr->transformations instead of in
3141     png_ptr->flags) (David Raymond).
3143 Version 1.5.0rc05 [December 31, 2010]
3144   Fixed typo in a comment in CMakeLists.txt (libpng14 => libpng15) (Cosmin)
3146 Version 1.5.0rc06 [January 4, 2011]
3147   Changed the new configure option "zprefix=string" to "zlib-prefix=string"
3149 Version 1.5.0rc07 [January 4, 2011]
3150   Updated copyright year.
3152 Version 1.5.0 [January 6, 2011]
3153   No changes.
3155 version 1.5.1beta01 [January 8, 2011]
3156   Added description of png_set_crc_action() to the manual.
3157   Added a note in the manual that the type of the iCCP profile was changed
3158     from png_charpp to png_bytepp in png_get_iCCP().  This change happened
3159     in version 1.5.0beta36 but is not noted in the CHANGES.  Similarly,
3160     it was changed from png_charpp to png_const_bytepp in png_set_iCCP().
3161   Ensure that png_rgb_to_gray ignores palette mapped images, if libpng
3162     internally happens to call it with one, and fixed a failure to handle
3163     palette mapped images correctly.  This fixes CVE-2690.
3165 Version 1.5.1beta02 [January 14, 2011]
3166   Fixed a bug in handling of interlaced images (bero at arklinux.org).
3167   Updated CMakeLists.txt (Clifford Yapp)
3169 Version 1.5.1beta03 [January 14, 2011]
3170   Fixed typecasting of some png_debug() statements (Cosmin)
3172 Version 1.5.1beta04 [January 16, 2011]
3173   Updated documentation of png_set|get_tRNS() (Thomas Klausner).
3174   Mentioned in the documentation that applications must #include "zlib.h"
3175     if they need access to anything in zlib.h, and that a number of
3176     macros such as png_memset() are no longer accessible by applications.
3177   Corrected pngvalid gamma test "sample" function to access all of the color
3178     samples of each pixel, instead of sampling the red channel three times.
3179   Prefixed variable names index, div, exp, gamma with "png_" to avoid "shadow"
3180     warnings, and (mistakenly) changed png_exp() to exp().
3182 Version 1.5.1beta05 [January 16, 2011]
3183   Changed variable names png_index, png_div, png_exp, and png_gamma to
3184     char_index, divisor, exp_b10, and gamma_val, respectively, and
3185     changed exp() back to png_exp().
3187 Version 1.5.1beta06 [January 20, 2011]
3188   Prevent png_push_crc_skip() from hanging while reading an unknown chunk
3189     or an over-large compressed zTXt chunk with the progressive reader.
3190   Eliminated more GCC "shadow" warnings.
3191   Revised png_fixed() in png.c to avoid compiler warning about reaching the
3192     end without returning anything.
3194 Version 1.5.1beta07 [January 22, 2011]
3195   In the manual, describe the png_get_IHDR() arguments in the correct order.
3196   Added const_png_structp and const_png_infop types, and used them in
3197     prototypes for most png_get_*() functions.
3199 Version 1.5.1beta08 [January 23, 2011]
3200   Added png_get_io_chunk_type() and deprecated png_get_io_chunk_name()
3201   Added synopses for the IO_STATE functions and other missing synopses
3202     to the manual. Removed the synopses from libpngpf.3 because they
3203     were out of date and no longer useful.  Better information can be
3204     obtained by reading the prototypes and comments in pngpriv.h
3205   Attempted to fix cpp on Solaris with S. Studio 12 cc, fix build
3206     Added a make macro DFNCPP that is a CPP that will accept the tokens in
3207     a .dfn file and adds configure stuff to test for such a CPP.  ./configure
3208     should fail if one is not available.
3209   Corrected const_png_ in png.h to png_const_ to avoid polluting the namespace.
3210   Added png_get_current_row_number and png_get_current_pass_number for the
3211     benefit of the user transform callback.
3212   Added png_process_data_pause and png_process_data_skip for the benefit of
3213     progressive readers that need to stop data processing or want to optimize
3214     skipping of unread data (e.g., if the reader marks a chunk to be skipped.)
3216 Version 1.5.1beta09 [January 24, 2011]
3217   Enhanced pngvalid, corrected an error in gray_to_rgb, corrected doc error.
3218     pngvalid contains tests of transforms, which tests are currently disabled
3219     because they are incompletely tested.  gray_to_rgb was failing to expand
3220     the bit depth for smaller bit depth images; this seems to be a long
3221     standing error and resulted, apparently, in invalid output
3222     (CVE-2011-0408, CERT VU#643140).  The documentation did not accurately
3223     describe what libpng really does when converting RGB to gray.
3225 Version 1.5.1beta10 [January 27, 2010]
3226   Fixed incorrect examples of callback prototypes in the manual, that were
3227     introduced in libpng-1.0.0.
3228   In addition the order of the png_get_uint macros with respect to the
3229     relevant function definitions has been reversed.  This helps the
3230     preprocessing of the symbol files be more robust.  Furthermore, the
3231     symbol file preprocessing now uses -DPNG_NO_USE_READ_MACROS even when
3232     the library may actually be built with PNG_USE_READ_MACROS; this stops
3233     the read macros interfering with the symbol file format.
3234   Made the manual, synopses, and function prototypes use the function
3235     argument names file_gamma, int_file_gamma, and srgb_intent consistently.
3237 Version 1.5.1beta11 [January 28, 2011]
3238   Changed PNG_UNUSED from "param=param;" to "{if(param){}}".
3239   Corrected local variable type in new API png_process_data_skip()
3240     The type was self-evidently incorrect but only causes problems on 64-bit
3241     architectures.
3242   Added transform tests to pngvalid and simplified the arguments.
3244 Version 1.5.1rc01 [January 29, 2011]
3245   No changes.
3247 Version 1.5.1rc02 [January 31, 2011]
3248   Added a request in the manual that applications do not use "png_" or
3249     "PNG_" to begin any of their own symbols.
3250   Changed PNG_UNUSED to "(void)param;" and updated the commentary in pngpriv.h
3252 Version 1.5.1 [February 3, 2011]
3253   No changes.
3255 Version 1.5.2beta01 [February 13, 2011]
3256   More -Wshadow fixes for older gcc compilers.  Older gcc versions apparently
3257     check formal parameters names in function declarations (as well as
3258     definitions) to see if they match a name in the global namespace.
3259   Revised PNG_EXPORTA macro to not use an empty parameter, to accommodate the
3260     old VisualC++ preprocessor.
3261   Turned on interlace handling in png_read_png().
3262   Fixed gcc pendantic warnings.
3263   Handle longjmp in Cygwin.
3264   Fixed png_get_current_row_number() in the interlaced case.
3265   Cleaned up ALPHA flags and transformations.
3266   Implemented expansion to 16 bits.
3268 Version 1.5.2beta02 [February 19, 2011]
3269   Fixed mistake in the descriptions of user read_transform and write_transform
3270     function prototypes in the manual.  The row_info struct is png_row_infop.
3271   Reverted png_get_current_row_number() to previous (1.5.2beta01) behavior.
3272   Corrected png_get_current_row_number documentation
3273   Fixed the read/write row callback documentation.
3274     This documents the current behavior, where the callback is called after
3275     every row with information pertaining to the next row.
3277 Version 1.5.2beta03 [March 3, 2011]
3278   Fixed scripts/makefile.vcwin32
3279   Updated contrib/pngsuite/README to add the word "modify".
3280   Define PNG_ALLOCATED to blank when _MSC_VER<1300.
3282 Version 1.5.2rc01 [March 19, 2011]
3283   Define remaining attributes to blank when MSC_VER<1300.
3284   ifdef out mask arrays in pngread.c when interlacing is not supported.
3286 Version 1.5.2rc02 [March 22, 2011]
3287   Added a hint to try CPP=/bin/cpp if "cpp -E" fails in scripts/pnglibconf.mak
3288     and in contrib/pngminim/*/makefile, eg., on SunOS 5.10, and removed "strip"
3289     from the makefiles.
3290   Fixed a bug (present since libpng-1.0.7) that makes png_handle_sPLT() fail
3291     to compile when PNG_NO_POINTER_INDEXING is defined (Chubanov Kirill)
3293 Version 1.5.2rc03 [March 24, 2011]
3294   Don't include standard header files in png.h while building the symbol table,
3295     to avoid cpp failure on SunOS (introduced PNG_BUILDING_SYMBOL_TABLE macro).
3297 Version 1.5.2 [March 31, 2011]
3298   No changes.
3300 Version 1.5.3beta01 [April 1, 2011]
3301   Re-initialize the zlib compressor before compressing non-IDAT chunks.
3302   Added API functions (png_set_text_compression_level() and four others) to
3303     set parameters for zlib compression of non-IDAT chunks.
3305 Version 1.5.3beta02 [April 3, 2011]
3306   Updated scripts/symbols.def with new API functions.
3307   Only compile the new zlib re-initializing code when text or iCCP is
3308     supported, using PNG_WRITE_COMPRESSED_TEXT_SUPPORTED macro.
3309   Improved the optimization of the zlib CMF byte (see libpng-1.2.6beta03).
3310   Optimize the zlib CMF byte in non-IDAT compressed chunks
3312 Version 1.5.3beta03 [April 16, 2011]
3313   Fixed gcc -ansi -pedantic compile. A strict ANSI system does not have
3314     snprintf, and the "__STRICT_ANSI__" detects that condition more reliably
3315     than __STDC__ (John Bowler).
3316   Removed the PNG_PTR_NORETURN attribute because it too dangerous. It tells
3317     the compiler that a user supplied callback (the error handler) does not
3318     return, yet there is no guarantee in practice that the application code
3319     will correctly implement the error handler because the compiler only
3320     issues a warning if there is a mistake (John Bowler).
3321   Removed the no-longer-used PNG_DEPSTRUCT macro.
3322   Updated the zlib version to 1.2.5 in the VStudio project.
3323   Fixed 64-bit builds where png_uint_32 is smaller than png_size_t in
3324     pngwutil.c (John Bowler).
3325   Fixed bug with stripping the filler or alpha channel when writing, that
3326     was introduced in libpng-1.5.2beta01 (bug report by Andrew Church).
3328 Version 1.5.3beta04 [April 27, 2011]
3329   Updated pngtest.png with the new zlib CMF optimization.
3330   Cleaned up conditional compilation code and of background/gamma handling
3331     Internal changes only except a new option to avoid compiling the
3332     png_build_grayscale_palette API (which is not used at all internally.)
3333     The main change is to move the transform tests (READ_TRANSFORMS,
3334     WRITE_TRANSFORMS) up one level to the caller of the APIs.  This avoids
3335     calls to spurious functions if all transforms are disabled and slightly
3336     simplifies those functions.  Pngvalid modified to handle this.
3337     A minor change is to stop the strip_16 and expand_16 interfaces from
3338     disabling each other; this allows the future alpha premultiplication
3339     code to use 16-bit intermediate values while still producing 8-bit output.
3340     png_do_background and png_do_gamma have been simplified to take a single
3341     pointer to the png_struct rather than pointers to every item required
3342     from the png_struct. This makes no practical difference to the internal
3343     code.
3344   A serious bug in the pngvalid internal routine 'standard_display_init' has
3345     been fixed - this failed to initialize the red channel and accidentally
3346     initialized the alpha channel twice.
3347   Changed png_struct jmp_buf member name from png_jmpbuf to tmp_jmpbuf to
3348     avoid a possible clash with the png_jmpbuf macro on some platforms.
3350 Version 1.5.3beta05 [May 6, 2011]
3351   Added the "_POSIX_SOURCE" feature test macro to ensure libpng sees the
3352     correct API. _POSIX_SOURCE is defined in pngpriv.h, pngtest.c and
3353     pngvalid.c to ensure that POSIX conformant systems disable non-POSIX APIs.
3354   Removed png_snprintf and added formatted warning messages.  This change adds
3355     internal APIs to allow png_warning messages to have parameters without
3356     requiring the host OS to implement snprintf.  As a side effect the
3357     dependency of the tIME-supporting RFC1132 code on stdio is removed and
3358     PNG_NO_WARNINGS does actually work now.
3359   Pass "" instead of '\0' to png_default_error() in png_err().  This mistake
3360     was introduced in libpng-1.2.20beta01.  This fixes CVE-2011-2691.
3361   Added PNG_WRITE_OPTIMIZE_CMF_SUPPORTED macro to make the zlib "CMF" byte
3362     optimization configureable.
3363   IDAT compression failed if preceded by a compressed text chunk (bug
3364     introduced in libpng-1.5.3beta01-02).  This was because the attempt to
3365     reset the zlib stream in png_write_IDAT happened after the first IDAT
3366     chunk had been deflated - much too late.  In this change internal
3367     functions were added to claim/release the z_stream and, hopefully, make
3368     the code more robust.  Also deflateEnd checking is added - previously
3369     libpng would ignore an error at the end of the stream.
3371 Version 1.5.3beta06 [May 8, 2011]
3372   Removed the -D_ALL_SOURCE from definitions for AIX in CMakeLists.txt
3373   Implemented premultiplied alpha support: png_set_alpha_mode API
3375 Version 1.5.3beta07 [May 11, 2011]
3376   Added expand_16 support to the high level interface.
3377   Added named value and 'flag' gamma support to png_set_gamma.  Made a minor
3378     change from the previous (unreleased) ABI/API to hide the exact value used
3379     for Macs - it's not a good idea to embed this in the ABI!
3380   Moved macro definitions for PNG_HAVE_IHDR, PNG_HAVE_PLTE, and PNG_AFTER_IDAT
3381     from pngpriv.h to png.h because they must be visible to applications
3382     that call png_set_unknown_chunks().
3383   Check for up->location !PNG_AFTER_IDAT when writing unknown chunks
3384     before IDAT.
3386 Version 1.5.3beta08 [May 16, 2011]
3387   Improved "pngvalid --speed" to exclude more of pngvalid from the time.
3388   Documented png_set_alpha_mode(), other changes in libpng.3/libpng-manual.txt
3389   The cHRM chunk now sets the defaults for png_set_rgb_to_gray() (when negative
3390     parameters are supplied by the caller), while in the absence of cHRM
3391     sRGB/Rec 709 values are still used.  This introduced a divide-by-zero
3392     bug in png_handle_cHRM().
3393   The bKGD chunk no longer overwrites the background value set by
3394     png_set_background(), allowing the latter to be used before the file
3395     header is read. It never performed any useful function to override
3396     the default anyway.
3397   Added memory overwrite and palette image checks to pngvalid.c
3398     Previously palette image code was poorly checked. Since the transformation
3399     code has a special palette path in most cases this was a severe weakness.
3400   Minor cleanup and some extra checking in pngrutil.c and pngrtran.c. When
3401     expanding an indexed image, always expand to RGBA if transparency is
3402     present.
3404 Version 1.5.3beta09 [May 17, 2011]
3405   Reversed earlier 1.5.3 change of transformation order; move png_expand_16
3406     back where it was.  The change doesn't work because it requires 16-bit
3407     gamma tables when the code only generates 8-bit ones.  This fails
3408     silently; the libpng code just doesn't do any gamma correction.  Moving
3409     the tests back leaves the old, inaccurate, 8-bit gamma calculations, but
3410     these are clearly better than none!
3412 Version 1.5.3beta10 [May 20, 2011]
3414   png_set_background() and png_expand_16() did not work together correctly.
3415     This problem is present in 1.5.2; if png_set_background is called with
3416     need_expand false and the matching 16 bit color libpng erroneously just
3417     treats it as an 8-bit color because of where png_do_expand_16 is in the
3418     transform list.  This simple fix reduces the supplied colour to 8-bits,
3419     so it gets smashed, but this is better than the current behavior.
3420   Added tests for expand16, more fixes for palette image tests to pngvalid.
3421     Corrects the code for palette image tests and disables attempts to
3422     validate palette colors.
3424 Version 1.5.3rc01 [June 3, 2011]
3425   No changes.
3427 Version 1.5.3rc02 [June 8, 2011]
3428   Fixed uninitialized memory read in png_format_buffer() (Bug report by
3429     Frank Busse, CVE-2011-2501, related to CVE-2004-0421).
3431 Version 1.5.3beta11 [June 11, 2011]
3432   Fixed png_handle_sCAL which is broken in 1.5. This fixes CVE 2011-2692.
3433   Added sCAL to pngtest.png
3434   Revised documentation about png_set_user_limits() to say that it also affects
3435     png writing.
3436   Revised handling of png_set_user_limits() so that it can increase the
3437     limit beyond the PNG_USER_WIDTH|HEIGHT_MAX; previously it could only
3438     reduce it.
3439   Make the 16-to-8 scaling accurate. Dividing by 256 with no rounding is
3440     wrong (high by one) 25% of the time. Dividing by 257 with rounding is
3441     wrong in 128 out of 65536 cases. Getting the right answer all the time
3442     without division is easy.
3444   Added projects/owatcom, an IDE project for OpenWatcom to replace
3445     scripts/makefile.watcom.  This project works with OpenWatcom 1.9. The
3446     IDE autogenerates appropriate makefiles (libpng.mk) for batch processing.
3447     The project is configurable, unlike the Visual Studio project, so long
3448     as the developer has an awk.
3449   Changed png_set_gAMA to limit the gamma value range so that the inverse
3450     of the stored value cannot overflow the fixed point representation,
3451     and changed other things OpenWatcom warns about.
3452   Revised pngvalid.c to test PNG_ALPHA_MODE_SUPPORTED correctly. This allows
3453     pngvalid to build when ALPHA_MODE is not supported, which is required if
3454     it is to build on libpng 1.4.
3455   Removed string/memory macros that are no longer used and are not
3456     necessarily fully supportable, particularly png_strncpy and png_snprintf.
3457   Added log option to pngvalid.c and attempted to improve gamma messages.
3459 Version 1.5.3 [omitted]
3460   People found the presence of a beta release following an rc release
3461     to be confusing; therefore we bump the version to libpng-1.5.4beta01
3462     and there will be no libpng-1.5.3 release.
3464 Version 1.5.4beta01 [June 14, 2011]
3465   Made it possible to undefine PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
3466     to get the same (inaccurate) output as libpng-1.5.2 and earlier.
3467   Moved definitions of PNG_HAVE_IHDR, PNG_AFTER_IDAT, and PNG_HAVE_PLTE
3468     outside of an unknown-chunk block in png.h because they are also
3469     needed for other uses.
3471 Version 1.5.4beta02 [June 14, 2011]
3472   Fixed and clarified LEGACY 16-to-8 scaling code.
3473   Added png_set_chop_16() API, to match inaccurate results from previous
3474     libpng versions.
3475   Removed the ACCURATE and LEGACY options (they are no longer useable)
3476   Use the old scaling method for background if png_set_chop_16() was
3477     called.
3478   Made png_set_chop_16() API removeable by disabling PNG_CHOP_16_TO_8_SUPPORTED
3480 Version 1.5.4beta03 [June 15, 2011]
3481   Fixed a problem in png_do_expand_palette() exposed by optimization in
3482     1.5.3beta06
3483   Also removed a spurious and confusing "trans" member ("trans") from png_info.
3484   The palette expand optimization prevented expansion to an intermediate RGBA
3485     form if tRNS was present but alpha was marked to be stripped; this exposed
3486     a check for tRNS in png_do_expand_palette() which is inconsistent with the
3487     code elsewhere in libpng.
3488   Correction to the expand_16 code; removed extra instance of
3489     png_set_scale_16_to_8 from pngpriv.h
3491 Version 1.5.4beta04 [June 16, 2011]
3492   Added a missing "#ifdef PNG_READ_BACKGROUND_SUPPORTED/#endif" in pngrtran.c
3493   Added PNG_TRANSFORM_CHOP_16 to the high-level read transforms.
3494   Made PNG_READ_16_TO_8_ACCURATE_SCALE configurable again.  If this is
3495     not enabled, png_set_strip_16() and png_do_scale_16_to_8() aren't built.
3496   Revised contrib/visupng, gregbook, and pngminim to demonstrate chop_16_to_8
3498 Version 1.5.4beta05 [June 16, 2011]
3499   Renamed png_set_strip_16() to png_set_scale_16() and renamed
3500     png_set_chop_16() to png_set_strip(16) in an attempt to minimize the
3501     behavior changes between libpng14 and libpng15.
3503 Version 1.5.4beta06 [June 18, 2011]
3504   Fixed new bug that was causing both strip_16 and scale_16 to be applied.
3506 Version 1.5.4beta07 [June 19, 2011]
3507   Fixed pngvalid, simplified macros, added checking for 0 in sCAL.
3508     The ACCURATE scale macro is no longer defined in 1.5 - call the
3509     png_scale_16_to_8 API.  Made sure that PNG_READ_16_TO_8 is still defined
3510     if the png_strip_16_to_8 API is present.  png_check_fp_number now
3511     maintains some state so that positive, negative and zero values are
3512     identified.  sCAL uses these to be strictly spec conformant.
3514 Version 1.5.4beta08 [June 23, 2011]
3515   Fixed pngvalid if ACCURATE_SCALE is defined.
3516   Updated scripts/pnglibconf.h.prebuilt.
3518 Version 1.5.4rc01 [June 30, 2011]
3519   Define PNG_ALLOCATED to "restrict" only if MSC_VER >= 1400.
3521 Version 1.5.4 [July 7, 2011]
3522   No changes.
3524 Version 1.5.5beta01 [July 13, 2011]
3525   Fixed some typos and made other minor changes in the manual.
3526   Updated contrib/pngminus/makefile.std (Samuli Souminen)
3528 Version 1.5.5beta02 [July 14, 2011]
3529   Revised Makefile.am and Makefile.in to look in the right directory for
3530     pnglibconf.h.prebuilt
3532 Version 1.5.5beta03 [July 27, 2011]
3533   Enabled compilation with g++ compiler.  This compiler does not recognize
3534     the file extension, so it always compiles with C++ rules.  Made minor
3535     changes to pngrutil.c to cast results where C++ expects it but C does not.
3536   Minor editing of libpng.3 and libpng-manual.txt.
3538 Version 1.5.5beta04 [July 29, 2011]
3539   Revised CMakeLists.txt (Clifford Yapp)
3540   Updated commentary about the png_rgb_to_gray() default coefficients
3541     in the manual and in pngrtran.c
3543 Version 1.5.5beta05 [August 17, 2011]
3544   Prevent unexpected API exports from non-libpng DLLs on Windows.  The "_DLL"
3545     is removed from the test of whether a DLL is being built (this erroneously
3546     caused the libpng APIs to be marked as DLL exports in static builds under
3547     Microsoft Visual Studio).  Almost all of the libpng building configuration
3548     is moved from pngconf.h to pngpriv.h, but PNG_DLL_EXPORT remains in
3549     pngconf.h, though, so that it is colocated with the import definition (it
3550     is no longer used anywhere in the installed headers).  The VStudio project
3551     definitions have been cleaned up: "_USRDLL" has been removed from the
3552     static library builds (this was incorrect), and PNG_USE_DLL has been added
3553     to pngvalid to test the functionality (pngtest does not supply it,
3554     deliberately).  The spurious "_EXPORTS" has been removed from the
3555     libpng build (all these errors were a result of copy/paste between project
3556     configurations.)
3557   Added new types and internal functions for CIE RGB end point handling to
3558     pngpriv.h (functions yet to be implemented).
3560 Version 1.5.5beta06 [August 26, 2011]
3561   Ensure the CMAKE_LIBRARY_OUTPUT_DIRECTORY is set in CMakeLists.txt
3562     (Clifford Yap)
3563   Fixes to rgb_to_gray and cHRM XYZ APIs (John Bowler):
3564     The rgb_to_gray code had errors when combined with gamma correction.
3565     Some pixels were treated as true grey when they weren't and such pixels
3566     and true grey ones were not gamma corrected (the original value of the
3567     red component was used instead).  APIs to get and set cHRM using color
3568     space end points have been added and the rgb_to_gray code that defaults
3569     based on cHRM, and the divide-by-zero bug in png_handle_cHRM (CERT
3570     VU#477046, CVE-2011-3328, introduced in 1.5.4) have been corrected.
3571   A considerable number of tests has been added to pngvalid for the
3572     rgb_to_gray transform.
3573   Arithmetic errors in rgb_to_gray whereby the calculated gray value was
3574     truncated to the bit depth rather than rounded have been fixed except in
3575     the 8-bit non-gamma-corrected case (where consistency seems more important
3576     than correctness.)  The code still has considerable inaccuracies in the
3577     8-bit case because 8-bit linear arithmetic is used.
3579 Version 1.5.5beta07 [September 7, 2011]
3580   Added "$(ARCH)" option to makefile.darwin
3581   Added SunOS support to configure.ac and Makefile.am
3582   Changed png_chunk_benign_error() to png_warning() in png.c, in
3583     png_XYZ_from_xy_checked().
3585 Version 1.5.5beta08 [September 10, 2011]
3586   Fixed 64-bit compilation errors (gcc). The errors fixed relate
3587     to conditions where types that are 32 bits in the GCC 32-bit
3588     world (uLong and png_size_t) become 64 bits in the 64-bit
3589     world.  This produces potential truncation errors which the
3590     compiler correctly flags.
3591   Relocated new HAVE_SOLARIS_LD definition in configure.ac
3592   Constant changes for 64-bit compatibility (removal of L suffixes). The
3593     16-bit cases still use "L" as we don't have a 16-bit test system.
3595 Version 1.5.5rc01 [September 15, 2011]
3596   Removed "L" suffixes in pngpriv.h
3598 Version 1.5.5 [September 22, 2011]
3599   No changes.
3601 Version 1.5.6beta01 [September 22, 2011]
3602   Fixed some 64-bit type conversion warnings in pngrtran.c
3603   Moved row_info from png_struct to a local variable.
3604   The various interlace mask arrays have been made into arrays of
3605     bytes and made PNG_CONST and static (previously some arrays were
3606     marked PNG_CONST and some weren't).
3607   Additional checks have been added to the transform code to validate the
3608     pixel depths after the transforms on both read and write.
3609   Removed some redundant code from pngwrite.c, in png_destroy_write_struct().
3610   Changed chunk reading/writing code to use png_uint_32 instead of png_byte[4].
3611     This removes the need to allocate temporary strings for chunk names on
3612     the stack in the read/write code.  Unknown chunk handling still uses the
3613     string form because this is exposed in the API.
3615 Version 1.5.6beta02 [September 26, 2011]
3616   Added a note in the manual the png_read_update_info() must be called only
3617     once with a particular info_ptr.
3618   Fixed a typo in the definition of the new PNG_STRING_FROM_CHUNK(s,c) macro.
3620 Version 1.5.6beta03 [September 28, 2011]
3621   Revised test-pngtest.sh to report FAIL when pngtest fails.
3622   Added "--strict" option to pngtest, to report FAIL when the failure is
3623     only because the resulting valid files are different.
3624   Revised CMakeLists.txt to work with mingw and removed some material from
3625     CMakeLists.txt that is no longer useful in libpng-1.5.
3627 Version 1.5.6beta04 [October 5, 2011]
3628   Fixed typo in Makefile.in and Makefile.am ("-M Wl" should be "-M -Wl")."
3630 Version 1.5.6beta05 [October 12, 2011]
3631   Speed up png_combine_row() for interlaced images. This reduces the generality
3632     of the code, allowing it to be optimized for Adam7 interlace.  The masks
3633     passed to png_combine_row() are now generated internally, avoiding
3634     some code duplication and localizing the interlace handling somewhat.
3635   Align png_struct::row_buf - previously it was always unaligned, caused by
3636     a bug in the code that attempted to align it; the code needs to subtract
3637     one from the pointer to take account of the filter byte prepended to
3638     each row.
3639   Optimized png_combine_row() when rows are aligned. This gains a small
3640     percentage for 16-bit and 32-bit pixels in the typical case where the
3641     output row buffers are appropriately aligned. The optimization was not
3642     previously possible because the png_struct buffer was always misaligned.
3643   Fixed bug in png_write_chunk_header() debug print, introduced in 1.5.6beta01.
3645 Version 1.5.6beta06 [October 17, 2011]
3646   Removed two redundant tests for unitialized row.
3647   Fixed a relatively harmless memory overwrite in compressed text writing
3648     with a 1 byte zlib buffer.
3649   Add ability to call png_read_update_info multiple times to pngvalid.c.
3650   Fixes for multiple calls to png_read_update_info. These fixes attend to
3651     most of the errors revealed in pngvalid, however doing the gamma work
3652     twice results in inaccuracies that can't be easily fixed.  There is now
3653     a warning in the code if this is going to happen.
3654   Turned on multiple png_read_update_info in pngvalid transform tests.
3655   Prevent libpng from overwriting unused bits at the end of the image when
3656     it is not byte aligned, while reading. Prior to libpng-1.5.6 libpng would
3657     overwrite the partial byte at the end of each row if the row width was not
3658     an exact multiple of 8 bits and the image is not interlaced.
3660 Version 1.5.6beta07 [October 21, 2011]
3661   Made png_ptr->prev_row an aligned pointer into png_ptr->big_prev_row
3662     (Mans Rullgard).
3664 Version 1.5.6rc01 [October 26, 2011]
3665   Changed misleading "Missing PLTE before cHRM" warning to "Out of place cHRM"
3667 Version 1.5.6rc02 [October 27, 2011]
3668   Added LSR() macro to defend against buggy compilers that evaluate non-taken
3669     code branches and complain about out-of-range shifts.
3671 Version 1.5.6rc03 [October 28, 2011]
3672   Renamed the LSR() macro to PNG_LSR() and added PNG_LSL() macro.
3673   Fixed compiler warnings with Intel and MSYS compilers. The logical shift
3674     fix for Microsoft Visual C is required by other compilers, so this
3675     enables that fix for all compilers when using compile-time constants.
3676     Under MSYS 'byte' is a name declared in a system header file, so we
3677     changed the name of a local variable to avoid the warnings that result.
3678   Added #define PNG_ALIGN_TYPE PNG_ALIGN_NONE to contrib/pngminim/*/pngusr.h
3680 Version 1.5.6 [November 3, 2011]
3681   No changes.
3683 Version 1.5.7beta01 [November 4, 2011]
3684   Added support for ARM processor, when decoding all PNG up-filtered rows
3685     and any other-filtered rows with 3 or 4 bytes per pixel (Mans Rullgard).
3686   Fixed bug in pngvalid on early allocation failure; fixed type cast in
3687     pngmem.c; pngvalid would attempt to call png_error() if the allocation
3688     of a png_struct or png_info failed. This would probably have led to a
3689     crash.  The pngmem.c implementation of png_malloc() included a cast
3690     to png_size_t which would fail on large allocations on 16-bit systems.
3691   Fix for the preprocessor of the Intel C compiler. The preprocessor
3692     splits adjacent @ signs with a space; this changes the concatentation
3693     token from @-@-@ to PNG_JOIN; that should work with all compiler
3694     preprocessors.
3695   Paeth filter speed improvements from work by Siarhei Siamashka. This
3696     changes the 'Paeth' reconstruction function to improve the GCC code
3697     generation on x86. The changes are only part of the suggested ones;
3698     just the changes that definitely improve speed and remain simple.
3699     The changes also slightly increase the clarity of the code.
3701 Version 1.5.7beta02 [November 11, 2011]
3702   Check compression_type parameter in png_get_iCCP and remove spurious
3703     casts. The compression_type parameter is always assigned to, so must
3704     be non-NULL. The cast of the profile length potentially truncated the
3705     value unnecessarily on a 16-bit int system, so the cast of the (byte)
3706     compression type to (int) is specified by ANSI-C anyway.
3707   Fixed FP division by zero in pngvalid.c; the 'test_pixel' code left
3708     the sBIT fields in the test pixel as 0, which resulted in a floating
3709     point division by zero which was irrelevant but causes systems where
3710     FP exceptions cause a crash. Added code to pngvalid to turn on FP
3711     exceptions if the appropriate glibc support is there to ensure this is
3712     tested in the future.
3713   Updated scripts/pnglibconf.mak and scripts/makefile.std to handle the
3714     new PNG_JOIN macro.
3715   Added versioning to pnglibconf.h comments.
3716   Simplified read/write API initial version; basic read/write tested on
3717     a variety of images, limited documentation (in the header file.)
3718   Installed more accurate linear to sRGB conversion tables. The slightly
3719     modified tables reduce the number of 16-bit values that
3720     convert to an off-by-one 8-bit value.  The "makesRGB.c" code that was used
3721     to generate the tables is now in a contrib/sRGBtables sub-directory.
3723 Version 1.5.7beta03 [November 17, 2011]
3724   Removed PNG_CONST from the sRGB table declarations in pngpriv.h and png.c
3725   Added run-time detection of NEON support.
3726   Added contrib/libtests; includes simplified API test and timing test and
3727     a color conversion utility for rapid checking of failed 'pngstest' results.
3728   Multiple transform bug fixes plus a work-round for double gamma correction.
3729     libpng does not support more than one transform that requires linear data
3730     at once - if this is tried typically the results is double gamma
3731     correction. Since the simplified APIs can need rgb to gray combined with
3732     a compose operation it is necessary to do one of these outside the main
3733     libpng transform code. This check-in also contains fixes to various bugs
3734     in the simplified APIs themselves and to some bugs in compose and rgb to
3735     gray (on palette) itself.
3736   Fixes for C++ compilation using g++ When libpng source is compiled
3737     using g++. The compiler imposes C++ rules on the C source; thus it
3738     is desireable to make the source work with either C or C++ rules
3739     without throwing away useful error information.  This change adds
3740     png_voidcast to allow C semantic (void*) cases or the corresponding
3741     C++ static_cast operation, as appropriate.
3742   Added --noexecstack to assembler file compilation. GCC does not set
3743     this on assembler compilation, even though it does on C compilation.
3744     This creates security issues if assembler code is enabled; the
3745     work-around is to set it by default in the flags for $(CCAS)
3746   Work around compilers that don't support declaration of const data. Some
3747     compilers fault 'extern const' data declarations (because the data is
3748     not initialized); this turns on const-ness only for compilers where
3749     this is known to work.
3751 Version 1.5.7beta04 [November 17, 2011]
3752   Since the gcc driver does not recognize the --noexecstack flag, we must
3753     use the -Wa prefix to have it passed through to the assembler.
3754     Also removed a duplicate setting of this flag.
3755   Added files that were omitted from the libpng-1.5.7beta03 zip distribution.
3757 Version 1.5.7beta05 [November 25, 2011]
3758   Removed "zTXt" from warning in generic chunk decompression function.
3759   Validate time settings passed to png_set_tIME() and png_convert_to_rfc1123()
3760     (Frank Busse). Note: This prevented CVE-2015-7981 from affecting
3761     libpng-1.5.7 and later.
3762   Added MINGW support to CMakeLists.txt
3763   Reject invalid compression flag or method when reading the iTXt chunk.
3764   Backed out 'simplified' API changes. The API seems too complex and there
3765     is a lack of consensus or enthusiasm for the proposals.  The API also
3766     reveals significant bugs inside libpng (double gamma correction and the
3767     known bug of being unable to retrieve a corrected palette). It seems
3768     better to wait until the bugs, at least, are corrected.
3769   Moved pngvalid.c into contrib/libtests
3770   Rebuilt Makefile.in, configure, etc., with autoconf-2.68
3772 Version 1.5.7rc01 [December 1, 2011]
3773   Replaced an "#if" with "#ifdef" in pngrtran.c
3774   Revised #if PNG_DO_BC block in png.c (use #ifdef and add #else)
3776 Version 1.5.7rc02 [December 5, 2011]
3777   Revised project files and contrib/pngvalid/pngvalid.c to account for
3778     the relocation of pngvalid into contrib/libtests.
3779   Revised pngconf.h to use " __declspec(restrict)" only when MSC_VER >= 1400,
3780     as in libpng-1.5.4.
3781   Put CRLF line endings in the owatcom project files.
3783 Version 1.5.7rc03 [December 7, 2011]
3784   Updated CMakeLists.txt to account for the relocation of pngvalid.c
3786 Version 1.5.7 [December 15, 2011]
3787   Minor fixes to pngvalid.c for gcc 4.6.2 compatibility to remove warnings
3788     reported by earlier versions.
3789   Fixed minor memset/sizeof errors in pngvalid.c.
3791 Version 1.6.0beta01 [December 15, 2011]
3792   Removed machine-generated configure files from the GIT repository (they will
3793     continue to appear in the tarball distributions and in the libpng15 and
3794     earlier GIT branches).
3795   Restored the new 'simplified' API, which was started in libpng-1.5.7beta02
3796     but later deleted from libpng-1.5.7beta05.
3797   Added example programs for the new 'simplified' API.
3798   Added ANSI-C (C90) headers and require them, and take advantage of the
3799     change. Also fixed some of the projects/* and contrib/* files that needed
3800     updates for libpng16 and the move of pngvalid.c.
3801     With this change the required ANSI-C header files are assumed to exist: the
3802     implementation must provide float.h, limits.h, stdarg.h and stddef.h and
3803     libpng relies on limits.h and stddef.h existing and behaving as defined
3804     (the other two required headers aren't used).  Non-ANSI systems that don't
3805     have stddef.h or limits.h will have to provide an appropriate fake
3806     containing the relevant types and #defines.
3807   Dropped support for 16-bit platforms. The use of FAR/far has been eliminated
3808     and the definition of png_alloc_size_t is now controlled by a flag so
3809     that 'small size_t' systems can select it if necessary.  Libpng 1.6 may
3810     not currently work on such systems -- it seems likely that it will
3811     ask 'malloc' for more than 65535 bytes with any image that has a
3812     sufficiently large row size (rather than simply failing to read such
3813     images).
3814   New tools directory containing tools used to generate libpng code.
3815   Fixed race conditions in parallel make builds. With higher degrees of
3816     parallelism during 'make' the use of the same temporary file names such
3817     as 'dfn*' can result in a race where a temporary file from one arm of the
3818     build is deleted or overwritten in another arm.  This changes the
3819     temporary files for suffix rules to always use $* and ensures that the
3820     non-suffix rules use unique file names.
3822 Version 1.6.0beta02 [December 21, 2011]
3823   Correct configure builds where build and source directories are separate.
3824     The include path of 'config.h' was erroneously made relative in pngvalid.c
3825     in libpng 1.5.7.
3827 Version 1.6.0beta03 [December 22, 2011]
3828   Start-up code size improvements, error handler flexibility. These changes
3829     alter how the tricky allocation of the initial png_struct and png_info
3830     structures are handled. png_info is now handled in pretty much the same
3831     way as everything else, except that the allocations handle NULL return
3832     silently.  png_struct is changed in a similar way on allocation and on
3833     deallocation a 'safety' error handler is put in place (which should never
3834     be required).  The error handler itself is changed to permit mismatches
3835     in the application and libpng error buffer size; however, this means a
3836     silent change to the API to return the jmp_buf if the size doesn't match
3837     the size from the libpng compilation; libpng now allocates the memory and
3838     this may fail.  Overall these changes result in slight code size
3839     reductions; however, this is a reduction in code that is always executed
3840     so is particularly valuable.  Overall on a 64-bit system the libpng DLL
3841     decreases in code size by 1733 bytes.  pngerror.o increases in size by
3842     about 465 bytes because of the new functionality.
3843   Added png_convert_to_rfc1123_buffer() and deprecated png_convert_to_rfc1123()
3844     to avoid including a spurious buffer in the png_struct.
3846 Version 1.6.0beta04 [December 30, 2011]
3847   Regenerated configure scripts with automake-1.11.2
3848   Eliminated png_info_destroy(). It is now used only in png.c and only calls
3849     one other internal function and memset().
3850   Enabled png_get_sCAL_fixed() if floating point APIs are enabled. Previously
3851     it was disabled whenever internal fixed point arithmetic was selected,
3852     which meant it didn't exist even on systems where FP was available but not
3853     preferred.
3854   Added pngvalid.c compile time checks for const APIs.
3855   Implemented 'restrict' for png_info and png_struct. Because of the way
3856     libpng works both png_info and png_struct are always accessed via a
3857     single pointer.  This means adding C99 'restrict' to the pointer gives
3858     the compiler some opportunity to optimize the code.  This change allows
3859     that.
3860   Moved AC_MSG_CHECKING([if libraries can be versioned]) later to the proper
3861     location in configure.ac (Gilles Espinasse).
3862   Changed png_memcpy to C assignment where appropriate. Changed all those
3863     uses of png_memcpy that were doing a simple assignment to assignments
3864     (all those cases where the thing being copied is a non-array C L-value).
3865   Added some error checking to png_set_*() routines.
3866   Removed the reference to the non-exported function png_memcpy() from
3867     example.c.
3868   Fixed the Visual C 64-bit build - it requires jmp_buf to be aligned, but
3869     it had become misaligned.
3870   Revised contrib/pngminus/pnm2png.c to avoid warnings when png_uint_32
3871     and unsigned long are of different sizes.
3873 Version 1.6.0beta05 [January 15, 2012]
3874   Updated manual with description of the simplified API (copied from png.h)
3875   Fix bug in pngerror.c: some long warnings were being improperly truncated
3876     (CVE-2011-3464, bug introduced in libpng-1.5.3beta05).
3878 Version 1.6.0beta06 [January 24, 2012]
3879   Added palette support to the simplified APIs. This commit
3880     changes some of the macro definitions in png.h, app code
3881     may need corresponding changes.
3882   Increased the formatted warning buffer to 192 bytes.
3883   Added color-map support to simplified API. This is an initial version for
3884     review; the documentation has not yet been updated.
3885   Fixed Min/GW uninstall to remove libpng.dll.a
3887 Version 1.6.0beta07 [January 28, 2012]
3888   Eliminated Intel icc/icl compiler warnings. The Intel (GCC derived)
3889     compiler issues slightly different warnings from those issued by the
3890     current vesions of GCC. This eliminates those warnings by
3891     adding/removing casts and small code rewrites.
3892   Updated configure.ac from autoupdate: added --enable-werror option.
3893     Also some layout regularization and removal of introduced tab characters
3894     (replaced with 3-character indentation).  Obsolete macros identified by
3895     autoupdate have been removed; the replacements are all in 2.59 so
3896     the pre-req hasn't been changed.  --enable-werror checks for support
3897     for -Werror (or the given argument) in the compiler.  This mimics the
3898     gcc configure option by allowing -Werror to be turned on safely; without
3899     the option the tests written in configure itself fail compilation because
3900     they cause compiler warnings.
3901   Rewrote autogen.sh to run autoreconf instead of running tools one-by-one.
3902   Conditionalize the install rules for MINGW and CYGWIN in CMakeLists.txt and
3903     set CMAKE_LIBRARY_OUTPUT_DIRECTORY to "lib" on all platforms (C. Yapp).
3904   Freeze libtool files in the 'scripts' directory. This version of autogen.sh
3905     attempts to dissuade people from running it when it is not, or should not,
3906     be necessary.  In fact, autogen.sh does not work when run in a libpng
3907     directory extracted from a tar distribution anymore. You must run it in
3908     a GIT clone instead.
3909   Added two images to contrib/pngsuite (1-bit and 2-bit transparent grayscale),
3910     and renamed three whose names were inconsistent with those in
3911     pngsuite/README.txt.
3913 Version 1.6.0beta08 [February 1, 2012]
3914   Fixed Image::colormap misalignment in pngstest.c
3915   Check libtool/libtoolize version number (2.4.2) in configure.ac
3916   Divide test-pngstest.sh into separate pngstest runs for basic and
3917     transparent images.
3918   Moved automake options to AM_INIT_AUTOMAKE in configure.ac
3919   Added color-tests, silent-rules (Not yet implemented in Makefile.am) and
3920     version checking to configure.ac
3921   Improved pngstest speed by not doing redundant tests and add const to
3922     the background parameter of png_image_finish_read. The --background
3923     option is now done automagically only when required, so that commandline
3924     option no longer exists.
3925   Cleaned up pngpriv.h to consistently declare all functions and data.
3926     Also eliminated PNG_CONST_DATA, which is apparently not needed but we
3927     can't be sure until it is gone.
3928   Added symbol prefixing that allows all the libpng external symbols
3929     to be prefixed (suggested by Reuben Hawkins).
3930   Updated "ftbb*.png" list in the owatcom and vstudio projects.
3931   Fixed 'prefix' builds on clean systems. The generation of pngprefix.h
3932     should not require itself.
3933   Updated INSTALL to explain that autogen.sh must be run in a GIT clone,
3934     not in a libpng directory extracted from a tar distribution.
3936 Version 1.6.0beta09 [February 1, 2012]
3937   Reverted the prebuilt configure files to libpng-1.6.0beta05 condition.
3939 Version 1.6.0beta10 [February 3, 2012]
3940   Added Z_SOLO for zlib-1.2.6+ and correct pngstest tests
3941   Updated list of test images in CMakeLists.txt
3942   Updated the prebuilt configure files to current condition.
3943   Revised INSTALL information about autogen.sh; it works in tar distributions.
3945 Version 1.6.0beta11 [February 16, 2012]
3946   Fix character count in pngstest command in projects/owatcom/pngstest.tgt
3947   Revised test-pngstest.sh to report PASS/FAIL for each image.
3948   Updated documentation about the simplified API.
3949   Corrected estimate of error in libpng png_set_rgb_to_gray API.  The API is
3950     extremely inaccurate for sRGB conversions because it uses an 8-bit
3951     intermediate linear value and it does not use the sRGB transform, so it
3952     suffers from the known instability in gamma transforms for values close
3953     to 0 (see Poynton).  The net result is that the calculation has a maximum
3954     error of 14.99/255; 0.5/255^(1/2.2).  pngstest now uses 15 for the
3955     permitted 8-bit error. This may still not be enough because of arithmetic
3956     error.
3957   Removed some unused arrays (with #ifdef) from png_read_push_finish_row().
3958   Fixed a memory overwrite bug in simplified read of RGB PNG with
3959     non-linear gamma Also bugs in the error checking in pngread.c and changed
3960     quite a lot of the checks in pngstest.c to be correct; either correctly
3961     written or not over-optimistic.  The pngstest changes are insufficient to
3962     allow all possible RGB transforms to be passed; pngstest cmppixel needs
3963     to be rewritten to make it clearer which errors it allows and then changed
3964     to permit known inaccuracies.
3965   Removed tests for no-longer-used *_EMPTY_PLTE_SUPPORTED from pngstruct.h
3966   Fixed fixed/float API export conditionals. 1) If FIXED_POINT or
3967     FLOATING_POINT options were switched off, png.h ended up with lone ';'
3968     characters.  This is not valid ANSI-C outside a function.  The ';'
3969     characters have been moved inside the definition of PNG_FP_EXPORT and
3970     PNG_FIXED_EXPORT. 2) If either option was switched off, the declaration
3971     of the corresponding functions were completely omitted, even though some
3972     of them are still used internally.  The result is still valid, but
3973     produces warnings from gcc with some warning options (including -Wall). The
3974     fix is to cause png.h to declare the functions with PNG_INTERNAL_FUNCTION
3975     when png.h is included from pngpriv.h.
3976   Check for invalid palette index while reading paletted PNG.  When one is
3977     found, issue a warning and increase png_ptr->num_palette accordingly.
3978     Apps are responsible for checking to see if that happened.
3980 Version 1.6.0beta12 [February 18, 2012]
3981   Do not increase num_palette on invalid_index.
3982   Relocated check for invalid palette index to pngrtran.c, after unpacking
3983     the sub-8-bit pixels.
3984   Fixed CVE-2011-3026 buffer overrun bug.  This bug was introduced when
3985     iCCP chunk support was added at libpng-1.0.6. Deal more correctly with the
3986     test on iCCP chunk length. Also removed spurious casts that may hide
3987     problems on 16-bit systems.
3989 Version 1.6.0beta13 [February 24, 2012]
3990   Eliminated redundant png_push_read_tEXt|zTXt|iTXt|unknown code from
3991     pngpread.c and use the sequential png_handle_tEXt, etc., in pngrutil.c;
3992     now that png_ptr->buffer is inaccessible to applications, the special
3993     handling is no longer useful.
3994   Added PNG_SAFE_LIMITS feature to pnglibconf.dfa, pngpriv.h, and new
3995     pngusr.dfa to reset the user limits to safe ones if PNG_SAFE_LIMITS is
3996     defined.  To enable, use "CPPFLAGS=-DPNG_SAFE_LIMITS_SUPPORTED=1" on the
3997     configure command or put #define PNG_SAFE_LIMITS_SUPPORTED in
3998     pnglibconf.h.prebuilt and pnglibconf.h.
4000 Version 1.6.0beta14 [February 27, 2012]
4001   Added information about the new limits in the manual.
4002   Updated Makefile.in
4004 Version 1.6.0beta15 [March 2, 2012]
4005   Removed unused "current_text" members of png_struct and the png_free()
4006     of png_ptr->current_text from pngread.c
4007   Rewrote pngstest.c for substantial speed improvement.
4008   Fixed transparent pixel and 16-bit rgb tests in pngstest and removed a
4009     spurious check in pngwrite.c
4010   Added PNG_IMAGE_FLAG_FAST for the benefit of applications that store
4011     intermediate files, or intermediate in-memory data, while processing
4012     image data with the simplified API.  The option makes the files larger
4013     but faster to write and read.  pngstest now uses this by default; this
4014     can be disabled with the --slow option.
4015   Improved pngstest fine tuning of error numbers, new test file generator.
4016     The generator generates images that test the full range of sample values,
4017     allow the error numbers in pngstest to be tuned and checked.  makepng
4018     also allows generation of images with extra chunks, although this is
4019     still work-in-progress.
4020   Added check for invalid palette index while reading.
4021   Fixed some bugs in ICC profile writing. The code should now accept
4022     all potentially valid ICC profiles and reject obviously invalid ones.
4023     It now uses png_error() to do so rather than casually writing a PNG
4024     without the necessary color data.
4025   Removed whitespace from the end of lines in all source files and scripts.
4027 Version 1.6.0beta16 [March 6, 2012]
4028   Relocated palette-index checking function from pngrutil.c to pngtrans.c
4029   Added palette-index checking while writing.
4030   Changed png_inflate() and calling routines to avoid overflow problems.
4031     This is an intermediate check-in that solves the immediate problems and
4032     introduces one performance improvement (avoiding a copy via png_ptr->zbuf.)
4033     Further changes will be made to make ICC profile handling more secure.
4034   Fixed build warnings (MSVC, GCC, GCC v3). Cygwin GCC with default options
4035     declares 'index' as a global, causing a warning if it is used as a local
4036     variable.  GCC 64-bit warns about assigning a (size_t) (unsigned 64-bit)
4037     to an (int) (signed 32-bit).  MSVC, however, warns about using the
4038     unary '-' operator on an unsigned value (even though it is well defined
4039     by ANSI-C to be ~x+1).  The padding calculation was changed to use a
4040     different method.  Removed the tests on png_ptr->pass.
4041   Added contrib/libtests/tarith.c to test internal arithmetic functions from
4042     png.c. This is a libpng maintainer program used to validate changes to the
4043     internal arithmetic functions.
4044   Made read 'inflate' handling like write 'deflate' handling. The read
4045     code now claims and releases png_ptr->zstream, like the write code.
4046     The bug whereby the progressive reader failed to release the zstream
4047     is now fixed, all initialization is delayed, and the code checks for
4048     changed parameters on deflate rather than always calling
4049     deflatedEnd/deflateInit.
4050   Validate the zTXt strings in pngvalid.
4051   Added code to validate the windowBits value passed to deflateInit2().
4052     If the call to deflateInit2() is wrong a png_warning will be issued
4053     (in fact this is harmless, but the PNG data produced may be sub-optimal).
4055 Version 1.6.0beta17 [March 10, 2012]
4056   Fixed PNG_LIBPNG_BUILD_BASE_TYPE definition. 
4057   Reject all iCCP chunks after the first, even if the first one is invalid.
4058   Deflate/inflate was reworked to move common zlib calls into single
4059     functions [rw]util.c.  A new shared keyword check routine was also added
4060     and the 'zbuf' is no longer allocated on progressive read.  It is now
4061     possible to call png_inflate() incrementally.  A warning is no longer
4062     issued if the language tag or translated keyword in the iTXt chunk
4063     has zero length.
4064   If benign errors are disabled use maximum window on ancilliary inflate.
4065     This works round a bug introduced in 1.5.4 where compressed ancillary
4066     chunks could end up with a too-small windowBits value in the deflate
4067     header.
4069 Version 1.6.0beta18 [March 16, 2012]
4070   Issue a png_benign_error() instead of png_warning() about bad palette index.
4071   In pngtest, treat benign errors as errors if "-strict" is present.
4072   Fixed an off-by-one error in the palette index checking function.
4073   Fixed a compiler warning under Cygwin (Windows-7, 32-bit system)
4074   Revised example.c to put text strings in a temporary character array
4075     instead of directly assigning string constants to png_textp members.
4076     This avoids compiler warnings when -Wwrite-strings is enabled.
4077   Added output flushing to aid debugging under Visual Studio. Unfortunately
4078     this is necessary because the VS2010 output window otherwise simply loses
4079     the error messages on error (they weren't flushed to the window before
4080     the process exited, apparently!)
4081   Added configuration support for benign errors and changed the read
4082     default. Also changed some warnings in the iCCP and sRGB handling
4083     from to benign errors. Configuration now makes read benign
4084     errors warnings and write benign errors to errors by default (thus
4085     changing the behavior on read).  The simplified API always forces
4086     read benign errors to warnings (regardless of the system default, unless
4087     this is disabled in which case the simplified API can't be built.)
4089 Version 1.6.0beta19 [March 18, 2012]
4090   Work around for duplicate row start calls; added warning messages.
4091     This turns on PNG_FLAG_DETECT_UNINITIALIZED to detect app code that
4092     fails to call one of the 'start' routines (not enabled in libpng-1.5
4093     because it is technically an API change, since it did normally work
4094     before.)  It also makes duplicate calls to png_read_start_row (an
4095     internal function called at the start of the image read) benign, as
4096     they were before changes to use png_inflate_claim. Somehow webkit is
4097     causing this to happen; this is probably a mis-feature in the zlib
4098     changes so this commit is only a work-round.
4099   Removed erroneous setting of DETECT_UNINITIALIZED and added more
4100     checks. The code now does a png_error if an attempt is made to do the
4101     row initialization twice; this is an application error and it has
4102     serious consequences because the transform data in png_struct is
4103     changed by each call.
4104   Added application error reporting and added chunk names to read
4105     benign errors; also added --strict to pngstest - not enabled
4106     yet because a warning is produced.
4107   Avoid the double gamma correction warning in the simplified API.
4108     This allows the --strict option to pass in the pngstest checks
4110 Version 1.6.0beta20 [March 29, 2012]
4111   Changed chunk handler warnings into benign errors, incrementally load iCCP
4112   Added checksum-icc.c to contrib/tools
4113   Prevent PNG_EXPAND+PNG_SHIFT doing the shift twice.
4114   Recognize known sRGB ICC profiles while reading; prefer writing the
4115     iCCP profile over writing the sRGB chunk, controlled by the
4116     PNG_sRGB_PROFILE_CHECKS option.
4117   Revised png_set_text_2() to avoid potential memory corruption (fixes
4118     CVE-2011-3048, also known as CVE-2012-3425).
4120 Version 1.6.0beta21 [April 27, 2012]
4121   Revised scripts/makefile.darwin: use system zlib; remove quotes around
4122     architecture list; add missing ppc architecture; add architecture options
4123     to shared library link; don't try to create a shared lib based on missing
4124     RELEASE variable.
4125   Enable png_set_check_for_invalid_index() for both read and write.
4126   Removed #ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED in pngpriv.h around
4127     declaration of png_handle_unknown().
4128   Added -lssp_nonshared in a comment in scripts/makefile.freebsd
4129     and changed deprecated NOOBJ and NOPROFILE to NO_OBJ and NO_PROFILE.
4131 Version 1.6.0beta22 [May 23, 2012]
4132   Removed need for -Wno-cast-align with clang.  clang correctly warns on
4133     alignment increasing pointer casts when -Wcast-align is passed. This
4134     fixes the cases that clang warns about either by eliminating the
4135     casts from png_bytep to png_uint_16p (pngread.c), or, for pngrutil.c
4136     where the cast is previously verified or pngstest.c where it is OK, by
4137     introducing new png_aligncast macros to do the cast in a way that clang
4138     accepts.
4140 Version 1.6.0beta23 [June 6, 2012]
4141   Revised CMakeLists.txt to not attempt to make a symlink under mingw.
4142   Made fixes for new optimization warnings from gcc 4.7.0. The compiler
4143     performs an optimization which is safe; however it then warns about it.
4144     Changing the type of 'palette_number' in pngvalid.c removes the warning.
4145   Do not depend upon a GCC feature macro being available for use in generating
4146     the linker mapfile symbol prefix.
4147   Improved performance of new do_check_palette_indexes() function (only
4148     update the value when it actually increases, move test for whether
4149     the check is wanted out of the function.
4151 Version 1.6.0beta24 [June 7, 2012]
4152   Don't check palette indexes if num_palette is 0 (as it can be in MNG files).
4154 Version 1.6.0beta25 [June 16, 2012]
4155   Revised png_set_keep_unknown_chunks() so num_chunks < 0 means ignore all
4156     unknown chunks and all known chunks except for IHDR, PLTE, tRNS, IDAT,
4157     and IEND.  Previously it only meant ignore all unknown chunks, the
4158     same as num_chunks == 0. Revised png_image_skip_unused_chunks() to
4159     provide a list of chunks to be processed instead of a list of chunks to
4160     ignore.  Revised contrib/gregbook/readpng2.c accordingly.
4162 Version 1.6.0beta26 [July 10, 2012]
4163   Removed scripts/makefile.cegcc from the *.zip and *.7z distributions; it
4164     depends on configure, which is not included in those archives.
4165   Moved scripts/chkfmt to contrib/tools.
4166   Changed "a+w" to "u+w" in Makefile.in to fix CVE-2012-3386.
4168 Version 1.6.0beta27 [August 11, 2012]
4169   Do not compile PNG_DEPRECATED, PNG_ALLOC and PNG_PRIVATE when __GNUC__ < 3.
4170   Do not use __restrict when GNUC is <= 3.1
4171   Removed references to png_zalloc() and png_zfree() from the manual.
4172   Fixed configurations where floating point is completely disabled.  Because
4173     of the changes to support symbol prefixing PNG_INTERNAL_FUNCTION declares
4174     floating point APIs during libpng builds even if they are completely
4175     disabled. This requires the png floating point types (png_double*) to be
4176     declared even though the functions are never actually defined.  This
4177     change provides a dummy definition so that the declarations work, yet any
4178     implementation will fail to compile because of an incomplete type.
4179   Re-eliminated the use of strcpy() in pngtest.c.  An unncessary use of
4180     strcpy() was accidentally re-introduced in libpng16; this change replaces
4181     it with strncpy().
4182   Eliminated use of png_sizeof(); use sizeof() instead.
4183   Use a consistent style for (sizeof type) and (sizeof (array))
4184   Cleanup of png_set_filler().  This function does very different things on
4185     read and write.  In libpng 1.6 the two cases can be distinguished and
4186     considerable code cleanup, and extra error checking, is possible.  This
4187     makes calls on the write side that have no effect be ignored with a
4188     png_app_error(), which can be disabled in the app using
4189     png_set_benign_errors(), and removes the spurious use of usr_channels
4190     on the read side.
4191   Insist on autotools 1.12.1 for git builds because there are security issues
4192     with 1.12 and insisting on anything less would allow 1.12 to be used.
4193   Removed info_ptr->signature[8] from WRITE-only builds.
4194   Add some conditions for compiling png_fixed().  This is a small function
4195     but it requires "-lm" on some platforms.
4196   Cause pngtest --strict to fail on any warning from libpng (not just errors)
4197     and cause it not to fail at the comparison step if libpng lacks support
4198     for writing chunks that it reads from the input (currently only implemented
4199     for compressed text chunks).
4200   Make all three "make check" test programs work without READ or WRITE support.
4201     Now "make check" will succeed even if libpng is compiled with -DPNG_NO_READ
4202     or -DPNG_NO_WRITE.  The tests performed are reduced, but the basic reading
4203     and writing of a PNG file is always tested by one or more of the tests.
4204   Consistently use strlen(), memset(), memcpy(), and memcmp() instead of the
4205     png_strlen(), png_memset(), png_memcpy(), and png_memcmp() macros.
4206   Removed the png_sizeof(), png_strlen(), png_memset(), png_memcpy(), and
4207     png_memcmp() macros.
4208   Work around gcc 3.x and Microsoft Visual Studio 2010 complaints. Both object
4209     to the split initialization of num_chunks.
4211 Version 1.6.0beta28 [August 29, 2012]
4212   Unknown handling fixes and clean up. This adds more correct option
4213     control of the unknown handling, corrects the pre-existing bug where
4214     the per-chunk 'keep' setting is ignored and makes it possible to skip
4215     IDAT chunks in the sequential reader (broken in earlier 1.6 versions).
4216     There is a new test program, test-unknown.c, which is a work in progress
4217     (not currently part of the test suite).  Comments in the header files now
4218     explain how the unknown handling works.
4219   Allow fine grain control of unknown chunk APIs. This change allows
4220     png_set_keep_unknown_chunks() to be turned off if not required and causes
4221     both read and write to behave appropriately (on read this is only possible
4222     if the user callback is used to handle unknown chunks).  The change
4223     also removes the support for storing unknown chunks in the info_struct
4224     if the only unknown handling enabled is via the callback, allowing libpng
4225     to be configured with callback reading and none of the unnecessary code.
4226   Corrected fix for unknown handling in pngtest. This reinstates the
4227     libpng handling of unknown chunks other than vpAg and sTER (including
4228     unsafe-to-copy chunks which were dropped before) and eliminates the
4229     repositioning of vpAg and sTER in pngtest.png by changing pngtest.png
4230     (so the chunks are where libpng would put them).
4231   Added "tunknown" test and corrected a logic error in png_handle_unknown()
4232     when SAVE support is absent.  Moved the shell test scripts for
4233     contrib/libtests from the libpng top directory to contrib/libtests.
4234     png_handle_unknown() must always read or skip the chunk, if
4235     SAVE_UNKNOWN_CHUNKS is turned off *and* the application does not set
4236     a user callback an unknown chunk will not be read, leading to a read
4237     error, which was revealed by the "tunknown" test.
4238   Cleaned up and corrected ICC profile handling.
4239     contrib/libtests/makepng: corrected 'rgb' and 'gray' cases.  profile_error
4240     messages could be truncated; made a correct buffer size calculation and
4241     adjusted pngerror.c appropriately. png_icc_check_* checking improved;
4242     changed the functions to receive the correct color type of the PNG on read
4243     or write and check that it matches the color space of the profile (despite
4244     what the comments said before, there is danger in assuming the app will
4245     cope correctly with an RGB profile on a grayscale image and, since it
4246     violates the PNG spec, allowing it is certain to produce inconsistent
4247     app behavior and might even cause app crashes.) Check that profiles
4248     contain the tags needed to process the PNG (tags all required by the ICC
4249     spec). Removed unused PNG_STATIC from pngpriv.h.
4251 Version 1.6.0beta29 [September 4, 2012]
4252   Fixed the simplified API example programs to add the *colormap parameter
4253     to several of he API and improved the error message if the version field
4254     is not set.
4255   Added contrib/examples/* to the *.zip and *.7z distributions.
4256   Updated simplified API synopses and description of the png_image structure
4257     in the manual.
4258   Made makepng and pngtest produce identical PNGs, add "--relaxed" option
4259     to pngtest. The "--relaxed" option turns off the benign errors that are
4260     enabled by default in pre-RC builds. makepng can now write ICC profiles
4261     where the length has not been extended to a multiple of 4, and pngtest
4262     now intercepts all libpng errors, allowing the previously-introduced
4263     "--strict test" on no warnings to actually work.
4264   Improved ICC profile handling including cHRM chunk generation and fixed
4265     Cygwin+MSVC build errors. The ICC profile handling now includes more
4266     checking.  Several errors that caused rejection of the profile are now
4267     handled with a warning in such a way that the invalid profiles will be
4268     read by default in release (but not pre-RC) builds but will not be
4269     written by default.  The easy part of handling the cHRM chunk is written,
4270     where the ICC profile contains the required data.  The more difficult
4271     part plus guessing a gAMA value requires code to pass selected RGB values
4272     through the profile.
4274 Version 1.6.0beta30 [October 24, 2012]
4275   Changed ICC profile matrix/vector types to not depend on array type rules.
4276     By the ANSI-C standard the new types should be identical to the previous
4277     versions, and all known versions of gcc tested with the previous versions
4278     except for GCC-4.2.1 work with this version.  The change makes the ANSI-C
4279     rule that const applied to an array of elements applies instead to the
4280     elements in the array moot by explicitly applying const to the base
4281     elements of the png_icc_matrix and png_icc_vector types. The accidental
4282     (harmless) 'const' previously applied to the parameters of two of the
4283     functions have also been removed.
4284   Added a work around for GCC 4.2 optimization bug.
4285   Marked the broken (bad white point) original HP sRGB profiles correctly and
4286     correct comments.
4287   Added -DZ_SOLO to contrib/pngminim/*/makefile to work with zlib-1.2.7
4288   Use /MDd for vstudio debug builds. Also added pngunkown to the vstudio
4289     builds, fixed build errors and corrected a minor exit code error in
4290     pngvalid if the 'touch' file name is invalid.
4291   Add updated WARNING file to projects/vstudio from libpng 1.5/vstudio
4292   Fixed build when using #define PNG_NO_READ_GAMMA in png_do_compose() in
4293     pngrtran.c (Domani Hannes).
4295 Version 1.6.0beta31 [November 1, 2012]
4296   Undid the erroneous change to vstudio/pngvalid build in libpng-1.6.0beta30.
4297   Made pngvalid so that it will build outside the libpng source tree.
4298   Made builds -DPNG_NO_READ_GAMMA compile (the unit tests still fail).
4299   Made PNG_NO_READ_GAMMA switch off interfaces that depend on READ_GAMMA.
4300     Prior to 1.6.0 switching off READ_GAMMA did unpredictable things to the
4301     interfaces that use it (specifically, png_do_background in 1.4 would
4302     simply display composite for grayscale images but do composition
4303     with the incorrect arithmetic for color ones). In 1.6 the semantic
4304     of -DPNG_NO_READ_GAMMA is changed to simply disable any interface that
4305     depends on it; this obliges people who set it to consider whether they
4306     really want it off if they happen to use any of the interfaces in
4307     question (typically most users who disable it won't).
4308   Fixed GUIDs in projects/vstudio. Some were duplicated or missing,
4309     resulting in VS2010 having to update the files.
4310   Removed non-working ICC profile support code that was mostly added to
4311     libpng-1.6.0beta29 and beta30. There was too much code for too little
4312     gain; implementing full ICC color correction may be desireable but is left
4313     up to applications.
4315 Version 1.6.0beta32 [November 25, 2012]
4316   Fixed an intermittent SEGV in pngstest due to an uninitialized array element.
4317   Added the ability for contrib/libtests/makepng.c to make a PNG with just one
4318     color. This is useful for debugging pngstest color inaccuracy reports.
4319   Fixed error checking in the simplified write API (Olaf van der Spek)
4320   Made png_user_version_check() ok to use with libpng version 1.10.x and later.
4322 Version 1.6.0beta33 [December 15, 2012]
4323   Fixed typo in png.c (PNG_SET_CHUNK_MALLOC_MAX should be PNG_CHUNK_MALLOC_MAX)
4324     that causes the MALLOC_MAX limit not to work (John Bowler)
4325   Change png_warning() to png_app_error() in pngwrite.c and comment the
4326     fall-through condition.
4327   Change png_warning() to png_app_warning() in png_write_tRNS().
4328   Rearranged the ARM-NEON optimizations: Isolated the machine specific code
4329     to the hardware subdirectory and added comments to pngrutil.c so that
4330     implementors of other optimizations know what to do.
4331   Fixed cases of unquoted DESTDIR in Makefile.am
4332   Rebuilt Makefile.in, etc., with autoconf-2.69 and automake-1.12.5.
4334 Version 1.6.0beta34 [December 19, 2012]
4335   Cleaned up whitespace in the synopsis portion of the manpage "libpng.3"
4336   Disassembled the version number in scripts/options.awk (necessary for
4337     building on SunOs).
4339 Version 1.6.0beta35 [December 23, 2012]
4340   Made default Zlib compression settings be configurable. This adds #defines to
4341     pnglibconf.h to control the defaults.
4342   Fixed Windows build issues, enabled ARM compilation. Various warnings issued
4343     by earlier versions of GCC fixed for Cygwin and Min/GW (which both use old
4344     GCCs.) ARM support is enabled by default in zlib.props (unsupported by
4345     Microsoft) and ARM compilation is made possible by deleting the check for
4346     x86. The test programs cannot be run because they are not signed.
4348 Version 1.6.0beta36 [January 2, 2013]
4349   Discontinued distributing libpng-1.x.x.tar.bz2.
4350   Discontinued distributing libpng-1.7.0-1.6.0-diff.txt and similar.
4351   Rebuilt configure with autoconf-2.69 (inadvertently not done in beta33)
4352   Fixed 'make distcheck' on SUN OS - libpng.so was not being removed
4354 Version 1.6.0beta37 [January 10, 2013]
4355   Fixed conceivable but difficult to repro overflow. Also added two test
4356     programs to generate and test a PNG which should have the problem.
4358 Version 1.6.0beta39 [January 19, 2013]
4359   Again corrected attempt at overflow detection in png_set_unknown_chunks()
4360   (CVE-2013-7353).  Added overflow detection in png_set_sPLT() and
4361   png_set_text_2() (CVE-2013-7354).
4363 Version 1.6.0beta40 [January 20, 2013]
4364   Use consistent handling of overflows in text, sPLT and unknown png_set_* APIs
4366 Version 1.6.0rc01 [January 26, 2013]
4367   No changes.
4369 Version 1.6.0rc02 [February 4, 2013]
4370   Added png_get_palette_max() function.
4372 Version 1.6.0rc03 [February 5, 2013]
4373   Fixed the png_get_palette_max API.
4375 Version 1.6.0rc04 [February 7, 2013]
4376   Turn serial tests back on (recently turned off by autotools upgrade).
4378 Version 1.6.0rc05 [February 8, 2013]
4379   Update manual about png_get_palette_max().
4381 Version 1.6.0rc06 [February 9, 2013]
4382   Fixed missing dependency in --prefix builds The intermediate
4383     internal 'prefix.h' file can only be generated correctly after
4384     pnglibconf.h, however the dependency was not in Makefile.am.  The
4385     symptoms are unpredictable depending on the order make chooses to
4386     build pngprefix.h and pnglibconf.h, often the error goes unnoticed
4387     because there is a system pnglibconf.h to use instead.
4389 Version 1.6.0rc07 [February 10, 2013]
4390   Enclosed the new png_get_palette_max in #ifdef PNG_GET_PALETTE_MAX_SUPPORTED
4391     block, and revised pnglibconf.h and pnglibconf.h.prebuilt accordingly.
4393 Version 1.6.0rc08 [February 10, 2013]
4394   Fix typo in png.h #ifdef
4396 Version 1.6.0 [February 14, 2013]
4397   No changes.
4399 Version 1.6.1beta01 [February 16, 2013]
4400   Made symbol prefixing work with the ARM neon optimizations. Also allow
4401     pngpriv.h to be included for preprocessor definitions only, so it can
4402     be used in non-C/C++ files. Back ported from libpng 1.7.
4403   Made sRGB check numbers consistent.
4404   Ported libpng 1.5 options.awk/dfn file handling to 1.6, fixed one bug.
4405   Removed cc -E workround, corrected png_get_palette_max API Tested on
4406     SUN OS cc 5.9, which demonstrates the tokenization problem previously
4407     avoided by using /lib/cpp.  Since all .dfn output is now protected in
4408     double quotes unless it is to be macro substituted the fix should
4409     work everywhere.
4410   Enabled parallel tests - back ported from libpng-1.7.
4411   scripts/pnglibconf.dfa formatting improvements back ported from libpng17.
4412   Fixed a race condition in the creation of the build 'scripts' directory
4413     while building with a parallel make.
4414   Use approved/supported Android method to check for NEON, use Linux/POSIX
4415     1003.1 API to check /proc/self/auxv avoiding buffer allocation and other
4416     library calls (ported from libpng15).
4418 Version 1.6.1beta02 [February 19, 2013]
4419   Use parentheses more consistently in "#if defined(MACRO)" tests.
4420   Folded long lines.
4421   Reenabled code to allow zero length PLTE chunks for MNG.
4423 Version 1.6.1beta03 [February 22, 2013]
4424   Fixed ALIGNED_MEMORY support.
4425   Added a new configure option:
4426     --enable-arm-neon=always will stop the run-time checks. New checks
4427     within arm/arm_init.c will cause the code not to be compiled unless
4428     __ARM_NEON__ is set. This should make it fail safe (if someone asks
4429     for it on then the build will fail if it can't be done.)
4430   Updated the INSTALL document.
4432 Version 1.6.1beta04 [February 27, 2013]
4433   Revised INSTALL to recommend using CPPFLAGS instead of INCLUDES.
4434   Revised scripts/makefile.freebsd to respect ZLIBLIB and ZLIBINC.
4435   Revised scripts/dfn.awk to work with the buggy MSYS awk that has trouble
4436     with CRLF line endings.
4438 Version 1.6.1beta05 [March 1, 2013]
4439   Avoid a possible memory leak in contrib/gregbook/readpng.c
4441 Version 1.6.1beta06 [March 4, 2013]
4442   Better documentation of unknown handling API interactions.
4443   Corrected Android builds and corrected libpng.vers with symbol
4444     prefixing.  It also makes those tests compile and link on Android.
4445   Added an API png_set_option() to set optimization options externally,
4446     providing an alternative and general solution for the non-portable
4447     run-time tests used by the ARM Neon code, using the PNG_ARM_NEON option.
4448   The order of settings vs options in pnglibconf.h is reversed to allow
4449     settings to depend on options and options can now set (or override) the
4450     defaults for settings.
4452 Version 1.6.1beta07 [March 7, 2013]
4453   Corrected simplified API default gamma for color-mapped output, added
4454     a flag to change default. In 1.6.0 when the simplified API was used
4455     to produce color-mapped output from an input image with no gamma
4456     information the gamma assumed for the input could be different from
4457     that assumed for non-color-mapped output.  In particular 16-bit depth
4458     input files were assumed to be sRGB encoded, whereas in the 'direct'
4459     case they were assumed to have linear data.  This was an error.  The
4460     fix makes the simplified API treat all input files the same way and
4461     adds a new flag to the png_image::flags member to allow the
4462     application/user to specify that 16-bit files contain sRGB data
4463     rather than the default linear.
4464   Fixed bugs in the pngpixel and makepng test programs.
4466 Version 1.6.1beta08 [March 7, 2013]
4467   Fixed CMakelists.txt to allow building a single variant of the library
4468     (Claudio Bley):
4469   Introduced a PNG_LIB_TARGETS variable that lists all activated library
4470     targets.  It is an error if this variable ends up empty, ie. you have
4471     to build at least one library variant.
4472   Made the *_COPY targets only depend on library targets actually being build.
4473   Use PNG_LIB_TARGETS to unify a code path.
4474   Changed the CREATE_SYMLINK macro to expect the full path to a file as the
4475     first argument. When symlinking the filename component of that path is
4476     determined and used as the link target.
4477   Use copy_if_different in the CREATE_SYMLINK macro.
4479 Version 1.6.1beta09 [March 13, 2013]
4480   Eliminated two warnings from the Intel C compiler. The warnings are
4481     technically valid, although a reasonable treatment of division would
4482     show it to be incorrect.
4484 Version 1.6.1rc01 [March 21, 2013]
4485   No changes.
4487 Version 1.6.1 [March 28, 2013]
4488   No changes.
4490 Version 1.6.2beta01 [April 14, 2013]
4491   Updated documentation of 1.5.x to 1.6.x changes in iCCP chunk handling.
4492   Fixed incorrect warning of excess deflate data. End condition - the
4493     warning would be produced if the end of the deflate stream wasn't read
4494     in the last row.  The warning is harmless.
4495   Corrected the test on user transform changes on read. It was in the
4496     png_set of the transform function, but that doesn't matter unless the
4497     transform function changes the rowbuf size, and that is only valid if
4498     transform_info is called.
4499   Corrected a misplaced closing bracket in contrib/libtests/pngvalid.c
4500     (Flavio Medeiros).
4501   Corrected length written to uncompressed iTXt chunks (Samuli Suominen).
4502     Bug was introduced in libpng-1.6.0.
4504 Version 1.6.2rc01 [April 18, 2013]
4505   Added contrib/tools/fixitxt.c, to repair the erroneous iTXt chunk length
4506     written by libpng-1.6.0 and 1.6.1.
4507   Disallow storing sRGB information when the sRGB is not supported.
4509 Version 1.6.2rc02 [April 18, 2013]
4510   Merge pngtest.c with libpng-1.7.0
4512 Version 1.6.2rc03 [April 22, 2013]
4513   Trivial spelling cleanup.
4515 Version 1.6.2rc04 and 1.6.2rc05 [omitted]
4517 Version 1.6.2rc06 [April 24, 2013]
4518   Reverted to version 1.6.2rc03.  Recent changes to arm/neon support
4519     have been ported to libpng-1.7.0beta09 and will reappear in version
4520     1.6.3beta01.
4522 Version 1.6.2 [April 25, 2013]
4523   No changes.
4525 Version 1.6.3beta01 [April 25, 2013]
4526   Revised stack marking in arm/filter_neon.S and configure.ac.
4527   Ensure that NEON filter stuff is completely disabled when switched 'off'.
4528     Previously the ARM NEON specific files were still built if the option
4529     was switched 'off' as opposed to being explicitly disabled.
4531 Version 1.6.3beta02 [April 26, 2013]
4532   Test for 'arm*' not just 'arm' in the host_cpu configure variable.
4533   Rebuilt the configure scripts.
4535 Version 1.6.3beta03 [April 30, 2013]
4536   Expanded manual paragraph about writing private chunks, particularly
4537     the need to call png_set_keep_unknown_chunks() when writing them.
4538   Avoid dereferencing NULL pointer possibly returned from
4539     png_create_write_struct() (Andrew Church).
4541 Version 1.6.3beta05 [May 9, 2013]
4542   Calculate our own zlib windowBits when decoding rather than trusting the
4543     CMF bytes in the PNG datastream.
4544   Added an option to force maximum window size for inflating, which was
4545     the behavior of libpng15 and earlier, via a new PNG_MAXIMUM_INFLATE_WINDOW
4546     option for png_set_options().
4547   Added png-fix-itxt and png-fix-too-far-back to the built programs and
4548     removed warnings from the source code and timepng that are revealed as
4549     a result.
4550   Detect wrong libpng versions linked to png-fix-too-far-back, which currently
4551     only works with libpng versions that can be made to reliably fail when
4552     the deflate data contains an out-of-window reference.  This means only
4553     1.6 and later.
4554   Fixed gnu issues: g++ needs a static_cast, gcc 4.4.7 has a broken warning
4555     message which it is easier to work round than ignore.
4556   Updated contrib/pngminus/pnm2png.c (Paul Stewart):
4557     Check for EOF
4558     Ignore "#" delimited comments in input file to pnm2png.c.
4559     Fixed whitespace handling
4560     Added a call to png_set_packing()
4561     Initialize dimension values so if sscanf fails at least we have known
4562       invalid values.
4563   Attempt to detect configuration issues with png-fix-too-far-back, which
4564     requires both the correct libpng and the correct zlib to function
4565     correctly.
4566   Check ZLIB_VERNUM for mismatches, enclose #error in quotes
4567   Added information in the documentation about problems with and fixes for
4568     the bad CRC and bad iTXt chunk situations.
4570 Version 1.6.3beta06 [May 12, 2013]
4571   Allow contrib/pngminus/pnm2png.c to compile without WRITE_INVERT and
4572     WRITE_PACK supported (writes error message that it can't read P1 or
4573     P4 PBM files).
4574   Improved png-fix-too-far-back usage message, added --suffix option.
4575   Revised contrib/pngminim/*/makefile to generate pnglibconf.h with the
4576     right zlib header files.
4577   Separated CPPFLAGS and CFLAGS in contrib/pngminim/*/makefile
4579 Version 1.6.3beta07 [June 8, 2013]
4580   Removed a redundant test in png_set_IHDR().
4581   Added set(CMAKE_CONFIGURATION_TYPES ...) to CMakeLists.txt (Andrew Hundt)
4582   Deleted set(CMAKE_BUILD_TYPE) block from CMakeLists.txt
4583   Enclose the prototypes for the simplified write API in
4584     #ifdef PNG_STDIO_SUPPORTED/#endif
4585   Make ARM NEON support work at compile time (not just configure time).
4586     This moves the test on __ARM_NEON__ into pngconf.h to avoid issues when
4587     using a compiler that compiles for multiple architectures at one time.
4589     pnglibconf.h, allowing more of the decisions to be made internally
4590     (pngpriv.h) during the compile.  Without this, symbol prefixing is broken
4591     under certain circumstances on ARM platforms.  Now only the API parts of
4592     the optimizations ('check' vs 'api') are exposed in the public header files
4593     except that the new setting PNG_ARM_NEON_OPT documents how libpng makes the
4594     decision about whether or not to use the optimizations.
4595   Protect symbol prefixing against CC/CPPFLAGS/CFLAGS useage.
4596     Previous iOS/Xcode fixes for the ARM NEON optimizations moved the test
4597     on __ARM_NEON__ from configure time to compile time.  This breaks symbol
4598     prefixing because the definition of the special png_init_filter_functions
4599     call was hidden at configure time if the relevant compiler arguments are
4600     passed in CFLAGS as opposed to CC.  This change attempts to avoid all
4601     the confusion that would result by declaring the init function even when
4602     it is not used, so that it will always get prefixed.
4604 Version 1.6.3beta08 [June 18, 2013]
4605   Revised libpng.3 so that "doclifter" can process it.
4607 Version 1.6.3beta09 [June 27, 2013]
4608   Revised example.c to illustrate use of PNG_DEFAULT_sRGB and PNG_GAMMA_MAC_18
4609     as parameters for png_set_gamma().  These have been available since
4610     libpng-1.5.4.
4611   Renamed contrib/tools/png-fix-too-far-back.c to pngfix.c and revised it
4612     to check all compressed chunks known to libpng.
4614 Version 1.6.3beta10 [July 5, 2013]
4615   Updated documentation to show default behavior of benign errors correctly.
4616   Only compile ARM code when PNG_READ_SUPPORTED is defined.
4617   Fixed undefined behavior in contrib/tools/pngfix.c and added new strip
4618     option. pngfix relied on undefined behavior and even a simple change from
4619     gcc to g++ caused it to fail.  The new strip option 'unsafe' has been
4620     implemented and is the default if --max is given.  Option names have
4621     been clarified, with --strip=transform now stripping the bKGD chunk,
4622     which was stripped previously with --strip=unused.
4623   Added all documented chunk types to pngpriv.h
4624   Unified pngfix.c source with libpng17.
4626 Version 1.6.3rc01 [July 11, 2013]
4627   No changes.
4629 Version 1.6.3 [July 18, 2013]
4630   Revised manual about changes in iTXt chunk handling made in libpng-1.6.0.
4631   Added "/* SAFE */" comments in pngrutil.c and pngrtran.c where warnings
4632     may be erroneously issued by code-checking applications.
4634 Version 1.6.4beta01 [August 21, 2013]
4635   Added information about png_set_options() to the manual.
4636   Delay calling png_init_filter_functions() until a row with nonzero filter
4637     is found.
4639 Version 1.6.4beta02 [August 30, 2013]
4640   Fixed inconsistent conditional compilation of png_chunk_unknown_handling()
4641     prototype, definition, and usage.  Made it depend on
4644 Version 1.6.4rc01 [September 5, 2013]
4645   No changes.
4647 Version 1.6.4 [September 12, 2013]
4648   No changes.
4650 Version 1.6.5 [September 14, 2013]
4651   Removed two stray lines of code from arm/arm_init.c.
4653 Version 1.6.6 [September 16, 2013]
4654   Removed two stray lines of code from arm/arm_init.c, again.
4656 Version 1.6.7beta01 [September 30, 2013]
4657   Revised unknown chunk code to correct several bugs in the NO_SAVE_/NO_WRITE
4658     combination
4659   Allow HANDLE_AS_UNKNOWN to work when other options are configured off. Also
4660     fixed the pngminim makefiles to work when $(MAKEFLAGS) contains stuff
4661     which terminates the make options (as by default in recent versions of
4662     Gentoo).
4663   Avoid up-cast warnings in pngvalid.c. On ARM the alignment requirements of
4664     png_modifier are greater than that of png_store and as a consequence
4665     compilation of pngvalid.c results in a warning about increased alignment
4666     requirements because of the bare cast to (png_modifier*). The code is safe,
4667     because the pointer is known to point to a stack allocated png_modifier,
4668     but this change avoids the warning.
4669   Fixed default behavior of ARM_NEON_API. If the ARM NEON API option was
4670     compiled without the CHECK option it defaulted to on, not off.
4671   Check user callback behavior in pngunknown.c. Previous versions compiled
4672     if SAVE_UNKNOWN was not available but did nothing since the callback
4673     was never implemented.
4674   Merged pngunknown.c with 1.7 version and back ported 1.7 improvements/fixes
4676 Version 1.6.7beta02 [October 12, 2013]
4677   Made changes for compatibility with automake 1.14:
4678     1) Added the 'compile' program to the list of programs that must be cleaned
4679        in autogen.sh
4680     2) Added 'subdir-objects' which causes .c files in sub-directories to be
4681        compiled such that the corresponding .o files are also in the
4682        sub-directory.  This is because automake 1.14 warns that the
4683        current behavior of compiling to the top level directory may be removed
4684        in the future.
4685     3) Updated dependencies on pnglibconf.h to match the new .o locations and
4686        added all the files in contrib/libtests and contrib/tools that depend
4687        on pnglibconf.h
4688     4) Added 'BUILD_SOURCES = pnglibconf.h'; this is the automake recommended
4689        way of handling the dependencies of sources that are machine generated;
4690        unfortunately it only works if the user does 'make all' or 'make check',
4691        so the dependencies (3) are still required.
4692   Cleaned up (char*) casts of zlib messages. The latest version of the Intel C
4693     compiler complains about casting a string literal as (char*), so copied the
4694     treatment of z_const from the library code into pngfix.c
4695   Simplified error message code in pngunknown. The simplification has the
4696     useful side effect of avoiding a bogus warning generated by the latest
4697     version of the Intel C compiler (it objects to
4698     condition ? string-literal : string-literal).
4699   Make autogen.sh work with automake 1.13 as well as 1.14. Do this by always
4700     removing the 1.14 'compile' script but never checking for it.
4702 Version 1.6.7beta03 [October 19, 2013]
4703   Added ARMv8 support (James Yu <james.yu at linaro.org>).  Added file
4704     arm/filter_neon_intrinsics.c; enable with -mfpu=neon.
4705   Revised pngvalid to generate size images with as many filters as it can
4706     manage, limited by the number of rows.
4707   Cleaned up ARM NEON compilation handling. The tests are now in pngpriv.h
4708     and detect the broken GCC compilers.
4710 Version 1.6.7beta04 [October 26, 2013]
4711   Allow clang derived from older GCC versions to use ARM intrinsics. This
4712     causes all clang builds that use -mfpu=neon to use the intrinsics code,
4713     not the assembler code.  This has only been tested on iOS 7. It may be
4714     necessary to exclude some earlier clang versions but this seems unlikely.
4715   Changed NEON implementation selection mechanism. This allows assembler
4716     or intrinsics to be turned on at compile time during the build by defining
4717     PNG_ARM_NEON_IMPLEMENTATION to the correct value (2 or 1).  This macro
4718     is undefined by default and the build type is selected in pngpriv.h.
4720 Version 1.6.7rc01 [November 2, 2013]
4721   No changes.
4723 Version 1.6.7rc02 [November 7, 2013]
4724   Fixed #include in filter_neon_intrinsics.c and ctype macros. The ctype char
4725     checking macros take an unsigned char argument, not a signed char.
4727 Version 1.6.7 [November 14, 2013]
4728   No changes.
4730 Version 1.6.8beta01 [November 24, 2013]
4731   Moved prototype for png_handle_unknown() in pngpriv.h outside of
4732     the #ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED/#endif block.
4733   Added "-Wall" to CFLAGS in contrib/pngminim/*/makefile
4734   Conditionally compile some unused functions reported by -Wall in
4735     pngminim.
4736   Fixed 'minimal' builds. Various obviously useful minimal configurations
4737     don't build because of missing contrib/libtests test programs and
4738     overly complex dependencies in scripts/pnglibconf.dfa. This change
4739     adds contrib/conftest/*.dfa files that can be used in automatic build
4740     scripts to ensure that these configurations continue to build.
4741   Enabled WRITE_INVERT and WRITE_PACK in contrib/pngminim/encoder.
4742   Fixed pngvalid 'fail' function declaration on the Intel C Compiler.
4743     This reverts to the previous 'static' implementation and works round
4744     the 'unused static function' warning by using PNG_UNUSED().
4746 Version 1.6.8beta02 [November 30, 2013]
4747   Removed or marked PNG_UNUSED some harmless "dead assignments" reported
4748     by clang scan-build.
4749   Changed tabs to 3 spaces in png_debug macros and changed '"%s"m'
4750     to '"%s" m' to improve portability among compilers.
4751   Changed png_free_default() to free() in pngtest.c
4753 Version 1.6.8rc01 [December 12, 2013]
4754   Tidied up pngfix inits and fixed pngtest no-write builds.
4756 Version 1.6.8rc02 [December 14, 2013]
4757   Handle zero-length PLTE chunk or NULL palette with png_error()
4758     instead of png_chunk_report(), which by default issues a warning
4759     rather than an error, leading to later reading from a NULL pointer
4760     (png_ptr->palette) in png_do_expand_palette(). This is CVE-2013-6954
4761     and VU#650142.  Libpng-1.6.1 through 1.6.7 are vulnerable.
4762     Libpng-1.6.0 and earlier do not have this bug.
4764 Version 1.6.8 [December 19, 2013]
4765   No changes.
4767 Version 1.6.9beta01 [December 26, 2013]
4768   Bookkeeping: Moved functions around (no changes). Moved transform
4769     function definitions before the place where they are called so that
4770     they can be made static. Move the intrapixel functions and the
4771     grayscale palette builder out of the png?tran.c files. The latter
4772     isn't a transform function and is no longer used internally, and the
4773     former MNG specific functions are better placed in pngread/pngwrite.c
4774   Made transform implementation functions static. This makes the internal
4775     functions called by png_do_{read|write}_transformations static. On an
4776     x86-64 DLL build (Gentoo Linux) this reduces the size of the text
4777     segment of the DLL by 1208 bytes, about 0.6%. It also simplifies
4778     maintenance by removing the declarations from pngpriv.h and allowing
4779     easier changes to the internal interfaces.
4780   Rebuilt configure scripts with automake-1.14.1 and autoconf-2.69
4781     in the tar distributions.
4783 Version 1.6.9beta02 [January 1, 2014]
4784   Added checks for libpng 1.5 to pngvalid.c.  This supports the use of
4785     this version of pngvalid in libpng 1.5
4786   Merged with pngvalid.c from libpng-1.7 changes to create a single
4787     pngvalid.c
4788   Removed #error macro from contrib/tools/pngfix.c (Thomas Klausner).
4789   Merged pngrio.c, pngtrans.c, pngwio.c, and pngerror.c with libpng-1.7.0
4790   Merged libpng-1.7.0 changes to make no-interlace configurations work
4791     with test programs.
4792   Revised pngvalid.c to support libpng 1.5, which does not support the
4793     PNG_MAXIMUM_INFLATE_WINDOW option, so #define it out when appropriate in
4794     pngvalid.c
4795   Allow unversioned links created on install to be disabled in configure.
4796     In configure builds 'make install' changes/adds links like png.h
4797     and libpng.a to point to the newly installed, versioned, files (e.g.
4798     libpng17/png.h and libpng17.a). Three new configure options and some
4799     rearrangement of Makefile.am allow creation of these links to be disabled.
4801 Version 1.6.9beta03 [January 10, 2014]
4802   Removed potentially misleading warning from png_check_IHDR().
4804 Version 1.6.9beta04 [January 20, 2014]
4805   Updated scripts/makefile.* to use CPPFLAGS (Cosmin).
4806   Added clang attribute support (Cosmin).
4808 Version 1.6.9rc01 [January 28, 2014]
4809   No changes.
4811 Version 1.6.9rc02 [January 30, 2014]
4812   Quiet an uninitialized memory warning from VC2013 in png_get_png().
4814 Version 1.6.9 [February 6, 2014]
4816 Version 1.6.10beta01 [February 9, 2014]
4817   Backported changes from libpng-1.7.0beta30 and beta31:
4818   Fixed a large number of instances where PNGCBAPI was omitted from
4819     function definitions.
4820   Added pngimage test program for png_read_png() and png_write_png()
4821     with two new test scripts.
4822   Removed dependence on !PNG_READ_EXPAND_SUPPORTED for calling
4823     png_set_packing() in png_read_png().
4824   Fixed combination of ~alpha with shift. On read invert alpha, processing
4825     occurred after shift processing, which causes the final values to be
4826     outside the range that should be produced by the shift. Reversing the
4827     order on read makes the two transforms work together correctly and mirrors
4828     the order used on write.
4829   Do not read invalid sBIT chunks. Previously libpng only checked sBIT
4830     values on write, so a malicious PNG writer could therefore cause
4831     the read code to return an invalid sBIT chunk, which might lead to
4832     application errors or crashes.  Such chunks are now skipped (with
4833     chunk_benign_error).
4834   Make png_read_png() and png_write_png() prototypes in png.h depend
4836   Support builds with unsupported PNG_TRANSFORM_* values.  All of the
4837     PNG_TRANSFORM_* values are always defined in png.h and, because they
4838     are used for both read and write in some cases, it is not reliable
4839     to #if out ones that are totally unsupported. This change adds error
4840     detection in png_read_image() and png_write_image() to do a
4841     png_app_error() if the app requests something that cannot be done
4842     and it adds corresponding code to pngimage.c to handle such options
4843     by not attempting to test them.
4845 Version 1.6.10beta02 [February 23, 2014]
4846   Moved redefines of png_error(), png_warning(), png_chunk_error(),
4847     and png_chunk_warning() from pngpriv.h to png.h to make them visible
4848     to libpng-calling applications.
4849   Moved OS dependent code from arm/arm_init.c, to allow the included
4850     implementation of the ARM NEON discovery function to be set at
4851     build-time and provide sample implementations from the current code in the
4852     contrib/arm-neon subdirectory. The __linux__ code has also been changed to
4853     compile and link on Android by using /proc/cpuinfo, and the old linux code
4854     is in contrib/arm-neon/linux-auxv.c.  The new code avoids POSIX and Linux
4855     dependencies apart from opening /proc/cpuinfo and is C90 compliant.
4856   Check for info_ptr == NULL early in png_read_end() so we don't need to
4857     run all the png_handle_*() and depend on them to return if info_ptr == NULL.
4858     This improves the performance of png_read_end(png_ptr, NULL) and makes
4859     it more robust against future programming errors.
4860   Check for __has_extension before using it in pngconf.h, to
4861     support older Clang versions (Jeremy Sequoia).
4862   Treat CRC error handling with png_set_crc_action(), instead of with
4863     png_set_benign_errors(), which has been the case since libpng-1.6.0beta18.
4864   Use a user warning handler in contrib/gregbook/readpng2.c instead of default,
4865     so warnings will be put on stderr even if libpng has CONSOLE_IO disabled.
4866   Added png_ptr->process_mode = PNG_READ_IDAT_MODE in png_push_read_chunk
4867     after recognizing the IDAT chunk, which avoids an infinite loop while
4868     reading a datastream whose first IDAT chunk is of zero-length.
4869     This fixes CERT VU#684412 and CVE-2014-0333.
4870   Don't recognize known sRGB profiles as sRGB if they have been hacked,
4871     but don't reject them and don't issue a copyright violation warning.
4873 Version 1.6.10beta03 [February 25, 2014]
4874   Moved some documentation from png.h to libpng.3 and libpng-manual.txt
4875   Minor editing of contrib/arm-neon/README and contrib/examples/*.c
4877 Version 1.6.10rc01 [February 27, 2014]
4878   Fixed typos in the manual and in scripts/pnglibconf.dfa (CFLAGS -> CPPFLAGS
4881 Version 1.6.10rc02 [February 28, 2014]
4882   Removed unreachable return statement after png_chunk_error()
4883     in pngrutil.c
4885 Version 1.6.10rc03 [March 4, 2014]
4886   Un-deprecated png_data_freer().
4888 Version 1.6.10 [March 6, 2014]
4889   No changes.
4891 Version 1.6.11beta01 [March 17, 2014]
4892   Use "if (value != 0)" instead of "if (value)" consistently.
4893   Changed ZlibSrcDir from 1.2.5 to 1.2.8 in projects/vstudio.
4894   Moved configuration information from the manual to the INSTALL file.
4896 Version 1.6.11beta02 [April 6, 2014]
4897   Removed #if/#else/#endif from inside two pow() calls in pngvalid.c because
4898     they were handled improperly by Portland Group's PGI-14.1 - PGI-14.3
4899     when using its "__builtin_pow()" function.
4900   Silence 'unused parameter' build warnings (Cosmin Truta).
4901   $(CP) is now used alongside $(RM_F).  Also, use 'copy' instead of 'cp'
4902     where applicable, and applied other minor makefile changes (Cosmin).
4903   Don't warn about invalid dimensions exceeding user limits (Cosmin).
4904   Allow an easy replacement of the default pre-built configuration
4905     header with a custom header, via the make PNGLIBCONF_H_PREBUILT
4906     macro (Cosmin).
4908 Version 1.6.11beta03 [April 6, 2014]
4909   Fixed a typo in pngrutil.c, introduced in libpng-1.5.6, that interferes
4910     with "blocky" expansion of sub-8-bit interlaced PNG files (Eric Huss).
4911   Optionally use  __builtin_bswap16() in png_do_swap().
4913 Version 1.6.11beta04 [April 19, 2014]
4914   Made progressive reading of interlaced images consistent with the
4915     behavior of the sequential reader and consistent with the manual, by
4916     moving some code out of the PNG_READ_INTERLACING_SUPPORTED blocks. The
4917     row_callback now receives the proper pass number and unexpanded rows, when
4918     png_combine_row() isn't built or used, and png_set_interlace_handling()
4919     is not called.
4920   Allow PNG_sRGB_PROFILE_CHECKING = (-1) to mean no sRGB profile checking.
4922 Version 1.6.11beta05 [April 26, 2014]
4923   Do not reject ICC V2 profiles that lack padding (Kai-Uwe Behrmann).
4924   Relocated closing bracket of the sRGB profile test loop to avoid getting
4925     "Not recognizing known sRGB profile that has been edited" warning for
4926     ICC V2 profiles that lack the MD5 signature in the profile header.
4928 Version 1.6.11beta06 [May 19, 2014]
4929   Added PNG_SKIP_sRGB_CHECK_PROFILE choice for png_set_option().
4931 Version 1.6.11rc01 [May 27, 2014]
4932   No changes.
4934 Version 1.6.11rc02 [June 3, 2014]
4935   Test ZLIB_VERNUM instead of PNG_ZLIB_VERNUM in contrib/tools/pngfix.c
4937 Version 1.6.11 [June 5, 2014]
4938   No changes.
4940 Version 1.6.12rc01 [June 6, 2014]
4941   Relocated new code from 1.6.11beta06 in png.c to a point after the
4942     declarations (Max Stepin).
4944 Version 1.6.12rc02 [June 7, 2014]
4945   Changed file permissions of contrib/tools/intgamma.sh,
4946     test-driver, and compile from 0644 to 0755 (Cosmin).
4948 Version 1.6.12rc03 [June 8, 2014]
4949   Ensure "__has_attribute()" macro exists before trying to use it with
4950     old clang compilers (MacPorts Ticket #43939).
4952 Version 1.6.12 [June 12, 2014]
4953   No changes.
4955 Version 1.6.13beta01 [July 4, 2014]
4956   Quieted -Wsign-compare and -Wclobber compiler warnings in
4957     contrib/pngminus/*.c
4958   Added "(void) png_ptr;" where needed in contrib/gregbook to quiet
4959     compiler complaints about unused pointers.
4960   Split a long output string in contrib/gregbook/rpng2-x.c.
4961   Added "PNG_SET_OPTION" requirement for sRGB chunk support to pnglibconf.dfa,
4962     Needed for write-only support (John Bowler).
4963   Changed "if defined(__ARM_NEON__)" to
4964     "if (defined(__ARM_NEON__) || defined(__ARM_NEON))" (James Wu).
4965   Fixed clang no-warning builds: png_digit was defined but never used.
4967 Version 1.6.13beta02 [July 21, 2014]
4968   Fixed an incorrect separator ("/" should be "\") in scripts/makefile.vcwin32
4969     (bug report from Wolfgang S. Kechel).  Bug was introduced in libpng-1.6.11.
4970     Also fixed makefile.bc32, makefile.bor, makefile.msc, makefile.intel, and
4971     makefile.tc3 similarly.
4973 Version 1.6.13beta03 [August 3, 2014]
4974   Removed scripts/makefile.elf. It has not worked since libpng-1.5.0beta14
4975     due to elimination of the PNG_FUNCTION_EXPORT and PNG_DATA_EXPORT
4976     definitions from pngconf.h.
4977   Ensure that CMakeLists.txt makes the target "lib" directory before making
4978     symbolic link into it (SourceForge bug report #226 by Rolf Timmermans).
4980 Version 1.6.13beta04 [August 8, 2014]
4981   Added opinion that the ECCN (Export Control Classification Number) for
4982     libpng is EAR99 to the README file.
4983   Eliminated use of "$<" in makefile explicit rules, when copying
4984     $PNGLIBCONF_H_PREBUILT.  This does not work on some versions of make;
4985     bug introduced in libpng version 1.6.11.
4987 Version 1.6.13rc01 [August 14, 2014]
4988   Made "ccopts" agree with "CFLAGS" in scripts/makefile.hp* and makefile.*sunu
4990 Version 1.6.13 [August 21, 2014]
4991   No changes.
4993 Version 1.6.14beta01 [September 14, 2014]
4994   Guard usage of png_ptr->options with #ifdef PNG_SET_OPTION_SUPPORTED.
4995   Do not build contrib/tools/pngfix.c when PNG_SETJMP_NOT_SUPPORTED,
4996     to allow "make" to complete without setjmp support (bug report by
4997     Claudio Fontana)
4998   Add "#include <setjmp.h>" to contrib/tools/pngfix.c (John Bowler)
5000 Version 1.6.14beta02 [September 18, 2014]
5001   Use nanosleep() instead of usleep() in contrib/gregbook/rpng2-x.c
5002     because usleep() is deprecated.
5003   Define usleep() in contrib/gregbook/rpng2-x.c if not already defined
5004     in unistd.h and nanosleep() is not available; fixes error introduced
5005     in libpng-1.6.13.
5006   Disable floating point exception handling in pngvalid.c when
5007     PNG_FLOATING_ARITHMETIC is not supported (bug report by "zootus
5008     at users.sourceforge.net").
5010 Version 1.6.14beta03 [September 19, 2014]
5011   Define FE_DIVBYZERO, FE_INVALID, and FE_OVERFLOW in pngvalid.c if not
5012     already defined.  Revert floating point exception handling in pngvalid.c
5013     to version 1.6.14beta01 behavior.
5015 Version 1.6.14beta04 [September 27, 2014]
5016   Fixed incorrect handling of the iTXt compression flag in pngrutil.c
5017     (bug report by Shunsaku Hirata).  Bug was introduced in libpng-1.6.0.
5019 Version 1.6.14beta05 [October 1, 2014]
5020   Added "option READ_iCCP enables READ_COMPRESSED_TEXT" to pnglibconf.dfa
5022 Version 1.6.14beta06 [October 5, 2014]
5023   Removed unused "text_len" parameter from private function png_write_zTXt().
5024   Conditionally compile some code in png_deflate_claim(), when
5026   Replaced repeated code in pngpread.c with PNG_PUSH_SAVE_BUFFER_IF_FULL.
5027   Added "chunk iTXt enables TEXT" and "chunk zTXt enables TEXT"
5028     to pnglibconf.dfa.
5029   Removed "option READ_COMPRESSED_TEXT enables READ_TEXT" from pnglibconf.dfa,
5030     to make it possible to configure a libpng that supports iCCP but not TEXT.
5032 Version 1.6.14beta07 [October 7, 2014]
5033   Removed "option WRITE_COMPRESSED_TEXT enables WRITE_TEXT" from pnglibconf.dfa
5034   Only mark text chunks as written after successfully writing them.
5036 Version 1.6.14rc01 [October 15, 2014]
5037   Fixed some typos in comments.
5039 Version 1.6.14rc02 [October 17, 2014]
5040   Changed png_convert_to_rfc_1123() to png_convert_to_rfc_1123_buffer()
5041     in the manual, to reflect the change made in libpng-1.6.0.
5042   Updated README file to explain that direct access to the png_struct
5043     and info_struct members has not been permitted since libpng-1.5.0.
5045 Version 1.6.14 [October 23, 2014]
5046   No changes.
5048 Version 1.6.15beta01 [October 29, 2014]
5049   Changed "if (!x)" to "if (x == 0)" and "if (x)" to "if (x != 0)"
5050   Simplified png_free_data().
5051   Added missing "ptr = NULL" after some instances of png_free().
5053 Version 1.6.15beta02 [November 1, 2014]
5054   Changed remaining "if (!x)" to "if (x == 0)" and "if (x)" to "if (x != 0)"
5056 Version 1.6.15beta03 [November 3, 2014]
5057   Added PNG_USE_ARM_NEON configuration flag (Marcin Juszkiewicz).
5059 Version 1.6.15beta04 [November 4, 2014]
5060   Removed new PNG_USE_ARM_NEON configuration flag and made a one-line
5061     revision to configure.ac to support ARM on aarch64 instead (John Bowler).
5063 Version 1.6.15beta05 [November 5, 2014]
5064   Use png_get_libpng_ver(NULL) instead of PNG_LIBPNG_VER_STRING in
5065     example.c, pngtest.c, and applications in the contrib directory.
5066   Fixed an out-of-range read in png_user_version_check() (Bug report from
5067     Qixue Xiao, CVE-2015-8540).
5068   Simplified and future-proofed png_user_version_check().
5069   Fixed GCC unsigned int->float warnings. Various versions of GCC
5070     seem to generate warnings when an unsigned value is implicitly
5071     converted to double. This is probably a GCC bug but this change
5072     avoids the issue by explicitly converting to (int) where safe.
5073   Free all allocated memory in pngimage. The file buffer cache was left
5074     allocated at the end of the program, harmless but it causes memory
5075     leak reports from clang.
5076   Fixed array size calculations to avoid warnings. At various points
5077     in the code the number of elements in an array is calculated using
5078     sizeof.  This generates a compile time constant of type (size_t) which
5079     is then typically assigned to an (unsigned int) or (int). Some versions
5080     of GCC on 64-bit systems warn about the apparent narrowing, even though
5081     the same compiler does apparently generate the correct, in-range,
5082     numeric constant.  This adds appropriate, safe, casts to make the
5083     warnings go away.
5085 Version 1.6.15beta06 [November 6, 2014]
5086   Reverted use png_get_libpng_ver(NULL) instead of PNG_LIBPNG_VER_STRING
5087     in the manual, example.c, pngtest.c, and applications in the contrib
5088     directory.  It was incorrect advice.
5090 Version 1.6.15beta07 [November 7, 2014]
5091   Removed #ifdef PNG_16BIT_SUPPORTED/#endif around png_product2(); it is
5092     needed by png_reciprocal2().
5093   Added #ifdef PNG_16BIT_SUPPORTED/#endif around png_log16bit() and
5094     png_do_swap().
5095   Changed all "#endif /* PNG_FEATURE_SUPPORTED */" to "#endif /* FEATURE */"
5097 Version 1.6.15beta08 [November 8, 2014]
5098   More housecleaning in *.h
5100 Version 1.6.15rc01 [November 13, 2014]
5102 Version 1.6.15rc02 [November 14, 2014]
5103   The macros passed in the command line to Borland make were ignored if
5104     similarly-named macros were already defined in makefiles. This behavior
5105     is different from POSIX make and other make programs.  Surround the
5106     macro definitions with ifndef guards (Cosmin).
5108 Version 1.6.15rc03 [November 16, 2014]
5109   Added "-D_CRT_SECURE_NO_WARNINGS" to CFLAGS in scripts/makefile.vcwin32.
5110   Removed the obsolete $ARCH variable from scripts/makefile.darwin.
5112 Version 1.6.15 [November 20, 2014]
5113   No changes.
5115 Version 1.6.16beta01 [December 14, 2014]
5116   Added ".align 2" to arm/filter_neon.S to support old GAS assemblers that
5117     don't do alignment correctly.
5118   Revised Makefile.am and scripts/symbols.dfn to work with MinGW/MSYS
5119     (Bob Friesenhahn).
5121 Version 1.6.16beta02 [December 15, 2014]
5122   Revised Makefile.am and scripts/*.dfn again to work with MinGW/MSYS;
5123     renamed scripts/*.dfn to scripts/*.c (John Bowler).
5125 Version 1.6.16beta03 [December 21, 2014]
5126   Quiet a "comparison always true" warning in pngstest.c (John Bowler).
5128 Version 1.6.16rc01 [December 21, 2014]
5129   Restored a test on width that was removed from png.c at libpng-1.6.9
5130     (Bug report by Alex Eubanks, CVE-2015-0973).
5132 Version 1.6.16rc02 [December 21, 2014]
5133   Undid the update to pngrutil.c in 1.6.16rc01.
5135 Version 1.6.16rc03 [December 21, 2014]
5136   Fixed an overflow in png_combine_row() with very wide interlaced images
5137     (Bug report and fix by John Bowler, CVE-2014-9495).
5139 Version 1.6.16 [December 22, 2014]
5140   No changes.
5142 Version 1.6.17beta01 [January 29, 2015]
5143   Removed duplicate PNG_SAFE_LIMITS_SUPPORTED handling from pngconf.h
5144   Corrected the width limit calculation in png_check_IHDR().
5145   Removed user limits from pngfix. Also pass NULL pointers to
5146     png_read_row to skip the unnecessary row de-interlace stuff.
5147   Added testing of png_set_packing() to pngvalid.c
5148   Regenerated configure scripts in the *.tar distributions with libtool-2.4.4
5149   Implement previously untested cases of libpng transforms in pngvalid.c
5150   Fixed byte order in png_do_read_filler() with 16-bit input. Previously
5151     the high and low bytes of the filler, from png_set_filler() or from
5152     png_set_add_alpha(), were read in the wrong order.
5153   Made the check for out-of-range values in png_set_tRNS() detect
5154     values that are exactly 2^bit_depth, and work on 16-bit platforms.
5155   Merged some parts of libpng-1.6.17beta01 and libpng-1.7.0beta47.
5156   Added #ifndef __COVERITY__ where needed in png.c, pngrutil.c and
5157     pngset.c to avoid warnings about dead code.
5158   Added "& 0xff" to many instances of expressions that are typecast
5159     to (png_byte), to avoid Coverity warnings.
5161 Version 1.6.17beta02 [February 7, 2015]
5162   Work around one more Coverity-scan dead-code warning.
5163   Do not build png_product2() when it is unused.
5165 Version 1.6.17beta03 [February 17, 2015]
5166   Display user limits in the output from pngtest.
5167   Eliminated the PNG_SAFE_LIMITS macro and restored the 1-million-column
5168     and 1-million-row default limits in pnglibconf.dfa, that can be reset
5169     by the user at build time or run time.  This provides a more robust
5170     defense against DOS and as-yet undiscovered overflows.
5172 Version 1.6.17beta04 [February 21, 2015]
5173   Added PNG_WRITE_CUSTOMIZE_COMPRESSION_SUPPORTED macro, on by default.
5174   Allow user to call png_get_IHDR() with NULL arguments (Reuben Hawkins).
5175   Rebuilt configure scripts with automake-1.15 and libtool-2.4.6
5177 Version 1.6.17beta05 [February 25, 2015]
5178   Restored compiling of png_reciprocal2 with PNG_NO_16BIT.
5180 Version 1.6.17beta06 [February 27, 2015]
5181   Moved png_set_filter() prototype into a PNG_WRITE_SUPPORTED block
5182     of png.h.
5183   Avoid runtime checks when converting integer to png_byte with
5184     Visual Studio (Sergey Kosarevsky)
5186 Version 1.6.17rc01 [March 4, 2015]
5187   No changes.
5189 Version 1.6.17rc02 [March 9, 2015]
5190   Removed some comments that the configure script did not handle
5191     properly from scripts/pnglibconf.dfa and pnglibconf.h.prebuilt.
5192   Free the unknown_chunks structure even when it contains no data.
5194 Version 1.6.17rc03 [March 12, 2015]
5195   Updated CMakeLists.txt to add OSX framework, change YES/NO to ON/OFF
5196     for consistency, and remove some useless tests (Alexey Petruchik).
5198 Version 1.6.17rc04 [March 16, 2015]
5199   Remove pnglibconf.h, pnglibconf.c, and pnglibconf.out instead of
5200     pnglibconf.* in "make clean" (Cosmin).
5201   Fix bug in calculation of maxbits, in png_write_sBIT, introduced
5202     in libpng-1.6.17beta01 (John Bowler).
5204 Version 1.6.17rc05 [March 21, 2015]
5205   Define PNG_FILTER_* and PNG_FILTER_VALUE_* in png.h even when WRITE
5206     is not supported (John Bowler).  This fixes an error introduced in
5207     libpng-1.6.17beta06.
5208   Reverted "& 0xff" additions of version 1.6.17beta01. Libpng passes
5209     the Coverity scan without them.
5211 Version 1.6.17rc06 [March 23, 2015]
5212   Remove pnglibconf.dfn and pnglibconf.pre with "make clean".
5213   Reformatted some "&0xff" instances to "& 0xff".
5214   Fixed simplified 8-bit-linear to sRGB alpha. The calculated alpha
5215     value was wrong.  It's not clear if this affected the final stored
5216     value; in the obvious code path the upper and lower 8-bits of the
5217     alpha value were identical and the alpha was truncated to 8-bits
5218     rather than dividing by 257 (John Bowler).
5220 Version 1.6.17 [March 26, 2015]
5221   No changes.
5223 Version 1.6.18beta01 [April 1, 2015]
5225     have been combined with PNG_SET_USER_LIMITS_SUPPORTED (resolves
5226     bug report by Andrew Church).
5227   Fixed rgb_to_gray checks and added tRNS checks to pngvalid.c.  This
5228     fixes some arithmetic errors that caused some tests to fail on
5229     some 32-bit platforms (Bug reports by Peter Breitenlohner [i686]
5230     and Petr Gajdos [i586]).
5232 Version 1.6.18beta02 [April 26, 2015]
5233   Suppressed some warnings from the Borland C++ 5.5.1/5.82 compiler
5234     (Bug report by Viktor Szakats).
5236 Version 1.6.18beta03 [May 6, 2015]
5237   Replaced "unexpected" with an integer (0xabadca11) in pngset.c
5238     where a long was expected, to avoid a compiler warning when PNG_DEBUG > 1.
5239   Added contrib/examples/simpleover.c, to demonstrate how to handle
5240     alpha compositing of multiple images, using the "simplified API"
5241     and an example PNG generation tool, contrib/examples/genpng.c
5242     (John Bowler).
5244 Version 1.6.18beta04 [May 20, 2015]
5245   PNG_RELEASE_BUILD replaces tests where the code depended on the build base
5246     type and can be defined on the command line, allowing testing in beta
5247     builds (John Bowler).
5248   Avoid Coverity issue 80858 (REVERSE NULL) in pngtest.c PNG_DEBUG builds.
5249   Avoid a harmless potential integer overflow in png_XYZ_from_xy() (Bug
5250     report from Christopher Ferris).
5252 Version 1.6.18beta05 [May 31, 2015]
5253   Backport filter selection code from libpng-1.7.0beta51, to combine
5254     sub_row, up_row, avg_row, and paeth_row into try_row and tst_row.
5255   Changed png_voidcast(), etc., to voidcast(), etc., in contrib/tools/pngfix.c
5256     to avoid confusion with the libpng private macros.
5257   Fixed old cut&paste bug in the weighted filter selection code in
5258     pngwutil.c, introduced in libpng-0.95, March 1997.
5260 Version 1.6.18beta06 [June 1, 2015]
5261   Removed WRITE_WEIGHTED_FILTERED code, to save a few kbytes of the
5262     compiled library size. It never worked properly and as far as we can
5263     tell, no one uses it. The png_set_filter_heuristics() and
5264     png_set_filter_heuristics_fixed() APIs are retained but deprecated
5265     and do nothing.
5267 Version 1.6.18beta07 [June 6, 2015]
5268   Removed non-working progressive reader 'skip' function. This
5269     function has apparently never been used. It was implemented
5270     to support back-door modification of png_struct in libpng-1.4.x
5271     but (because it does nothing and cannot do anything) was apparently
5272     never tested (John Bowler).
5273   Fixed cexcept.h in which GCC 5 now reports that one of the auto
5274     variables in the Try macro needs to be volatile to prevent value
5275     being lost over the setjmp (John Bowler).
5276   Fixed NO_WRITE_FILTER and -Wconversion build breaks (John Bowler).
5277   Fix g++ build breaks (John Bowler).
5278   Quieted some Coverity issues in pngfix.c, png-fix-itxt.c, pngvalid.c,
5279     pngstest.c, and pngimage.c. Most seem harmless, but png-fix-itxt
5280     would only work with iTXt chunks with length 255 or less.
5281   Added #ifdef's to contrib/examples programs so people don't try
5282     to compile them without the minimum required support enabled
5283     (suggested by Flavio Medeiros).
5285 Version 1.6.18beta08 [June 30, 2015]
5286   Eliminated the final two Coverity defects (insecure temporary file
5287     handling in contrib/libtests/pngstest.c; possible overflow of
5288     unsigned char in contrib/tools/png-fix-itxt.c). To use the "secure"
5289     file handling, define PNG_USE_MKSTEMP, otherwise "tmpfile()" will
5290     be used.
5291   Removed some unused WEIGHTED_FILTER macros from png.h and pngstruct.h
5293 Version 1.6.18beta09 [July 5, 2015]
5294   Removed some useless typecasts from contrib/tools/png-fix-itxt.c
5295   Fixed a new signed-unsigned comparison in pngrtran.c (Max Stepin).
5296   Replaced arbitrary use of 'extern' with #define PNG_LINKAGE_*.  To
5297     preserve API compatibility, the new defines all default to "extern"
5298     (requested by Jan Nijtmans).
5300 Version 1.6.18rc01 [July 9, 2015]
5301   Belatedly added Mans Rullgard and James Yu to the list of Contributing
5302     Authors.
5304 Version 1.6.18rc02 [July 12, 2015]
5305   Restored unused FILTER_HEURISTIC macros removed at libpng-1.6.18beta08
5306     to png.h to avoid compatibility warnings.
5308 Version 1.6.18rc03 [July 15, 2015]
5309   Minor changes to the man page
5311 Version 1.6.18 [July 23, 2015]
5312   No changes.
5314 Version 1.6.19beta01 [July 30, 2015]
5315   Updated obsolete information about the simplified API macros in the
5316     manual pages (Bug report by Arc Riley).
5317   Avoid potentially dereferencing NULL info_ptr in png_info_init_3().
5318   Rearranged png.h to put the major sections in the same order as
5319     in libpng17.
5321     PNG_WEIGHT_FACTOR macros.
5322   Suppressed some warnings from the Borland C++ 5.5.1/5.82 compiler
5323     (Bug report by Viktor Szakats).  Several warnings remain and are
5324     unavoidable, where we test for overflow.
5325   Fixed potential leak of png_pixels in contrib/pngminus/pnm2png.c
5326   Fixed uninitialized variable in contrib/gregbook/rpng2-x.c
5328 Version 1.6.19beta02 [August 19, 2015]
5329   Moved config.h.in~ from the "libpng_autotools_files" list to the
5330     "libpng_autotools_extra" list in autogen.sh because it was causing a
5331     false positive for missing files (bug report by Robert C. Seacord).
5332   Removed unreachable "break" statements in png.c, pngread.c, and pngrtran.c
5333     to suppress clang warnings (Bug report by Viktor Szakats).
5334   Fixed some bad links in the man page.
5335   Changed "n bit" to "n-bit" in comments.
5336   Added signed/unsigned 16-bit safety net. This removes the dubious
5337     0x8000 flag definitions on 16-bit systems. They aren't supported
5338     yet the defs *probably* work, however it seems much safer to do this
5339     and be advised if anyone, contrary to advice, is building libpng 1.6
5340     on a 16-bit system. It also adds back various switch default clauses
5341     for GCC; GCC errors out if they are not present (with an appropriately
5342     high level of warnings).
5343   Safely convert num_bytes to a png_byte in png_set_sig_bytes() (Robert
5344     Seacord).
5345   Fixed the recently reported 1's complement security issue by replacing
5346     the value that is illegal in the PNG spec, in both signed and unsigned
5347     values, with 0. Illegal unsigned values (anything greater than or equal
5348     to  0x80000000) can still pass through, but since these are not illegal
5349     in ANSI-C (unlike 0x80000000 in the signed case) the checking that
5350     occurs later can catch them (John Bowler).
5352 Version 1.6.19beta03 [September 26, 2015]
5353   Fixed png_save_int_32 when int is not 2's complement (John Bowler).
5354   Updated libpng16 with all the recent test changes from libpng17,
5355     including changes to pngvalid.c to ensure that the original,
5356     distributed, version of contrib/visupng/cexcept.h can be used
5357     (John Bowler).
5358   pngvalid contains the correction to the use of SAVE/STORE_
5359     UNKNOWN_CHUNKS; a bug revealed by changes in libpng 1.7. More
5360     tests contain the --strict option to detect warnings and the
5361     pngvalid-standard test has been corrected so that it does not
5362     turn on progressive-read. There is a separate test which does
5363     that. (John Bowler)
5364   Also made some signed/unsigned fixes.
5365   Make pngstest error limits version specific. Splitting the machine
5366     generated error structs out to a file allows the values to be updated
5367     without changing pngstest.c itself. Since libpng 1.6 and 1.7 have
5368     slightly different error limits this simplifies maintenance. The
5369     makepngs.sh script has also been updated to more accurately reflect
5370     current problems in libpng 1.7 (John Bowler).
5371   Incorporated new test PNG files into make check.  tests/pngstest-*
5372     are changed so that the new test files are divided into 8 groups by
5373     gamma and alpha channel.  These tests have considerably better code
5374     and pixel-value coverage than contrib/pngsuite; however,coverage is
5375     still incomplete (John Bowler).
5376   Removed the '--strict' in 1.6 because of the double-gamma-correction
5377     warning, updated pngstest-errors.h for the errors detected with the
5378     new contrib/testspngs PNG test files (John Bowler).
5380 Version 1.6.19beta04 [October 15, 2015]
5381   Worked around rgb-to-gray issues in libpng 1.6.  The previous
5382     attempts to ignore the errors in the code aren't quite enough to
5383     deal with the 'channel selection' encoding added to libpng 1.7; abort.
5384     pngvalid.c is changed to drop this encoding in prior versions.
5385   Fixed 'pow' macros in pngvalid.c. It is legal for 'pow' to be a
5386     macro, therefore the argument list cannot contain preprocessing
5387     directives.  Make sure pow is a function where this happens. This is
5388     a minimal safe fix, the issue only arises in non-performance-critical
5389     code (bug report by Curtis Leach, fix by John Bowler).
5390   Added sPLT support to pngtest.c
5392 Version 1.6.19rc01 [October 23, 2015]
5393   No changes.
5395 Version 1.6.19rc02 [October 31, 2015]
5396   Prevent setting or writing over-length PLTE chunk (Cosmin Truta).
5397   Silently truncate over-length PLTE chunk while reading.
5398   Libpng incorrectly calculated the output rowbytes when the application
5399     decreased either the number of channels or the bit depth (or both) in
5400     a user transform.  This was safe; libpng overallocated buffer space
5401    (potentially by quite a lot; up to 4 times the amount required) but,
5402    from 1.5.4 on, resulted in a png_error (John Bowler).
5404 Version 1.6.19rc03 [November 3, 2015]
5405   Fixed some inconsequential cut-and-paste typos in png_set_cHRM_XYZ_fixed().
5406   Clarified COPYRIGHT information to state explicitly that versions
5407     are derived from previous versions.
5408   Removed much of the long list of previous versions from png.h and
5409     libpng.3.
5411 Version 1.6.19rc04 [November 5, 2015]
5412   Fixed new bug with CRC error after reading an over-length palette
5413     (bug report by Cosmin Truta) (CVE-2015-8126).
5415 Version 1.6.19 [November 12, 2015]
5416   Cleaned up coding style in png_handle_PLTE().
5418 Version 1.6.20beta01 [November 20, 2015]
5419   Avoid potential pointer overflow/underflow in png_handle_sPLT() and
5420     png_handle_pCAL() (Bug report by John Regehr).
5422 Version 1.6.20beta02 [November 23, 2015]
5423   Fixed incorrect implementation of png_set_PLTE() that uses png_ptr
5424     not info_ptr, that left png_set_PLTE() open to the CVE-2015-8126
5425     vulnerability.  Fixes CVE-2015-8472.
5427 Version 1.6.20beta03 [November 24, 2015]
5428   Backported tests from libpng-1.7.0beta69.
5430 Version 1.6.20rc01 [November 26, 2015]
5431   Fixed an error in handling of bad zlib CMINFO field in pngfix, found by
5432     American Fuzzy Lop, reported by Brian Carpenter.  inflate() doesn't
5433     immediately fault a bad CMINFO field; instead a 'too far back' error
5434     happens later (at least some times).  pngfix failed to limit CMINFO to
5435     the allowed values but then assumed that window_bits was in range,
5436     triggering an assert. The bug is mostly harmless; the PNG file cannot
5437     be fixed.
5439 Version 1.6.20rc02 [November 29, 2015]
5440   In libpng 1.6 zlib initialization was changed to use the window size
5441     in the zlib stream, not a fixed value. This causes some invalid images,
5442     where CINFO is too large, to display 'correctly' if the rest of the
5443     data is valid.  This provides a workaround for zlib versions where the
5444     error arises (ones that support the API change to use the window size
5445     in the stream).
5447 Version 1.6.20 [December 3, 2015]
5448   No changes.
5450 Version 1.6.21beta01 [December 11, 2015]
5451   Fixed syntax "$(command)" in tests/pngstest that some shells other than
5452     bash could not parse (Bug report by Nelson Beebe). Use `command` instead.
5454 Version 1.6.21beta02 [December 14, 2015]
5455   Moved png_check_keyword() from pngwutil.c to pngset.c
5456   Removed LE/BE dependencies in pngvalid, to 'fix' the current problem
5457     in the BigEndian tests by not testing it, making the BE code the same 
5458     as the LE version.
5459   Fixes to pngvalid for various reduced build configurations (eliminate unused
5460     statics) and a fix for the case in rgb_to_gray when the digitize option
5461     reduces graylo to 0, producing a large error.
5463 Version 1.6.21beta03 [December 18, 2015]
5464   Widened the 'limit' check on the internally calculated error limits in
5465     the 'DIGITIZE' case (the code used prior to 1.7 for rgb_to_gray error
5466     checks) and changed the check to only operate in non-release builds
5467     (base build type not RC or RELEASE.)
5468   Fixed undefined behavior in pngvalid.c, undefined because
5469     (png_byte) << shift is undefined if it changes the signed bit
5470     (because png_byte is promoted to int). The libpng exported functions
5471     png_get_uint_32 and png_get_uint_16 handle this. (Bug reported by
5472     David Drysdale as a result of reports from UBSAN in clang 3.8).
5473   This changes pngvalid to use BE random numbers; this used to produce
5474     errors but these should not be fixed as a result of the previous changes.
5476 Version 1.6.21rc01 [January 4, 2016]
5477   In projects/vstudio, combined readme.txt and WARNING into README.txt
5479 Version 1.6.21rc02 [January 7, 2016]
5480   Relocated assert() in contrib/tools/pngfix.c, bug found by American
5481     Fuzzy Lop, reported by Brian Carpenter.
5482   Marked 'limit' UNUSED in transform_range_check().  This only affects
5483     release builds.
5485 Version 1.6.21 [January 15, 2016]
5486   Worked around a false-positive Coverity issue in pngvalid.c.
5488 Version 1.6.22beta01 [January 23, 2016]
5489   Changed PNG_USE_MKSTEMP to __COVERITY__ to select alternate
5490     "tmpfile()" implementation in contrib/libtests/pngstest.c
5491   Fixed NO_STDIO build of pngunknown.c to skip calling png_init_io()
5492     if there is no stdio.h support.
5493   Added a png_image_write_to_memory() API and a number of assist macros
5494     to allow an application that uses the simplified API write to bypass
5495     stdio and write directly to memory.
5496   Added some warnings (png.h) and some check code to detect *possible*
5497     overflow in the ROW_STRIDE and simplified image SIZE macros.  This
5498     disallows image width/height/format that *might* overflow.  This is
5499     a quiet API change that limits in-memory image size (uncompressed) to
5500     less than 4GByte and image row size (stride) to less than 2GByte.
5501   Revised workaround for false-positive Coverity issue in pngvalid.c.
5503 Version 1.6.22beta02 [February 8, 2016]
5504   Only use exit(77) in configure builds.
5505   Corrected error in PNG_IMAGE_PNG_SIZE_MAX. This new macro underreported
5506     the palette size because it failed to take into account that the memory
5507     palette has to be expanded to full RGB when it is written to PNG.
5508   Updated CMakeLists.txt, added supporting scripts/gen*.cmake.in
5509     and test.cmake.in (Roger Leigh).
5510   Relaxed limit checks on gamma values in pngrtran.c. As suggested in
5511     the comments gamma values outside the range currently permitted
5512     by png_set_alpha_mode are useful for HDR data encoding.  These values
5513     are already permitted by png_set_gamma so it is reasonable caution to
5514     extend the png_set_alpha_mode range as HDR imaging systems are starting
5515     to emerge.
5517 Version 1.6.22beta03 [March 9, 2016]
5518   Added a common-law trademark notice and export control information
5519     to the LICENSE file, png.h, and the man page.
5520   Restored "& 0xff" in png_save_uint_16() and png_save_uint_32() that
5521     were accidentally removed from libpng-1.6.17. 
5522   Changed PNG_INFO_cHNK and PNG_FREE_cHNK from 0xnnnn to 0xnnnnU in png.h
5523     (Robert C. Seacord).
5524   Removed dubious "#if INT_MAX" test from png.h that was added to
5525     libpng-1.6.19beta02 (John Bowler).
5526   Add ${INCLUDES} in scripts/genout.cmake.in (Bug report by Nixon Kwok).
5527   Updated LICENSE to say files in the contrib directory are not
5528     necessarily under the libpng license, and that some makefiles have
5529     other copyright owners.
5530   Added INTEL-SSE2 support (Mike Klein and Matt Sarett, Google, Inc.).
5531   Made contrib/libtests/timepng more robust.  The code no longer gives
5532     up/fails on invalid PNG data, it just skips it (with error messages).
5533     The code no longer fails on PNG files with data beyond IEND.  Options
5534     exist to use png_read_png (reading the whole image, not by row) and, in
5535     that case, to apply any of the supported transforms.  This makes for
5536     more realistic testing; the decoded data actually gets used in a
5537     meaningful fashion (John Bowler).
5538   Fixed some misleading indentation (Krishnaraj Bhat).
5540 Version 1.6.22beta04 [April 5, 2016]
5541   Force GCC compilation to C89 if needed (Dagobert Michelsen).
5542   SSE filter speed improvements for bpp=3:
5543     memcpy-free implementations of load3() / store3().
5544     call load3() only when needed at the end of a scanline.
5546 Version 1.6.22beta05 [April 27, 2016]
5547   Added PNG_FAST_FILTERS macro (defined as
5549   Various fixes for contrib/libtests/timepng.c
5550   Moved INTEL-SSE code from pngpriv.h into contrib/intel/intel_sse.patch.
5551   Fixed typo (missing underscore) in #define PNG_READ_16_TO_8_SUPPORTED
5552     (Bug report by Y.Ohashik).
5554 Version 1.6.22beta06 [May 5, 2016]
5555   Rebased contrib/intel_sse.patch.
5556   Quieted two Coverity issues in contrib/libtests/timepng.c.
5557   Fixed issues with scripts/genout.cmake.in (David Capello, Nixon Kwok):
5558     Added support to use multiple directories in ZLIBINCDIR variable,
5559     Fixed CMAKE_C_FLAGS with multiple values when genout is compiled on MSVC,
5560     Fixed pnglibconf.c compilation on OS X including the sysroot path.
5562 Version 1.6.22rc01 [May 14, 2016]
5563   No changes.
5565 Version 1.6.22rc02 [May 16, 2016]
5566   Removed contrib/timepng from default build; it does not build on platforms
5567     that don't supply clock_gettime().
5569 Version 1.6.22rc03 [May 17, 2016]
5570   Restored contrib/timepng to default build but check for the presence
5571     of clock_gettime() in configure.ac and Makefile.am.
5573 Version 1.6.22 [May 26, 2016]
5574   No changes.
5576 Version 1.6.23beta01 [May 29, 2016]
5577   Stop a potential memory leak in png_set_tRNS() (Bug report by Ted Ying).
5578   Fixed the progressive reader to handle empty first IDAT chunk properly
5579     (patch by Timothy Nikkel).  This bug was introduced in libpng-1.6.0 and
5580     only affected the libpng16 branch.
5581   Added tests in pngvalid.c to check zero-length IDAT chunks in various
5582     positions.  Fixed the sequential reader to handle these more robustly
5583     (John Bowler).
5585 Version 1.6.23rc01 [June 2, 2016]
5586   Corrected progressive read input buffer in pngvalid.c. The previous version
5587     the code invariably passed just one byte at a time to libpng.  The intent
5588     was to pass a random number of bytes in the range 0..511.
5589   Moved sse2 prototype from pngpriv.h to contrib/intel/intel_sse.patch.
5590   Added missing ")" in pngerror.c (Matt Sarrett).
5592 Version 1.6.23rc02 [June 4, 2016]
5593   Fixed undefined behavior in png_push_save_buffer(). Do not call
5594     memcpy() with a null source, even if count is zero (Leon Scroggins III).
5596 Version 1.6.23 [June 9, 2016]
5597   Fixed bad link to RFC2083 in png.5 (Nikola Forro).
5599 Version 1.6.24beta01 [June 11, 2016]
5600   Avoid potential overflow of the PNG_IMAGE_SIZE macro.  This macro
5601     is not used within libpng, but is used in some of the examples.
5603 Version 1.6.24beta02 [June 23, 2016]
5604   Correct filter heuristic overflow handling. This was broken when the
5605     write filter code was moved out-of-line; if there is a single filter and
5606     the heuristic sum overflows the calculation of the filtered line is not
5607     completed.  In versions prior to 1.6 the code was duplicated in-line
5608     and the check not performed, so the filter operation completed; however,
5609     in the multi-filter case where the sum is performed the 'none' filter would
5610     be selected if all the sums overflowed, even if it wasn't in the filter
5611     list.  The fix to the first problem is simply to provide PNG_SIZE_MAX as
5612     the current lmins sum value; this means the sum can never exceed it and
5613     overflows silently.  A reasonable compiler that does choose to inline
5614     the code will simply eliminate the sum check.
5615   The fix to the second problem is to use high precision arithmetic (this is
5616     implemented in 1.7), however a simple safe fix here is to chose the lowest
5617     numbered filter in the list from png_set_filter (this only works if the
5618     first problem is also fixed) (John Bowler).
5619   Use a more efficient absolute value calculation on SSE2 (Matthieu Darbois).
5620   Fixed the case where PNG_IMAGE_BUFFER_SIZE can overflow in the application
5621     as a result of the application using an increased 'row_stride'; previously
5622     png_image_finish_read only checked for overflow on the base calculation of
5623     components.  (I.e. it checked for overflow of a 32-bit number on the total
5624     number of pixel components in the output format, not the possibly padded row
5625     length and not the number of bytes, which for linear formats is twice the
5626     number of components.)
5627   MSVC does not like '-(unsigned)', so replaced it with 0U-(unsigned)
5628   MSVC does not like (uInt) = -(unsigned) (i.e. as an initializer), unless
5629     the conversion is explicitly invoked by a cast.
5630   Put the SKIP definition in the correct place. It needs to come after the
5631     png.h include (see all the other .c files in contrib/libtests) because it
5632     depends on PNG_LIBPNG_VER.
5633   Removed the three compile warning options from the individual project
5634     files into the zlib.props globals.  It increases the warning level from 4
5635     to All and adds a list of the warnings that need to be turned off.  This is
5636     semi-documentary; the intent is to tell libpng users which warnings have
5637     been examined and judged non-fixable at present.  The warning about
5638     structure padding is fixable, but it would be a signficant change (moving
5639     structure members around).
5641 Version 1.6.24beta03 [July 4, 2016]
5642   Optimized absolute value calculation in filter selection, similar to
5643     code in the PAETH decoder in pngrutil.c. Build with PNG_USE_ABS to
5644     use this.
5645   Added pngcp to the build together with a pngcp.dfa configuration test.
5646   Added high resolution timing to pngcp.
5647   Added "Common linking failures" section to INSTALL.
5648   Relocated misplaced #endif in png.c sRGB profile checking.
5649   Fixed two Coverity issues in pngcp.c.
5651 Version 1.6.24beta04 [July 8, 2016]
5652   Avoid filter-selection heuristic sum calculations in cases where only one
5653     filter is a candidate for selection. This trades off code size (added
5654     private png_setup_*_row_only() functions) for speed.
5656 Version 1.6.24beta05 [July 13, 2016]
5657   Fixed some indentation to comply with our coding style.
5658   Added contrib/tools/reindent.
5660 Version 1.6.24beta06 [July 18, 2016]
5661   Fixed more indentation to comply with our coding style.
5662   Eliminated unnecessary tests of boolean png_isaligned() vs 0.
5664 Version 1.6.24rc01 [July 25, 2016]
5665   No changes.
5667 Version 1.6.24rc02 [August 1, 2016]
5668   Conditionally compile SSE2 headers in contrib/intel/intel_sse.patch
5669   Conditionally compile png_decompress_chunk().
5671 Version 1.6.24rc03 [August 2, 2016]
5672   Conditionally compile ARM_NEON headers in pngpriv.h
5673   Updated contrib/intel/intel_sse.patch
5675 Version 1.6.24[August 4, 2016]
5676   No changes.
5678 Version 1.6.25beta01 [August 12, 2016]
5679   Reject oversized iCCP profile immediately.
5680   Cleaned up PNG_DEBUG compile of pngtest.c.
5681   Conditionally compile png_inflate().
5683 Version 1.6.25beta02 [August 18, 2016]
5684   Don't install pngcp; it conflicts with pngcp in the pngtools package.
5685   Minor editing of INSTALL, (whitespace, added copyright line)
5687 Version 1.6.25rc01 [August 24, 2016]
5688   No changes.
5690 Version 1.6.25rc02 [August 29, 2016]
5691   Added MIPS support (Mandar Sahastrabuddhe <Mandar.Sahastrabuddhe@imgtec.com>).
5692   Only the UP filter is currently implemented.
5694 Version 1.6.25rc03 [August 29, 2016]
5695   Rebased contrib/intel/intel_sse.patch after the MIPS implementation.
5697 Version 1.6.25rc04 [August 30, 2016]
5698   Added MIPS support for SUB, AVG, and PAETH filters (Mandar Sahastrabuddhe).
5700 Version 1.6.25rc05 [August 30, 2016]
5701   Rebased contrib/intel/intel_sse.patch after the MIPS implementation update..
5703 Version 1.6.25 [September 1, 2016]
5704   No changes.
5706 Version 1.6.26beta01 [September 26, 2016]
5707   Fixed handling zero length IDAT in pngfix (bug report by Agostino Sarubbo,
5708     bugfix by John Bowler).
5709   Do not issue a png_error() on read in png_set_pCAL() because png_handle_pCAL
5710     has allocated memory that libpng needs to free.
5711   Conditionally compile png_set_benign_errors() in pngread.c and pngtest.c
5712   Issue a png_benign_error instead of a png_error on ADLER32 mismatch
5713     while decoding compressed data chunks.
5714   Changed PNG_ZLIB_VERNUM to ZLIB_VERNUM in pngpriv.h, pngstruct.h, and
5715     pngrutil.c.
5716   If CRC handling of critical chunks has been set to PNG_CRC_QUIET_USE,
5717     ignore the ADLER32 checksum in the IDAT chunk as well as the chunk CRCs.
5718   Issue png_benign_error() on ADLER32 checksum mismatch instead of png_error().
5719   Add tests/badcrc.png and tests/badadler.png to tests/pngtest.
5720   Merged pngtest.c with libpng-1.7.0beta84/pngtest.c
5722 Version 1.6.26beta02 [October 1, 2016]
5723   Updated the documentation about CRC and ADLER32 handling.
5724   Quieted 117 warnings from clang-3.8 in pngtrans.c, pngread.c,
5725      pngwrite.c, pngunknown.c, and pngvalid.c.
5726   Quieted 58 (out of 144) -Wconversion compiler warnings by changing
5727     flag definitions in pngpriv.h from 0xnnnn to 0xnnnnU and trivial changes
5728     in png.c, pngread.c, and pngwutil.c.
5730 Version 1.6.26beta03 [October 2, 2016]
5731   Removed contrib/libtests/*.orig and *.rej that slipped into the tarballs.
5732   Quieted the 86 remaining -Wconversion compiler warnings by
5733     revising the png_isaligned() macro and trivial changes in png.c,
5734     pngerror.c, pngget.c, pngmem.c, pngset.c, pngrtran.c, pngrutil.c,
5735     pngwtran.c, pngwrite.c, and pngwutil.c.
5737 Version 1.6.26beta04 [October 3, 2016]
5738   Quieted (bogus?) clang warnings about "absolute value has no effect"
5739     when PNG_USE_ABS is defined.
5740   Fixed offsets in contrib/intel/intel_sse.patch
5742 Version 1.6.26beta05 [October 6, 2016]
5743   Changed integer constant 4294967294 to unsigned 4294967294U in pngconf.h
5744     to avoid a signed/unsigned compare in the preprocessor.
5746 Version 1.6.26beta06 [October 7, 2016]
5747   Use zlib- inflateValidate() instead of inflateReset2() to
5748     optionally avoid ADLER32 evaluation.
5750 Version 1.6.26rc01 [October 12, 2016]
5751   No changes.
5753 Version 1.6.26 [October 20, 2016]
5754   Cosmetic change, "ptr != 0" to "ptr != NULL" in png.c and pngrutil.c
5755   Despammed email addresses (replaced "@" with " at ").
5757 Version 1.6.27beta01 [November 2, 2016]
5758   Restrict the new ADLER32-skipping to IDAT chunks.  It broke iCCP chunk
5759     handling: an erroneous iCCP chunk would throw a png_error and reject the
5760     entire PNG image instead of rejecting just the iCCP chunk with a warning,
5761     if built with zlib-
5763 Version 1.6.27rc01 [December 27, 2016]
5764   Control ADLER32 checking with new PNG_IGNORE_ADLER32 option.
5765   Removed the use of a macro containing the pre-processor 'defined'
5766     operator.  It is unclear whether this is valid; a macro that
5767     "generates" 'defined' is not permitted, but the use of the word
5768     "generates" within the C90 standard seems to imply more than simple
5769     substitution of an expression itself containing a well-formed defined
5770     operation.
5771   Added ARM support to CMakeLists.txt (Andreas Franek).
5773 Version 1.6.27 [December 29, 2016]
5774   Fixed a potential null pointer dereference in png_set_text_2() (bug report
5775     and patch by Patrick Keshishian, CVE-2016-10087).
5777 Version 1.6.28rc01 [January 3, 2017]
5778   Fixed arm/aarch64 detection in CMakeLists.txt (Gianfranco Costamagna).
5779   Added option to Cmake build allowing a custom location of zlib to be
5780     specified in a scenario where libpng is being built as a subproject
5781     alongside zlib by another project (Sam Serrels).
5782   Changed png_ptr->options from a png_byte to png_uint_32, to accomodate
5783     up to 16 options.
5785 Version 1.6.28rc02 [January 4, 2017]
5786   Added "include(GNUInstallDirs)" to CMakeLists.txt (Gianfranco Costamagna).
5787   Moved SSE2 optimization code into the main libpng source directory.
5788     Configure libpng with "configure --enable-intel-sse" or compile
5789     libpng with "-DPNG_INTEL_SSE" in CPPFLAGS to enable it.
5791 Version 1.6.28rc03 [January 4, 2017]
5792   Backed out the SSE optimization and last CMakeLists.txt to allow time for QA.
5794 Version 1.6.28 [January 5, 2017]
5795   No changes.
5797 Send comments/corrections/commendations to png-mng-implement at lists.sf.net
5798 (subscription required; visit
5799 https://lists.sourceforge.net/lists/listinfo/png-mng-implement
5800 to subscribe)
5801 or to glennrp at users.sourceforge.net
5803 Glenn R-P
5804 #endif