.-.-.-.-.-

PKJM�\����CHANGESnu�[���#if 0
CHANGES - changes for libpng

version 0.1 [March 29, 1995]
  initial work-in-progress release

version 0.2 [April 1, 1995]
  added reader into png.h
  fixed small problems in stub file

version 0.3 [April 8, 1995]
  added pull reader
  split up pngwrite.c to several files
  added pnglib.txt
  added example.c
  cleaned up writer, adding a few new transformations
  fixed some bugs in writer
  interfaced with zlib 0.5
  added K&R support
  added check for 64 KB blocks for 16 bit machines

version 0.4 [April 26, 1995]
  cleaned up code and commented code
  simplified time handling into png_time
  created png_color_16 and png_color_8 to handle color needs
  cleaned up color type defines
  fixed various bugs
  made various names more consistent
  interfaced with zlib 0.71
  cleaned up zTXt reader and writer (using zlib's Reset functions)
  split transformations into pngrtran.c and pngwtran.c

version 0.5 [April 30, 1995]
  interfaced with zlib 0.8
  fixed many reading and writing bugs
  saved using 3 spaces instead of tabs

version 0.6 [May 1, 1995]
  first beta release
  added png_large_malloc() and png_large_free()
  added png_size_t
  cleaned up some compiler warnings
  added png_start_read_image()

version 0.7 [June 24, 1995]
  cleaned up lots of bugs
  finished dithering and other stuff
  added test program
  changed name from pnglib to libpng

version 0.71 [June 26, 1995]
  changed pngtest.png for zlib 0.93
  fixed error in libpng.txt and example.c

version 0.8 [August 20, 1995]
  cleaned up some bugs
  added png_set_filler()
  split up pngstub.c into pngmem.c, pngio.c, and pngerror.c
  added #define's to remove unwanted code
  moved png_info_init() to png.c
  added old_size into png_realloc()
  added functions to manually set filtering and compression info
  changed compression parameters based on image type
  optimized filter selection code
  added version info
  changed external functions passing floats to doubles (k&r problems?)
  put all the configurable stuff in pngconf.h
  enabled png_set_shift to work with paletted images on read
  added png_read_update_info() - updates info structure with transformations

Version 0.81 [August, 1995]
  incorporated Tim Wegner's medium model code (thanks, Tim)

Version 0.82 [September, 1995]
  [unspecified changes]

Version 0.85 [December, 1995]
  added more medium model code (almost everything's a far)
  added i/o, error, and memory callback functions
  fixed some bugs (16-bit, 4-bit interlaced, etc.)
  added first run progressive reader (barely tested)

Version 0.86 [January, 1996]
  fixed bugs
  improved documentation

Version 0.87 [January, 1996]
  fixed medium model bugs
  fixed other bugs introduced in 0.85 and 0.86
  added some minor documentation

Version 0.88 [January, 1996]
  fixed progressive bugs
  replaced tabs with spaces
  cleaned up documentation
  added callbacks for read/write and warning/error functions

Version 0.89 [June 5, 1996]
  Added new initialization API to make libpng work better with shared libs
    we now have png_create_read_struct(), png_create_write_struct(),
    png_create_info_struct(), png_destroy_read_struct(), and
    png_destroy_write_struct() instead of the separate calls to
    malloc and png_read_init(), png_info_init(), and png_write_init()
  Changed warning/error callback functions to fix bug - this means you
    should use the new initialization API if you were using the old
    png_set_message_fn() calls, and that the old API no longer exists
    so that people are aware that they need to change their code
  Changed filter selection API to allow selection of multiple filters
    since it didn't work in previous versions of libpng anyways
  Optimized filter selection code
  Fixed png_set_background() to allow using an arbitrary RGB color for
    paletted images
  Fixed gamma and background correction for paletted images, so
    png_correct_palette is not needed unless you are correcting an
    external palette (you will need to #define PNG_CORRECT_PALETTE_SUPPORTED
    in pngconf.h) - if nobody uses this, it may disappear in the future.
  Fixed bug with Borland 64K memory allocation (Alexander Lehmann)
  Fixed bug in interlace handling (Smarasderagd, I think)
  Added more error checking for writing and image to reduce invalid files
  Separated read and write functions so that they won't both be linked
    into a binary when only reading or writing functionality is used
  New pngtest image also has interlacing and zTXt
  Updated documentation to reflect new API

Version 0.89c [June 17, 1996]
  Bug fixes.

Version 0.90 [January, 1997]
  Made CRC errors/warnings on critical and ancillary chunks configurable
  libpng will use the zlib CRC routines by (compile-time) default
  Changed DOS small/medium model memory support - needs zlib 1.04 (Tim Wegner)
  Added external C++ wrapper statements to png.h (Gilles Dauphin)
  Allow PNG file to be read when some or all of file signature has already
    been read from the beginning of the stream.  ****This affects the size
    of info_struct and invalidates all programs that use a shared libpng****
  Fixed png_filler() declarations
  Fixed? background color conversions
  Fixed order of error function pointers to match documentation
  Current chunk name is now available in png_struct to reduce the number
    of nearly identical error messages (will simplify multi-lingual
    support when available)
  Try to get ready for unknown-chunk callback functions:
    - previously read critical chunks are flagged, so the chunk handling
      routines can determine if the chunk is in the right place
    - all chunk handling routines have the same prototypes, so we will
      be able to handle all chunks via a callback mechanism
  Try to fix Linux "setjmp" buffer size problems
  Removed png_large_malloc, png_large_free, and png_realloc functions.

Version 0.95 [March, 1997]
  Fixed bug in pngwutil.c allocating "up_row" twice and "avg_row" never
  Fixed bug in PNG file signature compares when start != 0
  Changed parameter type of png_set_filler(...filler...) from png_byte
    to png_uint_32
  Added test for MACOS to ensure that both math.h and fp.h are not #included
  Added macros for libpng to be compiled as a Windows DLL (Andreas Kupries)
  Added "packswap" transformation, which changes the endianness of
    packed-pixel bytes (Kevin Bracey)
  Added "strip_alpha" transformation, which removes the alpha channel of
    input images without using it (not necessarily a good idea)
  Added "swap_alpha" transformation, which puts the alpha channel in front
    of the color bytes instead of after
  Removed all implicit variable tests which assume NULL == 0 (I think)
  Changed several variables to "png_size_t" to show 16/32-bit limitations
  Added new pCAL chunk read/write support
  Added experimental filter selection weighting (Greg Roelofs)
  Removed old png_set_rgbx() and png_set_xrgb() functions that have been
    obsolete for about 2 years now (use png_set_filler() instead)
  Added macros to read 16- and 32-bit ints directly from buffer, to be
    used only on those systems that support it (namely PowerPC and 680x0)
    With some testing, this may become the default for MACOS/PPC systems.
  Only calculate CRC on data if we are going to use it
  Added macros for zTXt compression type PNG_zTXt_COMPRESSION_???
  Added macros for simple libpng debugging output selectable at compile time
  Removed PNG_READ_END_MODE in progressive reader (Smarasderagd)
  More description of info_struct in libpng.txt and png.h
  More instructions in example.c
  More chunk types tested in pngtest.c
  Renamed pngrcb.c to pngset.c, and all png_read_<chunk> functions to be
    png_set_<chunk>.  We now have corresponding png_get_<chunk>
    functions in pngget.c to get information in info_ptr.  This isolates
    the application from the internal organization of png_info_struct
    (good for shared library implementations).

Version 0.96 [May, 1997]
  Fixed serious bug with < 8bpp images introduced in 0.95
  Fixed 256-color transparency bug (Greg Roelofs)
  Fixed up documentation (Greg Roelofs, Laszlo Nyul)
  Fixed "error" in pngconf.h for Linux setjmp() behavior
  Fixed DOS medium model support (Tim Wegner)
  Fixed png_check_keyword() for case with error in static string text
  Added read of CRC after IEND chunk for embedded PNGs (Laszlo Nyul)
  Added typecasts to quiet compiler errors
  Added more debugging info

Version 0.97 [January, 1998]
  Removed PNG_USE_OWN_CRC capability
  Relocated png_set_crc_action from pngrutil.c to pngrtran.c
  Fixed typecasts of "new_key", etc. (Andreas Dilger)
  Added RFC 1152 [sic] date support
  Fixed bug in gamma handling of 4-bit grayscale
  Added 2-bit grayscale gamma handling (Glenn R-P)
  Added more typecasts. 65536L becomes (png_uint_32)65536L, etc. (Glenn R-P)
  Minor corrections in libpng.txt
  Added simple sRGB support (Glenn R-P)
  Easier conditional compiling, e.g.,
    define PNG_READ/WRITE_NOT_FULLY_SUPPORTED;
    all configurable options can be selected from command-line instead
    of having to edit pngconf.h (Glenn R-P)
  Fixed memory leak in pngwrite.c (free info_ptr->text) (Glenn R-P)
  Added more conditions for png_do_background, to avoid changing
    black pixels to background when a background is supplied and
    no pixels are transparent
  Repaired PNG_NO_STDIO behavior
  Tested NODIV support and made it default behavior (Greg Roelofs)
  Added "-m" option and PNGTEST_DEBUG_MEMORY to pngtest (John Bowler)
  Regularized version numbering scheme and bumped shared-library major
    version number to 2 to avoid problems with libpng 0.89 apps
    (Greg Roelofs)

Version 0.98 [January, 1998]
  Cleaned up some typos in libpng.txt and in code documentation
  Fixed memory leaks in pCAL chunk processing (Glenn R-P and John Bowler)
  Cosmetic change "display_gamma" to "screen_gamma" in pngrtran.c
  Changed recommendation about file_gamma for PC images to .51 from .45,
    in example.c and libpng.txt, added comments to distinguish between
    screen_gamma, viewing_gamma, and display_gamma.
  Changed all references to RFC1152 to read RFC1123 and changed the
    PNG_TIME_RFC1152_SUPPORTED macro to PNG_TIME_RFC1123_SUPPORTED
  Added png_invert_alpha capability (Glenn R-P -- suggestion by Jon Vincent)
  Changed srgb_intent from png_byte to int to avoid compiler bugs

Version 0.99 [January 30, 1998]
  Free info_ptr->text instead of end_info_ptr->text in pngread.c (John Bowler)
  Fixed a longstanding "packswap" bug in pngtrans.c
  Fixed some inconsistencies in pngconf.h that prevented compiling with
    PNG_READ_GAMMA_SUPPORTED and PNG_READ_hIST_SUPPORTED undefined
  Fixed some typos and made other minor rearrangement of libpng.txt (Andreas)
  Changed recommendation about file_gamma for PC images to .50 from .51 in
    example.c and libpng.txt, and changed file_gamma for sRGB images to .45
  Added a number of functions to access information from the png structure
    png_get_image_height(), etc. (Glenn R-P, suggestion by Brad Pettit)
  Added TARGET_MACOS similar to zlib-1.0.8
  Define PNG_ALWAYS_EXTERN when __MWERKS__ && WIN32 are defined
  Added type casting to all png_malloc() function calls

Version 0.99a [January 31, 1998]
  Added type casts and parentheses to all returns that return a value.(Tim W.)

Version 0.99b [February 4, 1998]
  Added type cast png_uint_32 on malloc function calls where needed.
  Changed type of num_hist from png_uint_32 to int (same as num_palette).
  Added checks for rowbytes overflow, in case png_size_t is less than 32 bits.
  Renamed makefile.elf to makefile.lnx.

Version 0.99c [February 7, 1998]
  More type casting.  Removed erroneous overflow test in pngmem.c.
  Added png_buffered_memcpy() and png_buffered_memset(), apply them to rowbytes.
  Added UNIX manual pages libpng.3 (incorporating libpng.txt) and  png.5.

Version 0.99d [February 11, 1998]
  Renamed "far_to_near()" "png_far_to_near()"
  Revised libpng.3
  Version 99c "buffered" operations didn't work as intended.  Replaced them
    with png_memcpy_check() and png_memset_check().
  Added many "if (png_ptr == NULL) return" to quell compiler warnings about
    unused png_ptr, mostly in pngget.c and pngset.c.
  Check for overlength tRNS chunk present when indexed-color PLTE is read.
  Cleaned up spelling errors in libpng.3/libpng.txt
  Corrected a problem with png_get_tRNS() which returned undefined trans array

Version 0.99e [February 28, 1998]
  Corrected png_get_tRNS() again.
  Add parentheses for easier reading of pngget.c, fixed "||" should be "&&".
  Touched up example.c to make more of it compileable, although the entire
    file still can't be compiled (Willem van Schaik)
  Fixed a bug in png_do_shift() (Bryan Tsai)
  Added a space in png.h prototype for png_write_chunk_start()
  Replaced pngtest.png with one created with zlib 1.1.1
  Changed pngtest to report PASS even when file size is different (Jean-loup G.)
  Corrected some logic errors in png_do_invert_alpha() (Chris Patterson)

Version 0.99f [March 5, 1998]
  Corrected a bug in pngpread() introduced in version 99c (Kevin Bracey)
  Moved makefiles into a "scripts" directory, and added INSTALL instruction file
  Added makefile.os2 and pngos2.def (A. Zabolotny) and makefile.s2x (W. Sebok)
  Added pointers to "note on libpng versions" in makefile.lnx and README
  Added row callback feature when reading and writing nonprogressive rows
    and added a test of this feature in pngtest.c
  Added user transform callbacks, with test of the feature in pngtest.c

Version 0.99g [March 6, 1998, morning]
  Minor changes to pngtest.c to suppress compiler warnings.
  Removed "beta" language from documentation.

Version 0.99h [March 6, 1998, evening]
  Minor changes to previous minor changes to pngtest.c
  Changed PNG_READ_NOT_FULLY_SUPPORTED to PNG_READ_TRANSFORMS_NOT_SUPPORTED
    and added PNG_PROGRESSIVE_READ_NOT_SUPPORTED macro
  Added user transform capability

Version 1.00 [March 7, 1998]
  Changed several typedefs in pngrutil.c
  Added makefile.wat (Pawel Mrochen), updated makefile.tc3 (Willem van Schaik)
  Replaced "while(1)" with "for(;;)"
  Added PNGARG() to prototypes in pngtest.c and removed some prototypes
  Updated some of the makefiles (Tom Lane)
  Changed some typedefs (s_start, etc.) in pngrutil.c
  Fixed dimensions of "short_months" array in pngwrite.c
  Replaced ansi2knr.c with the one from jpeg-v6

Version 1.0.0 [March 8, 1998]
  Changed name from 1.00 to 1.0.0 (Adam Costello)
  Added smakefile.ppc (with SCOPTIONS.ppc) for Amiga PPC (Andreas Kleinert)

Version 1.0.0a [March 9, 1998]
  Fixed three bugs in pngrtran.c to make gamma+background handling consistent
    (Greg Roelofs)
  Changed format of the PNG_LIBPNG_VER integer to xyyzz instead of xyz
    for major, minor, and bugfix releases.  This is 10001. (Adam Costello,
    Tom Lane)
  Make months range from 1-12 in png_convert_to_rfc1123

Version 1.0.0b [March 13, 1998]
  Quieted compiler complaints about two empty "for" loops in pngrutil.c
  Minor changes to makefile.s2x
  Removed #ifdef/#endif around a png_free() in pngread.c

Version 1.0.1 [March 14, 1998]
  Changed makefile.s2x to reduce security risk of using a relative pathname
  Fixed some typos in the documentation (Greg).
  Fixed a problem with value of "channels" returned by png_read_update_info()

Version 1.0.1a [April 21, 1998]
  Optimized Paeth calculations by replacing abs() function calls with intrinsics
  plus other loop optimizations. Improves avg decoding speed by about 20%.
  Commented out i386istic "align" compiler flags in makefile.lnx.
  Reduced the default warning level in some makefiles, to make them consistent.
  Removed references to IJG and JPEG in the ansi2knr.c copyright statement.
  Fixed a bug in png_do_strip_filler with XXRRGGBB => RRGGBB transformation.
  Added grayscale and 16-bit capability to png_do_read_filler().
  Fixed a bug in pngset.c, introduced in version 0.99c, that sets rowbytes
    too large when writing an image with bit_depth < 8 (Bob Dellaca).
  Corrected some bugs in the experimental weighted filtering heuristics.
  Moved a misplaced pngrutil code block that truncates tRNS if it has more
    than num_palette entries -- test was done before num_palette was defined.
  Fixed a png_convert_to_rfc1123() bug that converts day 31 to 0 (Steve Eddins).
  Changed compiler flags in makefile.wat for better optimization
    (Pawel Mrochen).

Version 1.0.1b [May 2, 1998]
  Relocated png_do_gray_to_rgb() within png_do_read_transformations() (Greg).
  Relocated the png_composite macros from pngrtran.c to png.h (Greg).
  Added makefile.sco (contributed by Mike Hopkirk).
  Fixed two bugs (missing definitions of "istop") introduced in libpng-1.0.1a.
  Fixed a bug in pngrtran.c that would set channels=5 under some circumstances.
  More work on the Paeth-filtering, achieving imperceptible speedup
    (A Kleinert).
  More work on loop optimization which may help when compiled with C++
    compilers.
  Added warnings when people try to use transforms they've defined out.
  Collapsed 4 "i" and "c" loops into single "i" loops in pngrtran and pngwtran.
  Revised paragraph about png_set_expand() in libpng.txt and libpng.3 (Greg)

Version 1.0.1c [May 11, 1998]
  Fixed a bug in pngrtran.c (introduced in libpng-1.0.1a) where the masks for
    filler bytes should have been 0xff instead of 0xf.
  Added max_pixel_depth=32 in pngrutil.c when using FILLER with palette images.
  Moved PNG_WRITE_WEIGHTED_FILTER_SUPPORTED and PNG_WRITE_FLUSH_SUPPORTED
    out of the PNG_WRITE_TRANSFORMS_NOT_SUPPORTED block of pngconf.h
  Added "PNG_NO_WRITE_TRANSFORMS" etc., as alternatives for *_NOT_SUPPORTED,
    for consistency, in pngconf.h
  Added individual "ifndef PNG_NO_[CAPABILITY]" in pngconf.h to make it easier
    to remove unwanted capabilities via the compile line
  Made some corrections to grammar (which, it's) in documentation (Greg).
  Corrected example.c, use of row_pointers in png_write_image().

Version 1.0.1d [May 24, 1998]
  Corrected several statements that used side effects illegally in pngrutil.c
    and pngtrans.c, that were introduced in version 1.0.1b
  Revised png_read_rows() to avoid repeated if-testing for NULL (A Kleinert)
  More corrections to example.c, use of row_pointers in png_write_image()
    and png_read_rows().
  Added pngdll.mak and pngdef.pas to scripts directory, contributed by
    Bob Dellaca, to make a png32bd.dll with Borland C++ 4.5
  Fixed error in example.c with png_set_text: num_text is 3, not 2 (Guido V.)
  Changed several loops from count-down to count-up, for consistency.

Version 1.0.1e [June 6, 1998]
  Revised libpng.txt and libpng.3 description of png_set_read|write_fn(), and
    added warnings when people try to set png_read_fn and png_write_fn in
    the same structure.
  Added a test such that png_do_gamma will be done when num_trans==0
    for truecolor images that have defined a background.  This corrects an
    error that was introduced in libpng-0.90 that can cause gamma processing
    to be skipped.
  Added tests in png.h to include "trans" and "trans_values" in structures
    when PNG_READ_BACKGROUND_SUPPORTED or PNG_READ_EXPAND_SUPPORTED is defined.
  Add png_free(png_ptr->time_buffer) in png_destroy_read_struct()
  Moved png_convert_to_rfc_1123() from pngwrite.c to png.c
  Added capability for user-provided malloc_fn() and free_fn() functions,
    and revised pngtest.c to demonstrate their use, replacing the
    PNGTEST_DEBUG_MEM feature.
  Added makefile.w32, for Microsoft C++ 4.0 and later (Tim Wegner).

Version 1.0.2 [June 14, 1998]
  Fixed two bugs in makefile.bor .

Version 1.0.2a [December 30, 1998]
  Replaced and extended code that was removed from png_set_filler() in 1.0.1a.
  Fixed a bug in png_do_filler() that made it fail to write filler bytes in
    the left-most pixel of each row (Kevin Bracey).
  Changed "static pngcharp tIME_string" to "static char tIME_string[30]"
    in pngtest.c (Duncan Simpson).
  Fixed a bug in pngtest.c that caused pngtest to try to write a tIME chunk
    even when no tIME chunk was present in the source file.
  Fixed a problem in pngrutil.c: gray_to_rgb didn't always work with 16-bit.
  Fixed a problem in png_read_push_finish_row(), which would not skip some
    passes that it should skip, for images that are less than 3 pixels high.
  Interchanged the order of calls to png_do_swap() and png_do_shift()
    in pngwtran.c (John Cromer).
  Added #ifdef PNG_DEBUG/#endif surrounding use of PNG_DEBUG in png.h .
  Changed "bad adaptive filter type" from error to warning in pngrutil.c .
  Fixed a documentation error about default filtering with 8-bit indexed-color.
  Separated the PNG_NO_STDIO macro into PNG_NO_STDIO and PNG_NO_CONSOLE_IO
    (L. Peter Deutsch).
  Added png_set_rgb_to_gray() and png_get_rgb_to_gray_status() functions.
  Added png_get_copyright() and png_get_header_version() functions.
  Revised comments on png_set_progressive_read_fn() in libpng.txt and example.c
  Added information about debugging in libpng.txt and libpng.3 .
  Changed "ln -sf" to "ln -s -f" in makefile.s2x, makefile.lnx, and
    makefile.sco.
  Removed lines after Dynamic Dependencies" in makefile.aco .
  Revised makefile.dec to make a shared library (Jeremie Petit).
  Removed trailing blanks from all files.

Version 1.0.2a [January 6, 1999]
  Removed misplaced #endif and #ifdef PNG_NO_EXTERN near the end of png.h
  Added "if" tests to silence complaints about unused png_ptr in png.h and png.c
  Changed "check_if_png" function in example.c to return true (nonzero) if PNG.
  Changed libpng.txt to demonstrate png_sig_cmp() instead of png_check_sig()
    which is obsolete.

Version 1.0.3 [January 14, 1999]
  Added makefile.hux, for Hewlett Packard HPUX 10.20 and 11.00 (Jim Rice)
  Added a statement of Y2K compliance in png.h, libpng.3, and Y2KINFO.

Version 1.0.3a [August 12, 1999]
  Added check for PNG_READ_INTERLACE_SUPPORTED in pngread.c; issue a warning
    if an attempt is made to read an interlaced image when it's not supported.
  Added check if png_ptr->trans is defined before freeing it in pngread.c
  Modified the Y2K statement to include versions back to version 0.71
  Fixed a bug in the check for valid IHDR bit_depth/color_types in pngrutil.c
  Modified makefile.wat (added -zp8 flag, ".symbolic", changed some comments)
  Replaced leading blanks with tab characters in makefile.hux
  Changed "dworkin.wustl.edu" to "ccrc.wustl.edu" in various documents.
  Changed (float)red and (float)green to (double)red, (double)green
    in png_set_rgb_to_gray() to avoid "promotion" problems in AIX.
  Fixed a bug in pngconf.h that omitted <stdio.h> when PNG_DEBUG==0 (K Bracey).
  Reformatted libpng.3 and libpngpf.3 with proper fonts (script by J. vanZandt).
  Updated documentation to refer to the PNG-1.2 specification.
  Removed ansi2knr.c and left pointers to the latest source for ansi2knr.c
    in makefile.knr, INSTALL, and README (L. Peter Deutsch)
  Fixed bugs in calculation of the length of rowbytes when adding alpha
    channels to 16-bit images, in pngrtran.c (Chris Nokleberg)
  Added function png_set_user_transform_info() to store user_transform_ptr,
    user_depth, and user_channels into the png_struct, and a function
    png_get_user_transform_ptr() to retrieve the pointer (Chris Nokleberg)
  Added function png_set_empty_plte_permitted() to make libpng useable
    in MNG applications.
  Corrected the typedef for png_free_ptr in png.h (Jesse Jones).
  Correct gamma with srgb is 45455 instead of 45000 in pngrutil.c, to be
    consistent with PNG-1.2, and allow variance of 500 before complaining.
  Added assembler code contributed by Intel in file pngvcrd.c and modified
    makefile.w32 to use it (Nirav Chhatrapati, INTEL Corporation,
    Gilles Vollant)
  Changed "ln -s -f" to "ln -f -s" in the makefiles to make Solaris happy.
  Added some aliases for png_set_expand() in pngrtran.c, namely
    png_set_expand_PLTE(), png_set_expand_depth(), and png_set_expand_tRNS()
    (Greg Roelofs, in "PNG: The Definitive Guide").
  Added makefile.beo for BEOS on X86, contributed by Sander Stok.

Version 1.0.3b [August 26, 1999]
  Replaced 2147483647L several places with PNG_MAX_UINT macro, defined in png.h
  Changed leading blanks to tabs in all makefiles.
  Define PNG_USE_PNGVCRD in makefile.w32, to get MMX assembler code.
  Made alternate versions of  png_set_expand() in pngrtran.c, namely
    png_set_gray_1_2_4_to_8, png_set_palette_to_rgb, and png_set_tRNS_to_alpha
    (Greg Roelofs, in "PNG: The Definitive Guide").  Deleted the 1.0.3a aliases.
  Relocated start of 'extern "C"' block in png.h so it doesn't include pngconf.h
  Revised calculation of num_blocks in pngmem.c to avoid a potentially
    negative shift distance, whose results are undefined in the C language.
  Added a check in pngset.c to prevent writing multiple tIME chunks.
  Added a check in pngwrite.c to detect invalid small window_bits sizes.

Version 1.0.3d [September 4, 1999]
  Fixed type casting of igamma in pngrutil.c
  Added new png_expand functions to scripts/pngdef.pas and pngos2.def
  Added a demo read_user_transform_fn that examines the row filters in pngtest.c

Version 1.0.4 [September 24, 1999, not distributed publicly]
  Define PNG_ALWAYS_EXTERN in pngconf.h if __STDC__ is defined
  Delete #define PNG_INTERNAL and include "png.h" from pngasmrd.h
  Made several minor corrections to pngtest.c
  Renamed the makefiles with longer but more user friendly extensions.
  Copied the PNG copyright and license to a separate LICENSE file.
  Revised documentation, png.h, and example.c to remove reference to
    "viewing_gamma" which no longer appears in the PNG specification.
  Revised pngvcrd.c to use MMX code for interlacing only on the final pass.
  Updated pngvcrd.c to use the faster C filter algorithms from libpng-1.0.1a
  Split makefile.win32vc into two versions, makefile.vcawin32 (uses MMX
    assembler code) and makefile.vcwin32 (doesn't).
  Added a CPU timing report to pngtest.c (enabled by defining PNGTEST_TIMING)
  Added a copy of pngnow.png to the distribution.

Version 1.0.4a [September 25, 1999]
  Increase max_pixel_depth in pngrutil.c if a user transform needs it.
  Changed several division operations to right-shifts in pngvcrd.c

Version 1.0.4b [September 30, 1999]
  Added parentheses in line 3732 of pngvcrd.c
  Added a comment in makefile.linux warning about buggy -O3 in pgcc 2.95.1

Version 1.0.4c [October 1, 1999]
  Added a "png_check_version" function in png.c and pngtest.c that will generate
    a helpful compiler error if an old png.h is found in the search path.
  Changed type of png_user_transform_depth|channels from int to png_byte.
  Added "Libpng is OSI Certified Open Source Software" statement to png.h

Version 1.0.4d [October 6, 1999]
  Changed 0.45 to 0.45455 in png_set_sRGB()
  Removed unused PLTE entries from pngnow.png
  Re-enabled some parts of pngvcrd.c (png_combine_row) that work properly.

Version 1.0.4e [October 10, 1999]
  Fixed sign error in pngvcrd.c (Greg Roelofs)
  Replaced some instances of memcpy with simple assignments in pngvcrd (GR-P)

Version 1.0.4f [October 15, 1999]
  Surrounded example.c code with #if 0 .. #endif to prevent people from
    inadvertently trying to compile it.
  Changed png_get_header_version() from a function to a macro in png.h
  Added type casting mostly in pngrtran.c and pngwtran.c
  Removed some pointless "ptr = NULL" in pngmem.c
  Added a "contrib" directory containing the source code from Greg's book.

Version 1.0.5 [October 15, 1999]
  Minor editing of the INSTALL and README files.

Version 1.0.5a [October 23, 1999]
  Added contrib/pngsuite and contrib/pngminus (Willem van Schaik)
  Fixed a typo in the png_set_sRGB() function call in example.c (Jan Nijtmans)
  Further optimization and bugfix of pngvcrd.c
  Revised pngset.c so that it does not allocate or free memory in the user's
    text_ptr structure.  Instead, it makes its own copy.
  Created separate write_end_info_struct in pngtest.c for a more severe test.
  Added code in pngwrite.c to free info_ptr->text[i].key to stop a memory leak.

Version 1.0.5b [November 23, 1999]
  Moved PNG_FLAG_HAVE_CHUNK_HEADER, PNG_FLAG_BACKGROUND_IS_GRAY and
    PNG_FLAG_WROTE_tIME from flags to mode.
  Added png_write_info_before_PLTE() function.
  Fixed some typecasting in contrib/gregbook/*.c
  Updated scripts/makevms.com and added makevms.com to contrib/gregbook
    and contrib/pngminus (Martin Zinser)

Version 1.0.5c [November 26, 1999]
  Moved png_get_header_version from png.h to png.c, to accommodate ansi2knr.
  Removed all global arrays (according to PNG_NO_GLOBAL_ARRAYS macro), to
    accommodate making DLL's: Moved usr_png_ver from global variable to function
    png_get_header_ver() in png.c.  Moved png_sig to png_sig_bytes in png.c and
    eliminated use of png_sig in pngwutil.c.  Moved the various png_CHNK arrays
    into pngtypes.h.  Eliminated use of global png_pass arrays.  Declared the
    png_CHNK and png_pass arrays to be "const".  Made the global arrays
    available to applications (although none are used in libpng itself) when
    PNG_NO_GLOBAL_ARRAYS is not defined or when PNG_GLOBAL_ARRAYS is defined.
  Removed some extraneous "-I" from contrib/pngminus/makefile.std
  Changed the PNG_sRGB_INTENT macros in png.h to be consistent with PNG-1.2.
  Change PNG_SRGB_INTENT to PNG_sRGB_INTENT in libpng.txt and libpng.3

Version 1.0.5d [November 29, 1999]
  Add type cast (png_const_charp) two places in png.c
  Eliminated pngtypes.h; use macros instead to declare PNG_CHNK arrays.
  Renamed "PNG_GLOBAL_ARRAYS" to "PNG_USE_GLOBAL_ARRAYS" and made available
    to applications a macro "PNG_USE_LOCAL_ARRAYS".
  comment out (with #ifdef) all the new declarations when
    PNG_USE_GLOBAL_ARRAYS is defined.
  Added PNG_EXPORT_VAR macro to accommodate making DLL's.

Version 1.0.5e [November 30, 1999]
  Added iCCP, iTXt, and sPLT support; added "lang" member to the png_text
    structure; refactored the inflate/deflate support to make adding new chunks
    with trailing compressed parts easier in the future, and added new functions
    png_free_iCCP, png_free_pCAL, png_free_sPLT, png_free_text, png_get_iCCP,
    png_get_spalettes, png_set_iCCP, png_set_spalettes (Eric S. Raymond).
    NOTE: Applications that write text chunks MUST define png_text->lang
    before calling png_set_text(). It must be set to NULL if you want to
    write tEXt or zTXt chunks.  If you want your application to be able to
    run with older versions of libpng, use

#ifdef PNG_iTXt_SUPPORTED
         png_text[i].lang = NULL;
      #endif

Changed png_get_oFFs() and png_set_oFFs() to use signed rather than unsigned
    offsets (Eric S. Raymond).
  Combined PNG_READ_cHNK_SUPPORTED and PNG_WRITE_cHNK_SUPPORTED macros into
    PNG_cHNK_SUPPORTED and combined the three types of PNG_text_SUPPORTED
    macros, leaving the separate macros also available.
  Removed comments on #endifs at the end of many short, non-nested #if-blocks.

Version 1.0.5f [December 6, 1999]
  Changed makefile.solaris to issue a warning about potential problems when
    the ucb "ld" is in the path ahead of the ccs "ld".
  Removed "- [date]" from the "synopsis" line in libpng.3 and libpngpf.3.
  Added sCAL chunk support (Eric S. Raymond).

Version 1.0.5g [December 7, 1999]
  Fixed "png_free_spallettes" typo in png.h
  Added code to handle new chunks in pngpread.c
  Moved PNG_CHNK string macro definitions outside of PNG_NO_EXTERN block
  Added "translated_key" to png_text structure and png_write_iTXt().
  Added code in pngwrite.c to work around a newly discovered zlib bug.

Version 1.0.5h [December 10, 1999]
  NOTE: regarding the note for version 1.0.5e, the following must also
    be included in your code:
        png_text[i].translated_key = NULL;
  Unknown chunk handling is now supported.
  Option to eliminate all floating point support was added.  Some new
    fixed-point functions such as png_set_gAMA_fixed() were added.
  Expanded tabs and removed trailing blanks in source files.

Version 1.0.5i [December 13, 1999]
  Added some type casts to silence compiler warnings.
  Renamed "png_free_spalette" to "png_free_spalettes" for consistency.
  Removed leading blanks from a #define in pngvcrd.c
  Added some parameters to the new png_set_keep_unknown_chunks() function.
  Added a test for up->location != 0 in the first instance of writing
    unknown chunks in pngwrite.c
  Changed "num" to "i" in png_free_spalettes() and png_free_unknowns() to
    prevent recursion.
  Added png_free_hIST() function.
  Various patches to fix bugs in the sCAL and integer cHRM processing,
    and to add some convenience macros for use with sCAL.

Version 1.0.5j [December 21, 1999]
  Changed "unit" parameter of png_write_sCAL from png_byte to int, to work
    around buggy compilers.
  Added new type "png_fixed_point" for integers that hold float*100000 values
  Restored backward compatibility of tEXt/zTXt chunk processing:
    Restored the first four members of png_text to the same order as v.1.0.5d.
    Added members "lang_key" and "itxt_length" to png_text struct.  Set
    text_length=0 when "text" contains iTXt data.  Use the "compression"
    member to distinguish among tEXt/zTXt/iTXt types.  Added
    PNG_ITXT_COMPRESSION_NONE (1) and PNG_ITXT_COMPRESSION_zTXt(2) macros.
    The "Note" above, about backward incompatibility of libpng-1.0.5e, no
    longer applies.
  Fixed png_read|write_iTXt() to read|write parameters in the right order,
    and to write the iTXt chunk after IDAT if it appears in the end_ptr.
  Added pnggccrd.c, version of pngvcrd.c Intel assembler for gcc (Greg Roelofs)
  Reversed the order of trying to write floating-point and fixed-point gAMA.

Version 1.0.5k [December 27, 1999]
  Added many parentheses, e.g., "if (a && b & c)" becomes "if (a && (b & c))"
  Added png_handle_as_unknown() function (Glenn)
  Added png_free_chunk_list() function and chunk_list and num_chunk_list members
    of png_ptr.
  Eliminated erroneous warnings about multiple sPLT chunks and sPLT-after-PLTE.
  Fixed a libpng-1.0.5h bug in pngrutil.c that was issuing erroneous warnings
    about ignoring incorrect gAMA with sRGB (gAMA was in fact not ignored)
  Added png_free_tRNS(); png_set_tRNS() now malloc's its own trans array (ESR).
  Define png_get_int_32 when oFFs chunk is supported as well as when pCAL is.
  Changed type of proflen from png_int_32 to png_uint_32 in png_get_iCCP().

Version 1.0.5l [January 1, 2000]
  Added functions png_set_read_user_chunk_fn() and png_get_user_chunk_ptr()
    for setting a callback function to handle unknown chunks and for
    retrieving the associated user pointer (Glenn).

Version 1.0.5m [January 7, 2000]
  Added high-level functions png_read_png(), png_write_png(), png_free_pixels().

Version 1.0.5n [January 9, 2000]
  Added png_free_PLTE() function, and modified png_set_PLTE() to malloc its
    own memory for info_ptr->palette.  This makes it safe for the calling
    application to free its copy of the palette any time after it calls
    png_set_PLTE().

Version 1.0.5o [January 20, 2000]
  Cosmetic changes only (removed some trailing blanks and TABs)

Version 1.0.5p [January 31, 2000]
  Renamed pngdll.mak to makefile.bd32
  Cosmetic changes in pngtest.c

Version 1.0.5q [February 5, 2000]
  Relocated the makefile.solaris warning about PATH problems.
  Fixed pngvcrd.c bug by pushing/popping registers in mmxsupport (Bruce Oberg)
  Revised makefile.gcmmx
  Added PNG_SETJMP_SUPPORTED, PNG_SETJMP_NOT_SUPPORTED, and PNG_ABORT() macros

Version 1.0.5r [February 7, 2000]
  Removed superfluous prototype for png_get_itxt from png.h
  Fixed a bug in pngrtran.c that improperly expanded the background color.
  Return *num_text=0 from png_get_text() when appropriate, and fix documentation
    of png_get_text() in libpng.txt/libpng.3.

Version 1.0.5s [February 18, 2000]
  Added "png_jmp_env()" macro to pngconf.h, to help people migrate to the
    new error handler that's planned for the next libpng release, and changed
    example.c, pngtest.c, and contrib programs to use this macro.
  Revised some of the DLL-export macros in pngconf.h (Greg Roelofs)
  Fixed a bug in png_read_png() that caused it to fail to expand some images
    that it should have expanded.
  Fixed some mistakes in the unused and undocumented INCH_CONVERSIONS functions
    in pngget.c
  Changed the allocation of palette, history, and trans arrays back to
    the version 1.0.5 method (linking instead of copying) which restores
    backward compatibility with version 1.0.5.  Added some remarks about
    that in example.c.  Added "free_me" member to info_ptr and png_ptr
    and added png_free_data() function.
  Updated makefile.linux and makefile.gccmmx to make directories conditionally.
  Made cosmetic changes to pngasmrd.h
  Added png_set_rows() and png_get_rows(), for use with png_read|write_png().
  Modified png_read_png() to allocate info_ptr->row_pointers only if it
    hasn't already been allocated.

Version 1.0.5t [March 4, 2000]
  Changed png_jmp_env() migration aiding macro to png_jmpbuf().
  Fixed "interlace" typo (should be "interlaced") in contrib/gregbook/read2-x.c
  Fixed bug with use of PNG_BEFORE_IHDR bit in png_ptr->mode, introduced when
    PNG_FLAG_HAVE_CHUNK_HEADER was moved into png_ptr->mode in version 1.0.5b
  Files in contrib/gregbook were revised to use png_jmpbuf() and to select
    a 24-bit visual if one is available, and to allow abbreviated options.
  Files in contrib/pngminus were revised to use the png_jmpbuf() macro.
  Removed spaces in makefile.linux and makefile.gcmmx, introduced in 1.0.5s

Version 1.0.5u [March 5, 2000]
  Simplified the code that detects old png.h in png.c and pngtest.c
  Renamed png_spalette (_p, _pp) to png_sPLT_t (_tp, _tpp)
  Increased precision of rgb_to_gray calculations from 8 to 15 bits and
    added png_set_rgb_to_gray_fixed() function.
  Added makefile.bc32 (32-bit Borland C++, C mode)

Version 1.0.5v [March 11, 2000]
  Added some parentheses to the png_jmpbuf macro definition.
  Updated references to the zlib home page, which has moved to freesoftware.com.
  Corrected bugs in documentation regarding png_read_row() and png_write_row().
  Updated documentation of png_rgb_to_gray calculations in libpng.3/libpng.txt.
  Renamed makefile.borland,turboc3 back to makefile.bor,tc3 as in version 1.0.3,
    revised borland makefiles; added makefile.ibmvac3 and makefile.gcc (Cosmin)

Version 1.0.6 [March 20, 2000]
  Minor revisions of makefile.bor, libpng.txt, and gregbook/rpng2-win.c
  Added makefile.sggcc (SGI IRIX with gcc)

Version 1.0.6d [April 7, 2000]
  Changed sprintf() to strcpy() in png_write_sCAL_s() to work without STDIO
  Added data_length parameter to png_decompress_chunk() function
  Revised documentation to remove reference to abandoned png_free_chnk functions
  Fixed an error in png_rgb_to_gray_fixed()
  Revised example.c, usage of png_destroy_write_struct().
  Renamed makefile.ibmvac3 to makefile.ibmc, added libpng.icc IBM project file
  Added a check for info_ptr->free_me&PNG_FREE_TEXT when freeing text in png.c
  Simplify png_sig_bytes() function to remove use of non-ISO-C strdup().

Version 1.0.6e [April 9, 2000]
  Added png_data_freer() function.
  In the code that checks for over-length tRNS chunks, added check of
    info_ptr->num_trans as well as png_ptr->num_trans (Matthias Benckmann)
  Minor revisions of libpng.txt/libpng.3.
  Check for existing data and free it if the free_me flag is set, in png_set_*()
    and png_handle_*().
  Only define PNG_WEIGHTED_FILTERS_SUPPORTED when PNG_FLOATING_POINT_SUPPORTED
    is defined.
  Changed several instances of PNG_NO_CONSOLE_ID to PNG_NO_STDIO in pngrutil.c
    and mentioned the purposes of the two macros in libpng.txt/libpng.3.

Version 1.0.6f [April 14, 2000]
  Revised png_set_iCCP() and png_set_rows() to avoid prematurely freeing data.
  Add checks in png_set_text() for NULL members of the input text structure.
  Revised libpng.txt/libpng.3.
  Removed superfluous prototype for png_set_iTXt from png.h
  Removed "else" from pngread.c, after png_error(), and changed "0" to "length".
  Changed several png_errors about malformed ancillary chunks to png_warnings.

Version 1.0.6g [April 24, 2000]
  Added png_pass-* arrays to pnggccrd.c when PNG_USE_LOCAL_ARRAYS is defined.
  Relocated paragraph about png_set_background() in libpng.3/libpng.txt
    and other revisions (Matthias Benckmann)
  Relocated info_ptr->free_me, png_ptr->free_me, and other info_ptr and
    png_ptr members to restore binary compatibility with libpng-1.0.5
    (breaks compatibility with libpng-1.0.6).

Version 1.0.6h [April 24, 2000]
  Changed shared library so-number pattern from 2.x.y.z to xy.z (this builds
    libpng.so.10 & libpng.so.10.6h instead of libpng.so.2 & libpng.so.2.1.0.6h)
    This is a temporary change for test purposes.

Version 1.0.6i [May 2, 2000]
  Rearranged some members at the end of png_info and png_struct, to put
    unknown_chunks_num and free_me within the original size of the png_structs
    and free_me, png_read_user_fn, and png_free_fn within the original png_info,
    because some old applications allocate the structs directly instead of
    using png_create_*().
  Added documentation of user memory functions in libpng.txt/libpng.3
  Modified png_read_png so that it will use user_allocated row_pointers
    if present, unless free_me directs that it be freed, and added description
    of the use of png_set_rows() and png_get_rows() in libpng.txt/libpng.3.
  Added PNG_LEGACY_SUPPORTED macro, and #ifdef out all new (since version
    1.00) members of png_struct and png_info, to regain binary compatibility
    when you define this macro.  Capabilities lost in this event
    are user transforms (new in version 1.0.0),the user transform pointer
    (new in version 1.0.2), rgb_to_gray (new in 1.0.5), iCCP, sCAL, sPLT,
    the high-level interface, and unknown chunks support (all new in 1.0.6).
    This was necessary because of old applications that allocate the structs
    directly as authors were instructed to do in libpng-0.88 and earlier,
    instead of using png_create_*().
  Added modes PNG_CREATED_READ_STRUCT and PNG_CREATED_WRITE_STRUCT which
    can be used to detect codes that directly allocate the structs, and
    code to check these modes in png_read_init() and png_write_init() and
    generate a libpng error if the modes aren't set and PNG_LEGACY_SUPPORTED
    was not defined.
  Added makefile.intel and updated makefile.watcom (Pawel Mrochen)

Version 1.0.6j [May 3, 2000]
  Overloaded png_read_init() and png_write_init() with macros that convert
    calls to png_read_init_2() or png_write_init_2() that check the version
    and structure sizes.

Version 1.0.7beta11 [May 7, 2000]
  Removed the new PNG_CREATED_READ_STRUCT and PNG_CREATED_WRITE_STRUCT modes
    which are no longer used.
  Eliminated the three new members of png_text when PNG_LEGACY_SUPPORTED is
    defined or when neither PNG_READ_iTXt_SUPPORTED nor PNG_WRITE_iTXt_SUPPORTED
    is defined.
  Made PNG_NO_READ|WRITE_iTXt the default setting, to avoid memory
    overrun when old applications fill the info_ptr->text structure directly.
  Added PNGAPI macro, and added it to the definitions of all exported functions.
  Relocated version macro definitions ahead of the includes of zlib.h and
    pngconf.h in png.h.

Version 1.0.7beta12 [May 12, 2000]
  Revised pngset.c to avoid a problem with expanding the png_debug macro.
  Deleted some extraneous defines from pngconf.h
  Made PNG_NO_CONSOLE_IO the default condition when PNG_BUILD_DLL is defined.
  Use MSC _RPTn debugging instead of fprintf if _MSC_VER is defined.
  Added png_access_version_number() function.
  Check for mask&PNG_FREE_CHNK (for TEXT, SCAL, PCAL) in png_free_data().
  Expanded libpng.3/libpng.txt information about png_data_freer().

Version 1.0.7beta14 [May 17, 2000] (beta13 was not published)
  Changed pnggccrd.c and pngvcrd.c to handle bad adaptive filter types as
    warnings instead of errors, as pngrutil.c does.
  Set the PNG_INFO_IDAT valid flag in png_set_rows() so png_write_png()
    will actually write IDATs.
  Made the default PNG_USE_LOCAL_ARRAYS depend on PNG_DLL instead of WIN32.
  Make png_free_data() ignore its final parameter except when freeing data
    that can have multiple instances (text, sPLT, unknowns).
  Fixed a new bug in png_set_rows().
  Removed info_ptr->valid tests from png_free_data(), as in version 1.0.5.
  Added png_set_invalid() function.
  Fixed incorrect illustrations of png_destroy_write_struct() in example.c.

Version 1.0.7beta15 [May 30, 2000]
  Revised the deliberately erroneous Linux setjmp code in pngconf.h to produce
    fewer error messages.
  Rearranged checks for Z_OK to check the most likely path first in pngpread.c
    and pngwutil.c.
  Added checks in pngtest.c for png_create_*() returning NULL, and mentioned
    in libpng.txt/libpng.3 the need for applications to check this.
  Changed names of png_default_*() functions in pngtest to pngtest_*().
  Changed return type of png_get_x|y_offset_*() from png_uint_32 to png_int_32.
  Fixed some bugs in the unused PNG_INCH_CONVERSIONS functions in pngget.c
  Set each pointer to NULL after freeing it in png_free_data().
  Worked around a problem in pngconf.h; AIX's strings.h defines an "index"
    macro that conflicts with libpng's png_color_16.index. (Dimitri
    Papadapoulos)
  Added "msvc" directory with MSVC++ project files (Simon-Pierre Cadieux).

Version 1.0.7beta16 [June 4, 2000]
  Revised the workaround of AIX string.h "index" bug.
  Added a check for overlength PLTE chunk in pngrutil.c.
  Added PNG_NO_POINTER_INDEXING macro to use array-indexing instead of pointer
    indexing in pngrutil.c and pngwutil.c to accommodate a buggy compiler.
  Added a warning in png_decompress_chunk() when it runs out of data, e.g.
    when it tries to read an erroneous PhotoShop iCCP chunk.
  Added PNG_USE_DLL macro.
  Revised the copyright/disclaimer/license notice.
  Added contrib/msvctest directory

Version 1.0.7rc1 [June 9, 2000]
  Corrected the definition of PNG_TRANSFORM_INVERT_ALPHA  (0x0400 not 0x0200)
  Added contrib/visupng directory (Willem van Schaik)

Version 1.0.7beta18 [June 23, 2000]
  Revised PNGAPI definition, and pngvcrd.c to work with __GCC__
    and do not redefine PNGAPI if it is passed in via a compiler directive.
  Revised visupng/PngFile.c to remove returns from within the Try block.
  Removed leading underscores from "_PNG_H" and "_PNG_SAVE_BSD_SOURCE" macros.
  Updated contrib/visupng/cexcept.h to version 1.0.0.
  Fixed bugs in pngwrite.c and pngwutil.c that prevented writing iCCP chunks.

Version 1.0.7rc2 [June 28, 2000]
  Updated license to include disclaimers required by UCITA.
  Fixed "DJBPP" typo in pnggccrd.c introduced in beta18.

Version 1.0.7 [July 1, 2000]
  Revised the definition of "trans_values" in libpng.3/libpng.txt

Version 1.0.8beta1 [July 8, 2000]
  Added png_free(png_ptr, key) two places in pngpread.c to stop memory leaks.
  Changed PNG_NO_STDIO to PNG_NO_CONSOLE_IO, several places in pngrutil.c and
    pngwutil.c.
  Changed PNG_EXPORT_VAR to use PNG_IMPEXP, in pngconf.h.
  Removed unused "#include <assert.h>" from png.c
  Added WindowsCE support.
  Revised pnggccrd.c to work with gcc-2.95.2 and in the Cygwin environment.

Version 1.0.8beta2 [July 10, 2000]
  Added project files to the wince directory and made further revisions
    of pngtest.c, pngrio.c, and pngwio.c in support of WindowsCE.

Version 1.0.8beta3 [July 11, 2000]
  Only set the PNG_FLAG_FREE_TRNS or PNG_FREE_TRNS flag in png_handle_tRNS()
    for indexed-color input files to avoid potential double-freeing trans array
    under some unusual conditions; problem was introduced in version 1.0.6f.
  Further revisions to pngtest.c and files in the wince subdirectory.

Version 1.0.8beta4 [July 14, 2000]
  Added the files pngbar.png and pngbar.jpg to the distribution.
  Added makefile.cygwin, and cygwin support in pngconf.h
  Added PNG_NO_ZALLOC_ZERO macro (makes png_zalloc skip zeroing memory)

Version 1.0.8rc1 [July 16, 2000]
  Revised png_debug() macros and statements to eliminate compiler warnings.

Version 1.0.8 [July 24, 2000]
  Added png_flush() in pngwrite.c, after png_write_IEND().
  Updated makefile.hpux to build a shared library.

Version 1.0.9beta1 [November 10, 2000]
  Fixed typo in scripts/makefile.hpux
  Updated makevms.com in scripts and contrib/* and contrib/* (Martin Zinser)
  Fixed seqence-point bug in contrib/pngminus/png2pnm (Martin Zinser)
  Changed "cdrom.com" in documentation to "libpng.org"
  Revised pnggccrd.c to get it all working, and updated makefile.gcmmx (Greg).
  Changed type of "params" from voidp to png_voidp in png_read|write_png().
  Make sure PNGAPI and PNG_IMPEXP are defined in pngconf.h.
  Revised the 3 instances of WRITEFILE in pngtest.c.
  Relocated "msvc" and "wince" project subdirectories into "dll" subdirectory.
  Updated png.rc in dll/msvc project
  Revised makefile.dec to define and use LIBPATH and INCPATH
  Increased size of global png_libpng_ver[] array from 12 to 18 chars.
  Made global png_libpng_ver[], png_sig[] and png_pass_*[] arrays const.
  Removed duplicate png_crc_finish() from png_handle_bKGD() function.
  Added a warning when application calls png_read_update_info() multiple times.
  Revised makefile.cygwin
  Fixed bugs in iCCP support in pngrutil.c and pngwutil.c.
  Replaced png_set_empty_plte_permitted() with png_permit_mng_features().

Version 1.0.9beta2 [November 19, 2000]
  Renamed the "dll" subdirectory "projects".
  Added borland project files to "projects" subdirectory.
  Set VS_FF_PRERELEASE and VS_FF_PATCHED flags in msvc/png.rc when appropriate.
  Add error message in png_set_compression_buffer_size() when malloc fails.

Version 1.0.9beta3 [November 23, 2000]
  Revised PNG_LIBPNG_BUILD_TYPE macro in png.h, used in the msvc project.
  Removed the png_flush() in pngwrite.c that crashes some applications
    that don't set png_output_flush_fn.
  Added makefile.macosx and makefile.aix to scripts directory.

Version 1.0.9beta4 [December 1, 2000]
  Change png_chunk_warning to png_warning in png_check_keyword().
  Increased the first part of msg buffer from 16 to 18 in png_chunk_error().

Version 1.0.9beta5 [December 15, 2000]
  Added support for filter method 64 (for PNG datastreams embedded in MNG).

Version 1.0.9beta6 [December 18, 2000]
  Revised png_set_filter() to accept filter method 64 when appropriate.
  Added new PNG_HAVE_PNG_SIGNATURE bit to png_ptr->mode and use it to
    help prevent applications from using MNG features in PNG datastreams.
  Added png_permit_mng_features() function.
  Revised libpng.3/libpng.txt.  Changed "filter type" to "filter method".

Version 1.0.9rc1 [December 23, 2000]
  Revised test for PNG_HAVE_PNG_SIGNATURE in pngrutil.c
  Fixed error handling of unknown compression type in png_decompress_chunk().
  In pngconf.h, define __cdecl when _MSC_VER is defined.

Version 1.0.9beta7 [December 28, 2000]
  Changed PNG_TEXT_COMPRESSION_zTXt to PNG_COMPRESSION_TYPE_BASE several places.
  Revised memory management in png_set_hIST and png_handle_hIST in a backward
    compatible manner.  PLTE and tRNS were revised similarly.
  Revised the iCCP chunk reader to ignore trailing garbage.

Version 1.0.9beta8 [January 12, 2001]
  Moved pngasmrd.h into pngconf.h.
  Improved handling of out-of-spec garbage iCCP chunks generated by PhotoShop.

Version 1.0.9beta9 [January 15, 2001]
  Added png_set_invalid, png_permit_mng_features, and png_mmx_supported to
    wince and msvc project module definition files.
  Minor revision of makefile.cygwin.
  Fixed bug with progressive reading of narrow interlaced images in pngpread.c

Version 1.0.9beta10 [January 16, 2001]
  Do not typedef png_FILE_p in pngconf.h when PNG_NO_STDIO is defined.
  Fixed "png_mmx_supported" typo in project definition files.

Version 1.0.9beta11 [January 19, 2001]
  Updated makefile.sgi to make shared library.
  Removed png_mmx_support() function and disabled PNG_MNG_FEATURES_SUPPORTED
    by default, for the benefit of DLL forward compatibility.  These will
    be re-enabled in version 1.2.0.

Version 1.0.9rc2 [January 22, 2001]
  Revised cygwin support.

Version 1.0.9 [January 31, 2001]
  Added check of cygwin's ALL_STATIC in pngconf.h
  Added "-nommx" parameter to contrib/gregbook/rpng2-win and rpng2-x demos.

Version 1.0.10beta1 [March 14, 2001]
  Revised makefile.dec, makefile.sgi, and makefile.sggcc; added makefile.hpgcc.
  Reformatted libpng.3 to eliminate bad line breaks.
  Added checks for _mmx_supported in the read_filter_row function of pnggccrd.c
  Added prototype for png_mmx_support() near the top of pnggccrd.c
  Moved some error checking from png_handle_IHDR to png_set_IHDR.
  Added PNG_NO_READ_SUPPORTED and PNG_NO_WRITE_SUPPORTED macros.
  Revised png_mmx_support() function in pnggccrd.c
  Restored version 1.0.8 PNG_WRITE_EMPTY_PLTE_SUPPORTED behavior in pngwutil.c
  Fixed memory leak in contrib/visupng/PngFile.c
  Fixed bugs in png_combine_row() in pnggccrd.c and pngvcrd.c (C version)
  Added warnings when retrieving or setting gamma=0.
  Increased the first part of msg buffer from 16 to 18 in png_chunk_warning().

Version 1.0.10rc1 [March 23, 2001]
  Changed all instances of memcpy, strcpy, and strlen to png_memcpy, png_strcpy,
    and png_strlen.
  Revised png_mmx_supported() function in pnggccrd.c to return proper value.
  Fixed bug in progressive reading (pngpread.c) with small images (height < 8).

Version 1.0.10 [March 30, 2001]
  Deleted extraneous space (introduced in 1.0.9) from line 42 of makefile.cygwin
  Added beos project files (Chris Herborth)

Version 1.0.11beta1 [April 3, 2001]
  Added type casts on several png_malloc() calls (Dimitri Papadapoulos).
  Removed a no-longer needed AIX work-around from pngconf.h
  Changed several "//" single-line comments to C-style in pnggccrd.c

Version 1.0.11beta2 [April 11, 2001]
  Removed PNGAPI from several functions whose prototypes did not have PNGAPI.
  Updated scripts/pngos2.def

Version 1.0.11beta3 [April 14, 2001]
  Added checking the results of many instances of png_malloc() for NULL

Version 1.0.11beta4 [April 20, 2001]
  Undid the changes from version 1.0.11beta3.  Added a check for NULL return
    from user's malloc_fn().
  Removed some useless type casts of the NULL pointer.
  Added makefile.netbsd

Version 1.0.11 [April 27, 2001]
  Revised makefile.netbsd

Version 1.0.12beta1 [May 14, 2001]
  Test for Windows platform in pngconf.h when including malloc.h (Emmanuel Blot)
  Updated makefile.cygwin and handling of Cygwin's ALL_STATIC in pngconf.h
  Added some never-to-be-executed code in pnggccrd.c to quiet compiler warnings.
  Eliminated the png_error about apps using png_read|write_init().  Instead,
    libpng will reallocate the png_struct and info_struct if they are too small.
    This retains future binary compatibility for old applications written for
    libpng-0.88 and earlier.

Version 1.2.0beta1 [May 6, 2001]
  Bumped DLLNUM to 2.
  Re-enabled PNG_MNG_FEATURES_SUPPORTED and enabled PNG_ASSEMBLER_CODE_SUPPORTED
    by default.
  Added runtime selection of MMX features.
  Added png_set_strip_error_numbers function and related macros.

Version 1.2.0beta2 [May 7, 2001]
  Finished merging 1.2.0beta1 with version 1.0.11
  Added a check for attempts to read or write PLTE in grayscale PNG datastreams.

Version 1.2.0beta3 [May 17, 2001]
  Enabled user memory function by default.
  Modified png_create_struct so it passes user mem_ptr to user memory allocator.
  Increased png_mng_features flag from png_byte to png_uint_32.
  Bumped shared-library (so-number) and dll-number to 3.

Version 1.2.0beta4 [June 23, 2001]
  Check for missing profile length field in iCCP chunk and free chunk_data
    in case of truncated iCCP chunk.
  Bumped shared-library number to 3 in makefile.sgi and makefile.sggcc
  Bumped dll-number from 2 to 3 in makefile.cygwin
  Revised contrib/gregbook/rpng*-x.c to avoid a memory leak and to exit cleanly
    if user attempts to run it on an 8-bit display.
  Updated contrib/gregbook
  Use png_malloc instead of png_zalloc to allocate palette in pngset.c
  Updated makefile.ibmc
  Added some typecasts to eliminate gcc 3.0 warnings.  Changed prototypes
    of png_write_oFFS width and height from png_uint_32 to png_int_32.
  Updated example.c
  Revised prototypes for png_debug_malloc and png_debug_free in pngtest.c

Version 1.2.0beta5 [August 8, 2001]
  Revised contrib/gregbook
  Revised makefile.gcmmx
  Revised pnggccrd.c to conditionally compile some thread-unsafe code only
    when PNG_THREAD_UNSAFE_OK is defined.
  Added tests to prevent pngwutil.c from writing a bKGD or tRNS chunk with
    value exceeding 2^bit_depth-1
  Revised makefile.sgi and makefile.sggcc
  Replaced calls to fprintf(stderr,...) with png_warning() in pnggccrd.c
  Removed restriction that do_invert_mono only operate on 1-bit opaque files

Version 1.2.0 [September 1, 2001]
  Changed a png_warning() to png_debug() in pnggccrd.c
  Fixed contrib/gregbook/rpng-x.c, rpng2-x.c to avoid crash with XFreeGC().

Version 1.2.1beta1 [October 19, 2001]
  Revised makefile.std in contrib/pngminus
  Include background_1 in png_struct regardless of gamma support.
  Revised makefile.netbsd and makefile.macosx, added makefile.darwin.
  Revised example.c to provide more details about using row_callback().

Version 1.2.1beta2 [October 25, 2001]
  Added type cast to each NULL appearing in a function call, except for
    WINCE functions.
  Added makefile.so9.

Version 1.2.1beta3 [October 27, 2001]
  Removed type casts from all NULLs.
  Simplified png_create_struct_2().

Version 1.2.1beta4 [November 7, 2001]
  Revised png_create_info_struct() and png_creat_struct_2().
  Added error message if png_write_info() was omitted.
  Type cast NULLs appearing in function calls when _NO_PROTO or
    PNG_TYPECAST_NULL is defined.

Version 1.2.1rc1 [November 24, 2001]
  Type cast NULLs appearing in function calls except when PNG_NO_TYPECAST_NULL
    is defined.
  Changed typecast of "size" argument to png_size_t in pngmem.c calls to
    the user malloc_fn, to agree with the prototype in png.h
  Added a pop/push operation to pnggccrd.c, to preserve Eflag (Maxim Sobolev)
  Updated makefile.sgi to recognize LIBPATH and INCPATH.
  Updated various makefiles so "make clean" does not remove previous major
    version of the shared library.

Version 1.2.1rc2 [December 4, 2001]
  Always allocate 256-entry internal palette, hist, and trans arrays, to
    avoid out-of-bounds memory reference caused by invalid PNG datastreams.
  Added a check for prefix_length > data_length in iCCP chunk handler.

Version 1.2.1 [December 7, 2001]
  None.

Version 1.2.2beta1 [February 22, 2002]
  Fixed a bug with reading the length of iCCP profiles (Larry Reeves).
  Revised makefile.linux, makefile.gcmmx, and makefile.sgi to generate
    libpng.a, libpng12.so (not libpng.so.3), and libpng12/png.h
  Revised makefile.darwin to remove "-undefined suppress" option.
  Added checks for gamma and chromaticity values over 21474.83, which exceed
    the limit for PNG unsigned 32-bit integers when encoded.
  Revised calls to png_create_read_struct() and png_create_write_struct()
    for simpler debugging.
  Revised png_zalloc() so zlib handles errors (uses PNG_FLAG_MALLOC_NULL_MEM_OK)

Version 1.2.2beta2 [February 23, 2002]
  Check chunk_length and idat_size for invalid (over PNG_MAX_UINT) lengths.
  Check for invalid image dimensions in png_get_IHDR.
  Added missing "fi;" in the install target of the SGI makefiles.
  Added install-static to all makefiles that make shared libraries.
  Always do gamma compensation when image is partially transparent.

Version 1.2.2beta3 [March 7, 2002]
  Compute background.gray and background_1.gray even when color_type is RGB
    in case image gets reduced to gray later.
  Modified shared-library makefiles to install pkgconfig/libpngNN.pc.
  Export (with PNGAPI) png_zalloc, png_zfree, and png_handle_as_unknown
  Removed unused png_write_destroy_info prototype from png.h
  Eliminated incorrect use of width_mmx from pnggccrd.c in pixel_bytes == 8 case
  Added install-shared target to all makefiles that make shared libraries.
  Stopped a double free of palette, hist, and trans when not using free_me.
  Added makefile.32sunu for Sun Ultra 32 and makefile.64sunu for Sun Ultra 64.

Version 1.2.2beta4 [March 8, 2002]
  Compute background.gray and background_1.gray even when color_type is RGB
    in case image gets reduced to gray later (Jason Summers).
  Relocated a misplaced /bin/rm in the "install-shared" makefile targets
  Added PNG_1_0_X macro which can be used to build a 1.0.x-compatible library.

Version 1.2.2beta5 [March 26, 2002]
  Added missing PNGAPI to several function definitions.
  Check for invalid bit_depth or color_type in png_get_IHDR(), and
    check for missing PLTE or IHDR in png_push_read_chunk() (Matthias Clasen).
  Revised iTXt support to accept NULL for lang and lang_key.
  Compute gamma for color components of background even when color_type is gray.
  Changed "()" to "{}" in scripts/libpng.pc.in.
  Revised makefiles to put png.h and pngconf.h only in $prefix/include/libpngNN
  Revised makefiles to make symlink to libpng.so.NN in addition to libpngNN.so

Version 1.2.2beta6 [March 31, 2002]

Version 1.0.13beta1 [March 31, 2002]
  Prevent png_zalloc() from trying to memset memory that it failed to acquire.
  Add typecasts of PNG_MAX_UINT in pngset_cHRM_fixed() (Matt Holgate).
  Ensure that the right function (user or default) is used to free the
    png_struct after an error in png_create_read_struct_2().

Version 1.2.2rc1 [April 7, 2002]

Version 1.0.13rc1 [April 7, 2002]
  Save the ebx register in pnggccrd.c (Sami Farin)
  Add "mem_ptr = png_ptr->mem_ptr" in png_destroy_write_struct() (Paul Gardner).
  Updated makefiles to put headers in include/libpng and remove old include/*.h.

Version 1.2.2 [April 15, 2002]

Version 1.0.13 [April 15, 2002]
  Revised description of png_set_filter() in libpng.3/libpng.txt.
  Revised makefile.netbsd and added makefile.neNNbsd and makefile.freebsd

Version 1.0.13patch01 [April 17, 2002]

Version 1.2.2patch01 [April 17, 2002]
  Changed ${PNGMAJ}.${PNGVER} bug to ${PNGVER} in makefile.sgi and
    makefile.sggcc
  Fixed VER -> PNGVER typo in makefile.macosx and added install-static to
    install
  Added install: target to makefile.32sunu and makefile.64sunu

Version 1.0.13patch03 [April 18, 2002]

Version 1.2.2patch03 [April 18, 2002]
  Revised 15 makefiles to link libpng.a to libpngNN.a and the include libpng
  subdirectory to libpngNN subdirectory without the full pathname.
  Moved generation of libpng.pc from "install" to "all" in 15 makefiles.

Version 1.2.3rc1 [April 28, 2002]
  Added install-man target to 15 makefiles (Dimitri Papadopolous-Orfanos).
  Added $(DESTDIR) feature to 24 makefiles (Tim Mooney)
  Fixed bug with $prefix, should be $(prefix) in makefile.hpux.
  Updated cygwin-specific portion of pngconf.h and revised makefile.cygwin
  Added a link from libpngNN.pc to libpng.pc in 15 makefiles.
  Added links from include/libpngNN/*.h to include/*.h in 24 makefiles.
  Revised makefile.darwin to make relative links without full pathname.
  Added setjmp() at the end of png_create_*_struct_2() in case user forgets
    to put one in their application.
  Restored png_zalloc() and png_zfree() prototypes to version 1.2.1 and
    removed them from module definition files.

Version 1.2.3rc2 [May 1, 2002]
  Fixed bug in reporting number of channels in pngget.c and pngset.c,
    that was introduced in version 1.2.2beta5.
  Exported png_zalloc(), png_zfree(), png_default_read(), png_default_write(),
    png_default_flush(), and png_push_fill_buffer() and included them in
    module definition files.
  Added "libpng.pc" dependency to the "install-shared" target in 15 makefiles.

Version 1.2.3rc3 [May 1, 2002]
  Revised prototype for png_default_flush()
  Remove old libpng.pc and libpngNN.pc before installing new ones.

Version 1.2.3rc4 [May 2, 2002]
  Typos in *.def files (png_default_read|write -> png_default_read|write_data)
  In makefiles, changed rm libpng.NN.pc to rm libpngNN.pc
  Added libpng-config and libpngNN-config and modified makefiles to install
    them.
  Changed $(MANPATH) to $(DESTDIR)$(MANPATH) in makefiles
  Added "Win32 DLL VB" configuration to projects/msvc/libpng.dsp

Version 1.2.3rc5 [May 11, 2002]
  Changed "error" and "message" in prototypes to "error_message" and
    "warning_message" to avoid namespace conflict.
  Revised 15 makefiles to build libpng-config from libpng-config-*.in
  Once more restored png_zalloc and png_zfree to regular nonexported form.
  Restored png_default_read|write_data, png_default_flush, png_read_fill_buffer
    to nonexported form, but with PNGAPI, and removed them from module def
    files.

Version 1.2.3rc6 [May 14, 2002]
  Removed "PNGAPI" from png_zalloc() and png_zfree() in png.c
  Changed "Gz" to "Gd" in projects/msvc/libpng.dsp and zlib.dsp.
  Removed leftover libpng-config "sed" script from four makefiles.
  Revised libpng-config creating script in 16 makefiles.

Version 1.2.3 [May 22, 2002]
  Revised libpng-config target in makefile.cygwin.
  Removed description of png_set_mem_fn() from documentation.
  Revised makefile.freebsd.
  Minor cosmetic changes to 15 makefiles, e.g., $(DI) = $(DESTDIR)/$(INCDIR).
  Revised projects/msvc/README.txt
  Changed -lpng to -lpngNN in LDFLAGS in several makefiles.

Version 1.2.4beta1 [May 24, 2002]
  Added libpng.pc and libpng-config to "all:" target in 16 makefiles.
  Fixed bug in 16 makefiles: $(DESTDIR)/$(LIBPATH) to $(DESTDIR)$(LIBPATH)
  Added missing "\" before closing double quote in makefile.gcmmx.
  Plugged various memory leaks; added png_malloc_warn() and png_set_text_2()
    functions.

Version 1.2.4beta2 [June 25, 2002]
  Plugged memory leak of png_ptr->current_text (Matt Holgate).
  Check for buffer overflow before reading CRC in pngpread.c (Warwick Allison)
  Added -soname to the loader flags in makefile.dec, makefile.sgi, and
    makefile.sggcc.
  Added "test-installed" target to makefile.linux, makefile.gcmmx,
    makefile.sgi, and makefile.sggcc.

Version 1.2.4beta3 [June 28, 2002]
  Plugged memory leak of row_buf in pngtest.c when there is a png_error().
  Detect buffer overflow in pngpread.c when IDAT is corrupted with extra data.
  Added "test-installed" target to makefile.32sunu, makefile.64sunu,
    makefile.beos, makefile.darwin, makefile.dec, makefile.macosx,
    makefile.solaris, makefile.hpux, makefile.hpgcc, and makefile.so9.

Version 1.2.4rc1 and 1.0.14rc1 [July 2, 2002]
  Added "test-installed" target to makefile.cygwin and makefile.sco.
  Revised pnggccrd.c to be able to back out version 1.0.x via PNG_1_0_X macro.

Version 1.2.4 and 1.0.14 [July 8, 2002]
  Changed png_warning() to png_error() when width is too large to process.

Version 1.2.4patch01 [July 20, 2002]
  Revised makefile.cygwin to use DLL number 12 instead of 13.

Version 1.2.5beta1 [August 6, 2002]
  Added code to contrib/gregbook/readpng2.c to ignore unused chunks.
  Replaced toucan.png in contrib/gregbook (it has been corrupt since 1.0.11)
  Removed some stray *.o files from contrib/gregbook.
  Changed png_error() to png_warning() about "Too much data" in pngpread.c
    and about "Extra compressed data" in pngrutil.c.
  Prevent png_ptr->pass from exceeding 7 in png_push_finish_row().
  Updated makefile.hpgcc
  Updated png.c and pnggccrd.c handling of return from png_mmx_support()

Version 1.2.5beta2 [August 15, 2002]
  Only issue png_warning() about "Too much data" in pngpread.c when avail_in
    is nonzero.
  Updated makefiles to install a separate libpng.so.3 with its own rpath.

Version 1.2.5rc1 and 1.0.15rc1 [August 24, 2002]
  Revised makefiles to not remove previous minor versions of shared libraries.

Version 1.2.5rc2 and 1.0.15rc2 [September 16, 2002]
  Revised 13 makefiles to remove "-lz" and "-L$(ZLIBLIB)", etc., from shared
    library loader directive.
  Added missing "$OBJSDLL" line to makefile.gcmmx.
  Added missing "; fi" to makefile.32sunu.

Version 1.2.5rc3 and 1.0.15rc3 [September 18, 2002]
  Revised libpng-config script.

Version 1.2.5 and 1.0.15 [October 3, 2002]
  Revised makefile.macosx, makefile.darwin, makefile.hpgcc, and makefile.hpux,
    and makefile.aix.
  Relocated two misplaced PNGAPI lines in pngtest.c

Version 1.2.6beta1 [October 22, 2002]
  Commented out warning about uninitialized mmx_support in pnggccrd.c.
  Changed "IBMCPP__" flag to "__IBMCPP__" in pngconf.h.
  Relocated two more misplaced PNGAPI lines in pngtest.c
  Fixed memory overrun bug in png_do_read_filler() with 16-bit datastreams,
    introduced in version 1.0.2.
  Revised makefile.macosx, makefile.dec, makefile.aix, and makefile.32sunu.

Version 1.2.6beta2 [November 1, 2002]
  Added libpng-config "--ldopts" output.
  Added "AR=ar" and "ARFLAGS=rc" and changed "ar rc" to "$(AR) $(ARFLAGS)"
    in makefiles.

Version 1.2.6beta3 [July 18, 2004]
  Reverted makefile changes from version 1.2.6beta2 and some of the changes
    from version 1.2.6beta1; these will be postponed until version 1.2.7.
    Version 1.2.6 is going to be a simple bugfix release.
  Changed the one instance of "ln -sf" to "ln -f -s" in each Sun makefile.
  Fixed potential overrun in pngerror.c by using strncpy instead of memcpy.
  Added "#!/bin/sh" at the top of configure, for recognition of the
    'x' flag under Cygwin (Cosmin).
  Optimized vacuous tests that silence compiler warnings, in png.c (Cosmin).
  Added support for PNG_USER_CONFIG, in pngconf.h (Cosmin).
  Fixed the special memory handler for Borland C under DOS, in pngmem.c
    (Cosmin).
  Removed some spurious assignments in pngrutil.c (Cosmin).
  Replaced 65536 with 65536L, and 0xffff with 0xffffL, to silence warnings
    on 16-bit platforms (Cosmin).
  Enclosed shift op expressions in parentheses, to silence warnings (Cosmin).
  Used proper type png_fixed_point, to avoid problems on 16-bit platforms,
    in png_handle_sRGB() (Cosmin).
  Added compression_type to png_struct, and optimized the window size
    inside the deflate stream (Cosmin).
  Fixed definition of isnonalpha(), in pngerror.c and pngrutil.c (Cosmin).
  Fixed handling of unknown chunks that come after IDAT (Cosmin).
  Allowed png_error() and png_warning() to work even if png_ptr == NULL
    (Cosmin).
  Replaced row_info->rowbytes with row_bytes in png_write_find_filter()
    (Cosmin).
  Fixed definition of PNG_LIBPNG_VER_DLLNUM (Simon-Pierre).
  Used PNG_LIBPNG_VER and PNG_LIBPNG_VER_STRING instead of the hardcoded
    values in png.c (Simon-Pierre, Cosmin).
  Initialized png_libpng_ver[] with PNG_LIBPNG_VER_STRING (Simon-Pierre).
  Replaced PNG_LIBPNG_VER_MAJOR with PNG_LIBPNG_VER_DLLNUM in png.rc
    (Simon-Pierre).
  Moved the definition of PNG_HEADER_VERSION_STRING near the definitions
    of the other PNG_LIBPNG_VER_... symbols in png.h (Cosmin).
  Relocated #ifndef PNGAPI guards in pngconf.h (Simon-Pierre, Cosmin).
  Updated scripts/makefile.vc(a)win32 (Cosmin).
  Updated the MSVC project (Simon-Pierre, Cosmin).
  Updated the Borland C++ Builder project (Cosmin).
  Avoided access to asm_flags in pngvcrd.c, if PNG_1_0_X is defined (Cosmin).
  Commented out warning about uninitialized mmx_support in pngvcrd.c (Cosmin).
  Removed scripts/makefile.bd32 and scripts/pngdef.pas (Cosmin).
  Added extra guard around inclusion of Turbo C memory headers, in pngconf.h
    (Cosmin).
  Renamed projects/msvc/ to projects/visualc6/, and projects/borland/ to
    projects/cbuilder5/ (Cosmin).
  Moved projects/visualc6/png32ms.def to scripts/pngw32.def,
    and projects/visualc6/png.rc to scripts/pngw32.rc (Cosmin).
  Added projects/visualc6/pngtest.dsp; removed contrib/msvctest/ (Cosmin).
  Changed line endings to DOS style in cbuilder5 and visualc6 files, even
    in the tar.* distributions (Cosmin).
  Updated contrib/visupng/VisualPng.dsp (Cosmin).
  Updated contrib/visupng/cexcept.h to version 2.0.0 (Cosmin).
  Added a separate distribution with "configure" and supporting files (Junichi).

Version 1.2.6beta4 [July 28, 2004]
  Added user ability to change png_size_t via a PNG_SIZE_T macro.
  Added png_sizeof() and png_convert_size() functions.
  Added PNG_SIZE_MAX (maximum value of a png_size_t variable.
  Added check in png_malloc_default() for (size_t)size != (png_uint_32)size
    which would indicate an overflow.
  Changed sPLT failure action from png_error to png_warning and abandon chunk.
  Changed sCAL and iCCP failures from png_error to png_warning and abandon.
  Added png_get_uint_31(png_ptr, buf) function.
  Added PNG_UINT_32_MAX macro.
  Renamed PNG_MAX_UINT to PNG_UINT_31_MAX.
  Made png_zalloc() issue a png_warning and return NULL on potential
    overflow.
  Turn on PNG_NO_ZALLOC_ZERO by default in version 1.2.x
  Revised "clobber list" in pnggccrd.c so it will compile under gcc-3.4.
  Revised Borland portion of png_malloc() to return NULL or issue
    png_error() according to setting of PNG_FLAG_MALLOC_NULL_MEM_OK.
  Added PNG_NO_SEQUENTIAL_READ_SUPPORTED macro to conditionally remove
    sequential read support.
  Added some "#if PNG_WRITE_SUPPORTED" blocks.
  Added #ifdef to remove some redundancy in png_malloc_default().
  Use png_malloc instead of png_zalloc to allocate the pallete.

Version 1.0.16rc1 and 1.2.6rc1 [August 4, 2004]
  Fixed buffer overflow vulnerability (CVE-2004-0597) in png_handle_tRNS().
  Fixed NULL dereference vulnerability (CVE-2004-0598) in png_handle_iCCP().
  Fixed integer overflow vulnerability (CVE-2004-0599) in png_read_png().
  Fixed some harmless bugs in png_handle_sBIT, etc, that would cause
    duplicate chunk types to go undetected.
  Fixed some timestamps in the -config version
  Rearranged order of processing of color types in png_handle_tRNS().
  Added ROWBYTES macro to calculate rowbytes without integer overflow.
  Updated makefile.darwin and removed makefile.macosx from scripts directory.
  Imposed default one million column, one-million row limits on the image
    dimensions, and added png_set_user_limits() function to override them.
  Revised use of PNG_SET_USER_LIMITS_SUPPORTED macro.
  Fixed wrong cast of returns from png_get_user_width|height_max().
  Changed some "keep the compiler happy" from empty statements to returns,
  Revised libpng.txt to remove 1.2.x stuff from the 1.0.x distribution

Version 1.0.16rc2 and 1.2.6rc2 [August 7, 2004]
  Revised makefile.darwin and makefile.solaris.  Removed makefile.macosx.
  Revised pngtest's png_debug_malloc() to use png_malloc() instead of
    png_malloc_default() which is not supposed to be exported.
  Fixed off-by-one error in one of the conversions to PNG_ROWBYTES() in
    pngpread.c.  Bug was introduced in 1.2.6rc1.
  Fixed bug in RGB to RGBX transformation introduced in 1.2.6rc1.
  Fixed old bug in RGB to Gray transformation.
  Fixed problem with 64-bit compilers by casting arguments to abs()
    to png_int_32.
  Changed "ln -sf" to "ln -f -s" in three makefiles (solaris, sco, so9).
  Changed "HANDLE_CHUNK_*" to "PNG_HANDLE_CHUNK_*" (Cosmin)
  Added "-@/bin/rm -f $(DL)/$(LIBNAME).so.$(PNGMAJ)" to 15 *NIX makefiles.
  Added code to update the row_info->colortype in png_do_read_filler() (MSB).

Version 1.0.16rc3 and 1.2.6rc3 [August 9, 2004]
  Eliminated use of "abs()" in testing cHRM and gAMA values, to avoid
    trouble with some 64-bit compilers.  Created PNG_OUT_OF_RANGE() macro.
  Revised documentation of png_set_keep_unknown_chunks().
  Check handle_as_unknown status in pngpread.c, as in pngread.c previously.
  Moved  "PNG_HANDLE_CHUNK_*" macros out of PNG_INTERNAL section of png.h
  Added "rim" definitions for CONST4 and CONST6 in pnggccrd.c

Version 1.0.16rc4 and 1.2.6rc4 [August 10, 2004]
  Fixed mistake in pngtest.c introduced in 1.2.6rc2 (declaration of
    "pinfo" was out of place).

Version 1.0.16rc5 and 1.2.6rc5 [August 10, 2004]
  Moved  "PNG_HANDLE_CHUNK_*" macros out of PNG_ASSEMBLER_CODE_SUPPORTED
    section of png.h where they were inadvertently placed in version rc3.

Version 1.2.6 and 1.0.16 [August 15, 2004]
  Revised pngtest so memory allocation testing is only done when PNG_DEBUG==1.

Version 1.2.7beta1 [August 26, 2004]
  Removed unused pngasmrd.h file.
  Removed references to uu.net for archived files.  Added references to
    PNG Spec (second edition) and the PNG ISO/IEC Standard.
  Added "test-dd" target in 15 makefiles, to run pngtest in DESTDIR.
  Fixed bug with "optimized window size" in the IDAT datastream, that
    causes libpng to write PNG files with incorrect zlib header bytes.

Version 1.2.7beta2 [August 28, 2004]
  Fixed bug with sCAL chunk and big-endian machines (David Munro).
  Undid new code added in 1.2.6rc2 to update the color_type in
    png_set_filler().
  Added png_set_add_alpha() that updates color type.

Version 1.0.17rc1 and 1.2.7rc1 [September 4, 2004]
  Revised png_set_strip_filler() to not remove alpha if color_type has alpha.

Version 1.2.7 and 1.0.17 [September 12, 2004]
  Added makefile.hp64
  Changed projects/msvc/png32ms.def to scripts/png32ms.def in makefile.cygwin

Version 1.2.8beta1 [November 1, 2004]
  Fixed bug in png_text_compress() that would fail to complete a large block.
  Fixed bug, introduced in libpng-1.2.7, that overruns a buffer during
    strip alpha operation in png_do_strip_filler().
  Added PNG_1_2_X definition in pngconf.h
  Use #ifdef to comment out png_info_init in png.c and png_read_init in
    pngread.c (as of 1.3.0)

Version 1.2.8beta2 [November 2, 2004]
  Reduce color_type to a nonalpha type after strip alpha operation in
    png_do_strip_filler().

Version 1.2.8beta3 [November 3, 2004]
  Revised definitions of PNG_MAX_UINT_32, PNG_MAX_SIZE, and PNG_MAXSUM

Version 1.2.8beta4 [November 12, 2004]
  Fixed (again) definition of PNG_LIBPNG_VER_DLLNUM in png.h (Cosmin).
  Added PNG_LIBPNG_BUILD_PRIVATE in png.h (Cosmin).
  Set png_ptr->zstream.data_type to Z_BINARY, to avoid unnecessary detection
    of data type in deflate (Cosmin).
  Deprecated but continue to support SPECIALBUILD and PRIVATEBUILD in favor of
    PNG_LIBPNG_BUILD_SPECIAL_STRING and PNG_LIBPNG_BUILD_PRIVATE_STRING.

Version 1.2.8beta5 [November 20, 2004]
  Use png_ptr->flags instead of png_ptr->transformations to pass
    PNG_STRIP_ALPHA info to png_do_strip_filler(), to preserve ABI
    compatibility.
  Revised handling of SPECIALBUILD, PRIVATEBUILD,
    PNG_LIBPNG_BUILD_SPECIAL_STRING and PNG_LIBPNG_BUILD_PRIVATE_STRING.

Version 1.2.8rc1 [November 24, 2004]
  Moved handling of BUILD macros from pngconf.h to png.h
  Added definition of PNG_LIBPNG_BASE_TYPE in png.h, inadvertently
    omitted from beta5.
  Revised scripts/pngw32.rc
  Despammed mailing addresses by masking "@" with "at".
  Inadvertently installed a supposedly faster test version of pngrutil.c

Version 1.2.8rc2 [November 26, 2004]
  Added two missing "\" in png.h
  Change tests in pngread.c and pngpread.c to
    if (png_ptr->transformations || (png_ptr->flags&PNG_FLAG_STRIP_ALPHA))
       png_do_read_transformations(png_ptr);

Version 1.2.8rc3 [November 28, 2004]
  Reverted pngrutil.c to version libpng-1.2.8beta5.
  Added scripts/makefile.elf with supporting code in pngconf.h for symbol
    versioning (John Bowler).

Version 1.2.8rc4 [November 29, 2004]
  Added projects/visualc7 (Simon-pierre).

Version 1.2.8rc5 [November 29, 2004]
  Fixed new typo in scripts/pngw32.rc

Version 1.2.8 [December 3, 2004]
  Removed projects/visualc7, added projects/visualc71.

Version 1.2.9beta1 [February 21, 2006]
  Initialized some structure members in pngwutil.c to avoid gcc-4.0.0 complaints
  Revised man page and libpng.txt to make it clear that one should not call
    png_read_end or png_write_end after png_read_png or png_write_png.
  Updated references to png-mng-implement mailing list.
  Fixed an incorrect typecast in pngrutil.c
  Added PNG_NO_READ_SUPPORTED conditional for making a write-only library.
  Added PNG_NO_WRITE_INTERLACING_SUPPORTED conditional.
  Optimized alpha-inversion loops in pngwtran.c
  Moved test for nonzero gamma outside of png_build_gamma_table() in pngrtran.c
  Make sure num_trans is <= 256 before copying data in png_set_tRNS().
  Make sure num_palette is <= 256 before copying data in png_set_PLTE().
  Interchanged order of write_swap_alpha and write_invert_alpha transforms.
  Added parentheses in the definition of PNG_LIBPNG_BUILD_TYPE (Cosmin).
  Optimized zlib window flag (CINFO) in contrib/pngsuite/*.png (Cosmin).
  Updated scripts/makefile.bc32 for Borland C++ 5.6 (Cosmin).
  Exported png_get_uint_32, png_save_uint_32, png_get_uint_16, png_save_uint_16,
    png_get_int_32, png_save_int_32, png_get_uint_31 (Cosmin).
  Added type cast (png_byte) in png_write_sCAL() (Cosmin).
  Fixed scripts/makefile.cygwin (Christian Biesinger, Cosmin).
  Default iTXt support was inadvertently enabled.

Version 1.2.9beta2 [February 21, 2006]
  Check for png_rgb_to_gray and png_gray_to_rgb read transformations before
    checking for png_read_dither in pngrtran.c
  Revised checking of chromaticity limits to accommodate extended RGB
    colorspace (John Denker).
  Changed line endings in some of the project files to CRLF, even in the
    "Unix" tar distributions (Cosmin).
  Made png_get_int_32 and png_save_int_32 always available (Cosmin).
  Updated scripts/pngos2.def, scripts/pngw32.def and projects/wince/png32ce.def
    with the newly exported functions.
  Eliminated distributions without the "configure" script.
  Updated INSTALL instructions.

Version 1.2.9beta3 [February 24, 2006]
  Fixed CRCRLF line endings in contrib/visupng/VisualPng.dsp
  Made libpng.pc respect EXEC_PREFIX (D. P. Kreil, J. Bowler)
  Removed reference to pngasmrd.h from Makefile.am
  Renamed CHANGES to ChangeLog.
  Renamed LICENSE to COPYING.
  Renamed ANNOUNCE to NEWS.
  Created AUTHORS file.

Version 1.2.9beta4 [March 3, 2006]
  Changed definition of PKGCONFIG from $prefix/lib to $libdir in configure.ac
  Reverted to filenames LICENSE and ANNOUNCE; removed AUTHORS and COPYING.
  Removed newline from the end of some error and warning messages.
  Removed test for sqrt() from configure.ac and configure.
  Made swap tables in pngtrans.c PNG_CONST (Carlo Bramix).
  Disabled default iTXt support that was inadvertently enabled in
    libpng-1.2.9beta1.
  Added "OS2" to list of systems that don't need underscores, in pnggccrd.c
  Removed libpng version and date from *.c files.

Version 1.2.9beta5 [March 4, 2006]
  Removed trailing blanks from source files.
  Put version and date of latest change in each source file, and changed
    copyright year accordingly.
  More cleanup of configure.ac, Makefile.am, and associated scripts.
  Restored scripts/makefile.elf which was inadvertently deleted.

Version 1.2.9beta6 [March 6, 2006]
  Fixed typo (RELEASE) in configuration files.

Version 1.2.9beta7 [March 7, 2006]
  Removed libpng.vers and libpng.sym from libpng12_la_SOURCES in Makefile.am
  Fixed inconsistent #ifdef's around png_sig_bytes() and png_set_sCAL_s()
    in png.h.
  Updated makefile.elf as suggested by debian.
  Made cosmetic changes to some makefiles, adding LN_SF and other macros.
  Made some makefiles accept "exec_prefix".

Version 1.2.9beta8 [March 9, 2006]
  Fixed some "#if defined (..." which should be "#if defined(..."
    Bug introduced in libpng-1.2.8.
  Fixed inconsistency in definition of png_default_read_data()
  Restored blank that was lost from makefile.sggcc "clean" target in beta7.
  Revised calculation of "current" and "major" for irix in ltmain.sh
  Changed "mkdir" to "MKDIR_P" in some makefiles.
  Separated PNG_EXPAND and PNG_EXPAND_tRNS.
  Added png_set_expand_gray_1_2_4_to_8() and deprecated
    png_set_gray_1_2_4_to_8() which also expands tRNS to alpha.

Version 1.2.9beta9 [March 10, 2006]
  Include "config.h" in pngconf.h when available.
  Added some checks for NULL png_ptr or NULL info_ptr (timeless)

Version 1.2.9beta10 [March 20, 2006]
  Removed extra CR from contrib/visualpng/VisualPng.dsw (Cosmin)
  Made pnggccrd.c PIC-compliant (Christian Aichinger).
  Added makefile.mingw (Wolfgang Glas).
  Revised pngconf.h MMX checking.

Version 1.2.9beta11 [March 22, 2006]
  Fixed out-of-order declaration in pngwrite.c that was introduced in beta9
  Simplified some makefiles by using LIBSO, LIBSOMAJ, and LIBSOVER macros.

Version 1.2.9rc1 [March 31, 2006]
  Defined PNG_USER_PRIVATEBUILD when including "pngusr.h" (Cosmin).
  Removed nonsensical assertion check from pngtest.c (Cosmin).

Version 1.2.9 [April 14, 2006]
  Revised makefile.beos and added "none" selector in ltmain.sh

Version 1.2.10beta1 [April 15, 2006]
  Renamed "config.h" to "png_conf.h" and revised Makefile.am to add
    -DPNG_BUILDING_LIBPNG to compile directive, and modified pngconf.h
    to include png_conf.h only when PNG_BUILDING_LIBPNG is defined.

Version 1.2.10beta2 [April 15, 2006]
  Manually updated Makefile.in and configure.  Changed png_conf.h.in
    back to config.h.

Version 1.2.10beta3 [April 15, 2006]
  Change png_conf.h back to config.h in pngconf.h.

Version 1.2.10beta4 [April 16, 2006]
  Change PNG_BUILDING_LIBPNG to PNG_CONFIGURE_LIBPNG in config/Makefile*.

Version 1.2.10beta5 [April 16, 2006]
  Added a configure check for compiling assembler code in pnggccrd.c

Version 1.2.10beta6 [April 17, 2006]
  Revised the configure check for pnggccrd.c
  Moved -DPNG_CONFIGURE_LIBPNG into @LIBPNG_DEFINES@
  Added @LIBPNG_DEFINES@ to arguments when building libpng.sym

Version 1.2.10beta7 [April 18, 2006]
  Change "exec_prefix=$prefix" to "exec_prefix=$(prefix)" in makefiles.

Version 1.2.10rc1 [April 19, 2006]
  Ensure pngconf.h doesn't define both PNG_USE_PNGGCCRD and PNG_USE_PNGVCRD
  Fixed "LN_FS" typo in makefile.sco and makefile.solaris.

Version 1.2.10rc2 [April 20, 2006]
  Added a backslash between -DPNG_CONFIGURE_LIBPNG and -DPNG_NO_ASSEMBLER_CODE
   in configure.ac and configure
  Made the configure warning about versioned symbols less arrogant.

Version 1.2.10rc3 [April 21, 2006]
  Added a note in libpng.txt that png_set_sig_bytes(8) can be used when
    writing an embedded PNG without the 8-byte signature.
  Revised makefiles and configure to avoid making links to libpng.so.*

Version 1.2.10 [April 23, 2006]
  Reverted configure to "rc2" state.

Version 1.2.11beta1 [May 31, 2006]
  scripts/libpng.pc.in contained "configure" style version info and would
    not work with makefiles.
  The shared-library makefiles were linking to libpng.so.0 instead of
    libpng.so.3 compatibility as the library.

Version 1.2.11beta2 [June 2, 2006]
  Increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
    buffer overflow.
  Fixed bug in example.c (png_set_palette_rgb -> png_set_palette_to_rgb)

Version 1.2.11beta3 [June 5, 2006]
  Prepended "#! /bin/sh" to ltmail.sh and contrib/pngminus/*.sh (Cosmin).
  Removed the accidental leftover Makefile.in~ (Cosmin).
  Avoided potential buffer overflow and optimized buffer in
    png_write_sCAL(), png_write_sCAL_s() (Cosmin).
  Removed the include directories and libraries from CFLAGS and LDFLAGS
    in scripts/makefile.gcc (Nelson A. de Oliveira, Cosmin).

Version 1.2.11beta4 [June 6, 2006]
  Allow zero-length IDAT chunks after the entire zlib datastream, but not
    after another intervening chunk type.

Version 1.0.19rc1, 1.2.11rc1 [June 13, 2006]
  Deleted extraneous square brackets from [config.h] in configure.ac

Version 1.0.19rc2, 1.2.11rc2 [June 14, 2006]
  Added prototypes for PNG_INCH_CONVERSIONS functions to png.h
  Revised INSTALL and autogen.sh
  Fixed typo in several makefiles (-W1 should be -Wl)
  Added typedef for png_int_32 and png_uint_32 on 64-bit systems.

Version 1.0.19rc3, 1.2.11rc3 [June 15, 2006]
  Removed the new typedefs for 64-bit systems (delay until version 1.4.0)
  Added one zero element to png_gamma_shift[] array in pngrtran.c to avoid
    reading out of bounds.

Version 1.0.19rc4, 1.2.11rc4 [June 15, 2006]
  Really removed the new typedefs for 64-bit systems.

Version 1.0.19rc5, 1.2.11rc5 [June 22, 2006]
  Removed png_sig_bytes entry from scripts/pngw32.def

Version 1.0.19, 1.2.11 [June 26, 2006]
  None.

Version 1.0.20, 1.2.12 [June 27, 2006]
  Really increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
    buffer overflow.

Version 1.2.13beta1 [October 2, 2006]
  Removed AC_FUNC_MALLOC from configure.ac
  Work around Intel-Mac compiler bug by setting PNG_NO_MMX_CODE in pngconf.h
  Change "logical" to "bitwise" throughout documentation.
  Detect and fix attempt to write wrong iCCP profile length (CVE-2006-7244)

Version 1.0.21, 1.2.13 [November 14, 2006]
  Fix potential buffer overflow in sPLT chunk handler.
  Fix Makefile.am to not try to link to noexistent files.
  Check all exported functions for NULL png_ptr.

Version 1.2.14beta1 [November 17, 2006]
  Relocated three misplaced tests for NULL png_ptr.
  Built Makefile.in with automake-1.9.6 instead of 1.9.2.
  Build configure with autoconf-2.60 instead of 2.59

Version 1.2.14beta2 [November 17, 2006]
  Added some typecasts in png_zalloc().

Version 1.2.14rc1 [November 20, 2006]
  Changed "strtod" to "png_strtod" in pngrutil.c

Version 1.0.22, 1.2.14    [November 27, 2006]
  Added missing "$(srcdir)" in Makefile.am and Makefile.in

Version 1.2.15beta1 [December 3, 2006]
  Generated configure with autoconf-2.61 instead of 2.60
  Revised configure.ac to update libpng.pc and libpng-config.

Version 1.2.15beta2 [December 3, 2006]
  Always export MMX asm functions, just stubs if not building pnggccrd.c

Version 1.2.15beta3 [December 4, 2006]
  Add "png_bytep" typecast to profile while calculating length in pngwutil.c

Version 1.2.15beta4 [December 7, 2006]
  Added scripts/CMakeLists.txt
  Changed PNG_NO_ASSEMBLER_CODE to PNG_NO_MMX_CODE in scripts, like 1.4.0beta

Version 1.2.15beta5 [December 7, 2006]
  Changed some instances of PNG_ASSEMBLER_* to PNG_MMX_* in pnggccrd.c
  Revised scripts/CMakeLists.txt

Version 1.2.15beta6 [December 13, 2006]
  Revised scripts/CMakeLists.txt and configure.ac

Version 1.2.15rc1 [December 18, 2006]
  Revised scripts/CMakeLists.txt

Version 1.2.15rc2 [December 21, 2006]
  Added conditional #undef jmpbuf in pngtest.c to undo #define in AIX headers.
  Added scripts/makefile.nommx

Version 1.2.15rc3 [December 25, 2006]
  Fixed shared library numbering error that was introduced in 1.2.15beta6.

Version 1.2.15rc4 [December 27, 2006]
  Fixed handling of rgb_to_gray when png_ptr->color.gray isn't set.

Version 1.2.15rc5 [December 31, 2006]
  Revised handling of rgb_to_gray.

Version 1.2.15 [January 5, 2007]
  Added some (unsigned long) typecasts in pngtest.c to avoid printing errors.

Version 1.2.16beta1 [January 6, 2007]
  Fix bugs in makefile.nommx

Version 1.2.16beta2 [January 16, 2007]
  Revised scripts/CMakeLists.txt

Version 1.2.16 [January 31, 2007]
  No changes.

Version 1.2.17beta1 [March 6, 2007]
  Revised scripts/CMakeLists.txt to install both shared and static libraries.
  Deleted a redundant line from pngset.c.

Version 1.2.17beta2 [April 26, 2007]
  Relocated misplaced test for png_ptr == NULL in pngpread.c
  Change "==" to "&" for testing PNG_RGB_TO_GRAY_ERR & PNG_RGB_TO_GRAY_WARN
    flags.
  Changed remaining instances of PNG_ASSEMBLER_* to PNG_MMX_*
  Added pngerror() when write_IHDR fails in deflateInit2().
  Added "const" to some array declarations.
  Mention examples of libpng usage in the libpng*.txt and libpng.3 documents.

Version 1.2.17rc1 [May 4, 2007]
  No changes.

Version 1.2.17rc2 [May 8, 2007]
  Moved several PNG_HAVE_* macros out of PNG_INTERNAL because applications
    calling set_unknown_chunk_location() need them.
  Changed transformation flag from PNG_EXPAND_tRNS to PNG_EXPAND in
    png_set_expand_gray_1_2_4_to_8().
  Added png_ptr->unknown_chunk to hold working unknown chunk data, so it
    can be free'ed in case of error.  Revised unknown chunk handling in
    pngrutil.c and pngpread.c to use this structure.

Version 1.2.17rc3 [May 8, 2007]
  Revised symbol-handling in configure script.

Version 1.2.17rc4 [May 10, 2007]
  Revised unknown chunk handling to avoid storing unknown critical chunks.

Version 1.0.25 [May 15, 2007]
Version 1.2.17 [May 15, 2007]
  Added "png_ptr->num_trans=0" before error return in png_handle_tRNS,
    to eliminate a vulnerability (CVE-2007-2445, CERT VU#684664)

Version 1.0.26 [May 15, 2007]
Version 1.2.18 [May 15, 2007]
  Reverted the libpng-1.2.17rc3 change to symbol-handling in configure script

Version 1.2.19beta1 [May 18, 2007]
  Changed "const static" to "static PNG_CONST" everywhere, mostly undoing
    change of libpng-1.2.17beta2.  Changed other "const" to "PNG_CONST"
  Changed some handling of unused parameters, to avoid compiler warnings.
    "if (unused == NULL) return;" becomes "unused = unused".

Version 1.2.19beta2 [May 18, 2007]
  Only use the valid bits of tRNS value in png_do_expand() (Brian Cartier)

Version 1.2.19beta3 [May 19, 2007]
  Add some "png_byte" typecasts in png_check_keyword() and write new_key
  instead of key in zTXt chunk (Kevin Ryde).

Version 1.2.19beta4 [May 21, 2007]
  Add png_snprintf() function and use it in place of sprint() for improved
    defense against buffer overflows.

Version 1.2.19beta5 [May 21, 2007]
  Fixed png_handle_tRNS() to only use the valid bits of tRNS value.
  Changed handling of more unused parameters, to avoid compiler warnings.
  Removed some PNG_CONST in pngwutil.c to avoid compiler warnings.

Version 1.2.19beta6 [May 22, 2007]
  Added some #ifdef PNG_MMX_CODE_SUPPORTED where needed in pngvcrd.c
  Added a special "_MSC_VER" case that defines png_snprintf to _snprintf

Version 1.2.19beta7 [May 22, 2007]
  Squelched png_squelch_warnings() in pnggccrd.c and added
    an #ifdef PNG_MMX_CODE_SUPPORTED block around the declarations that caused
    the warnings that png_squelch_warnings was squelching.

Version 1.2.19beta8 [May 22, 2007]
  Removed __MMX__ from test in pngconf.h.

Version 1.2.19beta9 [May 23, 2007]
  Made png_squelch_warnings() available via PNG_SQUELCH_WARNINGS macro.
  Revised png_squelch_warnings() so it might work.
  Updated makefile.sgcc and makefile.solaris; added makefile.solaris-x86.

Version 1.2.19beta10 [May 24, 2007]
  Resquelched png_squelch_warnings(), use "__attribute__((used))" instead.

Version 1.4.0beta1 [April 20, 2006]
  Enabled iTXt support (changes png_struct, thus requires so-number change).
  Cleaned up PNG_ASSEMBLER_CODE_SUPPORTED vs PNG_MMX_CODE_SUPPORTED
  Eliminated PNG_1_0_X and PNG_1_2_X macros.
  Removed deprecated functions png_read_init, png_write_init, png_info_init,
    png_permit_empty_plte, png_set_gray_1_2_4_to_8, png_check_sig, and
    removed the deprecated macro PNG_MAX_UINT.
  Moved "PNG_INTERNAL" parts of png.h and pngconf.h into pngintrn.h
  Removed many WIN32_WCE #ifdefs (Cosmin).
  Reduced dependency on C-runtime library when on Windows (Simon-Pierre)
  Replaced sprintf() with png_sprintf() (Simon-Pierre)

Version 1.4.0beta2 [April 20, 2006]
  Revised makefiles and configure to avoid making links to libpng.so.*
  Moved some leftover MMX-related defines from pngconf.h to pngintrn.h
  Updated scripts/pngos2.def, pngw32.def, and projects/wince/png32ce.def

Version 1.4.0beta3 [May 10, 2006]
  Updated scripts/pngw32.def to comment out MMX functions.
  Added PNG_NO_GET_INT_32 and PNG_NO_SAVE_INT_32 macros.
  Scripts/libpng.pc.in contained "configure" style version info and would
    not work with makefiles.
  Revised pngconf.h and added pngconf.h.in, so makefiles and configure can
    pass defines to libpng and applications.

Version 1.4.0beta4 [May 11, 2006]
  Revised configure.ac, Makefile.am, and many of the makefiles to write
    their defines in pngconf.h.

Version 1.4.0beta5 [May 15, 2006]
  Added a missing semicolon in Makefile.am and Makefile.in
  Deleted extraneous square brackets from configure.ac

Version 1.4.0beta6 [June 2, 2006]
  Increased sprintf buffer from 50 to 52 chars in pngrutil.c to avoid
    buffer overflow.
  Changed sonum from 0 to 1.
  Removed unused prototype for png_check_sig() from png.h

Version 1.4.0beta7 [June 16, 2006]
  Exported png_write_sig (Cosmin).
  Optimized buffer in png_handle_cHRM() (Cosmin).
  Set pHYs = 2835 x 2835 pixels per meter, and added
    sCAL = 0.352778e-3 x 0.352778e-3 meters, in pngtest.png (Cosmin).
  Added png_set_benign_errors(), png_benign_error(), png_chunk_benign_error().
  Added typedef for png_int_32 and png_uint_32 on 64-bit systems.
  Added "(unsigned long)" typecast on png_uint_32 variables in printf lists.

Version 1.4.0beta8 [June 22, 2006]
  Added demonstration of user chunk support in pngtest.c, to support the
    public sTER chunk and a private vpAg chunk.

Version 1.4.0beta9 [July 3, 2006]
  Removed ordinals from scripts/pngw32.def and removed png_info_int and
    png_set_gray_1_2_4_to_8 entries.
  Inline call of png_get_uint_32() in png_get_uint_31().
  Use png_get_uint_31() to get vpAg width and height in pngtest.c
  Removed WINCE and Netware projects.
  Removed standalone Y2KINFO file.

Version 1.4.0beta10 [July 12, 2006]
  Eliminated automatic copy of pngconf.h to pngconf.h.in from configure and
    some makefiles, because it was not working reliably.  Instead, distribute
    pngconf.h.in along with pngconf.h and cause configure and some of the
    makefiles to update pngconf.h from pngconf.h.in.
  Added pngconf.h to DEPENDENCIES in Makefile.am

Version 1.4.0beta11 [August 19, 2006]
  Removed AC_FUNC_MALLOC from configure.ac.
  Added a warning when writing iCCP profile with mismatched profile length.
  Patched pnggccrd.c to assemble on x86_64 platforms.
  Moved chunk header reading into a separate function png_read_chunk_header()
    in pngrutil.c.  The chunk header (len+sig) is now serialized in a single
    operation (Cosmin).
  Implemented support for I/O states. Added png_ptr member io_state, and
    functions png_get_io_chunk_name() and png_get_io_state() in pngget.c
    (Cosmin).
  Added png_get_io_chunk_name and png_get_io_state to scripts/*.def (Cosmin).
  Renamed scripts/pngw32.* to scripts/pngwin.* (Cosmin).
  Removed the include directories and libraries from CFLAGS and LDFLAGS
    in scripts/makefile.gcc (Cosmin).
  Used png_save_uint_32() to set vpAg width and height in pngtest.c (Cosmin).
  Cast to proper type when getting/setting vpAg units in pngtest.c (Cosmin).
  Added pngintrn.h to the Visual C++ projects (Cosmin).
  Removed scripts/list (Cosmin).
  Updated copyright year in scripts/pngwin.def (Cosmin).
  Removed PNG_TYPECAST_NULL and used standard NULL consistently (Cosmin).
  Disallowed the user to redefine png_size_t, and enforced a consistent use
    of png_size_t across libpng (Cosmin).
  Changed the type of png_ptr->rowbytes, PNG_ROWBYTES() and friends
    to png_size_t (Cosmin).
  Removed png_convert_size() and replaced png_sizeof with sizeof (Cosmin).
  Removed some unnecessary type casts (Cosmin).
  Changed prototype of png_get_compression_buffer_size() and
    png_set_compression_buffer_size() to work with png_size_t instead of
    png_uint_32 (Cosmin).
  Removed png_memcpy_check() and png_memset_check() (Cosmin).
  Fixed a typo (png_byte --> png_bytep) in libpng.3 and libpng.txt (Cosmin).
  Clarified that png_zalloc() does not clear the allocated memory,
    and png_zalloc() and png_zfree() cannot be PNGAPI (Cosmin).
  Renamed png_mem_size_t to png_alloc_size_t, fixed its definition in
    pngconf.h, and used it in all memory allocation functions (Cosmin).
  Renamed pngintrn.h to pngpriv.h, added a comment at the top of the file
    mentioning that the symbols declared in that file are private, and
    updated the scripts and the Visual C++ projects accordingly (Cosmin).
  Removed circular references between pngconf.h and pngconf.h.in in
    scripts/makefile.vc*win32 (Cosmin).
  Removing trailing '.' from the warning and error messages (Cosmin).
  Added pngdefs.h that is built by makefile or configure, instead of
    pngconf.h.in (Glenn).
  Detect and fix attempt to write wrong iCCP profile length.

Version 1.4.0beta12 [October 19, 2006]
  Changed "logical" to "bitwise" in the documentation.
  Work around Intel-Mac compiler bug by setting PNG_NO_MMX_CODE in pngconf.h
  Add a typecast to stifle compiler warning in pngrutil.c

Version 1.4.0beta13 [November 10, 2006]
  Fix potential buffer overflow in sPLT chunk handler.
  Fix Makefile.am to not try to link to noexistent files.

Version 1.4.0beta14 [November 15, 2006]
  Check all exported functions for NULL png_ptr.

Version 1.4.0beta15 [November 17, 2006]
  Relocated two misplaced tests for NULL png_ptr.
  Built Makefile.in with automake-1.9.6 instead of 1.9.2.
  Build configure with autoconf-2.60 instead of 2.59
  Add "install: all" in Makefile.am so "configure; make install" will work.

Version 1.4.0beta16 [November 17, 2006]
  Added a typecast in png_zalloc().

Version 1.4.0beta17 [December 4, 2006]
  Changed "new_key[79] = '\0';" to "(*new_key)[79] = '\0';" in pngwutil.c
  Add "png_bytep" typecast to profile while calculating length in pngwutil.c

Version 1.4.0beta18 [December 7, 2006]
  Added scripts/CMakeLists.txt

Version 1.4.0beta19 [May 16, 2007]
  Revised scripts/CMakeLists.txt
  Rebuilt configure and Makefile.in with newer tools.
  Added conditional #undef jmpbuf in pngtest.c to undo #define in AIX headers.
  Added scripts/makefile.nommx

Version 1.4.0beta20 [July 9, 2008]
  Moved several PNG_HAVE_* macros from pngpriv.h to png.h because applications
    calling set_unknown_chunk_location() need them.
  Moved several macro definitions from pngpriv.h to pngconf.h
  Merge with changes to the 1.2.X branch, as of 1.2.30beta04.
  Deleted all use of the MMX assembler code and Intel-licensed optimizations.
  Revised makefile.mingw

Version 1.4.0beta21 [July 21, 2008]
  Moved local array "chunkdata" from pngrutil.c to the png_struct, so
    it will be freed by png_read_destroy() in case of a read error (Kurt
    Christensen).

Version 1.4.0beta22 [July 21, 2008]
  Change "purpose" and "buffer" to png_ptr->chunkdata to avoid memory leaking.

Version 1.4.0beta23 [July 22, 2008]
  Change "chunkdata = NULL" to "png_ptr->chunkdata = NULL" several places in
    png_decompress_chunk().

Version 1.4.0beta24 [July 25, 2008]
  Change all remaining "chunkdata" to "png_ptr->chunkdata" in
    png_decompress_chunk(), and remove "chunkdata" from parameter list.
  Put a call to png_check_chunk_name() in png_read_chunk_header().
  Revised png_check_chunk_name() to reject a name with a lowercase 3rd byte.
  Removed two calls to png_check_chunk_name() occurring later in the process.
  Define PNG_NO_ERROR_NUMBERS by default in pngconf.h

Version 1.4.0beta25 [July 30, 2008]
  Added a call to png_check_chunk_name() in pngpread.c
  Reverted png_check_chunk_name() to accept a name with a lowercase 3rd byte.
  Added png_push_have_buffer() function to pngpread.c
  Eliminated PNG_BIG_ENDIAN_SUPPORTED and associated png_get_* macros.
  Made inline expansion of png_get_*() optional with PNG_USE_READ_MACROS.
  Eliminated all PNG_USELESS_TESTS and PNG_CORRECT_PALETTE_SUPPORTED code.
  Synced contrib directory and configure files with libpng-1.2.30beta06.
  Eliminated no-longer-used pngdefs.h (but it's still built in the makefiles)
  Relocated a misplaced "#endif /* PNG_NO_WRITE_FILTER */" in pngwutil.c

Version 1.4.0beta26 [August 4, 2008]
  Removed png_push_have_buffer() function in pngpread.c.  It increased the
    compiled library size slightly.
  Changed "-Wall" to "-W -Wall" in the CFLAGS in all makefiles (Cosmin Truta)
  Declared png_ptr "volatile" in pngread.c and pngwrite.c to avoid warnings.
  Updated contrib/visupng/cexcept.h to version 2.0.1
  Added PNG_LITERAL_CHARACTER macros for #, [, and ].

Version 1.4.0beta27 [August 5, 2008]
  Revised usage of PNG_LITERAL_SHARP in pngerror.c.
  Moved newline character from individual png_debug messages into the
    png_debug macros.
  Allow user to #define their own png_debug, png_debug1, and png_debug2.

Version 1.4.0beta28 [August 5, 2008]
  Revised usage of PNG_LITERAL_SHARP in pngerror.c.
  Added PNG_STRING_NEWLINE macro

Version 1.4.0beta29 [August 9, 2008]
  Revised usage of PNG_STRING_NEWLINE to work on non-ISO compilers.
  Added PNG_STRING_COPYRIGHT macro.
  Added non-ISO versions of png_debug macros.

Version 1.4.0beta30 [August 14, 2008]
  Added premultiplied alpha feature (Volker Wiendl).

Version 1.4.0beta31 [August 18, 2008]
  Moved png_set_premultiply_alpha from pngtrans.c to pngrtran.c
  Removed extra crc check at the end of png_handle_cHRM().  Bug introduced
    in libpng-1.4.0beta20.

Version 1.4.0beta32 [August 19, 2008]
  Added PNG_WRITE_FLUSH_SUPPORTED block around new png_flush() call.
  Revised PNG_NO_STDIO version of png_write_flush()

Version 1.4.0beta33 [August 20, 2008]
  Added png_get|set_chunk_cache_max() to limit the total number of sPLT,
    text, and unknown chunks that can be stored.

Version 1.4.0beta34 [September 6, 2008]
  Shortened tIME_string to 29 bytes in pngtest.c
  Fixed off-by-one error introduced in png_push_read_zTXt() function in
    libpng-1.2.30beta04/pngpread.c (Harald van Dijk)

Version 1.4.0beta35 [October 6, 2008]
  Changed "trans_values" to "trans_color".
  Changed so-number from 0 to 14.  Some OS do not like 0.
  Revised makefile.darwin to fix shared library numbering.
  Change png_set_gray_1_2_4_to_8() to png_set_expand_gray_1_2_4_to_8()
    in example.c (debian bug report)

Version 1.4.0beta36 [October 25, 2008]
  Sync with tEXt vulnerability fix in libpng-1.2.33rc02.

Version 1.4.0beta37 [November 13, 2008]
  Added png_check_cHRM in png.c and moved checking from pngget.c, pngrutil.c,
    and pngwrite.c

Version 1.4.0beta38 [November 22, 2008]
  Added check for zero-area RGB cHRM triangle in png_check_cHRM() and
    png_check_cHRM_fixed().

Version 1.4.0beta39 [November 23, 2008]
  Revised png_warning() to write its message on standard output by default
    when warning_fn is NULL.

Version 1.4.0beta40 [November 24, 2008]
  Eliminated png_check_cHRM().  Instead, always use png_check_cHRM_fixed().
  In png_check_cHRM_fixed(), ensure white_y is > 0, and removed redundant
    check for all-zero coordinates that is detected by the triangle check.

Version 1.4.0beta41 [November 26, 2008]
  Fixed string vs pointer-to-string error in png_check_keyword().
  Rearranged test expressions in png_check_cHRM_fixed() to avoid internal
    overflows.
  Added PNG_NO_CHECK_cHRM conditional.

Version 1.4.0beta42, 43 [December 1, 2008]
  Merge png_debug with version 1.2.34beta04.

Version 1.4.0beta44 [December 6, 2008]
  Removed redundant check for key==NULL before calling png_check_keyword()
    to ensure that new_key gets initialized and removed extra warning
    (Merge with version 1.2.34beta05 -- Arvan Pritchard).

Version 1.4.0beta45 [December 9, 2008]
  In png_write_png(), respect the placement of the filler bytes in an earlier
    call to png_set_filler() (Jim Barry).

Version 1.4.0beta46 [December 10, 2008]
  Undid previous change and added PNG_TRANSFORM_STRIP_FILLER_BEFORE and
    PNG_TRANSFORM_STRIP_FILLER_AFTER conditionals and deprecated
    PNG_TRANSFORM_STRIP_FILLER (Jim Barry).

Version 1.4.0beta47 [December 15, 2008]
  Support for dithering was disabled by default, because it has never
    been well tested and doesn't work very well.  The code has not
    been removed, however, and can be enabled by building libpng with
    PNG_READ_DITHER_SUPPORTED defined.

Version 1.4.0beta48 [February 14, 2009]
  Added new exported function png_calloc().
  Combined several instances of png_malloc(); png_memset() into png_calloc().
  Removed prototype for png_freeptr() that was added in libpng-1.4.0beta24
    but was never defined.

Version 1.4.0beta49 [February 28, 2009]
  Added png_fileno() macro to pngconf.h, used in pngwio.c
  Corrected order of #ifdef's in png_debug definition in png.h
  Fixed bug introduced in libpng-1.4.0beta48 with the memset arguments
    for pcal_params.
  Fixed order of #ifdef directives in the png_debug defines in png.h
    (bug introduced in libpng-1.2.34/1.4.0beta29).
  Revised comments in png_set_read_fn() and png_set_write_fn().

Version 1.4.0beta50 [March 18, 2009]
  Use png_calloc() instead of png_malloc() to allocate big_row_buf when
    reading an interlaced file, to avoid a possible UMR.
  Undid revision of PNG_NO_STDIO version of png_write_flush().  Users
    having trouble with fflush() can build with PNG_NO_WRITE_FLUSH defined
    or supply their own flush_fn() replacement.
  Revised libpng*.txt and png.h documentation about use of png_write_flush()
    and png_set_write_fn().
  Removed fflush() from pngtest.c.
  Added "#define PNG_NO_WRITE_FLUSH" to contrib/pngminim/encoder/pngusr.h

Version 1.4.0beta51 [March 21, 2009]
  Removed new png_fileno() macro from pngconf.h .

Version 1.4.0beta52 [March 27, 2009]
  Relocated png_do_chop() ahead of building gamma tables in pngrtran.c
    This avoids building 16-bit gamma tables unnecessarily.
  Removed fflush() from pngtest.c.
  Added "#define PNG_NO_WRITE_FLUSH" to contrib/pngminim/encoder/pngusr.h
  Added a section on differences between 1.0.x and 1.2.x to libpng.3/libpng.txt

Version 1.4.0beta53 [April 1, 2009]
  Removed some remaining MMX macros from pngpriv.h
  Fixed potential memory leak of "new_name" in png_write_iCCP() (Ralph Giles)

Version 1.4.0beta54 [April 13, 2009]
  Added "ifndef PNG_SKIP_SETJMP_CHECK" block in pngconf.h to allow
    application code writers to bypass the check for multiple inclusion
    of setjmp.h when they know that it is safe to ignore the situation.
  Eliminated internal use of setjmp() in pngread.c and pngwrite.c
  Reordered ancillary chunks in pngtest.png to be the same as what
    pngtest now produces, and made some cosmetic changes to pngtest output.
  Eliminated deprecated png_read_init_3() and png_write_init_3() functions.

Version 1.4.0beta55 [April 15, 2009]
  Simplified error handling in pngread.c and pngwrite.c by putting
    the new png_read_cleanup() and png_write_cleanup() functions inline.

Version 1.4.0beta56 [April 25, 2009]
  Renamed "user_chunk_data" to "my_user_chunk_data" in pngtest.c to suppress
    "shadowed declaration" warning from gcc-4.3.3.
  Renamed "gamma" to "png_gamma" in pngset.c to avoid "shadowed declaration"
    warning about a global "gamma" variable in math.h on some platforms.

Version 1.4.0beta57 [May 2, 2009]
  Removed prototype for png_freeptr() that was added in libpng-1.4.0beta24
    but was never defined (again).
  Rebuilt configure scripts with autoconf-2.63 instead of 2.62
  Removed pngprefs.h and MMX from makefiles

Version 1.4.0beta58 [May 14, 2009]
  Changed pngw32.def to pngwin.def in makefile.mingw (typo was introduced
    in beta57).
  Clarified usage of sig_bit versus sig_bit_p in example.c (Vincent Torri)

Version 1.4.0beta59 [May 15, 2009]
  Reformated sources in libpng style (3-space intentation, comment format)
  Fixed typo in libpng docs (PNG_FILTER_AVE should be PNG_FILTER_AVG)
  Added sections about the git repository and our coding style to the
    documentation
  Relocated misplaced #endif in pngwrite.c, sCAL chunk handler.

Version 1.4.0beta60 [May 19, 2009]
  Conditionally compile png_read_finish_row() which is not used by
    progressive readers.
  Added contrib/pngminim/preader to demonstrate building minimal progressive
    decoder, based on contrib/gregbook with embedded libpng and zlib.

Version 1.4.0beta61 [May 20, 2009]
  In contrib/pngminim/*, renamed "makefile.std" to "makefile", since there
    is only one makefile in those directories, and revised the README files
    accordingly.
  More reformatting of comments, mostly to capitalize sentences.

Version 1.4.0beta62 [June 2, 2009]
  Added "#define PNG_NO_WRITE_SWAP" to contrib/pngminim/encoder/pngusr.h
    and "define PNG_NO_READ_SWAP" to decoder/pngusr.h and preader/pngusr.h
  Reformatted several remaining "else statement" into two lines.
  Added a section to the libpng documentation about using png_get_io_ptr()
    in configure scripts to detect the presence of libpng.

Version 1.4.0beta63 [June 15, 2009]
  Revised libpng*.txt and libpng.3 to mention calling png_set_IHDR()
    multiple times and to specify the sample order in the tRNS chunk,
    because the ISO PNG specification has a typo in the tRNS table.
  Changed several PNG_UNKNOWN_CHUNK_SUPPORTED to
    PNG_HANDLE_AS_UNKNOWN_SUPPORTED, to make the png_set_keep mechanism
    available for ignoring known chunks even when not saving unknown chunks.
  Adopted preference for consistent use of "#ifdef" and "#ifndef" versus
    "#if defined()" and "if !defined()" where possible.

Version 1.4.0beta64 [June 24, 2009]
  Eliminated PNG_LEGACY_SUPPORTED code.
  Moved the various unknown chunk macro definitions outside of the
    PNG_READ|WRITE_ANCILLARY_CHUNK_SUPPORTED blocks.

Version 1.4.0beta65 [June 26, 2009]
  Added a reference to the libpng license in each file.

Version 1.4.0beta66 [June 27, 2009]
  Refer to the libpng license instead of the libpng license in each file.

Version 1.4.0beta67 [July 6, 2009]
  Relocated INVERT_ALPHA within png_read_png() and png_write_png().
  Added high-level API transform PNG_TRANSFORM_GRAY_TO_RGB.
  Added an "xcode" project to the projects directory (Alam Arias).

Version 1.4.0beta68 [July 19, 2009]
  Avoid some tests in filter selection in pngwutil.c

Version 1.4.0beta69 [July 25, 2009]
  Simplified the new filter-selection test.  This runs faster in the
    common "PNG_ALL_FILTERS" and PNG_FILTER_NONE cases.
  Removed extraneous declaration from the new call to png_read_gray_to_rgb()
    (bug introduced in libpng-1.4.0beta67).
  Fixed up xcode project (Alam Arias)
  Added a prototype for png_64bit_product() in png.c

Version 1.4.0beta70 [July 27, 2009]
  Avoid a possible NULL dereference in debug build, in png_set_text_2().
    (bug introduced in libpng-0.95, discovered by Evan Rouault)

Version 1.4.0beta71 [July 29, 2009]
  Rebuilt configure scripts with autoconf-2.64.

Version 1.4.0beta72 [August 1, 2009]
  Replaced *.tar.lzma with *.tar.xz in distribution.  Get the xz codec
    from <http://tukaani.org/xz>.

Version 1.4.0beta73 [August 1, 2009]
  Reject attempt to write iCCP chunk with negative embedded profile length
    (JD Chen) (CVE-2009-5063).

Version 1.4.0beta74 [August 8, 2009]
  Changed png_ptr and info_ptr member "trans" to "trans_alpha".

Version 1.4.0beta75 [August 21, 2009]
  Removed an extra png_debug() recently added to png_write_find_filter().
  Fixed incorrect #ifdef in pngset.c regarding unknown chunk support.

Version 1.4.0beta76 [August 22, 2009]
  Moved an incorrectly located test in png_read_row() in pngread.c

Version 1.4.0beta77 [August 27, 2009]
  Removed lpXYZ.tar.bz2 (with CRLF), KNOWNBUG, libpng-x.y.z-KNOWNBUG.txt,
    and the "noconfig" files from the distribution.
  Moved CMakeLists.txt from scripts into the main libpng directory.
  Various bugfixes and improvements to CMakeLists.txt (Philip Lowman)

Version 1.4.0beta78 [August 31, 2009]
  Converted all PNG_NO_* tests to PNG_*_SUPPORTED everywhere except pngconf.h
  Eliminated PNG_NO_FREE_ME and PNG_FREE_ME_SUPPORTED macros.
  Use png_malloc plus a loop instead of png_calloc() to initialize
    row_pointers in png_read_png().

Version 1.4.0beta79 [September 1, 2009]
  Eliminated PNG_GLOBAL_ARRAYS and PNG_LOCAL_ARRAYS; always use local arrays.
  Eliminated PNG_CALLOC_SUPPORTED macro and always provide png_calloc().

Version 1.4.0beta80 [September 17, 2009]
  Removed scripts/libpng.icc
  Changed typecast of filler from png_byte to png_uint_16 in png_set_filler().
    (Dennis Gustafsson)
  Fixed typo introduced in beta78 in pngtest.c ("#if def " should be "#ifdef ")

Version 1.4.0beta81 [September 23, 2009]
  Eliminated unused PNG_FLAG_FREE_* defines from pngpriv.h
  Expanded TAB characters in pngrtran.c
  Removed PNG_CONST from all "PNG_CONST PNG_CHNK" declarations to avoid
    compiler complaints about doubly declaring things "const".
  Changed all "#if [!]defined(X)" to "if[n]def X" where possible.
  Eliminated unused png_ptr->row_buf_size

Version 1.4.0beta82 [September 25, 2009]
  Moved redundant IHDR checking into new png_check_IHDR() in png.c
    and report all errors found in the IHDR data.
  Eliminated useless call to png_check_cHRM() from pngset.c

Version 1.4.0beta83 [September 25, 2009]
  Revised png_check_IHDR() to eliminate bogus complaint about filter_type.

Version 1.4.0beta84 [September 30, 2009]
  Fixed some inconsistent indentation in pngconf.h
  Revised png_check_IHDR() to add a test for width variable less than 32-bit.

Version 1.4.0beta85 [October 1, 2009]
  Revised png_check_IHDR() again, to check info_ptr members instead of
    the contents of the returned parameters.

Version 1.4.0beta86 [October 9, 2009]
  Updated the "xcode" project (Alam Arias).
  Eliminated a shadowed declaration of "pp" in png_handle_sPLT().

Version 1.4.0rc01 [October 19, 2009]
  Trivial cosmetic changes.

Version 1.4.0beta87 [October 30, 2009]
  Moved version 1.4.0 back into beta.

Version 1.4.0beta88 [October 30, 2009]
  Revised libpng*.txt section about differences between 1.2.x and 1.4.0
    because most of the new features have now been ported back to 1.2.41

Version 1.4.0beta89 [November 1, 2009]
  More bugfixes and improvements to CMakeLists.txt (Philip Lowman)
  Removed a harmless extra png_set_invert_alpha() from pngwrite.c
  Apply png_user_chunk_cache_max within png_decompress_chunk().
  Merged libpng-1.2.41.txt with libpng-1.4.0.txt where appropriate.

Version 1.4.0beta90 [November 2, 2009]
  Removed all remaining WIN32_WCE #ifdefs except those involving the
    time.h "tm" structure

Version 1.4.0beta91 [November 3, 2009]
  Updated scripts/pngw32.def and projects/wince/png32ce.def
  Copied projects/wince/png32ce.def to the scripts directory.
  Added scripts/makefile.wce
  Patched ltmain.sh for wince support.
  Added PNG_CONVERT_tIME_SUPPORTED macro.

Version 1.4.0beta92 [November 4, 2009]
  Make inclusion of time.h in pngconf.h depend on PNG_CONVERT_tIME_SUPPORTED
  Make #define PNG_CONVERT_tIME_SUPPORTED depend on PNG_WRITE_tIME_SUPPORTED
  Revised libpng*.txt to describe differences from 1.2.40 to 1.4.0 (instead
    of differences from 1.2.41 to 1.4.0)

Version 1.4.0beta93 [November 7, 2009]
  Added PNG_DEPSTRUCT, PNG_DEPRECATED, PNG_USE_RESULT, PNG_NORETURN, and
    PNG_ALLOCATED macros to detect deprecated direct access to the
    png_struct or info_struct members and other deprecated usage in
    applications (John Bowler).
  Updated scripts/makefile* to add "-DPNG_CONFIGURE_LIBPNG" to CFLAGS,
    to prevent warnings about direct access to png structs by libpng
    functions while building libpng.  They need to be tested, especially
    those using compilers other than gcc.
  Updated projects/visualc6 and visualc71 with "/d PNG_CONFIGURE_LIBPNG".
    They should work but still need to be updated to remove
    references to pnggccrd.c or pngvcrd.c and ASM building.
  Added README.txt to the beos, cbuilder5, netware, and xcode projects warning
    that they need to be updated, to remove references to pnggccrd.c and
    pngvcrd.c and to depend on pngpriv.h
  Removed three direct references to read_info_ptr members in pngtest.c
    that were detected by the new PNG_DEPSTRUCT macro.
  Moved the png_debug macro definitions and the png_read_destroy(),
    png_write_destroy() and png_far_to_near() prototypes from png.h
    to pngpriv.h (John Bowler)
  Moved the synopsis lines for png_read_destroy(), png_write_destroy()
    png_debug(), png_debug1(), and png_debug2() from libpng.3 to libpngpf.3.

Version 1.4.0beta94 [November 9, 2009]
  Removed the obsolete, unused pnggccrd.c and pngvcrd.c files.
  Updated CMakeLists.txt to add "-DPNG_CONFIGURE_LIBPNG" to the definitions.
  Removed dependency of pngtest.o on pngpriv.h in the makefiles.
  Only #define PNG_DEPSTRUCT, etc. in pngconf.h if not already defined.

Version 1.4.0beta95 [November 10, 2009]
  Changed png_check_sig() to !png_sig_cmp() in contrib programs.
  Added -DPNG_CONFIGURE_LIBPNG to contrib/pngminm/*/makefile
  Changed png_check_sig() to !png_sig_cmp() in contrib programs.
  Corrected the png_get_IHDR() call in contrib/gregbook/readpng2.c
  Changed pngminim/*/gather.sh to stop trying to remove pnggccrd.c and pngvcrd.c
  Added dependency on pngpriv.h in contrib/pngminim/*/makefile

Version 1.4.0beta96 [November 12, 2009]
  Renamed scripts/makefile.wce to scripts/makefile.cegcc
  Revised Makefile.am to use libpng.sys while building libpng.so
    so that only PNG_EXPORT functions are exported.
  Removed the deprecated png_check_sig() function/macro.
  Removed recently removed function names from scripts/*.def
  Revised pngtest.png to put chunks in the same order written by pngtest
    (evidently the same change made in libpng-1.0beta54 was lost).
  Added PNG_PRIVATE macro definition in pngconf.h for possible future use.

Version 1.4.0beta97 [November 13, 2009]
  Restored pngtest.png to the libpng-1.4.0beta7 version.
  Removed projects/beos and netware.txt; no one seems to be supporting them.
  Revised Makefile.in

Version 1.4.0beta98 [November 13, 2009]
  Added the "xcode" project to zip distributions,
  Fixed a typo in scripts/pngwin.def introduced in beta97.

Version 1.4.0beta99 [November 14, 2009]
  Moved libpng-config.in and libpng.pc-configure.in out of the scripts
    directory, to libpng-config.in and libpng-pc.in, respectively, and
    modified Makefile.am and configure.ac accordingly.  Now "configure"
    needs nothing from the "scripts" directory.
  Avoid redefining PNG_CONST in pngconf.h

Version 1.4.0beta100 [November 14, 2009]
  Removed ASM builds from projects/visualc6 and projects/visualc71
  Removed scripts/makefile.nommx and makefile.vcawin32
  Revised CMakeLists.txt to account for new location of libpng-config.in
    and libpng-pc.in
  Updated INSTALL to reflect removal and relocation of files.

Version 1.4.0beta101 [November 14, 2009]
  Restored the binary files (*.jpg, *.png, some project files) that were
    accidentally deleted from the zip and 7z distributions when the xcode
    project was added.

Version 1.4.0beta102 [November 18, 2009]
  Added libpng-config.in and libpng-pc.in to the zip and 7z distributions.
  Fixed a typo in projects/visualc6/pngtest.dsp, introduced in beta100.
  Moved descriptions of makefiles and other scripts out of INSTALL into
    scripts/README.txt
  Updated the copyright year in scripts/pngwin.rc from 2006 to 2009.

Version 1.4.0beta103 [November 21, 2009]
  Removed obsolete comments about ASM from projects/visualc71/README_zlib.txt
  Align row_buf on 16-byte boundary in memory.
  Restored the PNG_WRITE_FLUSH_AFTER_IEND_SUPPORTED guard around the call
    to png_flush() after png_write_IEND().  See 1.4.0beta32, 1.4.0beta50
    changes above and 1.2.30, 1.2.30rc01 and rc03 in 1.2.41 CHANGES.  Someone
    needs this feature.
  Make the 'png_jmpbuf' macro expand to a call that records the correct
    longjmp function as well as returning a pointer to the setjmp
    jmp_buf buffer, and marked direct access to jmpbuf 'deprecated'.
    (John Bowler)

Version 1.4.0beta104 [November 22, 2009]
  Removed png_longjmp_ptr from scripts/*.def and libpng.3
  Rebuilt configure scripts with autoconf-2.65

Version 1.4.0beta105 [November 25, 2009]
  Use fast integer PNG_DIVIDE_BY_255() or PNG_DIVIDE_BY_65535()
    to accomplish alpha premultiplication when
    PNG_READ_COMPOSITE_NODIV_SUPPORTED is defined.
  Changed "/255" to "/255.0" in background calculations to make it clear
    that the 255 is used as a double.

Version 1.4.0beta106 [November 27, 2009]
  Removed premultiplied alpha feature.

Version 1.4.0beta107 [December 4, 2009]
  Updated README
  Added "#define PNG_NO_PEDANTIC_WARNINGS" in the libpng source files.
  Removed "-DPNG_CONFIGURE_LIBPNG" from the makefiles and projects.
  Revised scripts/makefile.netbsd, makefile.openbsd, and makefile.sco
    to put png.h and pngconf.h in $prefix/include, like the other scripts,
    instead of in $prefix/include/libpng.  Also revised makefile.sco
    to put them in $prefix/include/libpng15 instead of in
    $prefix/include/libpng/libpng15.

Version 1.4.0beta108 [December 11, 2009]
  Removed leftover "-DPNG_CONFIGURE_LIBPNG" from contrib/pngminim/*/makefile
  Relocated png_do_chop() to its original position in pngrtran.c; the
    change in version 1.2.41beta08 caused transparency to be handled wrong
    in some 16-bit datastreams (Yusaku Sugai).

Version 1.4.0beta109 [December 13, 2009]
  Added "bit_depth" parameter to the private png_build_gamma_table() function.
  Pass bit_depth=8 to png_build_gamma_table() when bit_depth is 16 but the
    PNG_16_TO_8 transform has been set, to avoid unnecessary build of 16-bit
    tables.

Version 1.4.0rc02 [December 20, 2009]
  Declared png_cleanup_needed "volatile" in pngread.c and pngwrite.c

Version 1.4.0rc03 [December 22, 2009]
  Renamed libpng-pc.in back to libpng.pc.in and revised CMakeLists.txt
    (revising the change in 1.4.0beta99)

Version 1.4.0rc04 [December 25, 2009]
  Swapped PNG_UNKNOWN_CHUNKS_SUPPORTED and PNG_HANDLE_AS_UNKNOWN_SUPPORTED
    in pngset.c to be consistent with other changes in version 1.2.38.

Version 1.4.0rc05 [December 25, 2009]
  Changed "libpng-pc.in" to "libpng.pc.in" in configure.ac, configure, and
    Makefile.in to be consistent with changes in libpng-1.4.0rc03

Version 1.4.0rc06 [December 29, 2009]
  Reverted the gamma_table changes from libpng-1.4.0beta109.
  Fixed some indentation errors.

Version 1.4.0rc07 [January 1, 2010]
  Revised libpng*.txt and libpng.3 about 1.2.x->1.4.x differences.
  Use png_calloc() instead of png_malloc(); png_memset() in pngrutil.c
  Update copyright year to 2010.

Version 1.4.0rc08 [January 2, 2010]
  Avoid deprecated references to png_ptr-io_ptr and png_ptr->error_ptr
    in pngtest.c

Version 1.4.0 [January 3, 2010]
  No changes.

Version 1.4.1beta01 [January 8, 2010]
  Updated CMakeLists.txt for consistent indentation and to avoid an
    unclosed if-statement warning (Philip Lowman).
  Revised Makefile.am and Makefile.in to remove references to Y2KINFO,
    KNOWNBUG, and libpng.la (Robert Schwebel).
  Revised the makefiles to install the same files and symbolic
    links as configure, except for libpng.la and libpng14.la.
  Make png_set|get_compression_buffer_size() available even when
    PNG_WRITE_SUPPORTED is not enabled.
  Revised Makefile.am and Makefile.in to simplify their maintenance.
  Revised scripts/makefile.linux to install a link to libpng14.so.14.1

Version 1.4.1beta02 [January 9, 2010]
  Revised the rest of the makefiles to install a link to libpng14.so.14.1

Version 1.4.1beta03 [January 10, 2010]
  Removed png_set_premultiply_alpha() from scripts/*.def

Version 1.4.1rc01 [January 16, 2010]
  No changes.

Version 1.4.1beta04 [January 23, 2010]
  Revised png_decompress_chunk() to improve speed and memory usage when
    decoding large chunks.
  Added png_set|get_chunk_malloc_max() functions.

Version 1.4.1beta05 [January 26, 2010]
  Relocated "int k" declaration in pngtest.c to minimize its scope.

Version 1.4.1beta06 [January 28, 2010]
  Revised png_decompress_chunk() to use a two-pass method suggested by
    John Bowler.

Version 1.4.1beta07 [February 6, 2010]
  Folded some long lines in the source files.
  Added defineable PNG_USER_CHUNK_CACHE_MAX, PNG_USER_CHUNK_MALLOC_MAX,
    and a PNG_USER_LIMITS_SUPPORTED flag.
  Eliminated use of png_ptr->irowbytes and reused the slot in png_ptr as
    png_ptr->png_user_chunk_malloc_max.
  Revised png_push_save_buffer() to do fewer but larger png_malloc() calls.

Version 1.4.1beta08 [February 6, 2010]
  Minor cleanup and updating of dates and copyright year.

Version 1.5.0beta01 [February 7, 2010]
  Moved declaration of png_struct into private pngstruct.h and png_info
    into pnginfo.h

Version 1.4.1beta09 and 1.5.0beta02 [February 7, 2010]
  Reverted to original png_push_save_buffer() code.

Version 1.4.1beta10 and 1.5.0beta03 [February 8, 2010]
  Return allocated "old_buffer" in png_push_save_buffer() before
    calling png_error(), to avoid a potential memory leak.
  Updated configure script to use SO number 15.

Version 1.5.0beta04 [February 9, 2010]
  Removed malformed "incomplete struct declaration" of png_info from png.h

Version 1.5.0beta05 [February 12, 2010]
  Removed PNG_DEPSTRUCT markup in pngstruct.h and pnginfo.h, and undid the
    linewrapping that it entailed.
  Revised comments in pngstruct.h and pnginfo.h and added pointers to
    the libpng license.
  Changed PNG_INTERNAL to PNG_EXPOSE_INTERNAL_STRUCTURES
  Removed the cbuilder5 project, which has not been updated to 1.4.0.

Version 1.4.1beta12 and 1.5.0beta06 [February 14, 2010]
  Fixed type declaration of png_get_chunk_malloc_max() in pngget.c (Daisuke
    Nishikawa)

Version 1.5.0beta07 [omitted]

Version 1.5.0beta08 [February 19, 2010]
  Changed #ifdef PNG_NO_STDIO_SUPPORTED to #ifdef PNG_NO_CONSOLE_IO_SUPPORTED
    wherever png_snprintf() is used to construct error and warning messages.
  Noted in scripts/makefile.mingw that it expects to be run under MSYS.
  Removed obsolete unused MMX-querying support from contrib/gregbook
  Added exported png_longjmp() function.
  Removed the AIX redefinition of jmpbuf in png.h
  Added -D_ALLSOURCE in configure.ac, makefile.aix, and CMakeLists.txt
    when building on AIX.

Version 1.5.0beta09 [February 19, 2010]
  Removed -D_ALLSOURCE from configure.ac, makefile.aix, and CMakeLists.txt.
  Changed the name of png_ptr->jmpbuf to png_ptr->png_jmpbuf in pngstruct.h

Version 1.5.0beta10 [February 25, 2010]
  Removed unused gzio.c from contrib/pngminim gather and makefile scripts
  Removed replacement error handlers from contrib/gregbook.  Because of
    the new png_longjmp() function they are no longer needed.

Version 1.5.0beta11 [March 6, 2010]
  Removed checking for already-included setjmp.h from pngconf.h
  Fixed inconsistent indentations and made numerous cosmetic changes.
  Revised the "SEE ALSO" style of libpng.3, libpngpf.3, and png.5

Version 1.5.0beta12 [March 9, 2010]
  Moved "#include png.h" inside pngpriv.h and removed "#include png.h" from
    the source files, along with "#define PNG_EXPOSE_INTERNAL_STRUCTURES"
    and "#define PNG_NO_PEDANTIC_WARNINGS" (John Bowler).
  Created new pngdebug.h and moved debug definitions there.

Version 1.5.0beta13 [March 10, 2010]
  Protect pngstruct.h, pnginfo.h, and pngdebug.h from being included twice.
  Revise the "#ifdef" blocks in png_inflate() so it will compile when neither
    PNG_USER_CHUNK_MALLOC_MAX nor PNG_SET_CHUNK_MALLOC_LIMIT_SUPPORTED
    is defined.
  Removed unused png_measure_compressed_chunk() from pngpriv.h and libpngpf.3
  Moved the 'config.h' support from pngconf.h to pngpriv.h
  Removed PNGAPI from the png_longjmp_ptr typedef.
  Eliminated dependence of pngtest.c on the private pngdebug.h file.
  Make all png_debug macros into *unterminated* statements or
    expressions (i.e. a trailing ';' must always be added) and correct
    the format statements in various png_debug messages.

Version 1.5.0beta14 [March 14, 2010]
  Removed direct access to png_ptr->io_ptr from the Windows code in pngtest.c
  Revised Makefile.am to account for recent additions and replacements.
  Corrected CE and OS/2 DEF files (scripts/png*def) for symbols removed and
    added ordinal numbers to the Windows DEF file and corrected the duplicated
    ordinal numbers on CE symbols that are commented out.
  Added back in export symbols that can be present in the Windows build but
    are disabled by default.
  PNG_EXPORT changed to include an 'ordinal' field for DEF file generation.
    PNG_CALLBACK added to make callback definitions uniform.  PNGAPI split
    into PNGCAPI (base C form), PNGAPI (exports) and PNGCBAPI (callbacks),
    and appropriate changes made to all files.  Cygwin builds re-hinged to
    allow procedure call standard changes and to remove the need for the DEF
    file (fixes build on Cygwin).
  Enabled 'attribute' warnings that are relevant to library APIs and callbacks.
  Changed rules for generation of the various symbol files and added a new
    rule for a DEF file (which is also added to the distribution).
  Updated the symbol file generation to stop it adding spurious spaces
    to EOL (coming from preprocessor macro expansion).  Added a facility
    to join tokens in the output and rewrite *.dfn to use this.
  Eliminated scripts/*.def in favor of libpng.def; updated projects/visualc71
    and removed scripts/makefile.cygwin.
  Made PNG_BUILD_DLL safe: it can be set whenever a DLL is being built.
  Removed the include of sys/types.h - apparently unnecessary now on the
    platforms on which it happened (all but Mac OS and RISC OS).
  Moved the Mac OS test into pngpriv.h (the only place it is used.)

Version 1.5.0beta15 [March 17, 2010]
  Added symbols.chk target to Makefile.am to validate the symbols in png.h
    against the new DEF file scripts/symbols.def.
  Changed the default DEF file back to pngwin.def.
  Removed makefile.mingw.
  Eliminated PNG_NO_EXTERN and PNG_ALL_EXTERN

Version 1.5.0beta16 [April 1, 2010]
  Make png_text_struct independent of PNG_iTXt_SUPPORTED, so that
    fields are initialized in all configurations.  The READ/WRITE
    macros (PNG_(READ|WRITE)_iTXt_SUPPORTED) still function as
    before to disable code to actually read or write iTXt chunks
    and iTXt_SUPPORTED can be used to detect presence of either
    read or write support (but it is probably better to check for
    the one actually required - read or write.)
  Combined multiple png_warning() calls for a single error.
  Restored the macro definition of png_check_sig().

Version 1.5.0beta17 [April 17, 2010]
  Added some "(long)" typecasts to printf calls in png_handle_cHRM().
  Documented the fact that png_set_dither() was disabled since libpng-1.4.0.
  Reenabled png_set_dither() but renamed it to png_set_quantize() to reflect
    more accurately what it actually does.  At the same time, renamed
    the PNG_DITHER_[RED,GREEN_BLUE]_BITS macros to
    PNG_QUANTIZE_[RED,GREEN,BLUE]_BITS.
  Added some "(long)" typecasts to printf calls in png_handle_cHRM().
  Freeze build-time only configuration in the build.
    In all prior versions of libpng most configuration options
    controlled by compiler #defines had to be repeated by the
    application code that used libpng.  This patch changes this
    so that compilation options that can only be changed at build
    time are frozen in the build.  Options that are compiler
    dependent (and those that are system dependent) are evaluated
    each time - pngconf.h holds these.  Options that can be changed
    per-file in the application are in png.h.  Frozen options are
    in the new installed header file pnglibconf.h (John Bowler)
  Removed the xcode project because it has not been updated to work
    with libpng-1.5.0.
  Removed the ability to include optional pngusr.h

Version 1.5.0beta18 [April 17, 2010]
  Restored the ability to include optional pngusr.h
  Moved replacements for png_error() and png_warning() from the
    contrib/pngminim project to pngerror.c, for use when warnings or
    errors are disabled via PNG_NO_WARN or PNG_NO_ERROR_TEXT, to avoid
    storing unneeded error/warning text.
  Updated contrib/pngminim project to work with the new pnglibconf.h
  Added some PNG_NO_* defines to contrib/pngminim/*/pngusr.h to save space.

Version 1.5.0beta19 [April 24, 2010]
  Added PNG_{READ,WRITE}_INT_FUNCTIONS_SUPPORTED.  This allows the functions
    to read and write ints to be disabled independently of PNG_USE_READ_MACROS,
    which allows libpng to be built with the functions even though the default
    is to use the macros - this allows applications to choose at app build
    time whether or not to use macros (previously impossible because the
    functions weren't in the default build.)
  Changed Windows calling convention back to __cdecl for API functions.
    For Windows/x86 platforms only:
      __stdcall is no longer needed for Visual Basic, so libpng-1.5.0 uses
      __cdecl throughout (both API functions and callbacks) on Windows/x86
      platforms.
  Replaced visualc6 and visualc71 projects with new vstudio project
  Relaxed the overly-restrictive permissions of some files.

Version 1.5.0beta20 [April 24, 2010]
  Relaxed more overly-restrictive permissions of some files.

Version 1.5.0beta21 [April 27, 2010]
  Removed some unwanted binary bytes and changed CRLF to NEWLINE in the new
    vstudio project files, and some trivial editing of some files in the
    scripts directory.
  Set PNG_NO_READ_BGR, PNG_NO_IO_STATE, and PNG_NO_TIME_RFC1123 in
    contrib/pngminim/decoder/pngusr.h to make a smaller decoder application.

Version 1.5.0beta22 [April 28, 2010]
  Fixed dependencies of GET_INT_32 - it does not require READ_INT_FUNCTIONS
    because it has a macro equivalent.
  Improved the options.awk script; added an "everything off" option.
  Revised contrib/pngminim to use the "everything off" option in pngusr.dfa.

Version 1.5.0beta23 [April 29, 2010]
  Corrected PNG_REMOVED macro to take five arguments.
    The macro was documented with two arguments (name,ordinal), however
    the symbol checking .dfn files assumed five arguments.  The five
    argument form seems more useful so it is changed to that.
  Corrected PNG_UNKNOWN_CHUNKS_SUPPORTED to PNG_HANDLE_AS_UNKNOWN_SUPPORTED
    in gregbook/readpng2.c
  Corrected protection of png_get_user_transform_ptr. The API declaration in
    png.h is removed if both READ and WRITE USER_TRANSFORM are turned off
    but was left defined in pngtrans.c
  Added logunsupported=1 to cause pnglibconf.h to document disabled options.
    This makes the installed pnglibconf.h more readable but causes no
    other change.  The intention is that users of libpng will find it
    easier to understand if an API they need is missing.
  Include png_reset_zstream() in png.c only when PNG_READ_SUPPORTED is defined.
  Removed dummy_inflate.c from contrib/pngminim/encoder
  Removed contrib/pngminim/*/gather.sh; gathering is now done in the makefile.

Version 1.5.0beta24 [May 7, 2010]
  Use bitwise "&" instead of arithmetic mod in pngrutil.c calculation of the
    offset of the png_ptr->rowbuf pointer into png_ptr->big_row_buf.
  Added more blank lines for readability.

Version 1.5.0beta25 [June 18, 2010]
  In pngpread.c: png_push_have_row() add check for new_row > height
  Removed the now-redundant check for out-of-bounds new_row from example.c

Version 1.5.0beta26 [June 18, 2010]
  In pngpread.c: png_push_process_row() add check for too many rows.

Version 1.5.0beta27 [June 18, 2010]
  Removed the check added in beta25 as it is now redundant.

Version 1.5.0beta28 [June 20, 2010]
  Rewrote png_process_IDAT_data to consistently treat extra data as warnings
    and handle end conditions more cleanly.
  Removed the new (beta26) check in png_push_process_row().

Version 1.5.0beta29 [June 21, 2010]
  Revised scripts/options.awk to work on Sunos (but still doesn't work)
  Added comment to options.awk and contrib/pngminim/*/makefile to try nawk.

Version 1.5.0beta30 [June 22, 2010]
  Stop memory leak when reading a malformed sCAL chunk.

Version 1.5.0beta31 [June 26, 2010]
  Revised pngpread.c patch of beta28 to avoid an endless loop.
  Removed some trailing blanks.

Version 1.5.0beta32 [June 26, 2010]
  Removed leftover scripts/options.patch and scripts/options.rej

Version 1.5.0beta33 [July 6, 3010]
  Made FIXED and FLOATING options consistent in the APIs they enable and
    disable.  Corrected scripts/options.awk to handle both command line
    options and options specified in the .dfa files.
  Changed char *msg to PNG_CONST char *msg in pngrutil.c
  Make png_set_sRGB_gAMA_and_cHRM set values using either the fixed or
    floating point APIs, but not both.
  Reversed patch to remove error handler when the jmp_buf is stored in the
    main program structure, not the png_struct.
    The error handler is needed because the default handler in libpng will
    always use the jmp_buf in the library control structure; this is never
    set.  The gregbook code is a useful example because, even though it
    uses setjmp/longjmp, it shows how error handling can be implemented
    using control mechanisms not directly supported by libpng.  The
    technique will work correctly with mechanisms such as Microsoft
    Structure Exceptions or C++ exceptions (compiler willing - note that gcc
    does not by default support interworking of C and C++ error handling.)
  Reverted changes to call png_longjmp in contrib/gregbook where it is not
    appropriate.  If mainprog->jmpbuf is used by setjmp, then png_longjmp
    cannot be used.
  Changed "extern PNG_EXPORT" to "PNG_EXPORT" in png.h (Jan Nijtmans)
  Changed "extern" to "PNG_EXTERN" in pngpriv.h (except for the 'extern "C" {')

Version 1.5.0beta34 [July 12, 2010]
  Put #ifndef PNG_EXTERN, #endif around the define PNG_EXTERN in pngpriv.h

Version 1.5.0beta35 [July 24, 2010]
  Removed some newly-added TAB characters.
  Added -DNO_PNG_SNPRINTF to CFLAGS in scripts/makefile.dj2
  Moved the definition of png_snprintf() outside of the enclosing
    #ifdef blocks in pngconf.h

Version 1.5.0beta36 [July 29, 2010]
  Patches by John Bowler:
  Fixed point APIs are now supported throughout (no missing APIs).
  Internal fixed point arithmetic support exists for all internal floating
    point operations.
  sCAL validates the floating point strings it is passed.
  Safe, albeit rudimentary, Watcom support is provided by PNG_API_RULE==2
  Two new APIs exist to get the number of passes without turning on the
    PNG_INTERLACE transform and to get the number of rows in the current
    pass.
  A new test program, pngvalid.c, validates the gamma code.
  Errors in the 16-bit gamma correction (overflows) have been corrected.
  cHRM chunk testing is done consistently (previously the floating point
    API bypassed it, because the test really didn't work on FP, now the test
    is performed on the actual values to be stored in the PNG file so it
    works in the FP case too.)
  Most floating point APIs now simply call the fixed point APIs after
    converting the values to the fixed point form used in the PNG file.
  The standard headers no longer include zlib.h, which is currently only
    required for pngstruct.h and can therefore be internal.
  Revised png_get_int_32 to undo the PNG two's complement representation of
    negative numbers.

Version 1.5.0beta37 [July 30, 2010]
  Added a typecast in png_get_int_32() in png.h and pngrutil.h to avoid
    a compiler warning.
  Replaced oFFs 0,0 with oFFs -10,20 in pngtest.png

Version 1.5.0beta38 [July 31, 2010]
  Implemented remaining "_fixed" functions.
  Corrected a number of recently introduced warnings mostly resulting from
    safe but uncast assignments to shorter integers.  Also added a zlib
    VStudio release library project because the latest zlib Official Windows
    build does not include such a thing.
  Revised png_get_int_16() to be similar to png_get_int_32().
  Restored projects/visualc71.

Version 1.5.0beta39 [August 2, 2010]
  VisualC/GCC warning fixes, VisualC build fixes
  The changes include support for function attributes in VC in addition to
    those already present in GCC - necessary because without these some
    warnings are unavoidable.  Fixes include signed/unsigned fixes in
    pngvalid and checks with gcc -Wall -Wextra -Wunused.
  VC requires function attributes on function definitions as well as
    declarations, PNG_FUNCTION has been added to enable this and the
    relevant function definitions changed.

Version 1.5.0beta40 [August 6, 2010]
  Correct use of _WINDOWS_ in pngconf.h
  Removed png_mem_ #defines; they are no longer used.
  Added the sRGB chunk to pngtest.png

Version 1.5.0beta41 [August 11, 2010]
  Added the cHRM chunk to pngtest.png
  Don't try to use version-script with cygwin/mingw.
  Revised contrib/gregbook to work under cygwin/mingw.

Version 1.5.0beta42 [August 18, 2010]
  Add .dll.a to the list of extensions to be symlinked by Makefile.am (Yaakov)
  Made all API functions that have const arguments and constant string
    literal pointers declare them (John Bowler).

Version 1.5.0beta43 [August 20, 2010]
  Removed spurious tabs, shorten long lines (no source change)
    Also added scripts/chkfmt to validate the format of all the files that can
    reasonably be validated (it is suggested to run "make distclean" before
    checking, because some machine generated files have long lines.)
  Reformatted the CHANGES file to be more consistent throughout.
  Made changes to address various issues identified by GCC, mostly
    signed/unsigned and shortening problems on assignment but also a few
    difficult to optimize (for GCC) loops.
  Fixed non-GCC fixed point builds.  In png.c a declaration was misplaced
    in an earlier update.  Fixed to declare the auto variables at the head.
  Use cexcept.h in pngvalid.c.

Version 1.5.0beta44 [August 24, 2010]
  Updated CMakeLists.txt to use CMAKE_INSTALL_LIBDIR variable; useful for
    installing libpng in /usr/lib64 (Funda Wang).
  Revised CMakeLists.txt to put the man pages in share/man/man* not man/man*
  Revised CMakeLists.txt to make symlinks instead of copies when installing.
  Changed PNG_LIB_NAME from pngNN to libpngNN in CMakeLists.txt (Philip Lowman)
  Implemented memory checks within pngvalid
  Reformatted/rearranged pngvalid.c to assist use of progressive reader.
  Check interlaced images in pngvalid
  Clarified pngusr.h comments in pnglibconf.dfa
  Simplified the pngvalid error-handling code now that cexcept.h is in place.
  Implemented progressive reader in pngvalid.c for standard tests
  Implemented progressive read in pngvalid.c gamma tests
  Turn on progressive reader in pngvalid.c by default and tidy code.

Version 1.5.0beta45 [August 26, 2010]
  Added an explicit make step to projects/vstudio for pnglibconf.h
    Also corrected zlib.vcxproj into which Visual Studio had introduced
    what it calls an "authoring error".  The change to make pnglibconf.h
    simply copies the file; in the future it may actually generate the
    file from scripts/pnglibconf.dfa as the other build systems do.
  Changed pngvalid to work when floating point APIs are disabled
  Renamed the prebuilt scripts/pnglibconf.h to scripts/pnglibconf.h.prebuilt
  Supply default values for PNG_USER_PRIVATEBUILD and PNG_USER_DLLFNAME_POSTFIX
    in pngpriv.h in case the user neglected to define them in their pngusr.h

Version 1.5.0beta46 [August 28, 2010]
  Added new private header files to libpng_sources in CMakeLists.txt
  Added PNG_READ_16BIT, PNG_WRITE_16BIT, and PNG_16BIT options.
  Added reference to scripts/pnglibconf.h.prebuilt in the visualc71 project.

Version 1.5.0beta47 [September 11, 2010]
  Fixed a number of problems with 64-bit compilation reported by Visual
    Studio 2010 (John Bowler).

Version 1.5.0beta48 [October 4, 2010]
  Updated CMakeLists.txt (Philip Lowman).
  Revised autogen.sh to recognize and use $AUTOCONF, $AUTOMAKE, $AUTOHEADER,
    $AUTOPOINT, $ACLOCAL and $LIBTOOLIZE
  Fixed problem with symbols creation in Makefile.am which was assuming that
    all versions of ccp write to standard output by default (Martin Banky). The
    bug was introduced in libpng-1.2.9beta5.
  Removed unused mkinstalldirs.

Version 1.5.0beta49 [October 8, 2010]
  Undid Makefile.am revision of 1.5.0beta48.

Version 1.5.0beta50 [October 14, 2010]
  Revised Makefile.in to account for mkinstalldirs being removed.
  Added some "(unsigned long)" typecasts in printf statements in pngvalid.c.
  Suppressed a compiler warning in png_handle_sPLT().
  Check for out-of-range text compression mode in png_set_text().

Version 1.5.0beta51 [October 15, 2010]
  Changed embedded dates to "(PENDING RELEASE) in beta releases (and future
    rc releases) to minimize the difference between releases.

Version 1.5.0beta52 [October 16, 2010]
  Restored some of the embedded dates (in png.h, png.c, documentation, etc.)

Version 1.5.0beta53 [October 18, 2010]
  Updated INSTALL to mention using "make maintainer-clean" and to remove
    obsolete statement about a custom ltmain.sh
  Disabled "color-tests" by default in Makefile.am so it will work with
    automake versions earlier than 1.11.1
  Use document name "libpng-manual.txt" instead of "libpng-<version>.txt"
    to simplify version differences.
  Removed obsolete remarks about setjmp handling from INSTALL.
  Revised and renamed the typedef in png.h and png.c that was designed
    to catch library and header mismatch.

Version 1.5.0beta54 [November 10, 2010]
  Require 48 bytes, not 64 bytes, for big_row_buf in overflow checks.
  Used a consistent structure for the pngget.c functions.

Version 1.5.0beta55 [November 21, 2010]
  Revised png_get_uint_32, png_get_int_32, png_get_uint_16 (Cosmin)
  Moved reading of file signature into png_read_sig (Cosmin)
  Fixed atomicity of chunk header serialization (Cosmin)
  Added test for io_state in pngtest.c (Cosmin)
  Added "#!/bin/sh" at the top of contrib/pngminim/*/gather.sh scripts.
  Changes to remove gcc warnings (John Bowler)
    Certain optional gcc warning flags resulted in warnings in libpng code.
    With these changes only -Wconversion and -Wcast-qual cannot be turned on.
    Changes are trivial rearrangements of code.  -Wconversion is not possible
    for pngrutil.c (because of the widespread use of += et al on variables
    smaller than (int) or (unsigned int)) and -Wcast-qual is not possible
    with pngwio.c and pngwutil.c because the 'write' callback and zlib
    compression both fail to declare their input buffers with 'const'.

Version 1.5.0beta56 [December 7, 2010]
  Added the private PNG_UNUSED() macro definition in pngpriv.h.
  Added some commentary about PNG_EXPORT in png.h and pngconf.h
  Revised PNG_EXPORT() macro and added PNG_EXPORTA() macro, with the
    objective of simplifying and improving the cosmetic appearance of png.h.
  Fixed some incorrect "=" macro names in pnglibconf.dfa
  Included documentation of changes in 1.5.0 from 1.4.x in libpng-manual.txt

Version 1.5.0beta57 [December 9, 2010]
  Documented the pngvalid gamma error summary with additional comments and
    print statements.
  Improved missing symbol handling in checksym.awk; symbols missing in both
    the old and new files can now be optionally ignored, treated as errors
    or warnings.
  Removed references to pngvcrd.c and pnggccrd.c from the vstudio project.
  Updated "libpng14" to "libpng15" in the visualc71 project.
  Enabled the strip16 tests in pngvalid.`
  Don't display test results (except PASS/FAIL) when running "make test".
    Instead put them in pngtest-log.txt
  Added "--with-zprefix=<string>" to configure.ac
  Updated the prebuilt configuration files to autoconf version 2.68

Version 1.5.0beta58 [December 19, 2010]
  Fixed interlace image handling and add test cases (John Bowler)
  Fixed the clean rule in Makefile.am to remove pngtest-log.txt
  Made minor changes to work around warnings in gcc 3.4

Version 1.5.0rc01 [December 27, 2010]
  No changes.

Version 1.5.0rc02 [December 27, 2010]
  Eliminated references to the scripts/*.def files in project/visualc71.

Version 1.5.0rc03 [December 28, 2010]
  Eliminated scripts/*.def and revised Makefile.am accordingly

Version 1.5.0rc04 [December 29, 2010]
  Fixed bug in background transformation handling in pngrtran.c (it was
    looking for the flag in png_ptr->transformations instead of in
    png_ptr->flags) (David Raymond).

Version 1.5.0rc05 [December 31, 2010]
  Fixed typo in a comment in CMakeLists.txt (libpng14 => libpng15) (Cosmin)

Version 1.5.0rc06 [January 4, 2011]
  Changed the new configure option "zprefix=string" to "zlib-prefix=string"

Version 1.5.0rc07 [January 4, 2011]
  Updated copyright year.

Version 1.5.0 [January 6, 2011]
  No changes.

version 1.5.1beta01 [January 8, 2011]
  Added description of png_set_crc_action() to the manual.
  Added a note in the manual that the type of the iCCP profile was changed
    from png_charpp to png_bytepp in png_get_iCCP().  This change happened
    in version 1.5.0beta36 but is not noted in the CHANGES.  Similarly,
    it was changed from png_charpp to png_const_bytepp in png_set_iCCP().
  Ensure that png_rgb_to_gray ignores palette mapped images, if libpng
    internally happens to call it with one, and fixed a failure to handle
    palette mapped images correctly.  This fixes CVE-2690.

Version 1.5.1beta02 [January 14, 2011]
  Fixed a bug in handling of interlaced images (bero at arklinux.org).
  Updated CMakeLists.txt (Clifford Yapp)

Version 1.5.1beta03 [January 14, 2011]
  Fixed typecasting of some png_debug() statements (Cosmin)

Version 1.5.1beta04 [January 16, 2011]
  Updated documentation of png_set|get_tRNS() (Thomas Klausner).
  Mentioned in the documentation that applications must #include "zlib.h"
    if they need access to anything in zlib.h, and that a number of
    macros such as png_memset() are no longer accessible by applications.
  Corrected pngvalid gamma test "sample" function to access all of the color
    samples of each pixel, instead of sampling the red channel three times.
  Prefixed variable names index, div, exp, gamma with "png_" to avoid "shadow"
    warnings, and (mistakenly) changed png_exp() to exp().

Version 1.5.1beta05 [January 16, 2011]
  Changed variable names png_index, png_div, png_exp, and png_gamma to
    char_index, divisor, exp_b10, and gamma_val, respectively, and
    changed exp() back to png_exp().

Version 1.5.1beta06 [January 20, 2011]
  Prevent png_push_crc_skip() from hanging while reading an unknown chunk
    or an over-large compressed zTXt chunk with the progressive reader.
  Eliminated more GCC "shadow" warnings.
  Revised png_fixed() in png.c to avoid compiler warning about reaching the
    end without returning anything.

Version 1.5.1beta07 [January 22, 2011]
  In the manual, describe the png_get_IHDR() arguments in the correct order.
  Added const_png_structp and const_png_infop types, and used them in
    prototypes for most png_get_*() functions.

Version 1.5.1beta08 [January 23, 2011]
  Added png_get_io_chunk_type() and deprecated png_get_io_chunk_name()
  Added synopses for the IO_STATE functions and other missing synopses
    to the manual. Removed the synopses from libpngpf.3 because they
    were out of date and no longer useful.  Better information can be
    obtained by reading the prototypes and comments in pngpriv.h
  Attempted to fix cpp on Solaris with S. Studio 12 cc, fix build
    Added a make macro DFNCPP that is a CPP that will accept the tokens in
    a .dfn file and adds configure stuff to test for such a CPP.  ./configure
    should fail if one is not available.
  Corrected const_png_ in png.h to png_const_ to avoid polluting the namespace.
  Added png_get_current_row_number and png_get_current_pass_number for the
    benefit of the user transform callback.
  Added png_process_data_pause and png_process_data_skip for the benefit of
    progressive readers that need to stop data processing or want to optimize
    skipping of unread data (e.g., if the reader marks a chunk to be skipped.)

Version 1.5.1beta09 [January 24, 2011]
  Enhanced pngvalid, corrected an error in gray_to_rgb, corrected doc error.
    pngvalid contains tests of transforms, which tests are currently disabled
    because they are incompletely tested.  gray_to_rgb was failing to expand
    the bit depth for smaller bit depth images; this seems to be a long
    standing error and resulted, apparently, in invalid output
    (CVE-2011-0408, CERT VU#643140).  The documentation did not accurately
    describe what libpng really does when converting RGB to gray.

Version 1.5.1beta10 [January 27, 2010]
  Fixed incorrect examples of callback prototypes in the manual, that were
    introduced in libpng-1.0.0.
  In addition the order of the png_get_uint macros with respect to the
    relevant function definitions has been reversed.  This helps the
    preprocessing of the symbol files be more robust.  Furthermore, the
    symbol file preprocessing now uses -DPNG_NO_USE_READ_MACROS even when
    the library may actually be built with PNG_USE_READ_MACROS; this stops
    the read macros interfering with the symbol file format.
  Made the manual, synopses, and function prototypes use the function
    argument names file_gamma, int_file_gamma, and srgb_intent consistently.

Version 1.5.1beta11 [January 28, 2011]
  Changed PNG_UNUSED from "param=param;" to "{if(param){}}".
  Corrected local variable type in new API png_process_data_skip()
    The type was self-evidently incorrect but only causes problems on 64-bit
    architectures.
  Added transform tests to pngvalid and simplified the arguments.

Version 1.5.1rc01 [January 29, 2011]
  No changes.

Version 1.5.1rc02 [January 31, 2011]
  Added a request in the manual that applications do not use "png_" or
    "PNG_" to begin any of their own symbols.
  Changed PNG_UNUSED to "(void)param;" and updated the commentary in pngpriv.h

Version 1.5.1 [February 3, 2011]
  No changes.

Version 1.5.2beta01 [February 13, 2011]
  More -Wshadow fixes for older gcc compilers.  Older gcc versions apparently
    check formal parameters names in function declarations (as well as
    definitions) to see if they match a name in the global namespace.
  Revised PNG_EXPORTA macro to not use an empty parameter, to accommodate the
    old VisualC++ preprocessor.
  Turned on interlace handling in png_read_png().
  Fixed gcc pendantic warnings.
  Handle longjmp in Cygwin.
  Fixed png_get_current_row_number() in the interlaced case.
  Cleaned up ALPHA flags and transformations.
  Implemented expansion to 16 bits.

Version 1.5.2beta02 [February 19, 2011]
  Fixed mistake in the descriptions of user read_transform and write_transform
    function prototypes in the manual.  The row_info struct is png_row_infop.
  Reverted png_get_current_row_number() to previous (1.5.2beta01) behavior.
  Corrected png_get_current_row_number documentation
  Fixed the read/write row callback documentation.
    This documents the current behavior, where the callback is called after
    every row with information pertaining to the next row.

Version 1.5.2beta03 [March 3, 2011]
  Fixed scripts/makefile.vcwin32
  Updated contrib/pngsuite/README to add the word "modify".
  Define PNG_ALLOCATED to blank when _MSC_VER<1300.

Version 1.5.2rc01 [March 19, 2011]
  Define remaining attributes to blank when MSC_VER<1300.
  ifdef out mask arrays in pngread.c when interlacing is not supported.

Version 1.5.2rc02 [March 22, 2011]
  Added a hint to try CPP=/bin/cpp if "cpp -E" fails in scripts/pnglibconf.mak
    and in contrib/pngminim/*/makefile, eg., on SunOS 5.10, and removed "strip"
    from the makefiles.
  Fixed a bug (present since libpng-1.0.7) that makes png_handle_sPLT() fail
    to compile when PNG_NO_POINTER_INDEXING is defined (Chubanov Kirill)

Version 1.5.2rc03 [March 24, 2011]
  Don't include standard header files in png.h while building the symbol table,
    to avoid cpp failure on SunOS (introduced PNG_BUILDING_SYMBOL_TABLE macro).

Version 1.5.2 [March 31, 2011]
  No changes.

Version 1.5.3beta01 [April 1, 2011]
  Re-initialize the zlib compressor before compressing non-IDAT chunks.
  Added API functions (png_set_text_compression_level() and four others) to
    set parameters for zlib compression of non-IDAT chunks.

Version 1.5.3beta02 [April 3, 2011]
  Updated scripts/symbols.def with new API functions.
  Only compile the new zlib re-initializing code when text or iCCP is
    supported, using PNG_WRITE_COMPRESSED_TEXT_SUPPORTED macro.
  Improved the optimization of the zlib CMF byte (see libpng-1.2.6beta03).
  Optimize the zlib CMF byte in non-IDAT compressed chunks

Version 1.5.3beta03 [April 16, 2011]
  Fixed gcc -ansi -pedantic compile. A strict ANSI system does not have
    snprintf, and the "__STRICT_ANSI__" detects that condition more reliably
    than __STDC__ (John Bowler).
  Removed the PNG_PTR_NORETURN attribute because it too dangerous. It tells
    the compiler that a user supplied callback (the error handler) does not
    return, yet there is no guarantee in practice that the application code
    will correctly implement the error handler because the compiler only
    issues a warning if there is a mistake (John Bowler).
  Removed the no-longer-used PNG_DEPSTRUCT macro.
  Updated the zlib version to 1.2.5 in the VStudio project.
  Fixed 64-bit builds where png_uint_32 is smaller than png_size_t in
    pngwutil.c (John Bowler).
  Fixed bug with stripping the filler or alpha channel when writing, that
    was introduced in libpng-1.5.2beta01 (bug report by Andrew Church).

Version 1.5.3beta04 [April 27, 2011]
  Updated pngtest.png with the new zlib CMF optimization.
  Cleaned up conditional compilation code and of background/gamma handling
    Internal changes only except a new option to avoid compiling the
    png_build_grayscale_palette API (which is not used at all internally.)
    The main change is to move the transform tests (READ_TRANSFORMS,
    WRITE_TRANSFORMS) up one level to the caller of the APIs.  This avoids
    calls to spurious functions if all transforms are disabled and slightly
    simplifies those functions.  Pngvalid modified to handle this.
    A minor change is to stop the strip_16 and expand_16 interfaces from
    disabling each other; this allows the future alpha premultiplication
    code to use 16-bit intermediate values while still producing 8-bit output.
    png_do_background and png_do_gamma have been simplified to take a single
    pointer to the png_struct rather than pointers to every item required
    from the png_struct. This makes no practical difference to the internal
    code.
  A serious bug in the pngvalid internal routine 'standard_display_init' has
    been fixed - this failed to initialize the red channel and accidentally
    initialized the alpha channel twice.
  Changed png_struct jmp_buf member name from png_jmpbuf to tmp_jmpbuf to
    avoid a possible clash with the png_jmpbuf macro on some platforms.

Version 1.5.3beta05 [May 6, 2011]
  Added the "_POSIX_SOURCE" feature test macro to ensure libpng sees the
    correct API. _POSIX_SOURCE is defined in pngpriv.h, pngtest.c and
    pngvalid.c to ensure that POSIX conformant systems disable non-POSIX APIs.
  Removed png_snprintf and added formatted warning messages.  This change adds
    internal APIs to allow png_warning messages to have parameters without
    requiring the host OS to implement snprintf.  As a side effect the
    dependency of the tIME-supporting RFC1132 code on stdio is removed and
    PNG_NO_WARNINGS does actually work now.
  Pass "" instead of '\0' to png_default_error() in png_err().  This mistake
    was introduced in libpng-1.2.20beta01.  This fixes CVE-2011-2691.
  Added PNG_WRITE_OPTIMIZE_CMF_SUPPORTED macro to make the zlib "CMF" byte
    optimization configureable.
  IDAT compression failed if preceded by a compressed text chunk (bug
    introduced in libpng-1.5.3beta01-02).  This was because the attempt to
    reset the zlib stream in png_write_IDAT happened after the first IDAT
    chunk had been deflated - much too late.  In this change internal
    functions were added to claim/release the z_stream and, hopefully, make
    the code more robust.  Also deflateEnd checking is added - previously
    libpng would ignore an error at the end of the stream.

Version 1.5.3beta06 [May 8, 2011]
  Removed the -D_ALL_SOURCE from definitions for AIX in CMakeLists.txt
  Implemented premultiplied alpha support: png_set_alpha_mode API

Version 1.5.3beta07 [May 11, 2011]
  Added expand_16 support to the high level interface.
  Added named value and 'flag' gamma support to png_set_gamma.  Made a minor
    change from the previous (unreleased) ABI/API to hide the exact value used
    for Macs - it's not a good idea to embed this in the ABI!
  Moved macro definitions for PNG_HAVE_IHDR, PNG_HAVE_PLTE, and PNG_AFTER_IDAT
    from pngpriv.h to png.h because they must be visible to applications
    that call png_set_unknown_chunks().
  Check for up->location !PNG_AFTER_IDAT when writing unknown chunks
    before IDAT.

Version 1.5.3beta08 [May 16, 2011]
  Improved "pngvalid --speed" to exclude more of pngvalid from the time.
  Documented png_set_alpha_mode(), other changes in libpng.3/libpng-manual.txt
  The cHRM chunk now sets the defaults for png_set_rgb_to_gray() (when negative
    parameters are supplied by the caller), while in the absence of cHRM
    sRGB/Rec 709 values are still used.  This introduced a divide-by-zero
    bug in png_handle_cHRM().
  The bKGD chunk no longer overwrites the background value set by
    png_set_background(), allowing the latter to be used before the file
    header is read. It never performed any useful function to override
    the default anyway.
  Added memory overwrite and palette image checks to pngvalid.c
    Previously palette image code was poorly checked. Since the transformation
    code has a special palette path in most cases this was a severe weakness.
  Minor cleanup and some extra checking in pngrutil.c and pngrtran.c. When
    expanding an indexed image, always expand to RGBA if transparency is
    present.

Version 1.5.3beta09 [May 17, 2011]
  Reversed earlier 1.5.3 change of transformation order; move png_expand_16
    back where it was.  The change doesn't work because it requires 16-bit
    gamma tables when the code only generates 8-bit ones.  This fails
    silently; the libpng code just doesn't do any gamma correction.  Moving
    the tests back leaves the old, inaccurate, 8-bit gamma calculations, but
    these are clearly better than none!

Version 1.5.3beta10 [May 20, 2011]

png_set_background() and png_expand_16() did not work together correctly.
    This problem is present in 1.5.2; if png_set_background is called with
    need_expand false and the matching 16 bit color libpng erroneously just
    treats it as an 8-bit color because of where png_do_expand_16 is in the
    transform list.  This simple fix reduces the supplied colour to 8-bits,
    so it gets smashed, but this is better than the current behavior.
  Added tests for expand16, more fixes for palette image tests to pngvalid.
    Corrects the code for palette image tests and disables attempts to
    validate palette colors.

Version 1.5.3rc01 [June 3, 2011]
  No changes.

Version 1.5.3rc02 [June 8, 2011]
  Fixed uninitialized memory read in png_format_buffer() (Bug report by
    Frank Busse, CVE-2011-2501, related to CVE-2004-0421).

Version 1.5.3beta11 [June 11, 2011]
  Fixed png_handle_sCAL which is broken in 1.5. This fixes CVE 2011-2692.
  Added sCAL to pngtest.png
  Revised documentation about png_set_user_limits() to say that it also affects
    png writing.
  Revised handling of png_set_user_limits() so that it can increase the
    limit beyond the PNG_USER_WIDTH|HEIGHT_MAX; previously it could only
    reduce it.
  Make the 16-to-8 scaling accurate. Dividing by 256 with no rounding is
    wrong (high by one) 25% of the time. Dividing by 257 with rounding is
    wrong in 128 out of 65536 cases. Getting the right answer all the time
    without division is easy.
  Added "_SUPPORTED" to the PNG_WRITE_CUSTOMIZE_ZTXT_COMPRESSION macro.
  Added projects/owatcom, an IDE project for OpenWatcom to replace
    scripts/makefile.watcom.  This project works with OpenWatcom 1.9. The
    IDE autogenerates appropriate makefiles (libpng.mk) for batch processing.
    The project is configurable, unlike the Visual Studio project, so long
    as the developer has an awk.
  Changed png_set_gAMA to limit the gamma value range so that the inverse
    of the stored value cannot overflow the fixed point representation,
    and changed other things OpenWatcom warns about.
  Revised pngvalid.c to test PNG_ALPHA_MODE_SUPPORTED correctly. This allows
    pngvalid to build when ALPHA_MODE is not supported, which is required if
    it is to build on libpng 1.4.
  Removed string/memory macros that are no longer used and are not
    necessarily fully supportable, particularly png_strncpy and png_snprintf.
  Added log option to pngvalid.c and attempted to improve gamma messages.

Version 1.5.3 [omitted]
  People found the presence of a beta release following an rc release
    to be confusing; therefore we bump the version to libpng-1.5.4beta01
    and there will be no libpng-1.5.3 release.

Version 1.5.4beta01 [June 14, 2011]
  Made it possible to undefine PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
    to get the same (inaccurate) output as libpng-1.5.2 and earlier.
  Moved definitions of PNG_HAVE_IHDR, PNG_AFTER_IDAT, and PNG_HAVE_PLTE
    outside of an unknown-chunk block in png.h because they are also
    needed for other uses.

Version 1.5.4beta02 [June 14, 2011]
  Fixed and clarified LEGACY 16-to-8 scaling code.
  Added png_set_chop_16() API, to match inaccurate results from previous
    libpng versions.
  Removed the ACCURATE and LEGACY options (they are no longer useable)
  Use the old scaling method for background if png_set_chop_16() was
    called.
  Made png_set_chop_16() API removeable by disabling PNG_CHOP_16_TO_8_SUPPORTED

Version 1.5.4beta03 [June 15, 2011]
  Fixed a problem in png_do_expand_palette() exposed by optimization in
    1.5.3beta06
  Also removed a spurious and confusing "trans" member ("trans") from png_info.
  The palette expand optimization prevented expansion to an intermediate RGBA
    form if tRNS was present but alpha was marked to be stripped; this exposed
    a check for tRNS in png_do_expand_palette() which is inconsistent with the
    code elsewhere in libpng.
  Correction to the expand_16 code; removed extra instance of
    png_set_scale_16_to_8 from pngpriv.h

Version 1.5.4beta04 [June 16, 2011]
  Added a missing "#ifdef PNG_READ_BACKGROUND_SUPPORTED/#endif" in pngrtran.c
  Added PNG_TRANSFORM_CHOP_16 to the high-level read transforms.
  Made PNG_READ_16_TO_8_ACCURATE_SCALE configurable again.  If this is
    not enabled, png_set_strip_16() and png_do_scale_16_to_8() aren't built.
  Revised contrib/visupng, gregbook, and pngminim to demonstrate chop_16_to_8

Version 1.5.4beta05 [June 16, 2011]
  Renamed png_set_strip_16() to png_set_scale_16() and renamed
    png_set_chop_16() to png_set_strip(16) in an attempt to minimize the
    behavior changes between libpng14 and libpng15.

Version 1.5.4beta06 [June 18, 2011]
  Fixed new bug that was causing both strip_16 and scale_16 to be applied.

Version 1.5.4beta07 [June 19, 2011]
  Fixed pngvalid, simplified macros, added checking for 0 in sCAL.
    The ACCURATE scale macro is no longer defined in 1.5 - call the
    png_scale_16_to_8 API.  Made sure that PNG_READ_16_TO_8 is still defined
    if the png_strip_16_to_8 API is present.  png_check_fp_number now
    maintains some state so that positive, negative and zero values are
    identified.  sCAL uses these to be strictly spec conformant.

Version 1.5.4beta08 [June 23, 2011]
  Fixed pngvalid if ACCURATE_SCALE is defined.
  Updated scripts/pnglibconf.h.prebuilt.

Version 1.5.4rc01 [June 30, 2011]
  Define PNG_ALLOCATED to "restrict" only if MSC_VER >= 1400.

Version 1.5.4 [July 7, 2011]
  No changes.

Version 1.5.5beta01 [July 13, 2011]
  Fixed some typos and made other minor changes in the manual.
  Updated contrib/pngminus/makefile.std (Samuli Souminen)

Version 1.5.5beta02 [July 14, 2011]
  Revised Makefile.am and Makefile.in to look in the right directory for
    pnglibconf.h.prebuilt

Version 1.5.5beta03 [July 27, 2011]
  Enabled compilation with g++ compiler.  This compiler does not recognize
    the file extension, so it always compiles with C++ rules.  Made minor
    changes to pngrutil.c to cast results where C++ expects it but C does not.
  Minor editing of libpng.3 and libpng-manual.txt.

Version 1.5.5beta04 [July 29, 2011]
  Revised CMakeLists.txt (Clifford Yapp)
  Updated commentary about the png_rgb_to_gray() default coefficients
    in the manual and in pngrtran.c

Version 1.5.5beta05 [August 17, 2011]
  Prevent unexpected API exports from non-libpng DLLs on Windows.  The "_DLL"
    is removed from the test of whether a DLL is being built (this erroneously
    caused the libpng APIs to be marked as DLL exports in static builds under
    Microsoft Visual Studio).  Almost all of the libpng building configuration
    is moved from pngconf.h to pngpriv.h, but PNG_DLL_EXPORT remains in
    pngconf.h, though, so that it is colocated with the import definition (it
    is no longer used anywhere in the installed headers).  The VStudio project
    definitions have been cleaned up: "_USRDLL" has been removed from the
    static library builds (this was incorrect), and PNG_USE_DLL has been added
    to pngvalid to test the functionality (pngtest does not supply it,
    deliberately).  The spurious "_EXPORTS" has been removed from the
    libpng build (all these errors were a result of copy/paste between project
    configurations.)
  Added new types and internal functions for CIE RGB end point handling to
    pngpriv.h (functions yet to be implemented).

Version 1.5.5beta06 [August 26, 2011]
  Ensure the CMAKE_LIBRARY_OUTPUT_DIRECTORY is set in CMakeLists.txt
    (Clifford Yap)
  Fixes to rgb_to_gray and cHRM XYZ APIs (John Bowler):
    The rgb_to_gray code had errors when combined with gamma correction.
    Some pixels were treated as true grey when they weren't and such pixels
    and true grey ones were not gamma corrected (the original value of the
    red component was used instead).  APIs to get and set cHRM using color
    space end points have been added and the rgb_to_gray code that defaults
    based on cHRM, and the divide-by-zero bug in png_handle_cHRM (CERT
    VU#477046, CVE-2011-3328, introduced in 1.5.4) have been corrected.
  A considerable number of tests has been added to pngvalid for the
    rgb_to_gray transform.
  Arithmetic errors in rgb_to_gray whereby the calculated gray value was
    truncated to the bit depth rather than rounded have been fixed except in
    the 8-bit non-gamma-corrected case (where consistency seems more important
    than correctness.)  The code still has considerable inaccuracies in the
    8-bit case because 8-bit linear arithmetic is used.

Version 1.5.5beta07 [September 7, 2011]
  Added "$(ARCH)" option to makefile.darwin
  Added SunOS support to configure.ac and Makefile.am
  Changed png_chunk_benign_error() to png_warning() in png.c, in
    png_XYZ_from_xy_checked().

Version 1.5.5beta08 [September 10, 2011]
  Fixed 64-bit compilation errors (gcc). The errors fixed relate
    to conditions where types that are 32 bits in the GCC 32-bit
    world (uLong and png_size_t) become 64 bits in the 64-bit
    world.  This produces potential truncation errors which the
    compiler correctly flags.
  Relocated new HAVE_SOLARIS_LD definition in configure.ac
  Constant changes for 64-bit compatibility (removal of L suffixes). The
    16-bit cases still use "L" as we don't have a 16-bit test system.

Version 1.5.5rc01 [September 15, 2011]
  Removed "L" suffixes in pngpriv.h

Version 1.5.5 [September 22, 2011]
  No changes.

Version 1.5.6beta01 [September 22, 2011]
  Fixed some 64-bit type conversion warnings in pngrtran.c
  Moved row_info from png_struct to a local variable.
  The various interlace mask arrays have been made into arrays of
    bytes and made PNG_CONST and static (previously some arrays were
    marked PNG_CONST and some weren't).
  Additional checks have been added to the transform code to validate the
    pixel depths after the transforms on both read and write.
  Removed some redundant code from pngwrite.c, in png_destroy_write_struct().
  Changed chunk reading/writing code to use png_uint_32 instead of png_byte[4].
    This removes the need to allocate temporary strings for chunk names on
    the stack in the read/write code.  Unknown chunk handling still uses the
    string form because this is exposed in the API.

Version 1.5.6beta02 [September 26, 2011]
  Added a note in the manual the png_read_update_info() must be called only
    once with a particular info_ptr.
  Fixed a typo in the definition of the new PNG_STRING_FROM_CHUNK(s,c) macro.

Version 1.5.6beta03 [September 28, 2011]
  Revised test-pngtest.sh to report FAIL when pngtest fails.
  Added "--strict" option to pngtest, to report FAIL when the failure is
    only because the resulting valid files are different.
  Revised CMakeLists.txt to work with mingw and removed some material from
    CMakeLists.txt that is no longer useful in libpng-1.5.

Version 1.5.6beta04 [October 5, 2011]
  Fixed typo in Makefile.in and Makefile.am ("-M Wl" should be "-M -Wl")."

Version 1.5.6beta05 [October 12, 2011]
  Speed up png_combine_row() for interlaced images. This reduces the generality
    of the code, allowing it to be optimized for Adam7 interlace.  The masks
    passed to png_combine_row() are now generated internally, avoiding
    some code duplication and localizing the interlace handling somewhat.
  Align png_struct::row_buf - previously it was always unaligned, caused by
    a bug in the code that attempted to align it; the code needs to subtract
    one from the pointer to take account of the filter byte prepended to
    each row.
  Optimized png_combine_row() when rows are aligned. This gains a small
    percentage for 16-bit and 32-bit pixels in the typical case where the
    output row buffers are appropriately aligned. The optimization was not
    previously possible because the png_struct buffer was always misaligned.
  Fixed bug in png_write_chunk_header() debug print, introduced in 1.5.6beta01.

Version 1.5.6beta06 [October 17, 2011]
  Removed two redundant tests for unitialized row.
  Fixed a relatively harmless memory overwrite in compressed text writing
    with a 1 byte zlib buffer.
  Add ability to call png_read_update_info multiple times to pngvalid.c.
  Fixes for multiple calls to png_read_update_info. These fixes attend to
    most of the errors revealed in pngvalid, however doing the gamma work
    twice results in inaccuracies that can't be easily fixed.  There is now
    a warning in the code if this is going to happen.
  Turned on multiple png_read_update_info in pngvalid transform tests.
  Prevent libpng from overwriting unused bits at the end of the image when
    it is not byte aligned, while reading. Prior to libpng-1.5.6 libpng would
    overwrite the partial byte at the end of each row if the row width was not
    an exact multiple of 8 bits and the image is not interlaced.

Version 1.5.6beta07 [October 21, 2011]
  Made png_ptr->prev_row an aligned pointer into png_ptr->big_prev_row
    (Mans Rullgard).

Version 1.5.6rc01 [October 26, 2011]
  Changed misleading "Missing PLTE before cHRM" warning to "Out of place cHRM"

Version 1.5.6rc02 [October 27, 2011]
  Added LSR() macro to defend against buggy compilers that evaluate non-taken
    code branches and complain about out-of-range shifts.

Version 1.5.6rc03 [October 28, 2011]
  Renamed the LSR() macro to PNG_LSR() and added PNG_LSL() macro.
  Fixed compiler warnings with Intel and MSYS compilers. The logical shift
    fix for Microsoft Visual C is required by other compilers, so this
    enables that fix for all compilers when using compile-time constants.
    Under MSYS 'byte' is a name declared in a system header file, so we
    changed the name of a local variable to avoid the warnings that result.
  Added #define PNG_ALIGN_TYPE PNG_ALIGN_NONE to contrib/pngminim/*/pngusr.h

Version 1.5.6 [November 3, 2011]
  No changes.

Version 1.5.7beta01 [November 4, 2011]
  Added support for ARM processor, when decoding all PNG up-filtered rows
    and any other-filtered rows with 3 or 4 bytes per pixel (Mans Rullgard).
  Fixed bug in pngvalid on early allocation failure; fixed type cast in
    pngmem.c; pngvalid would attempt to call png_error() if the allocation
    of a png_struct or png_info failed. This would probably have led to a
    crash.  The pngmem.c implementation of png_malloc() included a cast
    to png_size_t which would fail on large allocations on 16-bit systems.
  Fix for the preprocessor of the Intel C compiler. The preprocessor
    splits adjacent @ signs with a space; this changes the concatentation
    token from @-@-@ to PNG_JOIN; that should work with all compiler
    preprocessors.
  Paeth filter speed improvements from work by Siarhei Siamashka. This
    changes the 'Paeth' reconstruction function to improve the GCC code
    generation on x86. The changes are only part of the suggested ones;
    just the changes that definitely improve speed and remain simple.
    The changes also slightly increase the clarity of the code.

Version 1.5.7beta02 [November 11, 2011]
  Check compression_type parameter in png_get_iCCP and remove spurious
    casts. The compression_type parameter is always assigned to, so must
    be non-NULL. The cast of the profile length potentially truncated the
    value unnecessarily on a 16-bit int system, so the cast of the (byte)
    compression type to (int) is specified by ANSI-C anyway.
  Fixed FP division by zero in pngvalid.c; the 'test_pixel' code left
    the sBIT fields in the test pixel as 0, which resulted in a floating
    point division by zero which was irrelevant but causes systems where
    FP exceptions cause a crash. Added code to pngvalid to turn on FP
    exceptions if the appropriate glibc support is there to ensure this is
    tested in the future.
  Updated scripts/pnglibconf.mak and scripts/makefile.std to handle the
    new PNG_JOIN macro.
  Added versioning to pnglibconf.h comments.
  Simplified read/write API initial version; basic read/write tested on
    a variety of images, limited documentation (in the header file.)
  Installed more accurate linear to sRGB conversion tables. The slightly
    modified tables reduce the number of 16-bit values that
    convert to an off-by-one 8-bit value.  The "makesRGB.c" code that was used
    to generate the tables is now in a contrib/sRGBtables sub-directory.

Version 1.5.7beta03 [November 17, 2011]
  Removed PNG_CONST from the sRGB table declarations in pngpriv.h and png.c
  Added run-time detection of NEON support.
  Added contrib/libtests; includes simplified API test and timing test and
    a color conversion utility for rapid checking of failed 'pngstest' results.
  Multiple transform bug fixes plus a work-round for double gamma correction.
    libpng does not support more than one transform that requires linear data
    at once - if this is tried typically the results is double gamma
    correction. Since the simplified APIs can need rgb to gray combined with
    a compose operation it is necessary to do one of these outside the main
    libpng transform code. This check-in also contains fixes to various bugs
    in the simplified APIs themselves and to some bugs in compose and rgb to
    gray (on palette) itself.
  Fixes for C++ compilation using g++ When libpng source is compiled
    using g++. The compiler imposes C++ rules on the C source; thus it
    is desireable to make the source work with either C or C++ rules
    without throwing away useful error information.  This change adds
    png_voidcast to allow C semantic (void*) cases or the corresponding
    C++ static_cast operation, as appropriate.
  Added --noexecstack to assembler file compilation. GCC does not set
    this on assembler compilation, even though it does on C compilation.
    This creates security issues if assembler code is enabled; the
    work-around is to set it by default in the flags for $(CCAS)
  Work around compilers that don't support declaration of const data. Some
    compilers fault 'extern const' data declarations (because the data is
    not initialized); this turns on const-ness only for compilers where
    this is known to work.

Version 1.5.7beta04 [November 17, 2011]
  Since the gcc driver does not recognize the --noexecstack flag, we must
    use the -Wa prefix to have it passed through to the assembler.
    Also removed a duplicate setting of this flag.
  Added files that were omitted from the libpng-1.5.7beta03 zip distribution.

Version 1.5.7beta05 [November 25, 2011]
  Removed "zTXt" from warning in generic chunk decompression function.
  Validate time settings passed to png_set_tIME() and png_convert_to_rfc1123()
    (Frank Busse). Note: This prevented CVE-2015-7981 from affecting
    libpng-1.5.7 and later.
  Added MINGW support to CMakeLists.txt
  Reject invalid compression flag or method when reading the iTXt chunk.
  Backed out 'simplified' API changes. The API seems too complex and there
    is a lack of consensus or enthusiasm for the proposals.  The API also
    reveals significant bugs inside libpng (double gamma correction and the
    known bug of being unable to retrieve a corrected palette). It seems
    better to wait until the bugs, at least, are corrected.
  Moved pngvalid.c into contrib/libtests
  Rebuilt Makefile.in, configure, etc., with autoconf-2.68

Version 1.5.7rc01 [December 1, 2011]
  Replaced an "#if" with "#ifdef" in pngrtran.c
  Revised #if PNG_DO_BC block in png.c (use #ifdef and add #else)

Version 1.5.7rc02 [December 5, 2011]
  Revised project files and contrib/pngvalid/pngvalid.c to account for
    the relocation of pngvalid into contrib/libtests.
  Revised pngconf.h to use " __declspec(restrict)" only when MSC_VER >= 1400,
    as in libpng-1.5.4.
  Put CRLF line endings in the owatcom project files.

Version 1.5.7rc03 [December 7, 2011]
  Updated CMakeLists.txt to account for the relocation of pngvalid.c

Version 1.5.7 [December 15, 2011]
  Minor fixes to pngvalid.c for gcc 4.6.2 compatibility to remove warnings
    reported by earlier versions.
  Fixed minor memset/sizeof errors in pngvalid.c.

Version 1.6.0beta01 [December 15, 2011]
  Removed machine-generated configure files from the GIT repository (they will
    continue to appear in the tarball distributions and in the libpng15 and
    earlier GIT branches).
  Restored the new 'simplified' API, which was started in libpng-1.5.7beta02
    but later deleted from libpng-1.5.7beta05.
  Added example programs for the new 'simplified' API.
  Added ANSI-C (C90) headers and require them, and take advantage of the
    change. Also fixed some of the projects/* and contrib/* files that needed
    updates for libpng16 and the move of pngvalid.c.
    With this change the required ANSI-C header files are assumed to exist: the
    implementation must provide float.h, limits.h, stdarg.h and stddef.h and
    libpng relies on limits.h and stddef.h existing and behaving as defined
    (the other two required headers aren't used).  Non-ANSI systems that don't
    have stddef.h or limits.h will have to provide an appropriate fake
    containing the relevant types and #defines.
  Dropped support for 16-bit platforms. The use of FAR/far has been eliminated
    and the definition of png_alloc_size_t is now controlled by a flag so
    that 'small size_t' systems can select it if necessary.  Libpng 1.6 may
    not currently work on such systems -- it seems likely that it will
    ask 'malloc' for more than 65535 bytes with any image that has a
    sufficiently large row size (rather than simply failing to read such
    images).
  New tools directory containing tools used to generate libpng code.
  Fixed race conditions in parallel make builds. With higher degrees of
    parallelism during 'make' the use of the same temporary file names such
    as 'dfn*' can result in a race where a temporary file from one arm of the
    build is deleted or overwritten in another arm.  This changes the
    temporary files for suffix rules to always use $* and ensures that the
    non-suffix rules use unique file names.

Version 1.6.0beta02 [December 21, 2011]
  Correct configure builds where build and source directories are separate.
    The include path of 'config.h' was erroneously made relative in pngvalid.c
    in libpng 1.5.7.

Version 1.6.0beta03 [December 22, 2011]
  Start-up code size improvements, error handler flexibility. These changes
    alter how the tricky allocation of the initial png_struct and png_info
    structures are handled. png_info is now handled in pretty much the same
    way as everything else, except that the allocations handle NULL return
    silently.  png_struct is changed in a similar way on allocation and on
    deallocation a 'safety' error handler is put in place (which should never
    be required).  The error handler itself is changed to permit mismatches
    in the application and libpng error buffer size; however, this means a
    silent change to the API to return the jmp_buf if the size doesn't match
    the size from the libpng compilation; libpng now allocates the memory and
    this may fail.  Overall these changes result in slight code size
    reductions; however, this is a reduction in code that is always executed
    so is particularly valuable.  Overall on a 64-bit system the libpng DLL
    decreases in code size by 1733 bytes.  pngerror.o increases in size by
    about 465 bytes because of the new functionality.
  Added png_convert_to_rfc1123_buffer() and deprecated png_convert_to_rfc1123()
    to avoid including a spurious buffer in the png_struct.

Version 1.6.0beta04 [December 30, 2011]
  Regenerated configure scripts with automake-1.11.2
  Eliminated png_info_destroy(). It is now used only in png.c and only calls
    one other internal function and memset().
  Enabled png_get_sCAL_fixed() if floating point APIs are enabled. Previously
    it was disabled whenever internal fixed point arithmetic was selected,
    which meant it didn't exist even on systems where FP was available but not
    preferred.
  Added pngvalid.c compile time checks for const APIs.
  Implemented 'restrict' for png_info and png_struct. Because of the way
    libpng works both png_info and png_struct are always accessed via a
    single pointer.  This means adding C99 'restrict' to the pointer gives
    the compiler some opportunity to optimize the code.  This change allows
    that.
  Moved AC_MSG_CHECKING([if libraries can be versioned]) later to the proper
    location in configure.ac (Gilles Espinasse).
  Changed png_memcpy to C assignment where appropriate. Changed all those
    uses of png_memcpy that were doing a simple assignment to assignments
    (all those cases where the thing being copied is a non-array C L-value).
  Added some error checking to png_set_*() routines.
  Removed the reference to the non-exported function png_memcpy() from
    example.c.
  Fixed the Visual C 64-bit build - it requires jmp_buf to be aligned, but
    it had become misaligned.
  Revised contrib/pngminus/pnm2png.c to avoid warnings when png_uint_32
    and unsigned long are of different sizes.

Version 1.6.0beta05 [January 15, 2012]
  Updated manual with description of the simplified API (copied from png.h)
  Fix bug in pngerror.c: some long warnings were being improperly truncated
    (CVE-2011-3464, bug introduced in libpng-1.5.3beta05).

Version 1.6.0beta06 [January 24, 2012]
  Added palette support to the simplified APIs. This commit
    changes some of the macro definitions in png.h, app code
    may need corresponding changes.
  Increased the formatted warning buffer to 192 bytes.
  Added color-map support to simplified API. This is an initial version for
    review; the documentation has not yet been updated.
  Fixed Min/GW uninstall to remove libpng.dll.a

Version 1.6.0beta07 [January 28, 2012]
  Eliminated Intel icc/icl compiler warnings. The Intel (GCC derived)
    compiler issues slightly different warnings from those issued by the
    current vesions of GCC. This eliminates those warnings by
    adding/removing casts and small code rewrites.
  Updated configure.ac from autoupdate: added --enable-werror option.
    Also some layout regularization and removal of introduced tab characters
    (replaced with 3-character indentation).  Obsolete macros identified by
    autoupdate have been removed; the replacements are all in 2.59 so
    the pre-req hasn't been changed.  --enable-werror checks for support
    for -Werror (or the given argument) in the compiler.  This mimics the
    gcc configure option by allowing -Werror to be turned on safely; without
    the option the tests written in configure itself fail compilation because
    they cause compiler warnings.
  Rewrote autogen.sh to run autoreconf instead of running tools one-by-one.
  Conditionalize the install rules for MINGW and CYGWIN in CMakeLists.txt and
    set CMAKE_LIBRARY_OUTPUT_DIRECTORY to "lib" on all platforms (C. Yapp).
  Freeze libtool files in the 'scripts' directory. This version of autogen.sh
    attempts to dissuade people from running it when it is not, or should not,
    be necessary.  In fact, autogen.sh does not work when run in a libpng
    directory extracted from a tar distribution anymore. You must run it in
    a GIT clone instead.
  Added two images to contrib/pngsuite (1-bit and 2-bit transparent grayscale),
    and renamed three whose names were inconsistent with those in
    pngsuite/README.txt.

Version 1.6.0beta08 [February 1, 2012]
  Fixed Image::colormap misalignment in pngstest.c
  Check libtool/libtoolize version number (2.4.2) in configure.ac
  Divide test-pngstest.sh into separate pngstest runs for basic and
    transparent images.
  Moved automake options to AM_INIT_AUTOMAKE in configure.ac
  Added color-tests, silent-rules (Not yet implemented in Makefile.am) and
    version checking to configure.ac
  Improved pngstest speed by not doing redundant tests and add const to
    the background parameter of png_image_finish_read. The --background
    option is now done automagically only when required, so that commandline
    option no longer exists.
  Cleaned up pngpriv.h to consistently declare all functions and data.
    Also eliminated PNG_CONST_DATA, which is apparently not needed but we
    can't be sure until it is gone.
  Added symbol prefixing that allows all the libpng external symbols
    to be prefixed (suggested by Reuben Hawkins).
  Updated "ftbb*.png" list in the owatcom and vstudio projects.
  Fixed 'prefix' builds on clean systems. The generation of pngprefix.h
    should not require itself.
  Updated INSTALL to explain that autogen.sh must be run in a GIT clone,
    not in a libpng directory extracted from a tar distribution.

Version 1.6.0beta09 [February 1, 2012]
  Reverted the prebuilt configure files to libpng-1.6.0beta05 condition.

Version 1.6.0beta10 [February 3, 2012]
  Added Z_SOLO for zlib-1.2.6+ and correct pngstest tests
  Updated list of test images in CMakeLists.txt
  Updated the prebuilt configure files to current condition.
  Revised INSTALL information about autogen.sh; it works in tar distributions.

Version 1.6.0beta11 [February 16, 2012]
  Fix character count in pngstest command in projects/owatcom/pngstest.tgt
  Revised test-pngstest.sh to report PASS/FAIL for each image.
  Updated documentation about the simplified API.
  Corrected estimate of error in libpng png_set_rgb_to_gray API.  The API is
    extremely inaccurate for sRGB conversions because it uses an 8-bit
    intermediate linear value and it does not use the sRGB transform, so it
    suffers from the known instability in gamma transforms for values close
    to 0 (see Poynton).  The net result is that the calculation has a maximum
    error of 14.99/255; 0.5/255^(1/2.2).  pngstest now uses 15 for the
    permitted 8-bit error. This may still not be enough because of arithmetic
    error.
  Removed some unused arrays (with #ifdef) from png_read_push_finish_row().
  Fixed a memory overwrite bug in simplified read of RGB PNG with
    non-linear gamma Also bugs in the error checking in pngread.c and changed
    quite a lot of the checks in pngstest.c to be correct; either correctly
    written or not over-optimistic.  The pngstest changes are insufficient to
    allow all possible RGB transforms to be passed; pngstest cmppixel needs
    to be rewritten to make it clearer which errors it allows and then changed
    to permit known inaccuracies.
  Removed tests for no-longer-used *_EMPTY_PLTE_SUPPORTED from pngstruct.h
  Fixed fixed/float API export conditionals. 1) If FIXED_POINT or
    FLOATING_POINT options were switched off, png.h ended up with lone ';'
    characters.  This is not valid ANSI-C outside a function.  The ';'
    characters have been moved inside the definition of PNG_FP_EXPORT and
    PNG_FIXED_EXPORT. 2) If either option was switched off, the declaration
    of the corresponding functions were completely omitted, even though some
    of them are still used internally.  The result is still valid, but
    produces warnings from gcc with some warning options (including -Wall). The
    fix is to cause png.h to declare the functions with PNG_INTERNAL_FUNCTION
    when png.h is included from pngpriv.h.
  Check for invalid palette index while reading paletted PNG.  When one is
    found, issue a warning and increase png_ptr->num_palette accordingly.
    Apps are responsible for checking to see if that happened.

Version 1.6.0beta12 [February 18, 2012]
  Do not increase num_palette on invalid_index.
  Relocated check for invalid palette index to pngrtran.c, after unpacking
    the sub-8-bit pixels.
  Fixed CVE-2011-3026 buffer overrun bug.  This bug was introduced when
    iCCP chunk support was added at libpng-1.0.6. Deal more correctly with the
    test on iCCP chunk length. Also removed spurious casts that may hide
    problems on 16-bit systems.

Version 1.6.0beta13 [February 24, 2012]
  Eliminated redundant png_push_read_tEXt|zTXt|iTXt|unknown code from
    pngpread.c and use the sequential png_handle_tEXt, etc., in pngrutil.c;
    now that png_ptr->buffer is inaccessible to applications, the special
    handling is no longer useful.
  Added PNG_SAFE_LIMITS feature to pnglibconf.dfa, pngpriv.h, and new
    pngusr.dfa to reset the user limits to safe ones if PNG_SAFE_LIMITS is
    defined.  To enable, use "CPPFLAGS=-DPNG_SAFE_LIMITS_SUPPORTED=1" on the
    configure command or put #define PNG_SAFE_LIMITS_SUPPORTED in
    pnglibconf.h.prebuilt and pnglibconf.h.

Version 1.6.0beta14 [February 27, 2012]
  Added information about the new limits in the manual.
  Updated Makefile.in

Version 1.6.0beta15 [March 2, 2012]
  Removed unused "current_text" members of png_struct and the png_free()
    of png_ptr->current_text from pngread.c
  Rewrote pngstest.c for substantial speed improvement.
  Fixed transparent pixel and 16-bit rgb tests in pngstest and removed a
    spurious check in pngwrite.c
  Added PNG_IMAGE_FLAG_FAST for the benefit of applications that store
    intermediate files, or intermediate in-memory data, while processing
    image data with the simplified API.  The option makes the files larger
    but faster to write and read.  pngstest now uses this by default; this
    can be disabled with the --slow option.
  Improved pngstest fine tuning of error numbers, new test file generator.
    The generator generates images that test the full range of sample values,
    allow the error numbers in pngstest to be tuned and checked.  makepng
    also allows generation of images with extra chunks, although this is
    still work-in-progress.
  Added check for invalid palette index while reading.
  Fixed some bugs in ICC profile writing. The code should now accept
    all potentially valid ICC profiles and reject obviously invalid ones.
    It now uses png_error() to do so rather than casually writing a PNG
    without the necessary color data.
  Removed whitespace from the end of lines in all source files and scripts.

Version 1.6.0beta16 [March 6, 2012]
  Relocated palette-index checking function from pngrutil.c to pngtrans.c
  Added palette-index checking while writing.
  Changed png_inflate() and calling routines to avoid overflow problems.
    This is an intermediate check-in that solves the immediate problems and
    introduces one performance improvement (avoiding a copy via png_ptr->zbuf.)
    Further changes will be made to make ICC profile handling more secure.
  Fixed build warnings (MSVC, GCC, GCC v3). Cygwin GCC with default options
    declares 'index' as a global, causing a warning if it is used as a local
    variable.  GCC 64-bit warns about assigning a (size_t) (unsigned 64-bit)
    to an (int) (signed 32-bit).  MSVC, however, warns about using the
    unary '-' operator on an unsigned value (even though it is well defined
    by ANSI-C to be ~x+1).  The padding calculation was changed to use a
    different method.  Removed the tests on png_ptr->pass.
  Added contrib/libtests/tarith.c to test internal arithmetic functions from
    png.c. This is a libpng maintainer program used to validate changes to the
    internal arithmetic functions.
  Made read 'inflate' handling like write 'deflate' handling. The read
    code now claims and releases png_ptr->zstream, like the write code.
    The bug whereby the progressive reader failed to release the zstream
    is now fixed, all initialization is delayed, and the code checks for
    changed parameters on deflate rather than always calling
    deflatedEnd/deflateInit.
  Validate the zTXt strings in pngvalid.
  Added code to validate the windowBits value passed to deflateInit2().
    If the call to deflateInit2() is wrong a png_warning will be issued
    (in fact this is harmless, but the PNG data produced may be sub-optimal).

Version 1.6.0beta17 [March 10, 2012]
  Fixed PNG_LIBPNG_BUILD_BASE_TYPE definition. 
  Reject all iCCP chunks after the first, even if the first one is invalid.
  Deflate/inflate was reworked to move common zlib calls into single
    functions [rw]util.c.  A new shared keyword check routine was also added
    and the 'zbuf' is no longer allocated on progressive read.  It is now
    possible to call png_inflate() incrementally.  A warning is no longer
    issued if the language tag or translated keyword in the iTXt chunk
    has zero length.
  If benign errors are disabled use maximum window on ancilliary inflate.
    This works round a bug introduced in 1.5.4 where compressed ancillary
    chunks could end up with a too-small windowBits value in the deflate
    header.

Version 1.6.0beta18 [March 16, 2012]
  Issue a png_benign_error() instead of png_warning() about bad palette index.
  In pngtest, treat benign errors as errors if "-strict" is present.
  Fixed an off-by-one error in the palette index checking function.
  Fixed a compiler warning under Cygwin (Windows-7, 32-bit system)
  Revised example.c to put text strings in a temporary character array
    instead of directly assigning string constants to png_textp members.
    This avoids compiler warnings when -Wwrite-strings is enabled.
  Added output flushing to aid debugging under Visual Studio. Unfortunately
    this is necessary because the VS2010 output window otherwise simply loses
    the error messages on error (they weren't flushed to the window before
    the process exited, apparently!)
  Added configuration support for benign errors and changed the read
    default. Also changed some warnings in the iCCP and sRGB handling
    from to benign errors. Configuration now makes read benign
    errors warnings and write benign errors to errors by default (thus
    changing the behavior on read).  The simplified API always forces
    read benign errors to warnings (regardless of the system default, unless
    this is disabled in which case the simplified API can't be built.)

Version 1.6.0beta19 [March 18, 2012]
  Work around for duplicate row start calls; added warning messages.
    This turns on PNG_FLAG_DETECT_UNINITIALIZED to detect app code that
    fails to call one of the 'start' routines (not enabled in libpng-1.5
    because it is technically an API change, since it did normally work
    before.)  It also makes duplicate calls to png_read_start_row (an
    internal function called at the start of the image read) benign, as
    they were before changes to use png_inflate_claim. Somehow webkit is
    causing this to happen; this is probably a mis-feature in the zlib
    changes so this commit is only a work-round.
  Removed erroneous setting of DETECT_UNINITIALIZED and added more
    checks. The code now does a png_error if an attempt is made to do the
    row initialization twice; this is an application error and it has
    serious consequences because the transform data in png_struct is
    changed by each call.
  Added application error reporting and added chunk names to read
    benign errors; also added --strict to pngstest - not enabled
    yet because a warning is produced.
  Avoid the double gamma correction warning in the simplified API.
    This allows the --strict option to pass in the pngstest checks

Version 1.6.0beta20 [March 29, 2012]
  Changed chunk handler warnings into benign errors, incrementally load iCCP
  Added checksum-icc.c to contrib/tools
  Prevent PNG_EXPAND+PNG_SHIFT doing the shift twice.
  Recognize known sRGB ICC profiles while reading; prefer writing the
    iCCP profile over writing the sRGB chunk, controlled by the
    PNG_sRGB_PROFILE_CHECKS option.
  Revised png_set_text_2() to avoid potential memory corruption (fixes
    CVE-2011-3048, also known as CVE-2012-3425).

Version 1.6.0beta21 [April 27, 2012]
  Revised scripts/makefile.darwin: use system zlib; remove quotes around
    architecture list; add missing ppc architecture; add architecture options
    to shared library link; don't try to create a shared lib based on missing
    RELEASE variable.
  Enable png_set_check_for_invalid_index() for both read and write.
  Removed #ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED in pngpriv.h around
    declaration of png_handle_unknown().
  Added -lssp_nonshared in a comment in scripts/makefile.freebsd
    and changed deprecated NOOBJ and NOPROFILE to NO_OBJ and NO_PROFILE.

Version 1.6.0beta22 [May 23, 2012]
  Removed need for -Wno-cast-align with clang.  clang correctly warns on
    alignment increasing pointer casts when -Wcast-align is passed. This
    fixes the cases that clang warns about either by eliminating the
    casts from png_bytep to png_uint_16p (pngread.c), or, for pngrutil.c
    where the cast is previously verified or pngstest.c where it is OK, by
    introducing new png_aligncast macros to do the cast in a way that clang
    accepts.

Version 1.6.0beta23 [June 6, 2012]
  Revised CMakeLists.txt to not attempt to make a symlink under mingw.
  Made fixes for new optimization warnings from gcc 4.7.0. The compiler
    performs an optimization which is safe; however it then warns about it.
    Changing the type of 'palette_number' in pngvalid.c removes the warning.
  Do not depend upon a GCC feature macro being available for use in generating
    the linker mapfile symbol prefix.
  Improved performance of new do_check_palette_indexes() function (only
    update the value when it actually increases, move test for whether
    the check is wanted out of the function.

Version 1.6.0beta24 [June 7, 2012]
  Don't check palette indexes if num_palette is 0 (as it can be in MNG files).

Version 1.6.0beta25 [June 16, 2012]
  Revised png_set_keep_unknown_chunks() so num_chunks < 0 means ignore all
    unknown chunks and all known chunks except for IHDR, PLTE, tRNS, IDAT,
    and IEND.  Previously it only meant ignore all unknown chunks, the
    same as num_chunks == 0. Revised png_image_skip_unused_chunks() to
    provide a list of chunks to be processed instead of a list of chunks to
    ignore.  Revised contrib/gregbook/readpng2.c accordingly.

Version 1.6.0beta26 [July 10, 2012]
  Removed scripts/makefile.cegcc from the *.zip and *.7z distributions; it
    depends on configure, which is not included in those archives.
  Moved scripts/chkfmt to contrib/tools.
  Changed "a+w" to "u+w" in Makefile.in to fix CVE-2012-3386.

Version 1.6.0beta27 [August 11, 2012]
  Do not compile PNG_DEPRECATED, PNG_ALLOC and PNG_PRIVATE when __GNUC__ < 3.
  Do not use __restrict when GNUC is <= 3.1
  Removed references to png_zalloc() and png_zfree() from the manual.
  Fixed configurations where floating point is completely disabled.  Because
    of the changes to support symbol prefixing PNG_INTERNAL_FUNCTION declares
    floating point APIs during libpng builds even if they are completely
    disabled. This requires the png floating point types (png_double*) to be
    declared even though the functions are never actually defined.  This
    change provides a dummy definition so that the declarations work, yet any
    implementation will fail to compile because of an incomplete type.
  Re-eliminated the use of strcpy() in pngtest.c.  An unncessary use of
    strcpy() was accidentally re-introduced in libpng16; this change replaces
    it with strncpy().
  Eliminated use of png_sizeof(); use sizeof() instead.
  Use a consistent style for (sizeof type) and (sizeof (array))
  Cleanup of png_set_filler().  This function does very different things on
    read and write.  In libpng 1.6 the two cases can be distinguished and
    considerable code cleanup, and extra error checking, is possible.  This
    makes calls on the write side that have no effect be ignored with a
    png_app_error(), which can be disabled in the app using
    png_set_benign_errors(), and removes the spurious use of usr_channels
    on the read side.
  Insist on autotools 1.12.1 for git builds because there are security issues
    with 1.12 and insisting on anything less would allow 1.12 to be used.
  Removed info_ptr->signature[8] from WRITE-only builds.
  Add some conditions for compiling png_fixed().  This is a small function
    but it requires "-lm" on some platforms.
  Cause pngtest --strict to fail on any warning from libpng (not just errors)
    and cause it not to fail at the comparison step if libpng lacks support
    for writing chunks that it reads from the input (currently only implemented
    for compressed text chunks).
  Make all three "make check" test programs work without READ or WRITE support.
    Now "make check" will succeed even if libpng is compiled with -DPNG_NO_READ
    or -DPNG_NO_WRITE.  The tests performed are reduced, but the basic reading
    and writing of a PNG file is always tested by one or more of the tests.
  Consistently use strlen(), memset(), memcpy(), and memcmp() instead of the
    png_strlen(), png_memset(), png_memcpy(), and png_memcmp() macros.
  Removed the png_sizeof(), png_strlen(), png_memset(), png_memcpy(), and
    png_memcmp() macros.
  Work around gcc 3.x and Microsoft Visual Studio 2010 complaints. Both object
    to the split initialization of num_chunks.

Version 1.6.0beta28 [August 29, 2012]
  Unknown handling fixes and clean up. This adds more correct option
    control of the unknown handling, corrects the pre-existing bug where
    the per-chunk 'keep' setting is ignored and makes it possible to skip
    IDAT chunks in the sequential reader (broken in earlier 1.6 versions).
    There is a new test program, test-unknown.c, which is a work in progress
    (not currently part of the test suite).  Comments in the header files now
    explain how the unknown handling works.
  Allow fine grain control of unknown chunk APIs. This change allows
    png_set_keep_unknown_chunks() to be turned off if not required and causes
    both read and write to behave appropriately (on read this is only possible
    if the user callback is used to handle unknown chunks).  The change
    also removes the support for storing unknown chunks in the info_struct
    if the only unknown handling enabled is via the callback, allowing libpng
    to be configured with callback reading and none of the unnecessary code.
  Corrected fix for unknown handling in pngtest. This reinstates the
    libpng handling of unknown chunks other than vpAg and sTER (including
    unsafe-to-copy chunks which were dropped before) and eliminates the
    repositioning of vpAg and sTER in pngtest.png by changing pngtest.png
    (so the chunks are where libpng would put them).
  Added "tunknown" test and corrected a logic error in png_handle_unknown()
    when SAVE support is absent.  Moved the shell test scripts for
    contrib/libtests from the libpng top directory to contrib/libtests.
    png_handle_unknown() must always read or skip the chunk, if
    SAVE_UNKNOWN_CHUNKS is turned off *and* the application does not set
    a user callback an unknown chunk will not be read, leading to a read
    error, which was revealed by the "tunknown" test.
  Cleaned up and corrected ICC profile handling.
    contrib/libtests/makepng: corrected 'rgb' and 'gray' cases.  profile_error
    messages could be truncated; made a correct buffer size calculation and
    adjusted pngerror.c appropriately. png_icc_check_* checking improved;
    changed the functions to receive the correct color type of the PNG on read
    or write and check that it matches the color space of the profile (despite
    what the comments said before, there is danger in assuming the app will
    cope correctly with an RGB profile on a grayscale image and, since it
    violates the PNG spec, allowing it is certain to produce inconsistent
    app behavior and might even cause app crashes.) Check that profiles
    contain the tags needed to process the PNG (tags all required by the ICC
    spec). Removed unused PNG_STATIC from pngpriv.h.

Version 1.6.0beta29 [September 4, 2012]
  Fixed the simplified API example programs to add the *colormap parameter
    to several of he API and improved the error message if the version field
    is not set.
  Added contrib/examples/* to the *.zip and *.7z distributions.
  Updated simplified API synopses and description of the png_image structure
    in the manual.
  Made makepng and pngtest produce identical PNGs, add "--relaxed" option
    to pngtest. The "--relaxed" option turns off the benign errors that are
    enabled by default in pre-RC builds. makepng can now write ICC profiles
    where the length has not been extended to a multiple of 4, and pngtest
    now intercepts all libpng errors, allowing the previously-introduced
    "--strict test" on no warnings to actually work.
  Improved ICC profile handling including cHRM chunk generation and fixed
    Cygwin+MSVC build errors. The ICC profile handling now includes more
    checking.  Several errors that caused rejection of the profile are now
    handled with a warning in such a way that the invalid profiles will be
    read by default in release (but not pre-RC) builds but will not be
    written by default.  The easy part of handling the cHRM chunk is written,
    where the ICC profile contains the required data.  The more difficult
    part plus guessing a gAMA value requires code to pass selected RGB values
    through the profile.

Version 1.6.0beta30 [October 24, 2012]
  Changed ICC profile matrix/vector types to not depend on array type rules.
    By the ANSI-C standard the new types should be identical to the previous
    versions, and all known versions of gcc tested with the previous versions
    except for GCC-4.2.1 work with this version.  The change makes the ANSI-C
    rule that const applied to an array of elements applies instead to the
    elements in the array moot by explicitly applying const to the base
    elements of the png_icc_matrix and png_icc_vector types. The accidental
    (harmless) 'const' previously applied to the parameters of two of the
    functions have also been removed.
  Added a work around for GCC 4.2 optimization bug.
  Marked the broken (bad white point) original HP sRGB profiles correctly and
    correct comments.
  Added -DZ_SOLO to contrib/pngminim/*/makefile to work with zlib-1.2.7
  Use /MDd for vstudio debug builds. Also added pngunkown to the vstudio
    builds, fixed build errors and corrected a minor exit code error in
    pngvalid if the 'touch' file name is invalid.
  Add updated WARNING file to projects/vstudio from libpng 1.5/vstudio
  Fixed build when using #define PNG_NO_READ_GAMMA in png_do_compose() in
    pngrtran.c (Domani Hannes).

Version 1.6.0beta31 [November 1, 2012]
  Undid the erroneous change to vstudio/pngvalid build in libpng-1.6.0beta30.
  Made pngvalid so that it will build outside the libpng source tree.
  Made builds -DPNG_NO_READ_GAMMA compile (the unit tests still fail).
  Made PNG_NO_READ_GAMMA switch off interfaces that depend on READ_GAMMA.
    Prior to 1.6.0 switching off READ_GAMMA did unpredictable things to the
    interfaces that use it (specifically, png_do_background in 1.4 would
    simply display composite for grayscale images but do composition
    with the incorrect arithmetic for color ones). In 1.6 the semantic
    of -DPNG_NO_READ_GAMMA is changed to simply disable any interface that
    depends on it; this obliges people who set it to consider whether they
    really want it off if they happen to use any of the interfaces in
    question (typically most users who disable it won't).
  Fixed GUIDs in projects/vstudio. Some were duplicated or missing,
    resulting in VS2010 having to update the files.
  Removed non-working ICC profile support code that was mostly added to
    libpng-1.6.0beta29 and beta30. There was too much code for too little
    gain; implementing full ICC color correction may be desireable but is left
    up to applications.

Version 1.6.0beta32 [November 25, 2012]
  Fixed an intermittent SEGV in pngstest due to an uninitialized array element.
  Added the ability for contrib/libtests/makepng.c to make a PNG with just one
    color. This is useful for debugging pngstest color inaccuracy reports.
  Fixed error checking in the simplified write API (Olaf van der Spek)
  Made png_user_version_check() ok to use with libpng version 1.10.x and later.

Version 1.6.0beta33 [December 15, 2012]
  Fixed typo in png.c (PNG_SET_CHUNK_MALLOC_MAX should be PNG_CHUNK_MALLOC_MAX)
    that causes the MALLOC_MAX limit not to work (John Bowler)
  Change png_warning() to png_app_error() in pngwrite.c and comment the
    fall-through condition.
  Change png_warning() to png_app_warning() in png_write_tRNS().
  Rearranged the ARM-NEON optimizations: Isolated the machine specific code
    to the hardware subdirectory and added comments to pngrutil.c so that
    implementors of other optimizations know what to do.
  Fixed cases of unquoted DESTDIR in Makefile.am
  Rebuilt Makefile.in, etc., with autoconf-2.69 and automake-1.12.5.

Version 1.6.0beta34 [December 19, 2012]
  Cleaned up whitespace in the synopsis portion of the manpage "libpng.3"
  Disassembled the version number in scripts/options.awk (necessary for
    building on SunOs).

Version 1.6.0beta35 [December 23, 2012]
  Made default Zlib compression settings be configurable. This adds #defines to
    pnglibconf.h to control the defaults.
  Fixed Windows build issues, enabled ARM compilation. Various warnings issued
    by earlier versions of GCC fixed for Cygwin and Min/GW (which both use old
    GCCs.) ARM support is enabled by default in zlib.props (unsupported by
    Microsoft) and ARM compilation is made possible by deleting the check for
    x86. The test programs cannot be run because they are not signed.

Version 1.6.0beta36 [January 2, 2013]
  Discontinued distributing libpng-1.x.x.tar.bz2.
  Discontinued distributing libpng-1.7.0-1.6.0-diff.txt and similar.
  Rebuilt configure with autoconf-2.69 (inadvertently not done in beta33)
  Fixed 'make distcheck' on SUN OS - libpng.so was not being removed

Version 1.6.0beta37 [January 10, 2013]
  Fixed conceivable but difficult to repro overflow. Also added two test
    programs to generate and test a PNG which should have the problem.

Version 1.6.0beta39 [January 19, 2013]
  Again corrected attempt at overflow detection in png_set_unknown_chunks()
  (CVE-2013-7353).  Added overflow detection in png_set_sPLT() and
  png_set_text_2() (CVE-2013-7354).

Version 1.6.0beta40 [January 20, 2013]
  Use consistent handling of overflows in text, sPLT and unknown png_set_* APIs

Version 1.6.0rc01 [January 26, 2013]
  No changes.

Version 1.6.0rc02 [February 4, 2013]
  Added png_get_palette_max() function.

Version 1.6.0rc03 [February 5, 2013]
  Fixed the png_get_palette_max API.

Version 1.6.0rc04 [February 7, 2013]
  Turn serial tests back on (recently turned off by autotools upgrade).

Version 1.6.0rc05 [February 8, 2013]
  Update manual about png_get_palette_max().

Version 1.6.0rc06 [February 9, 2013]
  Fixed missing dependency in --prefix builds The intermediate
    internal 'prefix.h' file can only be generated correctly after
    pnglibconf.h, however the dependency was not in Makefile.am.  The
    symptoms are unpredictable depending on the order make chooses to
    build pngprefix.h and pnglibconf.h, often the error goes unnoticed
    because there is a system pnglibconf.h to use instead.

Version 1.6.0rc07 [February 10, 2013]
  Enclosed the new png_get_palette_max in #ifdef PNG_GET_PALETTE_MAX_SUPPORTED
    block, and revised pnglibconf.h and pnglibconf.h.prebuilt accordingly.

Version 1.6.0rc08 [February 10, 2013]
  Fix typo in png.h #ifdef

Version 1.6.0 [February 14, 2013]
  No changes.

Version 1.6.1beta01 [February 16, 2013]
  Made symbol prefixing work with the ARM neon optimizations. Also allow
    pngpriv.h to be included for preprocessor definitions only, so it can
    be used in non-C/C++ files. Back ported from libpng 1.7.
  Made sRGB check numbers consistent.
  Ported libpng 1.5 options.awk/dfn file handling to 1.6, fixed one bug.
  Removed cc -E workround, corrected png_get_palette_max API Tested on
    SUN OS cc 5.9, which demonstrates the tokenization problem previously
    avoided by using /lib/cpp.  Since all .dfn output is now protected in
    double quotes unless it is to be macro substituted the fix should
    work everywhere.
  Enabled parallel tests - back ported from libpng-1.7.
  scripts/pnglibconf.dfa formatting improvements back ported from libpng17.
  Fixed a race condition in the creation of the build 'scripts' directory
    while building with a parallel make.
  Use approved/supported Android method to check for NEON, use Linux/POSIX
    1003.1 API to check /proc/self/auxv avoiding buffer allocation and other
    library calls (ported from libpng15).

Version 1.6.1beta02 [February 19, 2013]
  Use parentheses more consistently in "#if defined(MACRO)" tests.
  Folded long lines.
  Reenabled code to allow zero length PLTE chunks for MNG.

Version 1.6.1beta03 [February 22, 2013]
  Fixed ALIGNED_MEMORY support.
  Added a new configure option:
    --enable-arm-neon=always will stop the run-time checks. New checks
    within arm/arm_init.c will cause the code not to be compiled unless
    __ARM_NEON__ is set. This should make it fail safe (if someone asks
    for it on then the build will fail if it can't be done.)
  Updated the INSTALL document.

Version 1.6.1beta04 [February 27, 2013]
  Revised INSTALL to recommend using CPPFLAGS instead of INCLUDES.
  Revised scripts/makefile.freebsd to respect ZLIBLIB and ZLIBINC.
  Revised scripts/dfn.awk to work with the buggy MSYS awk that has trouble
    with CRLF line endings.

Version 1.6.1beta05 [March 1, 2013]
  Avoid a possible memory leak in contrib/gregbook/readpng.c

Version 1.6.1beta06 [March 4, 2013]
  Better documentation of unknown handling API interactions.
  Corrected Android builds and corrected libpng.vers with symbol
    prefixing.  It also makes those tests compile and link on Android.
  Added an API png_set_option() to set optimization options externally,
    providing an alternative and general solution for the non-portable
    run-time tests used by the ARM Neon code, using the PNG_ARM_NEON option.
  The order of settings vs options in pnglibconf.h is reversed to allow
    settings to depend on options and options can now set (or override) the
    defaults for settings.

Version 1.6.1beta07 [March 7, 2013]
  Corrected simplified API default gamma for color-mapped output, added
    a flag to change default. In 1.6.0 when the simplified API was used
    to produce color-mapped output from an input image with no gamma
    information the gamma assumed for the input could be different from
    that assumed for non-color-mapped output.  In particular 16-bit depth
    input files were assumed to be sRGB encoded, whereas in the 'direct'
    case they were assumed to have linear data.  This was an error.  The
    fix makes the simplified API treat all input files the same way and
    adds a new flag to the png_image::flags member to allow the
    application/user to specify that 16-bit files contain sRGB data
    rather than the default linear.
  Fixed bugs in the pngpixel and makepng test programs.

Version 1.6.1beta08 [March 7, 2013]
  Fixed CMakelists.txt to allow building a single variant of the library
    (Claudio Bley):
  Introduced a PNG_LIB_TARGETS variable that lists all activated library
    targets.  It is an error if this variable ends up empty, ie. you have
    to build at least one library variant.
  Made the *_COPY targets only depend on library targets actually being build.
  Use PNG_LIB_TARGETS to unify a code path.
  Changed the CREATE_SYMLINK macro to expect the full path to a file as the
    first argument. When symlinking the filename component of that path is
    determined and used as the link target.
  Use copy_if_different in the CREATE_SYMLINK macro.

Version 1.6.1beta09 [March 13, 2013]
  Eliminated two warnings from the Intel C compiler. The warnings are
    technically valid, although a reasonable treatment of division would
    show it to be incorrect.

Version 1.6.1rc01 [March 21, 2013]
  No changes.

Version 1.6.1 [March 28, 2013]
  No changes.

Version 1.6.2beta01 [April 14, 2013]
  Updated documentation of 1.5.x to 1.6.x changes in iCCP chunk handling.
  Fixed incorrect warning of excess deflate data. End condition - the
    warning would be produced if the end of the deflate stream wasn't read
    in the last row.  The warning is harmless.
  Corrected the test on user transform changes on read. It was in the
    png_set of the transform function, but that doesn't matter unless the
    transform function changes the rowbuf size, and that is only valid if
    transform_info is called.
  Corrected a misplaced closing bracket in contrib/libtests/pngvalid.c
    (Flavio Medeiros).
  Corrected length written to uncompressed iTXt chunks (Samuli Suominen).
    Bug was introduced in libpng-1.6.0.

Version 1.6.2rc01 [April 18, 2013]
  Added contrib/tools/fixitxt.c, to repair the erroneous iTXt chunk length
    written by libpng-1.6.0 and 1.6.1.
  Disallow storing sRGB information when the sRGB is not supported.

Version 1.6.2rc02 [April 18, 2013]
  Merge pngtest.c with libpng-1.7.0

Version 1.6.2rc03 [April 22, 2013]
  Trivial spelling cleanup.

Version 1.6.2rc04 and 1.6.2rc05 [omitted]

Version 1.6.2rc06 [April 24, 2013]
  Reverted to version 1.6.2rc03.  Recent changes to arm/neon support
    have been ported to libpng-1.7.0beta09 and will reappear in version
    1.6.3beta01.

Version 1.6.2 [April 25, 2013]
  No changes.

Version 1.6.3beta01 [April 25, 2013]
  Revised stack marking in arm/filter_neon.S and configure.ac.
  Ensure that NEON filter stuff is completely disabled when switched 'off'.
    Previously the ARM NEON specific files were still built if the option
    was switched 'off' as opposed to being explicitly disabled.

Version 1.6.3beta02 [April 26, 2013]
  Test for 'arm*' not just 'arm' in the host_cpu configure variable.
  Rebuilt the configure scripts.

Version 1.6.3beta03 [April 30, 2013]
  Expanded manual paragraph about writing private chunks, particularly
    the need to call png_set_keep_unknown_chunks() when writing them.
  Avoid dereferencing NULL pointer possibly returned from
    png_create_write_struct() (Andrew Church).

Version 1.6.3beta05 [May 9, 2013]
  Calculate our own zlib windowBits when decoding rather than trusting the
    CMF bytes in the PNG datastream.
  Added an option to force maximum window size for inflating, which was
    the behavior of libpng15 and earlier, via a new PNG_MAXIMUM_INFLATE_WINDOW
    option for png_set_options().
  Added png-fix-itxt and png-fix-too-far-back to the built programs and
    removed warnings from the source code and timepng that are revealed as
    a result.
  Detect wrong libpng versions linked to png-fix-too-far-back, which currently
    only works with libpng versions that can be made to reliably fail when
    the deflate data contains an out-of-window reference.  This means only
    1.6 and later.
  Fixed gnu issues: g++ needs a static_cast, gcc 4.4.7 has a broken warning
    message which it is easier to work round than ignore.
  Updated contrib/pngminus/pnm2png.c (Paul Stewart):
    Check for EOF
    Ignore "#" delimited comments in input file to pnm2png.c.
    Fixed whitespace handling
    Added a call to png_set_packing()
    Initialize dimension values so if sscanf fails at least we have known
      invalid values.
  Attempt to detect configuration issues with png-fix-too-far-back, which
    requires both the correct libpng and the correct zlib to function
    correctly.
  Check ZLIB_VERNUM for mismatches, enclose #error in quotes
  Added information in the documentation about problems with and fixes for
    the bad CRC and bad iTXt chunk situations.

Version 1.6.3beta06 [May 12, 2013]
  Allow contrib/pngminus/pnm2png.c to compile without WRITE_INVERT and
    WRITE_PACK supported (writes error message that it can't read P1 or
    P4 PBM files).
  Improved png-fix-too-far-back usage message, added --suffix option.
  Revised contrib/pngminim/*/makefile to generate pnglibconf.h with the
    right zlib header files.
  Separated CPPFLAGS and CFLAGS in contrib/pngminim/*/makefile

Version 1.6.3beta07 [June 8, 2013]
  Removed a redundant test in png_set_IHDR().
  Added set(CMAKE_CONFIGURATION_TYPES ...) to CMakeLists.txt (Andrew Hundt)
  Deleted set(CMAKE_BUILD_TYPE) block from CMakeLists.txt
  Enclose the prototypes for the simplified write API in
    #ifdef PNG_STDIO_SUPPORTED/#endif
  Make ARM NEON support work at compile time (not just configure time).
    This moves the test on __ARM_NEON__ into pngconf.h to avoid issues when
    using a compiler that compiles for multiple architectures at one time.
  Removed PNG_FILTER_OPTIMIZATIONS and PNG_ARM_NEON_SUPPORTED from
    pnglibconf.h, allowing more of the decisions to be made internally
    (pngpriv.h) during the compile.  Without this, symbol prefixing is broken
    under certain circumstances on ARM platforms.  Now only the API parts of
    the optimizations ('check' vs 'api') are exposed in the public header files
    except that the new setting PNG_ARM_NEON_OPT documents how libpng makes the
    decision about whether or not to use the optimizations.
  Protect symbol prefixing against CC/CPPFLAGS/CFLAGS useage.
    Previous iOS/Xcode fixes for the ARM NEON optimizations moved the test
    on __ARM_NEON__ from configure time to compile time.  This breaks symbol
    prefixing because the definition of the special png_init_filter_functions
    call was hidden at configure time if the relevant compiler arguments are
    passed in CFLAGS as opposed to CC.  This change attempts to avoid all
    the confusion that would result by declaring the init function even when
    it is not used, so that it will always get prefixed.

Version 1.6.3beta08 [June 18, 2013]
  Revised libpng.3 so that "doclifter" can process it.

Version 1.6.3beta09 [June 27, 2013]
  Revised example.c to illustrate use of PNG_DEFAULT_sRGB and PNG_GAMMA_MAC_18
    as parameters for png_set_gamma().  These have been available since
    libpng-1.5.4.
  Renamed contrib/tools/png-fix-too-far-back.c to pngfix.c and revised it
    to check all compressed chunks known to libpng.

Version 1.6.3beta10 [July 5, 2013]
  Updated documentation to show default behavior of benign errors correctly.
  Only compile ARM code when PNG_READ_SUPPORTED is defined.
  Fixed undefined behavior in contrib/tools/pngfix.c and added new strip
    option. pngfix relied on undefined behavior and even a simple change from
    gcc to g++ caused it to fail.  The new strip option 'unsafe' has been
    implemented and is the default if --max is given.  Option names have
    been clarified, with --strip=transform now stripping the bKGD chunk,
    which was stripped previously with --strip=unused.
  Added all documented chunk types to pngpriv.h
  Unified pngfix.c source with libpng17.

Version 1.6.3rc01 [July 11, 2013]
  No changes.

Version 1.6.3 [July 18, 2013]
  Revised manual about changes in iTXt chunk handling made in libpng-1.6.0.
  Added "/* SAFE */" comments in pngrutil.c and pngrtran.c where warnings
    may be erroneously issued by code-checking applications.

Version 1.6.4beta01 [August 21, 2013]
  Added information about png_set_options() to the manual.
  Delay calling png_init_filter_functions() until a row with nonzero filter
    is found.

Version 1.6.4beta02 [August 30, 2013]
  Fixed inconsistent conditional compilation of png_chunk_unknown_handling()
    prototype, definition, and usage.  Made it depend on
    PNG_HANDLE_AS_UNKNOWN_SUPPORTED everywhere.

Version 1.6.4rc01 [September 5, 2013]
  No changes.

Version 1.6.4 [September 12, 2013]
  No changes.

Version 1.6.5 [September 14, 2013]
  Removed two stray lines of code from arm/arm_init.c.

Version 1.6.6 [September 16, 2013]
  Removed two stray lines of code from arm/arm_init.c, again.

Version 1.6.7beta01 [September 30, 2013]
  Revised unknown chunk code to correct several bugs in the NO_SAVE_/NO_WRITE
    combination
  Allow HANDLE_AS_UNKNOWN to work when other options are configured off. Also
    fixed the pngminim makefiles to work when $(MAKEFLAGS) contains stuff
    which terminates the make options (as by default in recent versions of
    Gentoo).
  Avoid up-cast warnings in pngvalid.c. On ARM the alignment requirements of
    png_modifier are greater than that of png_store and as a consequence
    compilation of pngvalid.c results in a warning about increased alignment
    requirements because of the bare cast to (png_modifier*). The code is safe,
    because the pointer is known to point to a stack allocated png_modifier,
    but this change avoids the warning.
  Fixed default behavior of ARM_NEON_API. If the ARM NEON API option was
    compiled without the CHECK option it defaulted to on, not off.
  Check user callback behavior in pngunknown.c. Previous versions compiled
    if SAVE_UNKNOWN was not available but did nothing since the callback
    was never implemented.
  Merged pngunknown.c with 1.7 version and back ported 1.7 improvements/fixes

Version 1.6.7beta02 [October 12, 2013]
  Made changes for compatibility with automake 1.14:
    1) Added the 'compile' program to the list of programs that must be cleaned
       in autogen.sh
    2) Added 'subdir-objects' which causes .c files in sub-directories to be
       compiled such that the corresponding .o files are also in the
       sub-directory.  This is because automake 1.14 warns that the
       current behavior of compiling to the top level directory may be removed
       in the future.
    3) Updated dependencies on pnglibconf.h to match the new .o locations and
       added all the files in contrib/libtests and contrib/tools that depend
       on pnglibconf.h
    4) Added 'BUILD_SOURCES = pnglibconf.h'; this is the automake recommended
       way of handling the dependencies of sources that are machine generated;
       unfortunately it only works if the user does 'make all' or 'make check',
       so the dependencies (3) are still required.
  Cleaned up (char*) casts of zlib messages. The latest version of the Intel C
    compiler complains about casting a string literal as (char*), so copied the
    treatment of z_const from the library code into pngfix.c
  Simplified error message code in pngunknown. The simplification has the
    useful side effect of avoiding a bogus warning generated by the latest
    version of the Intel C compiler (it objects to
    condition ? string-literal : string-literal).
  Make autogen.sh work with automake 1.13 as well as 1.14. Do this by always
    removing the 1.14 'compile' script but never checking for it.

Version 1.6.7beta03 [October 19, 2013]
  Added ARMv8 support (James Yu <james.yu at linaro.org>).  Added file
    arm/filter_neon_intrinsics.c; enable with -mfpu=neon.
  Revised pngvalid to generate size images with as many filters as it can
    manage, limited by the number of rows.
  Cleaned up ARM NEON compilation handling. The tests are now in pngpriv.h
    and detect the broken GCC compilers.

Version 1.6.7beta04 [October 26, 2013]
  Allow clang derived from older GCC versions to use ARM intrinsics. This
    causes all clang builds that use -mfpu=neon to use the intrinsics code,
    not the assembler code.  This has only been tested on iOS 7. It may be
    necessary to exclude some earlier clang versions but this seems unlikely.
  Changed NEON implementation selection mechanism. This allows assembler
    or intrinsics to be turned on at compile time during the build by defining
    PNG_ARM_NEON_IMPLEMENTATION to the correct value (2 or 1).  This macro
    is undefined by default and the build type is selected in pngpriv.h.

Version 1.6.7rc01 [November 2, 2013]
  No changes.

Version 1.6.7rc02 [November 7, 2013]
  Fixed #include in filter_neon_intrinsics.c and ctype macros. The ctype char
    checking macros take an unsigned char argument, not a signed char.

Version 1.6.7 [November 14, 2013]
  No changes.

Version 1.6.8beta01 [November 24, 2013]
  Moved prototype for png_handle_unknown() in pngpriv.h outside of
    the #ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED/#endif block.
  Added "-Wall" to CFLAGS in contrib/pngminim/*/makefile
  Conditionally compile some unused functions reported by -Wall in
    pngminim.
  Fixed 'minimal' builds. Various obviously useful minimal configurations
    don't build because of missing contrib/libtests test programs and
    overly complex dependencies in scripts/pnglibconf.dfa. This change
    adds contrib/conftest/*.dfa files that can be used in automatic build
    scripts to ensure that these configurations continue to build.
  Enabled WRITE_INVERT and WRITE_PACK in contrib/pngminim/encoder.
  Fixed pngvalid 'fail' function declaration on the Intel C Compiler.
    This reverts to the previous 'static' implementation and works round
    the 'unused static function' warning by using PNG_UNUSED().

Version 1.6.8beta02 [November 30, 2013]
  Removed or marked PNG_UNUSED some harmless "dead assignments" reported
    by clang scan-build.
  Changed tabs to 3 spaces in png_debug macros and changed '"%s"m'
    to '"%s" m' to improve portability among compilers.
  Changed png_free_default() to free() in pngtest.c

Version 1.6.8rc01 [December 12, 2013]
  Tidied up pngfix inits and fixed pngtest no-write builds.

Version 1.6.8rc02 [December 14, 2013]
  Handle zero-length PLTE chunk or NULL palette with png_error()
    instead of png_chunk_report(), which by default issues a warning
    rather than an error, leading to later reading from a NULL pointer
    (png_ptr->palette) in png_do_expand_palette(). This is CVE-2013-6954
    and VU#650142.  Libpng-1.6.1 through 1.6.7 are vulnerable.
    Libpng-1.6.0 and earlier do not have this bug.

Version 1.6.8 [December 19, 2013]
  No changes.

Version 1.6.9beta01 [December 26, 2013]
  Bookkeeping: Moved functions around (no changes). Moved transform
    function definitions before the place where they are called so that
    they can be made static. Move the intrapixel functions and the
    grayscale palette builder out of the png?tran.c files. The latter
    isn't a transform function and is no longer used internally, and the
    former MNG specific functions are better placed in pngread/pngwrite.c
  Made transform implementation functions static. This makes the internal
    functions called by png_do_{read|write}_transformations static. On an
    x86-64 DLL build (Gentoo Linux) this reduces the size of the text
    segment of the DLL by 1208 bytes, about 0.6%. It also simplifies
    maintenance by removing the declarations from pngpriv.h and allowing
    easier changes to the internal interfaces.
  Rebuilt configure scripts with automake-1.14.1 and autoconf-2.69
    in the tar distributions.

Version 1.6.9beta02 [January 1, 2014]
  Added checks for libpng 1.5 to pngvalid.c.  This supports the use of
    this version of pngvalid in libpng 1.5
  Merged with pngvalid.c from libpng-1.7 changes to create a single
    pngvalid.c
  Removed #error macro from contrib/tools/pngfix.c (Thomas Klausner).
  Merged pngrio.c, pngtrans.c, pngwio.c, and pngerror.c with libpng-1.7.0
  Merged libpng-1.7.0 changes to make no-interlace configurations work
    with test programs.
  Revised pngvalid.c to support libpng 1.5, which does not support the
    PNG_MAXIMUM_INFLATE_WINDOW option, so #define it out when appropriate in
    pngvalid.c
  Allow unversioned links created on install to be disabled in configure.
    In configure builds 'make install' changes/adds links like png.h
    and libpng.a to point to the newly installed, versioned, files (e.g.
    libpng17/png.h and libpng17.a). Three new configure options and some
    rearrangement of Makefile.am allow creation of these links to be disabled.

Version 1.6.9beta03 [January 10, 2014]
  Removed potentially misleading warning from png_check_IHDR().

Version 1.6.9beta04 [January 20, 2014]
  Updated scripts/makefile.* to use CPPFLAGS (Cosmin).
  Added clang attribute support (Cosmin).

Version 1.6.9rc01 [January 28, 2014]
  No changes.

Version 1.6.9rc02 [January 30, 2014]
  Quiet an uninitialized memory warning from VC2013 in png_get_png().

Version 1.6.9 [February 6, 2014]

Version 1.6.10beta01 [February 9, 2014]
  Backported changes from libpng-1.7.0beta30 and beta31:
  Fixed a large number of instances where PNGCBAPI was omitted from
    function definitions.
  Added pngimage test program for png_read_png() and png_write_png()
    with two new test scripts.
  Removed dependence on !PNG_READ_EXPAND_SUPPORTED for calling
    png_set_packing() in png_read_png().
  Fixed combination of ~alpha with shift. On read invert alpha, processing
    occurred after shift processing, which causes the final values to be
    outside the range that should be produced by the shift. Reversing the
    order on read makes the two transforms work together correctly and mirrors
    the order used on write.
  Do not read invalid sBIT chunks. Previously libpng only checked sBIT
    values on write, so a malicious PNG writer could therefore cause
    the read code to return an invalid sBIT chunk, which might lead to
    application errors or crashes.  Such chunks are now skipped (with
    chunk_benign_error).
  Make png_read_png() and png_write_png() prototypes in png.h depend
    upon PNG_READ_SUPPORTED and PNG_WRITE_SUPPORTED.
  Support builds with unsupported PNG_TRANSFORM_* values.  All of the
    PNG_TRANSFORM_* values are always defined in png.h and, because they
    are used for both read and write in some cases, it is not reliable
    to #if out ones that are totally unsupported. This change adds error
    detection in png_read_image() and png_write_image() to do a
    png_app_error() if the app requests something that cannot be done
    and it adds corresponding code to pngimage.c to handle such options
    by not attempting to test them.

Version 1.6.10beta02 [February 23, 2014]
  Moved redefines of png_error(), png_warning(), png_chunk_error(),
    and png_chunk_warning() from pngpriv.h to png.h to make them visible
    to libpng-calling applications.
  Moved OS dependent code from arm/arm_init.c, to allow the included
    implementation of the ARM NEON discovery function to be set at
    build-time and provide sample implementations from the current code in the
    contrib/arm-neon subdirectory. The __linux__ code has also been changed to
    compile and link on Android by using /proc/cpuinfo, and the old linux code
    is in contrib/arm-neon/linux-auxv.c.  The new code avoids POSIX and Linux
    dependencies apart from opening /proc/cpuinfo and is C90 compliant.
  Check for info_ptr == NULL early in png_read_end() so we don't need to
    run all the png_handle_*() and depend on them to return if info_ptr == NULL.
    This improves the performance of png_read_end(png_ptr, NULL) and makes
    it more robust against future programming errors.
  Check for __has_extension before using it in pngconf.h, to
    support older Clang versions (Jeremy Sequoia).
  Treat CRC error handling with png_set_crc_action(), instead of with
    png_set_benign_errors(), which has been the case since libpng-1.6.0beta18.
  Use a user warning handler in contrib/gregbook/readpng2.c instead of default,
    so warnings will be put on stderr even if libpng has CONSOLE_IO disabled.
  Added png_ptr->process_mode = PNG_READ_IDAT_MODE in png_push_read_chunk
    after recognizing the IDAT chunk, which avoids an infinite loop while
    reading a datastream whose first IDAT chunk is of zero-length.
    This fixes CERT VU#684412 and CVE-2014-0333.
  Don't recognize known sRGB profiles as sRGB if they have been hacked,
    but don't reject them and don't issue a copyright violation warning.

Version 1.6.10beta03 [February 25, 2014]
  Moved some documentation from png.h to libpng.3 and libpng-manual.txt
  Minor editing of contrib/arm-neon/README and contrib/examples/*.c

Version 1.6.10rc01 [February 27, 2014]
  Fixed typos in the manual and in scripts/pnglibconf.dfa (CFLAGS -> CPPFLAGS
    and PNG_USR_CONFIG -> PNG_USER_CONFIG).

Version 1.6.10rc02 [February 28, 2014]
  Removed unreachable return statement after png_chunk_error()
    in pngrutil.c

Version 1.6.10rc03 [March 4, 2014]
  Un-deprecated png_data_freer().

Version 1.6.10 [March 6, 2014]
  No changes.

Version 1.6.11beta01 [March 17, 2014]
  Use "if (value != 0)" instead of "if (value)" consistently.
  Changed ZlibSrcDir from 1.2.5 to 1.2.8 in projects/vstudio.
  Moved configuration information from the manual to the INSTALL file.

Version 1.6.11beta02 [April 6, 2014]
  Removed #if/#else/#endif from inside two pow() calls in pngvalid.c because
    they were handled improperly by Portland Group's PGI-14.1 - PGI-14.3
    when using its "__builtin_pow()" function.
  Silence 'unused parameter' build warnings (Cosmin Truta).
  $(CP) is now used alongside $(RM_F).  Also, use 'copy' instead of 'cp'
    where applicable, and applied other minor makefile changes (Cosmin).
  Don't warn about invalid dimensions exceeding user limits (Cosmin).
  Allow an easy replacement of the default pre-built configuration
    header with a custom header, via the make PNGLIBCONF_H_PREBUILT
    macro (Cosmin).

Version 1.6.11beta03 [April 6, 2014]
  Fixed a typo in pngrutil.c, introduced in libpng-1.5.6, that interferes
    with "blocky" expansion of sub-8-bit interlaced PNG files (Eric Huss).
  Optionally use  __builtin_bswap16() in png_do_swap().

Version 1.6.11beta04 [April 19, 2014]
  Made progressive reading of interlaced images consistent with the
    behavior of the sequential reader and consistent with the manual, by
    moving some code out of the PNG_READ_INTERLACING_SUPPORTED blocks. The
    row_callback now receives the proper pass number and unexpanded rows, when
    png_combine_row() isn't built or used, and png_set_interlace_handling()
    is not called.
  Allow PNG_sRGB_PROFILE_CHECKING = (-1) to mean no sRGB profile checking.

Version 1.6.11beta05 [April 26, 2014]
  Do not reject ICC V2 profiles that lack padding (Kai-Uwe Behrmann).
  Relocated closing bracket of the sRGB profile test loop to avoid getting
    "Not recognizing known sRGB profile that has been edited" warning for
    ICC V2 profiles that lack the MD5 signature in the profile header.

Version 1.6.11beta06 [May 19, 2014]
  Added PNG_SKIP_sRGB_CHECK_PROFILE choice for png_set_option().

Version 1.6.11rc01 [May 27, 2014]
  No changes.

Version 1.6.11rc02 [June 3, 2014]
  Test ZLIB_VERNUM instead of PNG_ZLIB_VERNUM in contrib/tools/pngfix.c

Version 1.6.11 [June 5, 2014]
  No changes.

Version 1.6.12rc01 [June 6, 2014]
  Relocated new code from 1.6.11beta06 in png.c to a point after the
    declarations (Max Stepin).

Version 1.6.12rc02 [June 7, 2014]
  Changed file permissions of contrib/tools/intgamma.sh,
    test-driver, and compile from 0644 to 0755 (Cosmin).

Version 1.6.12rc03 [June 8, 2014]
  Ensure "__has_attribute()" macro exists before trying to use it with
    old clang compilers (MacPorts Ticket #43939).

Version 1.6.12 [June 12, 2014]
  No changes.

Version 1.6.13beta01 [July 4, 2014]
  Quieted -Wsign-compare and -Wclobber compiler warnings in
    contrib/pngminus/*.c
  Added "(void) png_ptr;" where needed in contrib/gregbook to quiet
    compiler complaints about unused pointers.
  Split a long output string in contrib/gregbook/rpng2-x.c.
  Added "PNG_SET_OPTION" requirement for sRGB chunk support to pnglibconf.dfa,
    Needed for write-only support (John Bowler).
  Changed "if defined(__ARM_NEON__)" to
    "if (defined(__ARM_NEON__) || defined(__ARM_NEON))" (James Wu).
  Fixed clang no-warning builds: png_digit was defined but never used.
    
Version 1.6.13beta02 [July 21, 2014]
  Fixed an incorrect separator ("/" should be "\") in scripts/makefile.vcwin32
    (bug report from Wolfgang S. Kechel).  Bug was introduced in libpng-1.6.11.
    Also fixed makefile.bc32, makefile.bor, makefile.msc, makefile.intel, and
    makefile.tc3 similarly.

Version 1.6.13beta03 [August 3, 2014]
  Removed scripts/makefile.elf. It has not worked since libpng-1.5.0beta14
    due to elimination of the PNG_FUNCTION_EXPORT and PNG_DATA_EXPORT
    definitions from pngconf.h.
  Ensure that CMakeLists.txt makes the target "lib" directory before making
    symbolic link into it (SourceForge bug report #226 by Rolf Timmermans).

Version 1.6.13beta04 [August 8, 2014]
  Added opinion that the ECCN (Export Control Classification Number) for
    libpng is EAR99 to the README file.
  Eliminated use of "$<" in makefile explicit rules, when copying
    $PNGLIBCONF_H_PREBUILT.  This does not work on some versions of make;
    bug introduced in libpng version 1.6.11.

Version 1.6.13rc01 [August 14, 2014]
  Made "ccopts" agree with "CFLAGS" in scripts/makefile.hp* and makefile.*sunu

Version 1.6.13 [August 21, 2014]
  No changes.

Version 1.6.14beta01 [September 14, 2014]
  Guard usage of png_ptr->options with #ifdef PNG_SET_OPTION_SUPPORTED.
  Do not build contrib/tools/pngfix.c when PNG_SETJMP_NOT_SUPPORTED,
    to allow "make" to complete without setjmp support (bug report by
    Claudio Fontana)
  Add "#include <setjmp.h>" to contrib/tools/pngfix.c (John Bowler)

Version 1.6.14beta02 [September 18, 2014]
  Use nanosleep() instead of usleep() in contrib/gregbook/rpng2-x.c
    because usleep() is deprecated.
  Define usleep() in contrib/gregbook/rpng2-x.c if not already defined
    in unistd.h and nanosleep() is not available; fixes error introduced
    in libpng-1.6.13.
  Disable floating point exception handling in pngvalid.c when
    PNG_FLOATING_ARITHMETIC is not supported (bug report by "zootus
    at users.sourceforge.net").

Version 1.6.14beta03 [September 19, 2014]
  Define FE_DIVBYZERO, FE_INVALID, and FE_OVERFLOW in pngvalid.c if not
    already defined.  Revert floating point exception handling in pngvalid.c
    to version 1.6.14beta01 behavior.

Version 1.6.14beta04 [September 27, 2014]
  Fixed incorrect handling of the iTXt compression flag in pngrutil.c
    (bug report by Shunsaku Hirata).  Bug was introduced in libpng-1.6.0.

Version 1.6.14beta05 [October 1, 2014]
  Added "option READ_iCCP enables READ_COMPRESSED_TEXT" to pnglibconf.dfa

Version 1.6.14beta06 [October 5, 2014]
  Removed unused "text_len" parameter from private function png_write_zTXt().
  Conditionally compile some code in png_deflate_claim(), when
    PNG_WARNINGS_SUPPORTED and PNG_ERROR_TEXT_SUPPORTED are disabled.
  Replaced repeated code in pngpread.c with PNG_PUSH_SAVE_BUFFER_IF_FULL.
  Added "chunk iTXt enables TEXT" and "chunk zTXt enables TEXT"
    to pnglibconf.dfa.
  Removed "option READ_COMPRESSED_TEXT enables READ_TEXT" from pnglibconf.dfa,
    to make it possible to configure a libpng that supports iCCP but not TEXT.

Version 1.6.14beta07 [October 7, 2014]
  Removed "option WRITE_COMPRESSED_TEXT enables WRITE_TEXT" from pnglibconf.dfa
  Only mark text chunks as written after successfully writing them.

Version 1.6.14rc01 [October 15, 2014]
  Fixed some typos in comments.

Version 1.6.14rc02 [October 17, 2014]
  Changed png_convert_to_rfc_1123() to png_convert_to_rfc_1123_buffer()
    in the manual, to reflect the change made in libpng-1.6.0.
  Updated README file to explain that direct access to the png_struct
    and info_struct members has not been permitted since libpng-1.5.0.

Version 1.6.14 [October 23, 2014]
  No changes.

Version 1.6.15beta01 [October 29, 2014]
  Changed "if (!x)" to "if (x == 0)" and "if (x)" to "if (x != 0)"
  Simplified png_free_data().
  Added missing "ptr = NULL" after some instances of png_free().

Version 1.6.15beta02 [November 1, 2014]
  Changed remaining "if (!x)" to "if (x == 0)" and "if (x)" to "if (x != 0)"

Version 1.6.15beta03 [November 3, 2014]
  Added PNG_USE_ARM_NEON configuration flag (Marcin Juszkiewicz).

Version 1.6.15beta04 [November 4, 2014]
  Removed new PNG_USE_ARM_NEON configuration flag and made a one-line
    revision to configure.ac to support ARM on aarch64 instead (John Bowler).

Version 1.6.15beta05 [November 5, 2014]
  Use png_get_libpng_ver(NULL) instead of PNG_LIBPNG_VER_STRING in
    example.c, pngtest.c, and applications in the contrib directory.
  Fixed an out-of-range read in png_user_version_check() (Bug report from
    Qixue Xiao, CVE-2015-8540).
  Simplified and future-proofed png_user_version_check().
  Fixed GCC unsigned int->float warnings. Various versions of GCC
    seem to generate warnings when an unsigned value is implicitly
    converted to double. This is probably a GCC bug but this change
    avoids the issue by explicitly converting to (int) where safe.
  Free all allocated memory in pngimage. The file buffer cache was left
    allocated at the end of the program, harmless but it causes memory
    leak reports from clang.
  Fixed array size calculations to avoid warnings. At various points
    in the code the number of elements in an array is calculated using
    sizeof.  This generates a compile time constant of type (size_t) which
    is then typically assigned to an (unsigned int) or (int). Some versions
    of GCC on 64-bit systems warn about the apparent narrowing, even though
    the same compiler does apparently generate the correct, in-range,
    numeric constant.  This adds appropriate, safe, casts to make the
    warnings go away.

Version 1.6.15beta06 [November 6, 2014]
  Reverted use png_get_libpng_ver(NULL) instead of PNG_LIBPNG_VER_STRING
    in the manual, example.c, pngtest.c, and applications in the contrib
    directory.  It was incorrect advice.

Version 1.6.15beta07 [November 7, 2014]
  Removed #ifdef PNG_16BIT_SUPPORTED/#endif around png_product2(); it is
    needed by png_reciprocal2().
  Added #ifdef PNG_16BIT_SUPPORTED/#endif around png_log16bit() and
    png_do_swap().
  Changed all "#endif /* PNG_FEATURE_SUPPORTED */" to "#endif /* FEATURE */"

Version 1.6.15beta08 [November 8, 2014]
  More housecleaning in *.h

Version 1.6.15rc01 [November 13, 2014]

Version 1.6.15rc02 [November 14, 2014]
  The macros passed in the command line to Borland make were ignored if
    similarly-named macros were already defined in makefiles. This behavior
    is different from POSIX make and other make programs.  Surround the
    macro definitions with ifndef guards (Cosmin).

Version 1.6.15rc03 [November 16, 2014]
  Added "-D_CRT_SECURE_NO_WARNINGS" to CFLAGS in scripts/makefile.vcwin32.
  Removed the obsolete $ARCH variable from scripts/makefile.darwin.

Version 1.6.15 [November 20, 2014]
  No changes.

Version 1.6.16beta01 [December 14, 2014]
  Added ".align 2" to arm/filter_neon.S to support old GAS assemblers that
    don't do alignment correctly.
  Revised Makefile.am and scripts/symbols.dfn to work with MinGW/MSYS
    (Bob Friesenhahn).

Version 1.6.16beta02 [December 15, 2014]
  Revised Makefile.am and scripts/*.dfn again to work with MinGW/MSYS;
    renamed scripts/*.dfn to scripts/*.c (John Bowler).

Version 1.6.16beta03 [December 21, 2014]
  Quiet a "comparison always true" warning in pngstest.c (John Bowler).

Version 1.6.16rc01 [December 21, 2014]
  Restored a test on width that was removed from png.c at libpng-1.6.9
    (Bug report by Alex Eubanks, CVE-2015-0973).

Version 1.6.16rc02 [December 21, 2014]
  Undid the update to pngrutil.c in 1.6.16rc01.

Version 1.6.16rc03 [December 21, 2014]
  Fixed an overflow in png_combine_row() with very wide interlaced images
    (Bug report and fix by John Bowler, CVE-2014-9495).

Version 1.6.16 [December 22, 2014]
  No changes.

Version 1.6.17beta01 [January 29, 2015]
  Removed duplicate PNG_SAFE_LIMITS_SUPPORTED handling from pngconf.h
  Corrected the width limit calculation in png_check_IHDR().
  Removed user limits from pngfix. Also pass NULL pointers to
    png_read_row to skip the unnecessary row de-interlace stuff.
  Added testing of png_set_packing() to pngvalid.c
  Regenerated configure scripts in the *.tar distributions with libtool-2.4.4
  Implement previously untested cases of libpng transforms in pngvalid.c
  Fixed byte order in png_do_read_filler() with 16-bit input. Previously
    the high and low bytes of the filler, from png_set_filler() or from
    png_set_add_alpha(), were read in the wrong order.
  Made the check for out-of-range values in png_set_tRNS() detect
    values that are exactly 2^bit_depth, and work on 16-bit platforms.
  Merged some parts of libpng-1.6.17beta01 and libpng-1.7.0beta47.
  Added #ifndef __COVERITY__ where needed in png.c, pngrutil.c and
    pngset.c to avoid warnings about dead code.
  Added "& 0xff" to many instances of expressions that are typecast
    to (png_byte), to avoid Coverity warnings.

Version 1.6.17beta02 [February 7, 2015]
  Work around one more Coverity-scan dead-code warning.
  Do not build png_product2() when it is unused.

Version 1.6.17beta03 [February 17, 2015]
  Display user limits in the output from pngtest.
  Eliminated the PNG_SAFE_LIMITS macro and restored the 1-million-column
    and 1-million-row default limits in pnglibconf.dfa, that can be reset
    by the user at build time or run time.  This provides a more robust
    defense against DOS and as-yet undiscovered overflows.

Version 1.6.17beta04 [February 21, 2015]
  Added PNG_WRITE_CUSTOMIZE_COMPRESSION_SUPPORTED macro, on by default.
  Allow user to call png_get_IHDR() with NULL arguments (Reuben Hawkins).
  Rebuilt configure scripts with automake-1.15 and libtool-2.4.6

Version 1.6.17beta05 [February 25, 2015]
  Restored compiling of png_reciprocal2 with PNG_NO_16BIT.

Version 1.6.17beta06 [February 27, 2015]
  Moved png_set_filter() prototype into a PNG_WRITE_SUPPORTED block
    of png.h.
  Avoid runtime checks when converting integer to png_byte with
    Visual Studio (Sergey Kosarevsky)

Version 1.6.17rc01 [March 4, 2015]
  No changes.

Version 1.6.17rc02 [March 9, 2015]
  Removed some comments that the configure script did not handle
    properly from scripts/pnglibconf.dfa and pnglibconf.h.prebuilt.
  Free the unknown_chunks structure even when it contains no data.

Version 1.6.17rc03 [March 12, 2015]
  Updated CMakeLists.txt to add OSX framework, change YES/NO to ON/OFF
    for consistency, and remove some useless tests (Alexey Petruchik).

Version 1.6.17rc04 [March 16, 2015]
  Remove pnglibconf.h, pnglibconf.c, and pnglibconf.out instead of
    pnglibconf.* in "make clean" (Cosmin).
  Fix bug in calculation of maxbits, in png_write_sBIT, introduced
    in libpng-1.6.17beta01 (John Bowler).

Version 1.6.17rc05 [March 21, 2015]
  Define PNG_FILTER_* and PNG_FILTER_VALUE_* in png.h even when WRITE
    is not supported (John Bowler).  This fixes an error introduced in
    libpng-1.6.17beta06.
  Reverted "& 0xff" additions of version 1.6.17beta01. Libpng passes
    the Coverity scan without them.

Version 1.6.17rc06 [March 23, 2015]
  Remove pnglibconf.dfn and pnglibconf.pre with "make clean".
  Reformatted some "&0xff" instances to "& 0xff".
  Fixed simplified 8-bit-linear to sRGB alpha. The calculated alpha
    value was wrong.  It's not clear if this affected the final stored
    value; in the obvious code path the upper and lower 8-bits of the
    alpha value were identical and the alpha was truncated to 8-bits
    rather than dividing by 257 (John Bowler).

Version 1.6.17 [March 26, 2015]
  No changes.

Version 1.6.18beta01 [April 1, 2015]
  Removed PNG_SET_CHUNK_[CACHE|MALLOC]_LIMIT_SUPPORTED macros.  They
    have been combined with PNG_SET_USER_LIMITS_SUPPORTED (resolves
    bug report by Andrew Church).
  Fixed rgb_to_gray checks and added tRNS checks to pngvalid.c.  This
    fixes some arithmetic errors that caused some tests to fail on
    some 32-bit platforms (Bug reports by Peter Breitenlohner [i686]
    and Petr Gajdos [i586]).

Version 1.6.18beta02 [April 26, 2015]
  Suppressed some warnings from the Borland C++ 5.5.1/5.82 compiler
    (Bug report by Viktor Szakats).

Version 1.6.18beta03 [May 6, 2015]
  Replaced "unexpected" with an integer (0xabadca11) in pngset.c
    where a long was expected, to avoid a compiler warning when PNG_DEBUG > 1.
  Added contrib/examples/simpleover.c, to demonstrate how to handle
    alpha compositing of multiple images, using the "simplified API"
    and an example PNG generation tool, contrib/examples/genpng.c
    (John Bowler).

Version 1.6.18beta04 [May 20, 2015]
  PNG_RELEASE_BUILD replaces tests where the code depended on the build base
    type and can be defined on the command line, allowing testing in beta
    builds (John Bowler).
  Avoid Coverity issue 80858 (REVERSE NULL) in pngtest.c PNG_DEBUG builds.
  Avoid a harmless potential integer overflow in png_XYZ_from_xy() (Bug
    report from Christopher Ferris).

Version 1.6.18beta05 [May 31, 2015]
  Backport filter selection code from libpng-1.7.0beta51, to combine
    sub_row, up_row, avg_row, and paeth_row into try_row and tst_row.
  Changed png_voidcast(), etc., to voidcast(), etc., in contrib/tools/pngfix.c
    to avoid confusion with the libpng private macros.
  Fixed old cut&paste bug in the weighted filter selection code in
    pngwutil.c, introduced in libpng-0.95, March 1997.

Version 1.6.18beta06 [June 1, 2015]
  Removed WRITE_WEIGHTED_FILTERED code, to save a few kbytes of the
    compiled library size. It never worked properly and as far as we can
    tell, no one uses it. The png_set_filter_heuristics() and
    png_set_filter_heuristics_fixed() APIs are retained but deprecated
    and do nothing.

Version 1.6.18beta07 [June 6, 2015]
  Removed non-working progressive reader 'skip' function. This
    function has apparently never been used. It was implemented
    to support back-door modification of png_struct in libpng-1.4.x
    but (because it does nothing and cannot do anything) was apparently
    never tested (John Bowler).
  Fixed cexcept.h in which GCC 5 now reports that one of the auto
    variables in the Try macro needs to be volatile to prevent value
    being lost over the setjmp (John Bowler).
  Fixed NO_WRITE_FILTER and -Wconversion build breaks (John Bowler).
  Fix g++ build breaks (John Bowler).
  Quieted some Coverity issues in pngfix.c, png-fix-itxt.c, pngvalid.c,
    pngstest.c, and pngimage.c. Most seem harmless, but png-fix-itxt
    would only work with iTXt chunks with length 255 or less.
  Added #ifdef's to contrib/examples programs so people don't try
    to compile them without the minimum required support enabled
    (suggested by Flavio Medeiros).

Version 1.6.18beta08 [June 30, 2015]
  Eliminated the final two Coverity defects (insecure temporary file
    handling in contrib/libtests/pngstest.c; possible overflow of
    unsigned char in contrib/tools/png-fix-itxt.c). To use the "secure"
    file handling, define PNG_USE_MKSTEMP, otherwise "tmpfile()" will
    be used.
  Removed some unused WEIGHTED_FILTER macros from png.h and pngstruct.h

Version 1.6.18beta09 [July 5, 2015]
  Removed some useless typecasts from contrib/tools/png-fix-itxt.c
  Fixed a new signed-unsigned comparison in pngrtran.c (Max Stepin).
  Replaced arbitrary use of 'extern' with #define PNG_LINKAGE_*.  To
    preserve API compatibility, the new defines all default to "extern"
    (requested by Jan Nijtmans).

Version 1.6.18rc01 [July 9, 2015]
  Belatedly added Mans Rullgard and James Yu to the list of Contributing
    Authors.

Version 1.6.18rc02 [July 12, 2015]
  Restored unused FILTER_HEURISTIC macros removed at libpng-1.6.18beta08
    to png.h to avoid compatibility warnings.

Version 1.6.18rc03 [July 15, 2015]
  Minor changes to the man page

Version 1.6.18 [July 23, 2015]
  No changes.

Version 1.6.19beta01 [July 30, 2015]
  Updated obsolete information about the simplified API macros in the
    manual pages (Bug report by Arc Riley).
  Avoid potentially dereferencing NULL info_ptr in png_info_init_3().
  Rearranged png.h to put the major sections in the same order as
    in libpng17.
  Eliminated unused PNG_COST_SHIFT, PNG_WEIGHT_SHIFT, PNG_COST_FACTOR, and
    PNG_WEIGHT_FACTOR macros.
  Suppressed some warnings from the Borland C++ 5.5.1/5.82 compiler
    (Bug report by Viktor Szakats).  Several warnings remain and are
    unavoidable, where we test for overflow.
  Fixed potential leak of png_pixels in contrib/pngminus/pnm2png.c
  Fixed uninitialized variable in contrib/gregbook/rpng2-x.c

Version 1.6.19beta02 [August 19, 2015]
  Moved config.h.in~ from the "libpng_autotools_files" list to the
    "libpng_autotools_extra" list in autogen.sh because it was causing a
    false positive for missing files (bug report by Robert C. Seacord).
  Removed unreachable "break" statements in png.c, pngread.c, and pngrtran.c
    to suppress clang warnings (Bug report by Viktor Szakats).
  Fixed some bad links in the man page.
  Changed "n bit" to "n-bit" in comments.
  Added signed/unsigned 16-bit safety net. This removes the dubious
    0x8000 flag definitions on 16-bit systems. They aren't supported
    yet the defs *probably* work, however it seems much safer to do this
    and be advised if anyone, contrary to advice, is building libpng 1.6
    on a 16-bit system. It also adds back various switch default clauses
    for GCC; GCC errors out if they are not present (with an appropriately
    high level of warnings).
  Safely convert num_bytes to a png_byte in png_set_sig_bytes() (Robert
    Seacord).
  Fixed the recently reported 1's complement security issue by replacing
    the value that is illegal in the PNG spec, in both signed and unsigned
    values, with 0. Illegal unsigned values (anything greater than or equal
    to  0x80000000) can still pass through, but since these are not illegal
    in ANSI-C (unlike 0x80000000 in the signed case) the checking that
    occurs later can catch them (John Bowler).

Version 1.6.19beta03 [September 26, 2015]
  Fixed png_save_int_32 when int is not 2's complement (John Bowler).
  Updated libpng16 with all the recent test changes from libpng17,
    including changes to pngvalid.c to ensure that the original,
    distributed, version of contrib/visupng/cexcept.h can be used
    (John Bowler).
  pngvalid contains the correction to the use of SAVE/STORE_
    UNKNOWN_CHUNKS; a bug revealed by changes in libpng 1.7. More
    tests contain the --strict option to detect warnings and the
    pngvalid-standard test has been corrected so that it does not
    turn on progressive-read. There is a separate test which does
    that. (John Bowler)
  Also made some signed/unsigned fixes.
  Make pngstest error limits version specific. Splitting the machine
    generated error structs out to a file allows the values to be updated
    without changing pngstest.c itself. Since libpng 1.6 and 1.7 have
    slightly different error limits this simplifies maintenance. The
    makepngs.sh script has also been updated to more accurately reflect
    current problems in libpng 1.7 (John Bowler).
  Incorporated new test PNG files into make check.  tests/pngstest-*
    are changed so that the new test files are divided into 8 groups by
    gamma and alpha channel.  These tests have considerably better code
    and pixel-value coverage than contrib/pngsuite; however,coverage is
    still incomplete (John Bowler).
  Removed the '--strict' in 1.6 because of the double-gamma-correction
    warning, updated pngstest-errors.h for the errors detected with the
    new contrib/testspngs PNG test files (John Bowler).

Version 1.6.19beta04 [October 15, 2015]
  Worked around rgb-to-gray issues in libpng 1.6.  The previous
    attempts to ignore the errors in the code aren't quite enough to
    deal with the 'channel selection' encoding added to libpng 1.7; abort.
    pngvalid.c is changed to drop this encoding in prior versions.
  Fixed 'pow' macros in pngvalid.c. It is legal for 'pow' to be a
    macro, therefore the argument list cannot contain preprocessing
    directives.  Make sure pow is a function where this happens. This is
    a minimal safe fix, the issue only arises in non-performance-critical
    code (bug report by Curtis Leach, fix by John Bowler).
  Added sPLT support to pngtest.c

Version 1.6.19rc01 [October 23, 2015]
  No changes.

Version 1.6.19rc02 [October 31, 2015]
  Prevent setting or writing over-length PLTE chunk (Cosmin Truta).
  Silently truncate over-length PLTE chunk while reading.
  Libpng incorrectly calculated the output rowbytes when the application
    decreased either the number of channels or the bit depth (or both) in
    a user transform.  This was safe; libpng overallocated buffer space
   (potentially by quite a lot; up to 4 times the amount required) but,
   from 1.5.4 on, resulted in a png_error (John Bowler).

Version 1.6.19rc03 [November 3, 2015]
  Fixed some inconsequential cut-and-paste typos in png_set_cHRM_XYZ_fixed().
  Clarified COPYRIGHT information to state explicitly that versions
    are derived from previous versions.
  Removed much of the long list of previous versions from png.h and
    libpng.3.

Version 1.6.19rc04 [November 5, 2015]
  Fixed new bug with CRC error after reading an over-length palette
    (bug report by Cosmin Truta) (CVE-2015-8126).

Version 1.6.19 [November 12, 2015]
  Cleaned up coding style in png_handle_PLTE().

Version 1.6.20beta01 [November 20, 2015]
  Avoid potential pointer overflow/underflow in png_handle_sPLT() and
    png_handle_pCAL() (Bug report by John Regehr).

Version 1.6.20beta02 [November 23, 2015]
  Fixed incorrect implementation of png_set_PLTE() that uses png_ptr
    not info_ptr, that left png_set_PLTE() open to the CVE-2015-8126
    vulnerability.  Fixes CVE-2015-8472.

Version 1.6.20beta03 [November 24, 2015]
  Backported tests from libpng-1.7.0beta69.

Version 1.6.20rc01 [November 26, 2015]
  Fixed an error in handling of bad zlib CMINFO field in pngfix, found by
    American Fuzzy Lop, reported by Brian Carpenter.  inflate() doesn't
    immediately fault a bad CMINFO field; instead a 'too far back' error
    happens later (at least some times).  pngfix failed to limit CMINFO to
    the allowed values but then assumed that window_bits was in range,
    triggering an assert. The bug is mostly harmless; the PNG file cannot
    be fixed.

Version 1.6.20rc02 [November 29, 2015]
  In libpng 1.6 zlib initialization was changed to use the window size
    in the zlib stream, not a fixed value. This causes some invalid images,
    where CINFO is too large, to display 'correctly' if the rest of the
    data is valid.  This provides a workaround for zlib versions where the
    error arises (ones that support the API change to use the window size
    in the stream).

Version 1.6.20 [December 3, 2015]
  No changes.

Version 1.6.21beta01 [December 11, 2015]
  Fixed syntax "$(command)" in tests/pngstest that some shells other than
    bash could not parse (Bug report by Nelson Beebe). Use `command` instead.

Version 1.6.21beta02 [December 14, 2015]
  Moved png_check_keyword() from pngwutil.c to pngset.c
  Removed LE/BE dependencies in pngvalid, to 'fix' the current problem
    in the BigEndian tests by not testing it, making the BE code the same 
    as the LE version.
  Fixes to pngvalid for various reduced build configurations (eliminate unused
    statics) and a fix for the case in rgb_to_gray when the digitize option
    reduces graylo to 0, producing a large error.

Version 1.6.21beta03 [December 18, 2015]
  Widened the 'limit' check on the internally calculated error limits in
    the 'DIGITIZE' case (the code used prior to 1.7 for rgb_to_gray error
    checks) and changed the check to only operate in non-release builds
    (base build type not RC or RELEASE.)
  Fixed undefined behavior in pngvalid.c, undefined because
    (png_byte) << shift is undefined if it changes the signed bit
    (because png_byte is promoted to int). The libpng exported functions
    png_get_uint_32 and png_get_uint_16 handle this. (Bug reported by
    David Drysdale as a result of reports from UBSAN in clang 3.8).
  This changes pngvalid to use BE random numbers; this used to produce
    errors but these should not be fixed as a result of the previous changes.

Version 1.6.21rc01 [January 4, 2016]
  In projects/vstudio, combined readme.txt and WARNING into README.txt

Version 1.6.21rc02 [January 7, 2016]
  Relocated assert() in contrib/tools/pngfix.c, bug found by American
    Fuzzy Lop, reported by Brian Carpenter.
  Marked 'limit' UNUSED in transform_range_check().  This only affects
    release builds.

Version 1.6.21 [January 15, 2016]
  Worked around a false-positive Coverity issue in pngvalid.c.

Version 1.6.22beta01 [January 23, 2016]
  Changed PNG_USE_MKSTEMP to __COVERITY__ to select alternate
    "tmpfile()" implementation in contrib/libtests/pngstest.c
  Fixed NO_STDIO build of pngunknown.c to skip calling png_init_io()
    if there is no stdio.h support.
  Added a png_image_write_to_memory() API and a number of assist macros
    to allow an application that uses the simplified API write to bypass
    stdio and write directly to memory.
  Added some warnings (png.h) and some check code to detect *possible*
    overflow in the ROW_STRIDE and simplified image SIZE macros.  This
    disallows image width/height/format that *might* overflow.  This is
    a quiet API change that limits in-memory image size (uncompressed) to
    less than 4GByte and image row size (stride) to less than 2GByte.
  Revised workaround for false-positive Coverity issue in pngvalid.c.

Version 1.6.22beta02 [February 8, 2016]
  Only use exit(77) in configure builds.
  Corrected error in PNG_IMAGE_PNG_SIZE_MAX. This new macro underreported
    the palette size because it failed to take into account that the memory
    palette has to be expanded to full RGB when it is written to PNG.
  Updated CMakeLists.txt, added supporting scripts/gen*.cmake.in
    and test.cmake.in (Roger Leigh).
  Relaxed limit checks on gamma values in pngrtran.c. As suggested in
    the comments gamma values outside the range currently permitted
    by png_set_alpha_mode are useful for HDR data encoding.  These values
    are already permitted by png_set_gamma so it is reasonable caution to
    extend the png_set_alpha_mode range as HDR imaging systems are starting
    to emerge.

Version 1.6.22beta03 [March 9, 2016]
  Added a common-law trademark notice and export control information
    to the LICENSE file, png.h, and the man page.
  Restored "& 0xff" in png_save_uint_16() and png_save_uint_32() that
    were accidentally removed from libpng-1.6.17. 
  Changed PNG_INFO_cHNK and PNG_FREE_cHNK from 0xnnnn to 0xnnnnU in png.h
    (Robert C. Seacord).
  Removed dubious "#if INT_MAX" test from png.h that was added to
    libpng-1.6.19beta02 (John Bowler).
  Add ${INCLUDES} in scripts/genout.cmake.in (Bug report by Nixon Kwok).
  Updated LICENSE to say files in the contrib directory are not
    necessarily under the libpng license, and that some makefiles have
    other copyright owners.
  Added INTEL-SSE2 support (Mike Klein and Matt Sarett, Google, Inc.).
  Made contrib/libtests/timepng more robust.  The code no longer gives
    up/fails on invalid PNG data, it just skips it (with error messages).
    The code no longer fails on PNG files with data beyond IEND.  Options
    exist to use png_read_png (reading the whole image, not by row) and, in
    that case, to apply any of the supported transforms.  This makes for
    more realistic testing; the decoded data actually gets used in a
    meaningful fashion (John Bowler).
  Fixed some misleading indentation (Krishnaraj Bhat).

Version 1.6.22beta04 [April 5, 2016]
  Force GCC compilation to C89 if needed (Dagobert Michelsen).
  SSE filter speed improvements for bpp=3:
    memcpy-free implementations of load3() / store3().
    call load3() only when needed at the end of a scanline.

Version 1.6.22beta05 [April 27, 2016]
  Added PNG_FAST_FILTERS macro (defined as
    PNG_FILTER_NONE|PNG_FILTER_SUB|PNG_FILTER_UP).
  Various fixes for contrib/libtests/timepng.c
  Moved INTEL-SSE code from pngpriv.h into contrib/intel/intel_sse.patch.
  Fixed typo (missing underscore) in #define PNG_READ_16_TO_8_SUPPORTED
    (Bug report by Y.Ohashik).

Version 1.6.22beta06 [May 5, 2016]
  Rebased contrib/intel_sse.patch.
  Quieted two Coverity issues in contrib/libtests/timepng.c.
  Fixed issues with scripts/genout.cmake.in (David Capello, Nixon Kwok):
    Added support to use multiple directories in ZLIBINCDIR variable,
    Fixed CMAKE_C_FLAGS with multiple values when genout is compiled on MSVC,
    Fixed pnglibconf.c compilation on OS X including the sysroot path.

Version 1.6.22rc01 [May 14, 2016]
  No changes.

Version 1.6.22rc02 [May 16, 2016]
  Removed contrib/timepng from default build; it does not build on platforms
    that don't supply clock_gettime().

Version 1.6.22rc03 [May 17, 2016]
  Restored contrib/timepng to default build but check for the presence
    of clock_gettime() in configure.ac and Makefile.am.

Version 1.6.22 [May 26, 2016]
  No changes.

Version 1.6.23beta01 [May 29, 2016]
  Stop a potential memory leak in png_set_tRNS() (Bug report by Ted Ying).
  Fixed the progressive reader to handle empty first IDAT chunk properly
    (patch by Timothy Nikkel).  This bug was introduced in libpng-1.6.0 and
    only affected the libpng16 branch.
  Added tests in pngvalid.c to check zero-length IDAT chunks in various
    positions.  Fixed the sequential reader to handle these more robustly
    (John Bowler).

Version 1.6.23rc01 [June 2, 2016]
  Corrected progressive read input buffer in pngvalid.c. The previous version
    the code invariably passed just one byte at a time to libpng.  The intent
    was to pass a random number of bytes in the range 0..511.
  Moved sse2 prototype from pngpriv.h to contrib/intel/intel_sse.patch.
  Added missing ")" in pngerror.c (Matt Sarrett).

Version 1.6.23rc02 [June 4, 2016]
  Fixed undefined behavior in png_push_save_buffer(). Do not call
    memcpy() with a null source, even if count is zero (Leon Scroggins III).

Version 1.6.23 [June 9, 2016]
  Fixed bad link to RFC2083 in png.5 (Nikola Forro).

Version 1.6.24beta01 [June 11, 2016]
  Avoid potential overflow of the PNG_IMAGE_SIZE macro.  This macro
    is not used within libpng, but is used in some of the examples.

Version 1.6.24beta02 [June 23, 2016]
  Correct filter heuristic overflow handling. This was broken when the
    write filter code was moved out-of-line; if there is a single filter and
    the heuristic sum overflows the calculation of the filtered line is not
    completed.  In versions prior to 1.6 the code was duplicated in-line
    and the check not performed, so the filter operation completed; however,
    in the multi-filter case where the sum is performed the 'none' filter would
    be selected if all the sums overflowed, even if it wasn't in the filter
    list.  The fix to the first problem is simply to provide PNG_SIZE_MAX as
    the current lmins sum value; this means the sum can never exceed it and
    overflows silently.  A reasonable compiler that does choose to inline
    the code will simply eliminate the sum check.
  The fix to the second problem is to use high precision arithmetic (this is
    implemented in 1.7), however a simple safe fix here is to chose the lowest
    numbered filter in the list from png_set_filter (this only works if the
    first problem is also fixed) (John Bowler).
  Use a more efficient absolute value calculation on SSE2 (Matthieu Darbois).
  Fixed the case where PNG_IMAGE_BUFFER_SIZE can overflow in the application
    as a result of the application using an increased 'row_stride'; previously
    png_image_finish_read only checked for overflow on the base calculation of
    components.  (I.e. it checked for overflow of a 32-bit number on the total
    number of pixel components in the output format, not the possibly padded row
    length and not the number of bytes, which for linear formats is twice the
    number of components.)
  MSVC does not like '-(unsigned)', so replaced it with 0U-(unsigned)
  MSVC does not like (uInt) = -(unsigned) (i.e. as an initializer), unless
    the conversion is explicitly invoked by a cast.
  Put the SKIP definition in the correct place. It needs to come after the
    png.h include (see all the other .c files in contrib/libtests) because it
    depends on PNG_LIBPNG_VER.
  Removed the three compile warning options from the individual project
    files into the zlib.props globals.  It increases the warning level from 4
    to All and adds a list of the warnings that need to be turned off.  This is
    semi-documentary; the intent is to tell libpng users which warnings have
    been examined and judged non-fixable at present.  The warning about
    structure padding is fixable, but it would be a signficant change (moving
    structure members around).

Version 1.6.24beta03 [July 4, 2016]
  Optimized absolute value calculation in filter selection, similar to
    code in the PAETH decoder in pngrutil.c. Build with PNG_USE_ABS to
    use this.
  Added pngcp to the build together with a pngcp.dfa configuration test.
  Added high resolution timing to pngcp.
  Added "Common linking failures" section to INSTALL.
  Relocated misplaced #endif in png.c sRGB profile checking.
  Fixed two Coverity issues in pngcp.c.

Version 1.6.24beta04 [July 8, 2016]
  Avoid filter-selection heuristic sum calculations in cases where only one
    filter is a candidate for selection. This trades off code size (added
    private png_setup_*_row_only() functions) for speed.

Version 1.6.24beta05 [July 13, 2016]
  Fixed some indentation to comply with our coding style.
  Added contrib/tools/reindent.

Version 1.6.24beta06 [July 18, 2016]
  Fixed more indentation to comply with our coding style.
  Eliminated unnecessary tests of boolean png_isaligned() vs 0.

Version 1.6.24rc01 [July 25, 2016]
  No changes.

Version 1.6.24rc02 [August 1, 2016]
  Conditionally compile SSE2 headers in contrib/intel/intel_sse.patch
  Conditionally compile png_decompress_chunk().

Version 1.6.24rc03 [August 2, 2016]
  Conditionally compile ARM_NEON headers in pngpriv.h
  Updated contrib/intel/intel_sse.patch

Version 1.6.24[August 4, 2016]
  No changes.

Version 1.6.25beta01 [August 12, 2016]
  Reject oversized iCCP profile immediately.
  Cleaned up PNG_DEBUG compile of pngtest.c.
  Conditionally compile png_inflate().

Version 1.6.25beta02 [August 18, 2016]
  Don't install pngcp; it conflicts with pngcp in the pngtools package.
  Minor editing of INSTALL, (whitespace, added copyright line)

Version 1.6.25rc01 [August 24, 2016]
  No changes.

Version 1.6.25rc02 [August 29, 2016]
  Added MIPS support (Mandar Sahastrabuddhe <Mandar.Sahastrabuddhe@imgtec.com>).
  Only the UP filter is currently implemented.

Version 1.6.25rc03 [August 29, 2016]
  Rebased contrib/intel/intel_sse.patch after the MIPS implementation.

Version 1.6.25rc04 [August 30, 2016]
  Added MIPS support for SUB, AVG, and PAETH filters (Mandar Sahastrabuddhe).

Version 1.6.25rc05 [August 30, 2016]
  Rebased contrib/intel/intel_sse.patch after the MIPS implementation update..

Version 1.6.25 [September 1, 2016]
  No changes.

Version 1.6.26beta01 [September 26, 2016]
  Fixed handling zero length IDAT in pngfix (bug report by Agostino Sarubbo,
    bugfix by John Bowler).
  Do not issue a png_error() on read in png_set_pCAL() because png_handle_pCAL
    has allocated memory that libpng needs to free.
  Conditionally compile png_set_benign_errors() in pngread.c and pngtest.c
  Issue a png_benign_error instead of a png_error on ADLER32 mismatch
    while decoding compressed data chunks.
  Changed PNG_ZLIB_VERNUM to ZLIB_VERNUM in pngpriv.h, pngstruct.h, and
    pngrutil.c.
  If CRC handling of critical chunks has been set to PNG_CRC_QUIET_USE,
    ignore the ADLER32 checksum in the IDAT chunk as well as the chunk CRCs.
  Issue png_benign_error() on ADLER32 checksum mismatch instead of png_error().
  Add tests/badcrc.png and tests/badadler.png to tests/pngtest.
  Merged pngtest.c with libpng-1.7.0beta84/pngtest.c

Version 1.6.26beta02 [October 1, 2016]
  Updated the documentation about CRC and ADLER32 handling.
  Quieted 117 warnings from clang-3.8 in pngtrans.c, pngread.c,
     pngwrite.c, pngunknown.c, and pngvalid.c.
  Quieted 58 (out of 144) -Wconversion compiler warnings by changing
    flag definitions in pngpriv.h from 0xnnnn to 0xnnnnU and trivial changes
    in png.c, pngread.c, and pngwutil.c.

Version 1.6.26beta03 [October 2, 2016]
  Removed contrib/libtests/*.orig and *.rej that slipped into the tarballs.
  Quieted the 86 remaining -Wconversion compiler warnings by
    revising the png_isaligned() macro and trivial changes in png.c,
    pngerror.c, pngget.c, pngmem.c, pngset.c, pngrtran.c, pngrutil.c,
    pngwtran.c, pngwrite.c, and pngwutil.c.

Version 1.6.26beta04 [October 3, 2016]
  Quieted (bogus?) clang warnings about "absolute value has no effect"
    when PNG_USE_ABS is defined.
  Fixed offsets in contrib/intel/intel_sse.patch

Version 1.6.26beta05 [October 6, 2016]
  Changed integer constant 4294967294 to unsigned 4294967294U in pngconf.h
    to avoid a signed/unsigned compare in the preprocessor.

Version 1.6.26beta06 [October 7, 2016]
  Use zlib-1.2.8.1 inflateValidate() instead of inflateReset2() to
    optionally avoid ADLER32 evaluation.

Version 1.6.26rc01 [October 12, 2016]
  No changes.

Version 1.6.26 [October 20, 2016]
  Cosmetic change, "ptr != 0" to "ptr != NULL" in png.c and pngrutil.c
  Despammed email addresses (replaced "@" with " at ").

Version 1.6.27beta01 [November 2, 2016]
  Restrict the new ADLER32-skipping to IDAT chunks.  It broke iCCP chunk
    handling: an erroneous iCCP chunk would throw a png_error and reject the
    entire PNG image instead of rejecting just the iCCP chunk with a warning,
    if built with zlib-1.2.8.1.

Version 1.6.27rc01 [December 27, 2016]
  Control ADLER32 checking with new PNG_IGNORE_ADLER32 option. Fixes
    an endless loop when handling erroneous ADLER32 checksums; bug
    introduced in libpng-1.6.26.
  Removed the use of a macro containing the pre-processor 'defined'
    operator.  It is unclear whether this is valid; a macro that
    "generates" 'defined' is not permitted, but the use of the word
    "generates" within the C90 standard seems to imply more than simple
    substitution of an expression itself containing a well-formed defined
    operation.
  Added ARM support to CMakeLists.txt (Andreas Franek).

Version 1.6.27 [December 29, 2016]
  Fixed a potential null pointer dereference in png_set_text_2() (bug report
    and patch by Patrick Keshishian, CVE-2016-10087).

Version 1.6.28rc01 [January 3, 2017]
  Fixed arm/aarch64 detection in CMakeLists.txt (Gianfranco Costamagna).
  Added option to Cmake build allowing a custom location of zlib to be
    specified in a scenario where libpng is being built as a subproject
    alongside zlib by another project (Sam Serrels).
  Changed png_ptr->options from a png_byte to png_uint_32, to accomodate
    up to 16 options.

Version 1.6.28rc02 [January 4, 2017]
  Added "include(GNUInstallDirs)" to CMakeLists.txt (Gianfranco Costamagna).
  Moved SSE2 optimization code into the main libpng source directory.
    Configure libpng with "configure --enable-intel-sse" or compile
    libpng with "-DPNG_INTEL_SSE" in CPPFLAGS to enable it.

Version 1.6.28rc03 [January 4, 2017]
  Backed out the SSE optimization and last CMakeLists.txt to allow time for QA.

Version 1.6.28 [January 5, 2017]
  No changes.

Version 1.6.29beta01 [January 12, 2017]
  Readded "include(GNUInstallDirs)" to CMakeLists.txt (Gianfranco Costamagna).
  Moved SSE2 optimization code into the main libpng source directory.
    Configure libpng with "configure --enable-intel-sse" or compile
    libpng with "-DPNG_INTEL_SSE" in CPPFLAGS to enable it.
  Simplified conditional compilation in pngvalid.c, for AIX (Michael Felt).

Version 1.6.29beta02 [February 22, 2017]
  Avoid conditional directives that break statements in pngrutil.c (Romero
    Malaquias)
  The contrib/examples/pngtopng.c recovery code was in the wrong "if"
    branches; the comments were correct.
  Added code for PowerPC VSX optimisation (Vadim Barkov).

Version 1.6.29beta03 [March 1, 2017]
  Avoid potential overflow of shift operations in png_do_expand() (Aaron Boxer).
  Change test ZLIB_VERNUM >= 0x1281 to ZLIB_VERNUM >= 0x1290 in pngrutil.c
    because Solaris 11 distributes zlib-1.2.8.f that is older than 1.2.8.1,
    as suggested in zlib FAQ, item 24.
  Suppress clang warnings about implicit sign changes in png.c

Version 1.6.29 [March 16, 2017]
  No changes.

Version 1.6.30beta01 [April 1, 2017]
  Added missing "$(CPPFLAGS)" to the compile line for c.pic.o in
    makefile.linux and makefile.solaris-x86 (Cosmin).
  Revised documentation of png_get_error_ptr() in the libpng manual.
  Silence clang -Wcomma and const drop warnings (Viktor Szakats).
  Update Sourceforge URLs in documentation (https instead of http).

Version 1.6.30beta02 [April 22, 2017]
  Document need to check for integer overflow when allocating a pixel
    buffer for multiple rows in contrib/gregbook, contrib/pngminus,
    example.c, and in the manual (suggested by Jaeseung Choi). This
    is similar to the bug reported against pngquant in CVE-2016-5735.
  Removed reference to the obsolete PNG_SAFE_LIMITS macro in the documentation.

Version 1.6.30beta03 [May 22, 2017]
  Check for integer overflow in contrib/visupng and contrib/tools/genpng.
  Do not double evaluate CMAKE_SYSTEM_PROCESSOR in CMakeLists.txt.
  Test CMAKE_HOST_WIN32 instead of WIN32 in CMakeLists.txt.
  Fix some URL in documentation.

Version 1.6.30beta04 [June 7, 2017]
  Avoid writing an empty IDAT when the last IDAT exactly fills the
    compression buffer (bug report by Brian Baird). This bug was
    introduced in libpng-1.6.0.

Version 1.6.30rc01 [June 14, 2017]
  No changes.

Version 1.6.30rc02 [June 25, 2017]
  Update copyright year in pnglibconf.h, make ltmain.sh executable.
  Add a reference to the libpng.download site in README.

Version 1.6.30 [June 28, 2017]
  No changes.

Version 1.6.31beta01 [July 5, 2017]
  Guard the definition of _POSIX_SOURCE in pngpriv.h (AIX already defines it;
    bug report by Michael Felt).
  Revised pngpriv.h to work around failure to compile arm/filter_neon.S
    ("typedef" directive is unrecognized by the assembler). The problem
    was introduced in libpng-1.6.30beta01.
  Added "Requires: zlib" to libpng.pc.in (Pieter Neerincx).
  Added special case for FreeBSD in arm/filter_neon.S (Maya Rashish).

Version 1.6.31beta02 [July 8, 2017]
  Added instructions for disabling hardware optimizations in INSTALL.
  Added "--enable-hardware-optimizations" configuration flag to enable
    or disable all hardware optimizations with one flag.

Version 1.6.31beta03 [July 9, 2017]
  Updated CMakeLists.txt to add INTEL_SSE and MIPS_MSA platforms.
  Changed "int" to "png_size_t" in intel/filter_sse2.c to prevent
    possible integer overflow (Bug report by John Bowler).
  Quieted "declaration after statement" warnings in intel/filter_sse2.c.
  Added scripts/makefile-linux-opt, which has hardware optimizations enabled.

Version 1.6.31beta04 [July 11, 2017]
  Removed one of the GCC-7.1.0 'strict-overflow' warnings that result when
    integers appear on both sides of a compare.  Worked around the others by
    forcing the strict-overflow setting in the relevant functions to a level
    where they are not reported (John Bowler).
  Changed "FALL THROUGH" comments to "FALLTHROUGH" because GCC doesn't like
    the space.
  Worked around some C-style casts from (void*) because g++ 5.4.0 objects
    to them.
  Increased the buffer size for 'sprint' to pass the gcc 7.1.0 'sprint
    overflow' check that is on by default with -Wall -Wextra.

Version 1.6.31beta05 [July 13, 2017]
  Added eXIf chunk support.

Version 1.6.31beta06 [July 17, 2017]
  Added a minimal eXIf chunk (with Orientation and FocalLengthIn35mmFilm
    tags) to pngtest.png.

Version 1.6.31beta07 [July 18, 2017]
  Revised the eXIf chunk in pngtest.png to fix "Bad IFD1 Directory" warning.

Version 1.6.31rc01 [July 19, 2017]
  No changes.

Version 1.6.31rc02 [July 25, 2017]
  Fixed typo in example.c (png_free_image should be png_image_free) (Bug
    report by John Smith)

Version 1.6.31 [July 27, 2017]
  No changes.

Version 1.6.32beta01 [July 31, 2017]
  Avoid possible NULL dereference in png_handle_eXIf when benign_errors
    are allowed. Avoid leaking the input buffer "eXIf_buf".
  Eliminated png_ptr->num_exif member from pngstruct.h and added num_exif
    to arguments for png_get_eXIf() and png_set_eXIf().
  Added calls to png_handle_eXIf(() in pngread.c and png_write_eXIf() in
    pngwrite.c, and made various other fixes to png_write_eXIf().
  Changed name of png_get_eXIF and png_set_eXIf() to png_get_eXIf_1() and
    png_set_eXIf_1(), respectively, to avoid breaking API compatibility
    with libpng-1.6.31.

Version 1.6.32beta02 [August 1, 2017]
  Updated contrib/libtests/pngunknown.c with eXIf chunk.

Version 1.6.32beta03 [August 2, 2017]
  Initialized btoa[] in pngstest.c
  Stop memory leak when returning from png_handle_eXIf() with an error
    (Bug report from the OSS-fuzz project).

Version 1.6.32beta04 [August 2, 2017]
  Replaced local eXIf_buf with info_ptr-eXIf_buf in png_handle_eXIf().
  Update libpng.3 and libpng-manual.txt about eXIf functions.

Version 1.6.32beta05 [August 2, 2017]
  Restored png_get_eXIf() and png_set_eXIf() to maintain API compatability.

Version 1.6.32beta06 [August 2, 2017]
  Removed png_get_eXIf_1() and png_set_eXIf_1().

Version 1.6.32beta07 [August 3, 2017]
  Check length of all chunks except IDAT against user limit to fix an
    OSS-fuzz issue (Fixes CVE-2017-12652).

Version 1.6.32beta08 [August 3, 2017]
  Check length of IDAT against maximum possible IDAT size, accounting
    for height, rowbytes, interlacing and zlib/deflate overhead.
  Restored png_get_eXIf_1() and png_set_eXIf_1(), because strlen(eXIf_buf)
    does not work (the eXIf chunk data can contain zeroes).

Version 1.6.32beta09 [August 3, 2017]
  Require cmake-2.8.8 in CMakeLists.txt. Revised symlink creation,
    no longer using deprecated cmake LOCATION feature (Clifford Yapp).
  Fixed five-byte error in the calculation of IDAT maximum possible size.
  
Version 1.6.32beta10 [August 5, 2017]
  Moved chunk-length check into a png_check_chunk_length() private
    function (Suggested by Max Stepin).
  Moved bad pngs from tests to contrib/libtests/crashers
  Moved testing of bad pngs into a separate tests/pngtest-badpngs script
  Added the --xfail (expected FAIL) option to pngtest.c. It writes XFAIL
    in the output but PASS for the libpng test.
  Require cmake-3.0.2 in CMakeLists.txt (Clifford Yapp).
  Fix "const" declaration info_ptr argument to png_get_eXIf_1() and the
    num_exif argument to png_get_eXIf_1() (Github Issue 171).

Version 1.6.32beta11 [August 7, 2017]
  Added "eXIf" to "chunks_to_ignore[]" in png_set_keep_unknown_chunks().
  Added huge_IDAT.png and empty_ancillary_chunks.png to testpngs/crashers.
  Make pngtest --strict, --relax, --xfail options imply -m (multiple).
  Removed unused chunk_name parameter from png_check_chunk_length().
  Relocated setting free_me for eXIf data, to stop an OSS-fuzz leak.
  Initialize profile_header[] in png_handle_iCCP() to fix OSS-fuzz issue.
  Initialize png_ptr->row_buf[0] to 255 in png_read_row() to fix OSS-fuzz UMR.
  Attempt to fix a UMR in png_set_text_2() to fix OSS-fuzz issue.
  Increase minimum zlib stream from 9 to 14 in png_handle_iCCP(), to account
    for the minimum 'deflate' stream, and relocate the test to a point
    after the keyword has been read.
  Check that the eXIf chunk has at least 2 bytes and begins with "II" or "MM".

Version 1.6.32rc01 [August 18, 2017]
  Added a set of "huge_xxxx_chunk.png" files to contrib/testpngs/crashers,
    one for each known chunk type, with length = 2GB-1.
  Check for 0 return from png_get_rowbytes() and added some (size_t) typecasts
    in contrib/pngminus/*.c to stop some Coverity issues (162705, 162706,
    and 162707).
  Renamed chunks in contrib/testpngs/crashers to avoid having files whose
    names differ only in case; this causes problems with some platforms
    (github issue #172).

Version 1.6.32rc02 [August 22, 2017]
  Added contrib/oss-fuzz directory which contains files used by the oss-fuzz
    project (https://github.com/google/oss-fuzz/tree/master/projects/libpng).

Version 1.6.32 [August 24, 2017]
  No changes.

Version 1.6.33beta01 [August 28, 2017]
  Added PNGMINUS_UNUSED macro to contrib/pngminus/p*.c and added missing
    parenthesis in contrib/pngminus/pnm2png.c (bug report by Christian Hesse).
  Fixed off-by-one error in png_do_check_palette_indexes() (Bug report
    by Mick P., Source Forge Issue #269).

Version 1.6.33beta02 [September 3, 2017]
  Initialize png_handler.row_ptr in contrib/oss-fuzz/libpng_read_fuzzer.cc
    to fix shortlived oss-fuzz issue 3234.
  Compute a larger limit on IDAT because some applications write a deflate
    buffer for each row (Bug report by Andrew Church).
  Use current date (DATE) instead of release-date (RDATE) in last
    changed date of contrib/oss-fuzz files.
  Enabled ARM support in CMakeLists.txt (Bernd Kuhls).

Version 1.6.33beta03 [September 14, 2017]
  Fixed incorrect typecast of some arguments to png_malloc() and
    png_calloc() that were png_uint_32 instead of png_alloc_size_t
    (Bug report by "irwir" in Github libpng issue #175).
  Use pnglibconf.h.prebuilt when building for ANDROID with cmake (Github
    issue 162, by rcdailey).

Version 1.6.33rc01 [September 20, 2017]
  Initialize memory allocated by png_inflate to zero, using memset, to
    stop an oss-fuzz "use of uninitialized value" detection in png_set_text_2()
    due to truncated iTXt or zTXt chunk.
  Initialize memory allocated by png_read_buffer to zero, using memset, to
    stop an oss-fuzz "use of uninitialized value" detection in
    png_icc_check_tag_table() due to truncated iCCP chunk.
  Removed a redundant test (suggested by "irwir" in Github issue #180).

Version 1.6.33rc02 [September 23, 2017]
  Added an interlaced version of each file in contrib/pngsuite.
  Relocate new memset() call in pngrutil.c.
  Removed more redundant tests (suggested by "irwir" in Github issue #180).
  Add support for loading images with associated alpha in the Simplified
    API (Samuel Williams).

Version 1.6.33 [September 28, 2017]
  Revert contrib/oss-fuzz/libpng_read_fuzzer.cc to libpng-1.6.32 state.
  Initialize png_handler.row_ptr in contrib/oss-fuzz/libpng_read_fuzzer.cc
  Add end_info structure and png_read_end() to the libpng fuzzer.

Version 1.6.34 [September 29, 2017]
  Removed contrib/pngsuite/i*.png; some of these were incorrect and caused
    test failures.

Send comments/corrections/commendations to png-mng-implement at lists.sf.net
(subscription required; visit
https://lists.sourceforge.net/lists/listinfo/png-mng-implement
to subscribe)
or to glennrp at users.sourceforge.net

Glenn R-P
#endif
PKJM�\V���TODOnu�[���/*
TODO - list of things to do for libpng:

Final bug fixes.
Better C++ wrapper/full C++ implementation?
Fix problem with C++ and EXTERN "C".
cHRM transformation.
Remove setjmp/longjmp usage in favor of returning error codes. As a start on
  this, minimize the use of png_error(), replacing them with
  png_warning(); return(0); or similar.
Palette creation.
Add "grayscale->palette" transformation and "palette->grayscale" detection.
Improved dithering.
Multi-lingual error and warning message support.
Complete sRGB transformation (presently it simply uses gamma=0.45455).
Man pages for function calls.
Better documentation.
Better filter selection
   (counting huffman bits/precompression?  filter inertia?  filter costs?).
Histogram creation.
Text conversion between different code pages (Latin-1 -> Mac and DOS).
Avoid building gamma tables whenever possible.
Use greater precision when changing to linear gamma for compositing against
  background and doing rgb-to-gray transformation.
Investigate pre-incremented loop counters and other loop constructions.
Add interpolated method of handling interlacing.
Extend pngvalid.c to validate more of the libpng transformations.
Refactor preprocessor conditionals to compile entire statements

*/
PKJM�\��]��	example.cnu�[���
#if 0 /* in case someone actually tries to compile this */

/* example.c - an example of using libpng
 * Last changed in libpng 1.6.24 [August 4, 2016]
 * Maintained 1998-2016 Glenn Randers-Pehrson
 * Maintained 1996, 1997 Andreas Dilger)
 * Written 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
 * To the extent possible under law, the authors have waived
 * all copyright and related or neighboring rights to this file.
 * This work is published from: United States.
 */

/* This is an example of how to use libpng to read and write PNG files.
 * The file libpng-manual.txt is much more verbose then this.  If you have not
 * read it, do so first.  This was designed to be a starting point of an
 * implementation.  This is not officially part of libpng, is hereby placed
 * in the public domain, and therefore does not require a copyright notice.
 *
 * This file does not currently compile, because it is missing certain
 * parts, like allocating memory to hold an image.  You will have to
 * supply these parts to get it to compile.  For an example of a minimal
 * working PNG reader/writer, see pngtest.c, included in this distribution;
 * see also the programs in the contrib directory.
 */

/* The simple, but restricted, approach to reading a PNG file or data stream
 * just requires two function calls, as in the following complete program.
 * Writing a file just needs one function call, so long as the data has an
 * appropriate layout.
 *
 * The following code reads PNG image data from a file and writes it, in a
 * potentially new format, to a new file.  While this code will compile there is
 * minimal (insufficient) error checking; for a more realistic version look at
 * contrib/examples/pngtopng.c
 */
#include <stddef.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <png.h>
#include <zlib.h>

int main(int argc, const char **argv)
{
   if (argc == 3)
   {
      png_image image; /* The control structure used by libpng */

/* Initialize the 'png_image' structure. */
      memset(&image, 0, (sizeof image));
      image.version = PNG_IMAGE_VERSION;

/* The first argument is the file to read: */
      if (png_image_begin_read_from_file(&image, argv[1]) != 0)
      {
         png_bytep buffer;

/* Set the format in which to read the PNG file; this code chooses a
          * simple sRGB format with a non-associated alpha channel, adequate to
          * store most images.
          */
         image.format = PNG_FORMAT_RGBA;

/* Now allocate enough memory to hold the image in this format; the
          * PNG_IMAGE_SIZE macro uses the information about the image (width,
          * height and format) stored in 'image'.
          */
         buffer = malloc(PNG_IMAGE_SIZE(image));

/* If enough memory was available read the image in the desired format
          * then write the result out to the new file.  'background' is not
          * necessary when reading the image because the alpha channel is
          * preserved; if it were to be removed, for example if we requested
          * PNG_FORMAT_RGB, then either a solid background color would have to
          * be supplied or the output buffer would have to be initialized to the
          * actual background of the image.
          *
          * The fourth argument to png_image_finish_read is the 'row_stride' -
          * this is the number of components allocated for the image in each
          * row.  It has to be at least as big as the value returned by
          * PNG_IMAGE_ROW_STRIDE, but if you just allocate space for the
          * default, minimum, size using PNG_IMAGE_SIZE as above you can pass
          * zero.
          *
          * The final argument is a pointer to a buffer for the colormap;
          * colormaps have exactly the same format as a row of image pixels (so
          * you choose what format to make the colormap by setting
          * image.format).  A colormap is only returned if
          * PNG_FORMAT_FLAG_COLORMAP is also set in image.format, so in this
          * case NULL is passed as the final argument.  If you do want to force
          * all images into an index/color-mapped format then you can use:
          *
          *    PNG_IMAGE_COLORMAP_SIZE(image)
          *
          * to find the maximum size of the colormap in bytes.
          */
         if (buffer != NULL &&
            png_image_finish_read(&image, NULL/*background*/, buffer,
                0/*row_stride*/, NULL/*colormap*/) != 0)
         {
            /* Now write the image out to the second argument.  In the write
             * call 'convert_to_8bit' allows 16-bit data to be squashed down to
             * 8 bits; this isn't necessary here because the original read was
             * to the 8-bit format.
             */
            if (png_image_write_to_file(&image, argv[2], 0/*convert_to_8bit*/,
                buffer, 0/*row_stride*/, NULL/*colormap*/) != 0)
            {
               /* The image has been written successfully. */
               exit(0);
            }
         }

else
         {
            /* Calling png_image_free is optional unless the simplified API was
             * not run to completion.  In this case if there wasn't enough
             * memory for 'buffer' we didn't complete the read, so we must free
             * the image:
             */
            if (buffer == NULL)
               png_image_free(&image);

else
               free(buffer);
      }

/* Something went wrong reading or writing the image.  libpng stores a
       * textual message in the 'png_image' structure:
       */
      fprintf(stderr, "pngtopng: error: %s\n", image.message);
      exit (1);
   }

fprintf(stderr, "pngtopng: usage: pngtopng input-file output-file\n");
   exit(1);
}

/* That's it ;-)  Of course you probably want to do more with PNG files than
 * just converting them all to 32-bit RGBA PNG files; you can do that between
 * the call to png_image_finish_read and png_image_write_to_file.  You can also
 * ask for the image data to be presented in a number of different formats.  You
 * do this by simply changing the 'format' parameter set before allocating the
 * buffer.
 *
 * The format parameter consists of five flags that define various aspects of
 * the image, you can simply add these together to get the format or you can use
 * one of the predefined macros from png.h (as above):
 *
 * PNG_FORMAT_FLAG_COLOR: if set the image will have three color components per
 *    pixel (red, green and blue), if not set the image will just have one
 *    luminance (grayscale) component.
 *
 * PNG_FORMAT_FLAG_ALPHA: if set each pixel in the image will have an additional
 *    alpha value; a linear value that describes the degree the image pixel
 *    covers (overwrites) the contents of the existing pixel on the display.
 *
 * PNG_FORMAT_FLAG_LINEAR: if set the components of each pixel will be returned
 *    as a series of 16-bit linear values, if not set the components will be
 *    returned as a series of 8-bit values encoded according to the 'sRGB'
 *    standard.  The 8-bit format is the normal format for images intended for
 *    direct display, because almost all display devices do the inverse of the
 *    sRGB transformation to the data they receive.  The 16-bit format is more
 *    common for scientific data and image data that must be further processed;
 *    because it is linear simple math can be done on the component values.
 *    Regardless of the setting of this flag the alpha channel is always linear,
 *    although it will be 8 bits or 16 bits wide as specified by the flag.
 *
 * PNG_FORMAT_FLAG_BGR: if set the components of a color pixel will be returned
 *    in the order blue, then green, then red.  If not set the pixel components
 *    are in the order red, then green, then blue.
 *
 * PNG_FORMAT_FLAG_AFIRST: if set the alpha channel (if present) precedes the
 *    color or grayscale components.  If not set the alpha channel follows the
 *    components.
 *
 * You do not have to read directly from a file.  You can read from memory or,
 * on systems that support it, from a <stdio.h> FILE*.  This is controlled by
 * the particular png_image_read_from_ function you call at the start.  Likewise
 * on write you can write to a FILE* if your system supports it.  Check the
 * macro PNG_STDIO_SUPPORTED to see if stdio support has been included in your
 * libpng build.
 *
 * If you read 16-bit (PNG_FORMAT_FLAG_LINEAR) data you may need to write it in
 * the 8-bit format for display.  You do this by setting the convert_to_8bit
 * flag to 'true'.
 *
 * Don't repeatedly convert between the 8-bit and 16-bit forms.  There is
 * significant data loss when 16-bit data is converted to the 8-bit encoding and
 * the current libpng implementation of conversion to 16-bit is also
 * significantly lossy.  The latter will be fixed in the future, but the former
 * is unavoidable - the 8-bit format just doesn't have enough resolution.
 */

/* If your program needs more information from the PNG data it reads, or if you
 * need to do more complex transformations, or minimize transformations, on the
 * data you read, then you must use one of the several lower level libpng
 * interfaces.
 *
 * All these interfaces require that you do your own error handling - your
 * program must be able to arrange for control to return to your own code any
 * time libpng encounters a problem.  There are several ways to do this, but the
 * standard way is to use the ANSI-C (C90) <setjmp.h> interface to establish a
 * return point within your own code.  You must do this if you do not use the
 * simplified interface (above).
 *
 * The first step is to include the header files you need, including the libpng
 * header file.  Include any standard headers and feature test macros your
 * program requires before including png.h:
 */
#include <png.h>

/* The png_jmpbuf() macro, used in error handling, became available in
  * libpng version 1.0.6.  If you want to be able to run your code with older
  * versions of libpng, you must define the macro yourself (but only if it
  * is not already defined by libpng!).
  */

#ifndef png_jmpbuf
#  define png_jmpbuf(png_ptr) ((png_ptr)->png_jmpbuf)
#endif

/* Check to see if a file is a PNG file using png_sig_cmp().  png_sig_cmp()
 * returns zero if the image is a PNG and nonzero if it isn't a PNG.
 *
 * The function check_if_png() shown here, but not used, returns nonzero (true)
 * if the file can be opened and is a PNG, 0 (false) otherwise.
 *
 * If this call is successful, and you are going to keep the file open,
 * you should call png_set_sig_bytes(png_ptr, PNG_BYTES_TO_CHECK); once
 * you have created the png_ptr, so that libpng knows your application
 * has read that many bytes from the start of the file.  Make sure you
 * don't call png_set_sig_bytes() with more than 8 bytes read or give it
 * an incorrect number of bytes read, or you will either have read too
 * many bytes (your fault), or you are telling libpng to read the wrong
 * number of magic bytes (also your fault).
 *
 * Many applications already read the first 2 or 4 bytes from the start
 * of the image to determine the file type, so it would be easiest just
 * to pass the bytes to png_sig_cmp() or even skip that if you know
 * you have a PNG file, and call png_set_sig_bytes().
 */
#define PNG_BYTES_TO_CHECK 4
int check_if_png(char *file_name, FILE **fp)
{
   char buf[PNG_BYTES_TO_CHECK];

/* Open the prospective PNG file. */
   if ((*fp = fopen(file_name, "rb")) == NULL)
      return 0;

/* Read in some of the signature bytes */
   if (fread(buf, 1, PNG_BYTES_TO_CHECK, *fp) != PNG_BYTES_TO_CHECK)
      return 0;

/* Compare the first PNG_BYTES_TO_CHECK bytes of the signature.
      Return nonzero (true) if they match */

return(!png_sig_cmp(buf, (png_size_t)0, PNG_BYTES_TO_CHECK));
}

/* Read a PNG file.  You may want to return an error code if the read
 * fails (depending upon the failure).  There are two "prototypes" given
 * here - one where we are given the filename, and we need to open the
 * file, and the other where we are given an open file (possibly with
 * some or all of the magic bytes read - see comments above).
 */
#ifdef open_file /* prototype 1 */
void read_png(char *file_name)  /* We need to open the file */
{
   png_structp png_ptr;
   png_infop info_ptr;
   int sig_read = 0;
   png_uint_32 width, height;
   int bit_depth, color_type, interlace_type;
   FILE *fp;

if ((fp = fopen(file_name, "rb")) == NULL)
      return (ERROR);

#else no_open_file /* prototype 2 */
void read_png(FILE *fp, int sig_read)  /* File is already open */
{
   png_structp png_ptr;
   png_infop info_ptr;
   png_uint_32 width, height;
   int bit_depth, color_type, interlace_type;
#endif no_open_file /* Only use one prototype! */

/* Create and initialize the png_struct with the desired error handler
    * functions.  If you want to use the default stderr and longjump method,
    * you can supply NULL for the last three parameters.  We also supply the
    * the compiler header file version, so that we know if the application
    * was compiled with a compatible version of the library.  REQUIRED
    */
   png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,
       png_voidp user_error_ptr, user_error_fn, user_warning_fn);

if (png_ptr == NULL)
   {
      fclose(fp);
      return (ERROR);
   }

/* Allocate/initialize the memory for image information.  REQUIRED. */
   info_ptr = png_create_info_struct(png_ptr);
   if (info_ptr == NULL)
   {
      fclose(fp);
      png_destroy_read_struct(&png_ptr, NULL, NULL);
      return (ERROR);
   }

/* Set error handling if you are using the setjmp/longjmp method (this is
    * the normal method of doing things with libpng).  REQUIRED unless you
    * set up your own error handlers in the png_create_read_struct() earlier.
    */

if (setjmp(png_jmpbuf(png_ptr)))
   {
      /* Free all of the memory associated with the png_ptr and info_ptr */
      png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
      fclose(fp);
      /* If we get here, we had a problem reading the file */
      return (ERROR);
   }

/* One of the following I/O initialization methods is REQUIRED */
#ifdef streams /* PNG file I/O method 1 */
   /* Set up the input control if you are using standard C streams */
   png_init_io(png_ptr, fp);

#else no_streams /* PNG file I/O method 2 */
   /* If you are using replacement read functions, instead of calling
    * png_init_io() here you would call:
    */
   png_set_read_fn(png_ptr, (void *)user_io_ptr, user_read_fn);
   /* where user_io_ptr is a structure you want available to the callbacks */
#endif no_streams /* Use only one I/O method! */

/* If we have already read some of the signature */
   png_set_sig_bytes(png_ptr, sig_read);

#ifdef hilevel
   /*
    * If you have enough memory to read in the entire image at once,
    * and you need to specify only transforms that can be controlled
    * with one of the PNG_TRANSFORM_* bits (this presently excludes
    * quantizing, filling, setting background, and doing gamma
    * adjustment), then you can read the entire image (including
    * pixels) into the info structure with this call:
    */
   png_read_png(png_ptr, info_ptr, png_transforms, NULL);

#else
   /* OK, you're doing it the hard way, with the lower-level functions */

/* The call to png_read_info() gives us all of the information from the
    * PNG file before the first IDAT (image data chunk).  REQUIRED
    */
   png_read_info(png_ptr, info_ptr);

png_get_IHDR(png_ptr, info_ptr, &width, &height, &bit_depth, &color_type,
       &interlace_type, NULL, NULL);

/* Set up the data transformations you want.  Note that these are all
    * optional.  Only call them if you want/need them.  Many of the
    * transformations only work on specific types of images, and many
    * are mutually exclusive.
    */

/* Tell libpng to strip 16 bits/color files down to 8 bits/color.
    * Use accurate scaling if it's available, otherwise just chop off the
    * low byte.
    */
#ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
   png_set_scale_16(png_ptr);
#else
   png_set_strip_16(png_ptr);
#endif

/* Strip alpha bytes from the input data without combining with the
    * background (not recommended).
    */
   png_set_strip_alpha(png_ptr);

/* Extract multiple pixels with bit depths of 1, 2, and 4 from a single
    * byte into separate bytes (useful for paletted and grayscale images).
    */
   png_set_packing(png_ptr);

/* Change the order of packed pixels to least significant bit first
    * (not useful if you are using png_set_packing). */
   png_set_packswap(png_ptr);

/* Expand paletted colors into true RGB triplets */
   if (color_type == PNG_COLOR_TYPE_PALETTE)
      png_set_palette_to_rgb(png_ptr);

/* Expand grayscale images to the full 8 bits from 1, 2, or 4 bits/pixel */
   if (color_type == PNG_COLOR_TYPE_GRAY && bit_depth < 8)
      png_set_expand_gray_1_2_4_to_8(png_ptr);

/* Expand paletted or RGB images with transparency to full alpha channels
    * so the data will be available as RGBA quartets.
    */
   if (png_get_valid(png_ptr, info_ptr, PNG_INFO_tRNS) != 0)
      png_set_tRNS_to_alpha(png_ptr);

/* Set the background color to draw transparent and alpha images over.
    * It is possible to set the red, green, and blue components directly
    * for paletted images instead of supplying a palette index.  Note that
    * even if the PNG file supplies a background, you are not required to
    * use it - you should use the (solid) application background if it has one.
    */

png_color_16 my_background, *image_background;

if (png_get_bKGD(png_ptr, info_ptr, &image_background) != 0)
      png_set_background(png_ptr, image_background,
          PNG_BACKGROUND_GAMMA_FILE, 1, 1.0);
   else
      png_set_background(png_ptr, &my_background,
          PNG_BACKGROUND_GAMMA_SCREEN, 0, 1.0);

/* Some suggestions as to how to get a screen gamma value
    *
    * Note that screen gamma is the display_exponent, which includes
    * the CRT_exponent and any correction for viewing conditions
    */
   if (/* We have a user-defined screen gamma value */)
   {
      screen_gamma = user-defined screen_gamma;
   }
   /* This is one way that applications share the same screen gamma value */
   else if ((gamma_str = getenv("SCREEN_GAMMA")) != NULL)
   {
      screen_gamma = atof(gamma_str);
   }
   /* If we don't have another value */
   else
   {
      screen_gamma = PNG_DEFAULT_sRGB;  /* A good guess for a PC monitor
                                           in a dimly lit room */
      screen_gamma = PNG_GAMMA_MAC_18 or 1.0; /* Good guesses for Mac systems */
   }

/* Tell libpng to handle the gamma conversion for you.  The final call
    * is a good guess for PC generated images, but it should be configurable
    * by the user at run time by the user.  It is strongly suggested that
    * your application support gamma correction.
    */

int intent;

if (png_get_sRGB(png_ptr, info_ptr, &intent) != 0)
      png_set_gamma(png_ptr, screen_gamma, PNG_DEFAULT_sRGB);
   else
   {
      double image_gamma;
      if (png_get_gAMA(png_ptr, info_ptr, &image_gamma) != 0)
         png_set_gamma(png_ptr, screen_gamma, image_gamma);
      else
         png_set_gamma(png_ptr, screen_gamma, 0.45455);
   }

#ifdef PNG_READ_QUANTIZE_SUPPORTED
   /* Quantize RGB files down to 8-bit palette or reduce palettes
    * to the number of colors available on your screen.
    */
   if ((color_type & PNG_COLOR_MASK_COLOR) != 0)
   {
      int num_palette;
      png_colorp palette;

/* This reduces the image to the application supplied palette */
      if (/* We have our own palette */)
      {
         /* An array of colors to which the image should be quantized */
         png_color std_color_cube[MAX_SCREEN_COLORS];

png_set_quantize(png_ptr, std_color_cube, MAX_SCREEN_COLORS,
             MAX_SCREEN_COLORS, NULL, 0);
      }
      /* This reduces the image to the palette supplied in the file */
      else if (png_get_PLTE(png_ptr, info_ptr, &palette, &num_palette) != 0)
      {
         png_uint_16p histogram = NULL;

png_get_hIST(png_ptr, info_ptr, &histogram);

png_set_quantize(png_ptr, palette, num_palette,
             max_screen_colors, histogram, 0);
      }
   }
#endif /* READ_QUANTIZE */

/* Invert monochrome files to have 0 as white and 1 as black */
   png_set_invert_mono(png_ptr);

/* If you want to shift the pixel values from the range [0,255] or
    * [0,65535] to the original [0,7] or [0,31], or whatever range the
    * colors were originally in:
    */
   if (png_get_valid(png_ptr, info_ptr, PNG_INFO_sBIT) != 0)
   {
      png_color_8p sig_bit_p;

png_get_sBIT(png_ptr, info_ptr, &sig_bit_p);
      png_set_shift(png_ptr, sig_bit_p);
   }

/* Flip the RGB pixels to BGR (or RGBA to BGRA) */
   if ((color_type & PNG_COLOR_MASK_COLOR) != 0)
      png_set_bgr(png_ptr);

/* Swap the RGBA or GA data to ARGB or AG (or BGRA to ABGR) */
   png_set_swap_alpha(png_ptr);

/* Swap bytes of 16-bit files to least significant byte first */
   png_set_swap(png_ptr);

/* Add filler (or alpha) byte (before/after each RGB triplet) */
   png_set_filler(png_ptr, 0xffff, PNG_FILLER_AFTER);

#ifdef PNG_READ_INTERLACING_SUPPORTED
   /* Turn on interlace handling.  REQUIRED if you are not using
    * png_read_image().  To see how to handle interlacing passes,
    * see the png_read_row() method below:
    */
   number_passes = png_set_interlace_handling(png_ptr);
#else /* !READ_INTERLACING */
   number_passes = 1;
#endif /* READ_INTERLACING */

/* Optional call to gamma correct and add the background to the palette
    * and update info structure.  REQUIRED if you are expecting libpng to
    * update the palette for you (ie you selected such a transform above).
    */
   png_read_update_info(png_ptr, info_ptr);

/* Allocate the memory to hold the image using the fields of info_ptr. */

/* The easiest way to read the image: */
   png_bytep row_pointers[height];

/* Clear the pointer array */
   for (row = 0; row < height; row++)
      row_pointers[row] = NULL;

for (row = 0; row < height; row++)
      row_pointers[row] = png_malloc(png_ptr, png_get_rowbytes(png_ptr,
          info_ptr));

/* Now it's time to read the image.  One of these methods is REQUIRED */
#ifdef entire /* Read the entire image in one go */
   png_read_image(png_ptr, row_pointers);

#else no_entire /* Read the image one or more scanlines at a time */
   /* The other way to read images - deal with interlacing: */

for (pass = 0; pass < number_passes; pass++)
   {
#ifdef single /* Read the image a single row at a time */
      for (y = 0; y < height; y++)
      {
         png_read_rows(png_ptr, &row_pointers[y], NULL, 1);
      }

#else no_single /* Read the image several rows at a time */
      for (y = 0; y < height; y += number_of_rows)
      {
#ifdef sparkle /* Read the image using the "sparkle" effect. */
         png_read_rows(png_ptr, &row_pointers[y], NULL,
             number_of_rows);
#else no_sparkle /* Read the image using the "rectangle" effect */
         png_read_rows(png_ptr, NULL, &row_pointers[y],
             number_of_rows);
#endif no_sparkle /* Use only one of these two methods */
      }

/* If you want to display the image after every pass, do so here */
#endif no_single /* Use only one of these two methods */
   }
#endif no_entire /* Use only one of these two methods */

/* Read rest of file, and get additional chunks in info_ptr - REQUIRED */
   png_read_end(png_ptr, info_ptr);
#endif hilevel

/* At this point you have read the entire image */

/* Clean up after the read, and free any memory allocated - REQUIRED */
   png_destroy_read_struct(&png_ptr, &info_ptr, NULL);

/* Close the file */
   fclose(fp);

/* That's it */
   return (OK);
}

/* Progressively read a file */

int
initialize_png_reader(png_structp *png_ptr, png_infop *info_ptr)
{
   /* Create and initialize the png_struct with the desired error handler
    * functions.  If you want to use the default stderr and longjump method,
    * you can supply NULL for the last three parameters.  We also check that
    * the library version is compatible in case we are using dynamically
    * linked libraries.
    */
   *png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,
        png_voidp user_error_ptr, user_error_fn, user_warning_fn);

if (*png_ptr == NULL)
   {
      *info_ptr = NULL;
      return (ERROR);
   }

*info_ptr = png_create_info_struct(png_ptr);

if (*info_ptr == NULL)
   {
      png_destroy_read_struct(png_ptr, info_ptr, NULL);
      return (ERROR);
   }

if (setjmp(png_jmpbuf((*png_ptr))))
   {
      png_destroy_read_struct(png_ptr, info_ptr, NULL);
      return (ERROR);
   }

/* This one's new.  You will need to provide all three
    * function callbacks, even if you aren't using them all.
    * If you aren't using all functions, you can specify NULL
    * parameters.  Even when all three functions are NULL,
    * you need to call png_set_progressive_read_fn().
    * These functions shouldn't be dependent on global or
    * static variables if you are decoding several images
    * simultaneously.  You should store stream specific data
    * in a separate struct, given as the second parameter,
    * and retrieve the pointer from inside the callbacks using
    * the function png_get_progressive_ptr(png_ptr).
    */
   png_set_progressive_read_fn(*png_ptr, (void *)stream_data,
       info_callback, row_callback, end_callback);

return (OK);
}

int
process_data(png_structp *png_ptr, png_infop *info_ptr,
    png_bytep buffer, png_uint_32 length)
{
   if (setjmp(png_jmpbuf((*png_ptr))))
   {
      /* Free the png_ptr and info_ptr memory on error */
      png_destroy_read_struct(png_ptr, info_ptr, NULL);
      return (ERROR);
   }

/* This one's new also.  Simply give it chunks of data as
    * they arrive from the data stream (in order, of course).
    * On segmented machines, don't give it any more than 64K.
    * The library seems to run fine with sizes of 4K, although
    * you can give it much less if necessary (I assume you can
    * give it chunks of 1 byte, but I haven't tried with less
    * than 256 bytes yet).  When this function returns, you may
    * want to display any rows that were generated in the row
    * callback, if you aren't already displaying them there.
    */
   png_process_data(*png_ptr, *info_ptr, buffer, length);
   return (OK);
}

info_callback(png_structp png_ptr, png_infop info)
{
   /* Do any setup here, including setting any of the transformations
    * mentioned in the Reading PNG files section.  For now, you _must_
    * call either png_start_read_image() or png_read_update_info()
    * after all the transformations are set (even if you don't set
    * any).  You may start getting rows before png_process_data()
    * returns, so this is your last chance to prepare for that.
    */
}

row_callback(png_structp png_ptr, png_bytep new_row,
    png_uint_32 row_num, int pass)
{
   /*
    * This function is called for every row in the image.  If the
    * image is interlaced, and you turned on the interlace handler,
    * this function will be called for every row in every pass.
    *
    * In this function you will receive a pointer to new row data from
    * libpng called new_row that is to replace a corresponding row (of
    * the same data format) in a buffer allocated by your application.
    *
    * The new row data pointer "new_row" may be NULL, indicating there is
    * no new data to be replaced (in cases of interlace loading).
    *
    * If new_row is not NULL then you need to call
    * png_progressive_combine_row() to replace the corresponding row as
    * shown below:
    */

/* Get pointer to corresponding row in our
    * PNG read buffer.
    */
   png_bytep old_row = ((png_bytep *)our_data)[row_num];

#ifdef PNG_READ_INTERLACING_SUPPORTED
   /* If both rows are allocated then copy the new row
    * data to the corresponding row data.
    */
   if ((old_row != NULL) && (new_row != NULL))
   png_progressive_combine_row(png_ptr, old_row, new_row);

/*
    * The rows and passes are called in order, so you don't really
    * need the row_num and pass, but I'm supplying them because it
    * may make your life easier.
    *
    * For the non-NULL rows of interlaced images, you must call
    * png_progressive_combine_row() passing in the new row and the
    * old row, as demonstrated above.  You can call this function for
    * NULL rows (it will just return) and for non-interlaced images
    * (it just does the memcpy for you) if it will make the code
    * easier.  Thus, you can just do this for all cases:
    */

png_progressive_combine_row(png_ptr, old_row, new_row);

/* where old_row is what was displayed for previous rows.  Note
    * that the first pass (pass == 0 really) will completely cover
    * the old row, so the rows do not have to be initialized.  After
    * the first pass (and only for interlaced images), you will have
    * to pass the current row as new_row, and the function will combine
    * the old row and the new row.
    */
#endif /* READ_INTERLACING */
}

end_callback(png_structp png_ptr, png_infop info)
{
   /* This function is called when the whole image has been read,
    * including any chunks after the image (up to and including
    * the IEND).  You will usually have the same info chunk as you
    * had in the header, although some data may have been added
    * to the comments and time fields.
    *
    * Most people won't do much here, perhaps setting a flag that
    * marks the image as finished.
    */
}

/* Write a png file */
void write_png(char *file_name /* , ... other image information ... */)
{
   FILE *fp;
   png_structp png_ptr;
   png_infop info_ptr;
   png_colorp palette;

/* Open the file */
   fp = fopen(file_name, "wb");
   if (fp == NULL)
      return (ERROR);

/* Create and initialize the png_struct with the desired error handler
    * functions.  If you want to use the default stderr and longjump method,
    * you can supply NULL for the last three parameters.  We also check that
    * the library version is compatible with the one used at compile time,
    * in case we are using dynamically linked libraries.  REQUIRED.
    */
   png_ptr = png_create_write_struct(PNG_LIBPNG_VER_STRING,
       png_voidp user_error_ptr, user_error_fn, user_warning_fn);

if (png_ptr == NULL)
   {
      fclose(fp);
      return (ERROR);
   }

/* Allocate/initialize the image information data.  REQUIRED */
   info_ptr = png_create_info_struct(png_ptr);
   if (info_ptr == NULL)
   {
      fclose(fp);
      png_destroy_write_struct(&png_ptr,  NULL);
      return (ERROR);
   }

/* Set error handling.  REQUIRED if you aren't supplying your own
    * error handling functions in the png_create_write_struct() call.
    */
   if (setjmp(png_jmpbuf(png_ptr)))
   {
      /* If we get here, we had a problem writing the file */
      fclose(fp);
      png_destroy_write_struct(&png_ptr, &info_ptr);
      return (ERROR);
   }

/* One of the following I/O initialization functions is REQUIRED */

#ifdef streams /* I/O initialization method 1 */
   /* Set up the output control if you are using standard C streams */
   png_init_io(png_ptr, fp);

#else no_streams /* I/O initialization method 2 */
   /* If you are using replacement write functions, instead of calling
    * png_init_io() here you would call
    */
   png_set_write_fn(png_ptr, (void *)user_io_ptr, user_write_fn,
       user_IO_flush_function);
   /* where user_io_ptr is a structure you want available to the callbacks */
#endif no_streams /* Only use one initialization method */

#ifdef hilevel
   /* This is the easy way.  Use it if you already have all the
    * image info living in the structure.  You could "|" many
    * PNG_TRANSFORM flags into the png_transforms integer here.
    */
   png_write_png(png_ptr, info_ptr, png_transforms, NULL);

#else
   /* This is the hard way */

/* Set the image information here.  Width and height are up to 2^31,
    * bit_depth is one of 1, 2, 4, 8, or 16, but valid values also depend on
    * the color_type selected. color_type is one of PNG_COLOR_TYPE_GRAY,
    * PNG_COLOR_TYPE_GRAY_ALPHA, PNG_COLOR_TYPE_PALETTE, PNG_COLOR_TYPE_RGB,
    * or PNG_COLOR_TYPE_RGB_ALPHA.  interlace is either PNG_INTERLACE_NONE or
    * PNG_INTERLACE_ADAM7, and the compression_type and filter_type MUST
    * currently be PNG_COMPRESSION_TYPE_BASE and PNG_FILTER_TYPE_BASE. REQUIRED
    */
   png_set_IHDR(png_ptr, info_ptr, width, height, bit_depth, PNG_COLOR_TYPE_???,
       PNG_INTERLACE_????, PNG_COMPRESSION_TYPE_BASE, PNG_FILTER_TYPE_BASE);

/* Set the palette if there is one.  REQUIRED for indexed-color images */
   palette = (png_colorp)png_malloc(png_ptr, PNG_MAX_PALETTE_LENGTH
             * (sizeof (png_color)));
   /* ... Set palette colors ... */
   png_set_PLTE(png_ptr, info_ptr, palette, PNG_MAX_PALETTE_LENGTH);
   /* You must not free palette here, because png_set_PLTE only makes a link to
    * the palette that you malloced.  Wait until you are about to destroy
    * the png structure.
    */

/* Optional significant bit (sBIT) chunk */
   png_color_8 sig_bit;

/* If we are dealing with a grayscale image then */
   sig_bit.gray = true_bit_depth;

/* Otherwise, if we are dealing with a color image then */
   sig_bit.red = true_red_bit_depth;
   sig_bit.green = true_green_bit_depth;
   sig_bit.blue = true_blue_bit_depth;

/* If the image has an alpha channel then */
   sig_bit.alpha = true_alpha_bit_depth;

png_set_sBIT(png_ptr, info_ptr, &sig_bit);

/* Optional gamma chunk is strongly suggested if you have any guess
    * as to the correct gamma of the image.
    */
   png_set_gAMA(png_ptr, info_ptr, gamma);

/* Optionally write comments into the image */
   {
      png_text text_ptr[3];

char key0[]="Title";
      char text0[]="Mona Lisa";
      text_ptr[0].key = key0;
      text_ptr[0].text = text0;
      text_ptr[0].compression = PNG_TEXT_COMPRESSION_NONE;
      text_ptr[0].itxt_length = 0;
      text_ptr[0].lang = NULL;
      text_ptr[0].lang_key = NULL;

char key1[]="Author";
      char text1[]="Leonardo DaVinci";
      text_ptr[1].key = key1;
      text_ptr[1].text = text1;
      text_ptr[1].compression = PNG_TEXT_COMPRESSION_NONE;
      text_ptr[1].itxt_length = 0;
      text_ptr[1].lang = NULL;
      text_ptr[1].lang_key = NULL;

char key2[]="Description";
      char text2[]="<long text>";
      text_ptr[2].key = key2;
      text_ptr[2].text = text2;
      text_ptr[2].compression = PNG_TEXT_COMPRESSION_zTXt;
      text_ptr[2].itxt_length = 0;
      text_ptr[2].lang = NULL;
      text_ptr[2].lang_key = NULL;

png_set_text(write_ptr, write_info_ptr, text_ptr, 3);
   }

/* Other optional chunks like cHRM, bKGD, tRNS, tIME, oFFs, pHYs */

/* Note that if sRGB is present the gAMA and cHRM chunks must be ignored
    * on read and, if your application chooses to write them, they must
    * be written in accordance with the sRGB profile
    */

/* Write the file header information.  REQUIRED */
   png_write_info(png_ptr, info_ptr);

/* If you want, you can write the info in two steps, in case you need to
    * write your private chunk ahead of PLTE:
    *
    *   png_write_info_before_PLTE(write_ptr, write_info_ptr);
    *   write_my_chunk();
    *   png_write_info(png_ptr, info_ptr);
    *
    * However, given the level of known- and unknown-chunk support in 1.2.0
    * and up, this should no longer be necessary.
    */

/* Once we write out the header, the compression type on the text
    * chunk gets changed to PNG_TEXT_COMPRESSION_NONE_WR or
    * PNG_TEXT_COMPRESSION_zTXt_WR, so it doesn't get written out again
    * at the end.
    */

/* Set up the transformations you want.  Note that these are
    * all optional.  Only call them if you want them.
    */

/* Invert monochrome pixels */
   png_set_invert_mono(png_ptr);

/* Shift the pixels up to a legal bit depth and fill in
    * as appropriate to correctly scale the image.
    */
   png_set_shift(png_ptr, &sig_bit);

/* Pack pixels into bytes */
   png_set_packing(png_ptr);

/* Swap location of alpha bytes from ARGB to RGBA */
   png_set_swap_alpha(png_ptr);

/* Get rid of filler (OR ALPHA) bytes, pack XRGB/RGBX/ARGB/RGBA into
    * RGB (4 channels -> 3 channels). The second parameter is not used.
    */
   png_set_filler(png_ptr, 0, PNG_FILLER_BEFORE);

/* Flip BGR pixels to RGB */
   png_set_bgr(png_ptr);

/* Swap bytes of 16-bit files to most significant byte first */
   png_set_swap(png_ptr);

/* Swap bits of 1-bit, 2-bit, 4-bit packed pixel formats */
   png_set_packswap(png_ptr);

/* Turn on interlace handling if you are not using png_write_image() */
   if (interlacing != 0)
      number_passes = png_set_interlace_handling(png_ptr);

else
      number_passes = 1;

/* The easiest way to write the image (you may have a different memory
    * layout, however, so choose what fits your needs best).  You need to
    * use the first method if you aren't handling interlacing yourself.
    */
   png_uint_32 k, height, width;

/* In this example, "image" is a one-dimensional array of bytes */

/* Guard against integer overflow */
   if (height > PNG_SIZE_MAX/(width*bytes_per_pixel)) {
      png_error(png_ptr, "Image_data buffer would be too large");
   }
   png_byte image[height*width*bytes_per_pixel];

png_bytep row_pointers[height];

if (height > PNG_UINT_32_MAX/(sizeof (png_bytep)))
      png_error (png_ptr, "Image is too tall to process in memory");

/* Set up pointers into your "image" byte array */
   for (k = 0; k < height; k++)
      row_pointers[k] = image + k*width*bytes_per_pixel;

/* One of the following output methods is REQUIRED */

#ifdef entire /* Write out the entire image data in one call */
   png_write_image(png_ptr, row_pointers);

/* The other way to write the image - deal with interlacing */

#else no_entire /* Write out the image data by one or more scanlines */

/* The number of passes is either 1 for non-interlaced images,
    * or 7 for interlaced images.
    */
   for (pass = 0; pass < number_passes; pass++)
   {
      /* Write a few rows at a time. */
      png_write_rows(png_ptr, &row_pointers[first_row], number_of_rows);

/* If you are only writing one row at a time, this works */
      for (y = 0; y < height; y++)
         png_write_rows(png_ptr, &row_pointers[y], 1);
   }
#endif no_entire /* Use only one output method */

/* You can write optional chunks like tEXt, zTXt, and tIME at the end
    * as well.  Shouldn't be necessary in 1.2.0 and up as all the public
    * chunks are supported and you can use png_set_unknown_chunks() to
    * register unknown chunks into the info structure to be written out.
    */

/* It is REQUIRED to call this to finish writing the rest of the file */
   png_write_end(png_ptr, info_ptr);
#endif hilevel

/* If you png_malloced a palette, free it here (don't free info_ptr->palette,
    * as recommended in versions 1.0.5m and earlier of this example; if
    * libpng mallocs info_ptr->palette, libpng will free it).  If you
    * allocated it with malloc() instead of png_malloc(), use free() instead
    * of png_free().
    */
   png_free(png_ptr, palette);
   palette = NULL;

/* Similarly, if you png_malloced any data that you passed in with
    * png_set_something(), such as a hist or trans array, free it here,
    * when you can be sure that libpng is through with it.
    */
   png_free(png_ptr, trans);
   trans = NULL;
   /* Whenever you use png_free() it is a good idea to set the pointer to
    * NULL in case your application inadvertently tries to png_free() it
    * again.  When png_free() sees a NULL it returns without action, thus
    * avoiding the double-free security problem.
    */

/* Clean up after the write, and free any memory allocated */
   png_destroy_write_struct(&png_ptr, &info_ptr);

/* Close the file */
   fclose(fp);

/* That's it */
   return (OK);
}

#endif /* if 0 */
PKJM�\����S�S�libpng-manual.txtnu�[���libpng-manual.txt - A description on how to use and modify libpng

libpng version 1.6.34 - September 29, 2017
 Updated and distributed by Glenn Randers-Pehrson
 <glennrp at users.sourceforge.net>
 Copyright (c) 1998-2017 Glenn Randers-Pehrson

This document is released under the libpng license.
 For conditions of distribution and use, see the disclaimer
 and license in png.h

Based on:

libpng versions 0.97, January 1998, through 1.6.34 - September 29, 2017
 Updated and distributed by Glenn Randers-Pehrson
 Copyright (c) 1998-2017 Glenn Randers-Pehrson

libpng 1.0 beta 6 - version 0.96 - May 28, 1997
 Updated and distributed by Andreas Dilger
 Copyright (c) 1996, 1997 Andreas Dilger

libpng 1.0 beta 2 - version 0.88 - January 26, 1996
 For conditions of distribution and use, see copyright
 notice in png.h. Copyright (c) 1995, 1996 Guy Eric
 Schalnat, Group 42, Inc.

Updated/rewritten per request in the libpng FAQ
 Copyright (c) 1995, 1996 Frank J. T. Wojcik
 December 18, 1995 & January 20, 1996

TABLE OF CONTENTS

I. Introduction
   II. Structures
  III. Reading
   IV. Writing
    V. Simplified API
   VI. Modifying/Customizing libpng
  VII. MNG support
 VIII. Changes to Libpng from version 0.88
   IX. Changes to Libpng from version 1.0.x to 1.2.x
    X. Changes to Libpng from version 1.0.x/1.2.x to 1.4.x
   XI. Changes to Libpng from version 1.4.x to 1.5.x
  XII. Changes to Libpng from version 1.5.x to 1.6.x
 XIII. Detecting libpng
  XIV. Source code repository
   XV. Coding style
  XVI. Y2K Compliance in libpng

I. Introduction

This file describes how to use and modify the PNG reference library
(known as libpng) for your own use.  In addition to this
file, example.c is a good starting point for using the library, as
it is heavily commented and should include everything most people
will need.  We assume that libpng is already installed; see the
INSTALL file for instructions on how to configure and install libpng.

For examples of libpng usage, see the files "example.c", "pngtest.c",
and the files in the "contrib" directory, all of which are included in
the libpng distribution.

Libpng was written as a companion to the PNG specification, as a way
of reducing the amount of time and effort it takes to support the PNG
file format in application programs.

The PNG specification (second edition), November 2003, is available as
a W3C Recommendation and as an ISO Standard (ISO/IEC 15948:2004 (E)) at
<https://www.w3.org/TR/2003/REC-PNG-20031110/
The W3C and ISO documents have identical technical content.

The PNG-1.2 specification is available at
<https://png-mng.sourceforge.io/pub/png/spec/1.2/>.
It is technically equivalent
to the PNG specification (second edition) but has some additional material.

The PNG-1.0 specification is available as RFC 2083
<https://png-mng.sourceforge.io/pub/png/spec/1.0/> and as a
W3C Recommendation <https://www.w3.org/TR/REC-png-961001>.

Some additional chunks are described in the special-purpose public chunks
documents at <http://www.libpng.org/pub/png/spec/register/>

Other information
about PNG, and the latest version of libpng, can be found at the PNG home
page, <http://www.libpng.org/pub/png/>.

Most users will not have to modify the library significantly; advanced
users may want to modify it more.  All attempts were made to make it as
complete as possible, while keeping the code easy to understand.
Currently, this library only supports C.  Support for other languages
is being considered.

Libpng has been designed to handle multiple sessions at one time,
to be easily modifiable, to be portable to the vast majority of
machines (ANSI, K&R, 16-, 32-, and 64-bit) available, and to be easy
to use.  The ultimate goal of libpng is to promote the acceptance of
the PNG file format in whatever way possible.  While there is still
work to be done (see the TODO file), libpng should cover the
majority of the needs of its users.

Libpng uses zlib for its compression and decompression of PNG files.
Further information about zlib, and the latest version of zlib, can
be found at the zlib home page, <https://zlib.net/>.
The zlib compression utility is a general purpose utility that is
useful for more than PNG files, and can be used without libpng.
See the documentation delivered with zlib for more details.
You can usually find the source files for the zlib utility wherever you
find the libpng source files.

Libpng is thread safe, provided the threads are using different
instances of the structures.  Each thread should have its own
png_struct and png_info instances, and thus its own image.
Libpng does not protect itself against two threads using the
same instance of a structure.

II. Structures

There are two main structures that are important to libpng, png_struct
and png_info.  Both are internal structures that are no longer exposed
in the libpng interface (as of libpng 1.5.0).

The png_info structure is designed to provide information about the
PNG file.  At one time, the fields of png_info were intended to be
directly accessible to the user.  However, this tended to cause problems
with applications using dynamically loaded libraries, and as a result
a set of interface functions for png_info (the png_get_*() and png_set_*()
functions) was developed, and direct access to the png_info fields was
deprecated..

The png_struct structure is the object used by the library to decode a
single image.  As of 1.5.0 this structure is also not exposed.

Almost all libpng APIs require a pointer to a png_struct as the first argument.
Many (in particular the png_set and png_get APIs) also require a pointer
to png_info as the second argument.  Some application visible macros
defined in png.h designed for basic data access (reading and writing
integers in the PNG format) don't take a png_info pointer, but it's almost
always safe to assume that a (png_struct*) has to be passed to call an API
function.

You can have more than one png_info structure associated with an image,
as illustrated in pngtest.c, one for information valid prior to the
IDAT chunks and another (called "end_info" below) for things after them.

The png.h header file is an invaluable reference for programming with libpng.
And while I'm on the topic, make sure you include the libpng header file:

#include <png.h>

and also (as of libpng-1.5.0) the zlib header file, if you need it:

#include <zlib.h>

Types

The png.h header file defines a number of integral types used by the
APIs.  Most of these are fairly obvious; for example types corresponding
to integers of particular sizes and types for passing color values.

One exception is how non-integral numbers are handled.  For application
convenience most APIs that take such numbers have C (double) arguments;
however, internally PNG, and libpng, use 32 bit signed integers and encode
the value by multiplying by 100,000.  As of libpng 1.5.0 a convenience
macro PNG_FP_1 is defined in png.h along with a type (png_fixed_point)
which is simply (png_int_32).

All APIs that take (double) arguments also have a matching API that
takes the corresponding fixed point integer arguments.  The fixed point
API has the same name as the floating point one with "_fixed" appended.
The actual range of values permitted in the APIs is frequently less than
the full range of (png_fixed_point) (-21474 to +21474).  When APIs require
a non-negative argument the type is recorded as png_uint_32 above.  Consult
the header file and the text below for more information.

Special care must be take with sCAL chunk handling because the chunk itself
uses non-integral values encoded as strings containing decimal floating point
numbers.  See the comments in the header file.

Configuration

The main header file function declarations are frequently protected by C
preprocessing directives of the form:

#ifdef PNG_feature_SUPPORTED
    declare-function
    #endif
    ...
    #ifdef PNG_feature_SUPPORTED
    use-function
    #endif

The library can be built without support for these APIs, although a
standard build will have all implemented APIs.  Application programs
should check the feature macros before using an API for maximum
portability.  From libpng 1.5.0 the feature macros set during the build
of libpng are recorded in the header file "pnglibconf.h" and this file
is always included by png.h.

If you don't need to change the library configuration from the default, skip to
the next section ("Reading").

Notice that some of the makefiles in the 'scripts' directory and (in 1.5.0) all
of the build project files in the 'projects' directory simply copy
scripts/pnglibconf.h.prebuilt to pnglibconf.h.  This means that these build
systems do not permit easy auto-configuration of the library - they only
support the default configuration.

The easiest way to make minor changes to the libpng configuration when
auto-configuration is supported is to add definitions to the command line
using (typically) CPPFLAGS.  For example:

CPPFLAGS=-DPNG_NO_FLOATING_ARITHMETIC

will change the internal libpng math implementation for gamma correction and
other arithmetic calculations to fixed point, avoiding the need for fast
floating point support.  The result can be seen in the generated pnglibconf.h -
make sure it contains the changed feature macro setting.

If you need to make more extensive configuration changes - more than one or two
feature macro settings - you can either add -DPNG_USER_CONFIG to the build
command line and put a list of feature macro settings in pngusr.h or you can set
DFA_XTRA (a makefile variable) to a file containing the same information in the
form of 'option' settings.

A. Changing pnglibconf.h

A variety of methods exist to build libpng.  Not all of these support
reconfiguration of pnglibconf.h.  To reconfigure pnglibconf.h it must either be
rebuilt from scripts/pnglibconf.dfa using awk or it must be edited by hand.

Hand editing is achieved by copying scripts/pnglibconf.h.prebuilt to
pnglibconf.h and changing the lines defining the supported features, paying
very close attention to the 'option' information in scripts/pnglibconf.dfa
that describes those features and their requirements.  This is easy to get
wrong.

B. Configuration using DFA_XTRA

Rebuilding from pnglibconf.dfa is easy if a functioning 'awk', or a later
variant such as 'nawk' or 'gawk', is available.  The configure build will
automatically find an appropriate awk and build pnglibconf.h.
The scripts/pnglibconf.mak file contains a set of make rules for doing the
same thing if configure is not used, and many of the makefiles in the scripts
directory use this approach.

When rebuilding simply write a new file containing changed options and set
DFA_XTRA to the name of this file.  This causes the build to append the new file
to the end of scripts/pnglibconf.dfa.  The pngusr.dfa file should contain lines
of the following forms:

everything = off

This turns all optional features off.  Include it at the start of pngusr.dfa to
make it easier to build a minimal configuration.  You will need to turn at least
some features on afterward to enable either reading or writing code, or both.

option feature on
option feature off

Enable or disable a single feature.  This will automatically enable other
features required by a feature that is turned on or disable other features that
require a feature which is turned off.  Conflicting settings will cause an error
message to be emitted by awk.

setting feature default value

Changes the default value of setting 'feature' to 'value'.  There are a small
number of settings listed at the top of pnglibconf.h, they are documented in the
source code.  Most of these values have performance implications for the library
but most of them have no visible effect on the API.  Some can also be overridden
from the API.

This method of building a customized pnglibconf.h is illustrated in
contrib/pngminim/*.  See the "$(PNGCONF):" target in the makefile and
pngusr.dfa in these directories.

C. Configuration using PNG_USER_CONFIG

If -DPNG_USER_CONFIG is added to the CPPFLAGS when pnglibconf.h is built,
the file pngusr.h will automatically be included before the options in
scripts/pnglibconf.dfa are processed.  Your pngusr.h file should contain only
macro definitions turning features on or off or setting settings.

Apart from the global setting "everything = off" all the options listed above
can be set using macros in pngusr.h:

#define PNG_feature_SUPPORTED

is equivalent to:

option feature on

#define PNG_NO_feature

is equivalent to:

option feature off

#define PNG_feature value

is equivalent to:

setting feature default value

Notice that in both cases, pngusr.dfa and pngusr.h, the contents of the
pngusr file you supply override the contents of scripts/pnglibconf.dfa

If confusing or incomprehensible behavior results it is possible to
examine the intermediate file pnglibconf.dfn to find the full set of
dependency information for each setting and option.  Simply locate the
feature in the file and read the C comments that precede it.

This method is also illustrated in the contrib/pngminim/* makefiles and
pngusr.h.

III. Reading

We'll now walk you through the possible functions to call when reading
in a PNG file sequentially, briefly explaining the syntax and purpose
of each one.  See example.c and png.h for more detail.  While
progressive reading is covered in the next section, you will still
need some of the functions discussed in this section to read a PNG
file.

Setup

You will want to do the I/O initialization(*) before you get into libpng,
so if it doesn't work, you don't have much to undo.  Of course, you
will also want to insure that you are, in fact, dealing with a PNG
file.  Libpng provides a simple check to see if a file is a PNG file.
To use it, pass in the first 1 to 8 bytes of the file to the function
png_sig_cmp(), and it will return 0 (false) if the bytes match the
corresponding bytes of the PNG signature, or nonzero (true) otherwise.
Of course, the more bytes you pass in, the greater the accuracy of the
prediction.

If you are intending to keep the file pointer open for use in libpng,
you must ensure you don't read more than 8 bytes from the beginning
of the file, and you also have to make a call to png_set_sig_bytes()
with the number of bytes you read from the beginning.  Libpng will
then only check the bytes (if any) that your program didn't read.

(*): If you are not using the standard I/O functions, you will need
to replace them with custom functions.  See the discussion under
Customizing libpng.

FILE *fp = fopen(file_name, "rb");
    if (!fp)
    {
       return (ERROR);
    }

if (fread(header, 1, number, fp) != number)
    {
       return (ERROR);
    }

is_png = !png_sig_cmp(header, 0, number);
    if (!is_png)
    {
       return (NOT_PNG);
    }

Next, png_struct and png_info need to be allocated and initialized.  In
order to ensure that the size of these structures is correct even with a
dynamically linked libpng, there are functions to initialize and
allocate the structures.  We also pass the library version, optional
pointers to error handling functions, and a pointer to a data struct for
use by the error functions, if necessary (the pointer and functions can
be NULL if the default error handlers are to be used).  See the section
on Changes to Libpng below regarding the old initialization functions.
The structure allocation functions quietly return NULL if they fail to
create the structure, so your application should check for that.

png_structp png_ptr = png_create_read_struct
        (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
        user_error_fn, user_warning_fn);

if (!png_ptr)
       return (ERROR);

png_infop info_ptr = png_create_info_struct(png_ptr);

if (!info_ptr)
    {
       png_destroy_read_struct(&png_ptr,
           (png_infopp)NULL, (png_infopp)NULL);
       return (ERROR);
    }

If you want to use your own memory allocation routines,
use a libpng that was built with PNG_USER_MEM_SUPPORTED defined, and use
png_create_read_struct_2() instead of png_create_read_struct():

png_structp png_ptr = png_create_read_struct_2
        (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
        user_error_fn, user_warning_fn, (png_voidp)
        user_mem_ptr, user_malloc_fn, user_free_fn);

The error handling routines passed to png_create_read_struct()
and the memory alloc/free routines passed to png_create_struct_2()
are only necessary if you are not using the libpng supplied error
handling and memory alloc/free functions.

When libpng encounters an error, it expects to longjmp back
to your routine.  Therefore, you will need to call setjmp and pass
your png_jmpbuf(png_ptr).  If you read the file from different
routines, you will need to update the longjmp buffer every time you enter
a new routine that will call a png_*() function.

See your documentation of setjmp/longjmp for your compiler for more
information on setjmp/longjmp.  See the discussion on libpng error
handling in the Customizing Libpng section below for more information
on the libpng error handling.  If an error occurs, and libpng longjmp's
back to your setjmp, you will want to call png_destroy_read_struct() to
free any memory.

if (setjmp(png_jmpbuf(png_ptr)))
    {
       png_destroy_read_struct(&png_ptr, &info_ptr,
           &end_info);
       fclose(fp);
       return (ERROR);
    }

Pass (png_infopp)NULL instead of &end_info if you didn't create
an end_info structure.

If you would rather avoid the complexity of setjmp/longjmp issues,
you can compile libpng with PNG_NO_SETJMP, in which case
errors will result in a call to PNG_ABORT() which defaults to abort().

You can #define PNG_ABORT() to a function that does something
more useful than abort(), as long as your function does not
return.

Now you need to set up the input code.  The default for libpng is to
use the C function fread().  If you use this, you will need to pass a
valid FILE * in the function png_init_io().  Be sure that the file is
opened in binary mode.  If you wish to handle reading data in another
way, you need not call the png_init_io() function, but you must then
implement the libpng I/O methods discussed in the Customizing Libpng
section below.

png_init_io(png_ptr, fp);

If you had previously opened the file and read any of the signature from
the beginning in order to see if this was a PNG file, you need to let
libpng know that there are some bytes missing from the start of the file.

png_set_sig_bytes(png_ptr, number);

You can change the zlib compression buffer size to be used while
reading compressed data with

png_set_compression_buffer_size(png_ptr, buffer_size);

where the default size is 8192 bytes.  Note that the buffer size
is changed immediately and the buffer is reallocated immediately,
instead of setting a flag to be acted upon later.

If you want CRC errors to be handled in a different manner than
the default, use

png_set_crc_action(png_ptr, crit_action, ancil_action);

The values for png_set_crc_action() say how libpng is to handle CRC errors in
ancillary and critical chunks, and whether to use the data contained
therein. Starting with libpng-1.6.26, this also governs how an ADLER32 error
is handled while reading the IDAT chunk. Note that it is impossible to
"discard" data in a critical chunk.

Choices for (int) crit_action are
   PNG_CRC_DEFAULT      0  error/quit
   PNG_CRC_ERROR_QUIT   1  error/quit
   PNG_CRC_WARN_USE     3  warn/use data
   PNG_CRC_QUIET_USE    4  quiet/use data
   PNG_CRC_NO_CHANGE    5  use the current value

Choices for (int) ancil_action are
   PNG_CRC_DEFAULT      0  error/quit
   PNG_CRC_ERROR_QUIT   1  error/quit
   PNG_CRC_WARN_DISCARD 2  warn/discard data
   PNG_CRC_WARN_USE     3  warn/use data
   PNG_CRC_QUIET_USE    4  quiet/use data
   PNG_CRC_NO_CHANGE    5  use the current value

When the setting for crit_action is PNG_CRC_QUIET_USE, the CRC and ADLER32
checksums are not only ignored, but they are not evaluated.

Setting up callback code

You can set up a callback function to handle any unknown chunks in the
input stream. You must supply the function

read_chunk_callback(png_structp png_ptr,
         png_unknown_chunkp chunk);
    {
       /* The unknown chunk structure contains your
          chunk data, along with similar data for any other
          unknown chunks: */

png_byte name[5];
           png_byte *data;
           png_size_t size;

/* Note that libpng has already taken care of
          the CRC handling */

/* put your code here.  Search for your chunk in the
          unknown chunk structure, process it, and return one
          of the following: */

return (-n); /* chunk had an error */
       return (0); /* did not recognize */
       return (n); /* success */
    }

(You can give your function another name that you like instead of
"read_chunk_callback")

To inform libpng about your function, use

png_set_read_user_chunk_fn(png_ptr, user_chunk_ptr,
        read_chunk_callback);

This names not only the callback function, but also a user pointer that
you can retrieve with

png_get_user_chunk_ptr(png_ptr);

If you call the png_set_read_user_chunk_fn() function, then all unknown
chunks which the callback does not handle will be saved when read.  You can
cause them to be discarded by returning '1' ("handled") instead of '0'.  This
behavior will change in libpng 1.7 and the default handling set by the
png_set_keep_unknown_chunks() function, described below, will be used when the
callback returns 0.  If you want the existing behavior you should set the global
default to PNG_HANDLE_CHUNK_IF_SAFE now; this is compatible with all current
versions of libpng and with 1.7.  Libpng 1.6 issues a warning if you keep the
default, or PNG_HANDLE_CHUNK_NEVER, and the callback returns 0.

At this point, you can set up a callback function that will be
called after each row has been read, which you can use to control
a progress meter or the like.  It's demonstrated in pngtest.c.
You must supply a function

void read_row_callback(png_structp png_ptr,
       png_uint_32 row, int pass);
    {
      /* put your code here */
    }

(You can give it another name that you like instead of "read_row_callback")

To inform libpng about your function, use

png_set_read_status_fn(png_ptr, read_row_callback);

When this function is called the row has already been completely processed and
the 'row' and 'pass' refer to the next row to be handled.  For the
non-interlaced case the row that was just handled is simply one less than the
passed in row number, and pass will always be 0.  For the interlaced case the
same applies unless the row value is 0, in which case the row just handled was
the last one from one of the preceding passes.  Because interlacing may skip a
pass you cannot be sure that the preceding pass is just 'pass-1'; if you really
need to know what the last pass is record (row,pass) from the callback and use
the last recorded value each time.

As with the user transform you can find the output row using the
PNG_ROW_FROM_PASS_ROW macro.

Unknown-chunk handling

Now you get to set the way the library processes unknown chunks in the
input PNG stream. Both known and unknown chunks will be read.  Normal
behavior is that known chunks will be parsed into information in
various info_ptr members while unknown chunks will be discarded. This
behavior can be wasteful if your application will never use some known
chunk types. To change this, you can call:

png_set_keep_unknown_chunks(png_ptr, keep,
        chunk_list, num_chunks);

keep       - 0: default unknown chunk handling
                 1: ignore; do not keep
                 2: keep only if safe-to-copy
                 3: keep even if unsafe-to-copy

You can use these definitions:
                 PNG_HANDLE_CHUNK_AS_DEFAULT   0
                 PNG_HANDLE_CHUNK_NEVER        1
                 PNG_HANDLE_CHUNK_IF_SAFE      2
                 PNG_HANDLE_CHUNK_ALWAYS       3

chunk_list - list of chunks affected (a byte string,
                 five bytes per chunk, NULL or '\0' if
                 num_chunks is positive; ignored if
                 numchunks <= 0).

num_chunks - number of chunks affected; if 0, all
                 unknown chunks are affected.  If positive,
                 only the chunks in the list are affected,
                 and if negative all unknown chunks and
                 all known chunks except for the IHDR,
                 PLTE, tRNS, IDAT, and IEND chunks are
                 affected.

Unknown chunks declared in this way will be saved as raw data onto a
list of png_unknown_chunk structures.  If a chunk that is normally
known to libpng is named in the list, it will be handled as unknown,
according to the "keep" directive.  If a chunk is named in successive
instances of png_set_keep_unknown_chunks(), the final instance will
take precedence.  The IHDR and IEND chunks should not be named in
chunk_list; if they are, libpng will process them normally anyway.
If you know that your application will never make use of some particular
chunks, use PNG_HANDLE_CHUNK_NEVER (or 1) as demonstrated below.

Here is an example of the usage of png_set_keep_unknown_chunks(),
where the private "vpAg" chunk will later be processed by a user chunk
callback function:

png_byte vpAg[5]={118, 112,  65, 103, (png_byte) '\0'};

#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
      png_byte unused_chunks[]=
      {
        104,  73,  83,  84, (png_byte) '\0',   /* hIST */
        105,  84,  88, 116, (png_byte) '\0',   /* iTXt */
        112,  67,  65,  76, (png_byte) '\0',   /* pCAL */
        115,  67,  65,  76, (png_byte) '\0',   /* sCAL */
        115,  80,  76,  84, (png_byte) '\0',   /* sPLT */
        116,  73,  77,  69, (png_byte) '\0',   /* tIME */
      };
    #endif

...

#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
      /* ignore all unknown chunks
       * (use global setting "2" for libpng16 and earlier):
       */
      png_set_keep_unknown_chunks(read_ptr, 2, NULL, 0);

/* except for vpAg: */
      png_set_keep_unknown_chunks(read_ptr, 2, vpAg, 1);

/* also ignore unused known chunks: */
      png_set_keep_unknown_chunks(read_ptr, 1, unused_chunks,
         (int)(sizeof unused_chunks)/5);
    #endif

User limits

The PNG specification allows the width and height of an image to be as
large as 2^31-1 (0x7fffffff), or about 2.147 billion rows and columns.
For safety, libpng imposes a default limit of 1 million rows and columns.
Larger images will be rejected immediately with a png_error() call. If
you wish to change these limits, you can use

png_set_user_limits(png_ptr, width_max, height_max);

to set your own limits (libpng may reject some very wide images
anyway because of potential buffer overflow conditions).

You should put this statement after you create the PNG structure and
before calling png_read_info(), png_read_png(), or png_process_data().

When writing a PNG datastream, put this statement before calling
png_write_info() or png_write_png().

If you need to retrieve the limits that are being applied, use

width_max = png_get_user_width_max(png_ptr);
   height_max = png_get_user_height_max(png_ptr);

The PNG specification sets no limit on the number of ancillary chunks
allowed in a PNG datastream.  By default, libpng imposes a limit of
a total of 1000 sPLT, tEXt, iTXt, zTXt, and unknown chunks to be stored.
If you have set up both info_ptr and end_info_ptr, the limit applies
separately to each.  You can change the limit on the total number of such
chunks that will be stored, with

png_set_chunk_cache_max(png_ptr, user_chunk_cache_max);

where 0x7fffffffL means unlimited.  You can retrieve this limit with

chunk_cache_max = png_get_chunk_cache_max(png_ptr);

Libpng imposes a limit of 8 Megabytes (8,000,000 bytes) on the amount of
memory that any chunk other than IDAT can occupy, originally or when
decompressed (prior to libpng-1.6.32 the limit was only applied to compressed
chunks after decompression). You can change this limit with

png_set_chunk_malloc_max(png_ptr, user_chunk_malloc_max);

and you can retrieve the limit with

chunk_malloc_max = png_get_chunk_malloc_max(png_ptr);

Any chunks that would cause either of these limits to be exceeded will
be ignored.

Information about your system

If you intend to display the PNG or to incorporate it in other image data you
need to tell libpng information about your display or drawing surface so that
libpng can convert the values in the image to match the display.

From libpng-1.5.4 this information can be set before reading the PNG file
header.  In earlier versions png_set_gamma() existed but behaved incorrectly if
called before the PNG file header had been read and png_set_alpha_mode() did not
exist.

If you need to support versions prior to libpng-1.5.4 test the version number
as illustrated below using "PNG_LIBPNG_VER >= 10504" and follow the procedures
described in the appropriate manual page.

You give libpng the encoding expected by your system expressed as a 'gamma'
value.  You can also specify a default encoding for the PNG file in
case the required information is missing from the file.  By default libpng
assumes that the PNG data matches your system, to keep this default call:

png_set_gamma(png_ptr, screen_gamma, output_gamma);

or you can use the fixed point equivalent:

png_set_gamma_fixed(png_ptr, PNG_FP_1*screen_gamma,
      PNG_FP_1*output_gamma);

If you don't know the gamma for your system it is probably 2.2 - a good
approximation to the IEC standard for display systems (sRGB).  If images are
too contrasty or washed out you got the value wrong - check your system
documentation!

Many systems permit the system gamma to be changed via a lookup table in the
display driver, a few systems, including older Macs, change the response by
default.  As of 1.5.4 three special values are available to handle common
situations:

PNG_DEFAULT_sRGB: Indicates that the system conforms to the
                     IEC 61966-2-1 standard.  This matches almost
                     all systems.
   PNG_GAMMA_MAC_18: Indicates that the system is an older
                     (pre Mac OS 10.6) Apple Macintosh system with
                     the default settings.
   PNG_GAMMA_LINEAR: Just the fixed point value for 1.0 - indicates
                     that the system expects data with no gamma
                     encoding.

You would use the linear (unencoded) value if you need to process the pixel
values further because this avoids the need to decode and re-encode each
component value whenever arithmetic is performed.  A lot of graphics software
uses linear values for this reason, often with higher precision component values
to preserve overall accuracy.

The output_gamma value expresses how to decode the output values, not how
they are encoded.  The values used correspond to the normal numbers used to
describe the overall gamma of a computer display system; for example 2.2 for
an sRGB conformant system.  The values are scaled by 100000 in the _fixed
version of the API (so 220000 for sRGB.)

The inverse of the value is always used to provide a default for the PNG file
encoding if it has no gAMA chunk and if png_set_gamma() has not been called
to override the PNG gamma information.

When the ALPHA_OPTIMIZED mode is selected the output gamma is used to encode
opaque pixels however pixels with lower alpha values are not encoded,
regardless of the output gamma setting.

When the standard Porter Duff handling is requested with mode 1 the output
encoding is set to be linear and the output_gamma value is only relevant
as a default for input data that has no gamma information.  The linear output
encoding will be overridden if png_set_gamma() is called - the results may be
highly unexpected!

The following numbers are derived from the sRGB standard and the research
behind it.  sRGB is defined to be approximated by a PNG gAMA chunk value of
0.45455 (1/2.2) for PNG.  The value implicitly includes any viewing
correction required to take account of any differences in the color
environment of the original scene and the intended display environment; the
value expresses how to *decode* the image for display, not how the original
data was *encoded*.

sRGB provides a peg for the PNG standard by defining a viewing environment.
sRGB itself, and earlier TV standards, actually use a more complex transform
(a linear portion then a gamma 2.4 power law) than PNG can express.  (PNG is
limited to simple power laws.)  By saying that an image for direct display on
an sRGB conformant system should be stored with a gAMA chunk value of 45455
(11.3.3.2 and 11.3.3.5 of the ISO PNG specification) the PNG specification
makes it possible to derive values for other display systems and
environments.

The Mac value is deduced from the sRGB based on an assumption that the actual
extra viewing correction used in early Mac display systems was implemented as
a power 1.45 lookup table.

Any system where a programmable lookup table is used or where the behavior of
the final display device characteristics can be changed requires system
specific code to obtain the current characteristic.  However this can be
difficult and most PNG gamma correction only requires an approximate value.

By default, if png_set_alpha_mode() is not called, libpng assumes that all
values are unencoded, linear, values and that the output device also has a
linear characteristic.  This is only very rarely correct - it is invariably
better to call png_set_alpha_mode() with PNG_DEFAULT_sRGB than rely on the
default if you don't know what the right answer is!

The special value PNG_GAMMA_MAC_18 indicates an older Mac system (pre Mac OS
10.6) which used a correction table to implement a somewhat lower gamma on an
otherwise sRGB system.

Both these values are reserved (not simple gamma values) in order to allow
more precise correction internally in the future.

NOTE: the values can be passed to either the fixed or floating
point APIs, but the floating point API will also accept floating point
values.

The second thing you may need to tell libpng about is how your system handles
alpha channel information.  Some, but not all, PNG files contain an alpha
channel.  To display these files correctly you need to compose the data onto a
suitable background, as described in the PNG specification.

Libpng only supports composing onto a single color (using png_set_background;
see below).  Otherwise you must do the composition yourself and, in this case,
you may need to call png_set_alpha_mode:

#if PNG_LIBPNG_VER >= 10504
      png_set_alpha_mode(png_ptr, mode, screen_gamma);
   #else
      png_set_gamma(png_ptr, screen_gamma, 1.0/screen_gamma);
   #endif

The screen_gamma value is the same as the argument to png_set_gamma; however,
how it affects the output depends on the mode.  png_set_alpha_mode() sets the
file gamma default to 1/screen_gamma, so normally you don't need to call
png_set_gamma.  If you need different defaults call png_set_gamma() before
png_set_alpha_mode() - if you call it after it will override the settings made
by png_set_alpha_mode().

The mode is as follows:

PNG_ALPHA_PNG: The data is encoded according to the PNG
specification.  Red, green and blue, or gray, components are
gamma encoded color values and are not premultiplied by the
alpha value.  The alpha value is a linear measure of the
contribution of the pixel to the corresponding final output pixel.

You should normally use this format if you intend to perform
color correction on the color values; most, maybe all, color
correction software has no handling for the alpha channel and,
anyway, the math to handle pre-multiplied component values is
unnecessarily complex.

Before you do any arithmetic on the component values you need
to remove the gamma encoding and multiply out the alpha
channel.  See the PNG specification for more detail.  It is
important to note that when an image with an alpha channel is
scaled, linear encoded, pre-multiplied component values must
be used!

The remaining modes assume you don't need to do any further color correction or
that if you do, your color correction software knows all about alpha (it
probably doesn't!).  They 'associate' the alpha with the color information by
storing color channel values that have been scaled by the alpha.  The
advantage is that the color channels can be resampled (the image can be
scaled) in this form.  The disadvantage is that normal practice is to store
linear, not (gamma) encoded, values and this requires 16-bit channels for
still images rather than the 8-bit channels that are just about sufficient if
gamma encoding is used.  In addition all non-transparent pixel values,
including completely opaque ones, must be gamma encoded to produce the final
image.  These are the 'STANDARD', 'ASSOCIATED' or 'PREMULTIPLIED' modes
described below (the latter being the two common names for associated alpha
color channels). Note that PNG files always contain non-associated color
channels; png_set_alpha_mode() with one of the modes causes the decoder to
convert the pixels to an associated form before returning them to your
application.

Since it is not necessary to perform arithmetic on opaque color values so
long as they are not to be resampled and are in the final color space it is
possible to optimize the handling of alpha by storing the opaque pixels in
the PNG format (adjusted for the output color space) while storing partially
opaque pixels in the standard, linear, format.  The accuracy required for
standard alpha composition is relatively low, because the pixels are
isolated, therefore typically the accuracy loss in storing 8-bit linear
values is acceptable.  (This is not true if the alpha channel is used to
simulate transparency over large areas - use 16 bits or the PNG mode in
this case!)  This is the 'OPTIMIZED' mode.  For this mode a pixel is
treated as opaque only if the alpha value is equal to the maximum value.

PNG_ALPHA_STANDARD:  The data libpng produces is encoded in the
standard way assumed by most correctly written graphics software.
The gamma encoding will be removed by libpng and the
linear component values will be pre-multiplied by the
alpha channel.

With this format the final image must be re-encoded to
match the display gamma before the image is displayed.
If your system doesn't do that, yet still seems to
perform arithmetic on the pixels without decoding them,
it is broken - check out the modes below.

With PNG_ALPHA_STANDARD libpng always produces linear
component values, whatever screen_gamma you supply.  The
screen_gamma value is, however, used as a default for
the file gamma if the PNG file has no gamma information.

If you call png_set_gamma() after png_set_alpha_mode() you
will override the linear encoding.  Instead the
pre-multiplied pixel values will be gamma encoded but
the alpha channel will still be linear.  This may
actually match the requirements of some broken software,
but it is unlikely.

While linear 8-bit data is often used it has
insufficient precision for any image with a reasonable
dynamic range.  To avoid problems, and if your software
supports it, use png_set_expand_16() to force all
components to 16 bits.

PNG_ALPHA_OPTIMIZED: This mode is the same as PNG_ALPHA_STANDARD
except that completely opaque pixels are gamma encoded according to
the screen_gamma value.  Pixels with alpha less than 1.0
will still have linear components.

Use this format if you have control over your
compositing software and so don't do other arithmetic
(such as scaling) on the data you get from libpng.  Your
compositing software can simply copy opaque pixels to
the output but still has linear values for the
non-opaque pixels.

In normal compositing, where the alpha channel encodes
partial pixel coverage (as opposed to broad area
translucency), the inaccuracies of the 8-bit
representation of non-opaque pixels are irrelevant.

You can also try this format if your software is broken;
it might look better.

PNG_ALPHA_BROKEN: This is PNG_ALPHA_STANDARD; however, all component
values, including the alpha channel are gamma encoded.  This is
broken because, in practice, no implementation that uses this choice
correctly undoes the encoding before handling alpha composition.  Use this
choice only if other serious errors in the software or hardware you use
mandate it.  In most cases of broken software or hardware the bug in the
final display manifests as a subtle halo around composited parts of the
image.  You may not even perceive this as a halo; the composited part of
the image may simply appear separate from the background, as though it had
been cut out of paper and pasted on afterward.

If you don't have to deal with bugs in software or hardware, or if you can fix
them, there are three recommended ways of using png_set_alpha_mode():

png_set_alpha_mode(png_ptr, PNG_ALPHA_PNG,
       screen_gamma);

You can do color correction on the result (libpng does not currently
support color correction internally).  When you handle the alpha channel
you need to undo the gamma encoding and multiply out the alpha.

png_set_alpha_mode(png_ptr, PNG_ALPHA_STANDARD,
       screen_gamma);
   png_set_expand_16(png_ptr);

If you are using the high level interface, don't call png_set_expand_16();
instead pass PNG_TRANSFORM_EXPAND_16 to the interface.

With this mode you can't do color correction, but you can do arithmetic,
including composition and scaling, on the data without further processing.

png_set_alpha_mode(png_ptr, PNG_ALPHA_OPTIMIZED,
       screen_gamma);

You can avoid the expansion to 16-bit components with this mode, but you
lose the ability to scale the image or perform other linear arithmetic.
All you can do is compose the result onto a matching output.  Since this
mode is libpng-specific you also need to write your own composition
software.

The following are examples of calls to png_set_alpha_mode to achieve the
required overall gamma correction and, where necessary, alpha
premultiplication.

png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_DEFAULT_sRGB);

Choices for the alpha_mode are

PNG_ALPHA_PNG           0 /* according to the PNG standard */
    PNG_ALPHA_STANDARD      1 /* according to Porter/Duff */
    PNG_ALPHA_ASSOCIATED    1 /* as above; this is the normal practice */
    PNG_ALPHA_PREMULTIPLIED 1 /* as above */
    PNG_ALPHA_OPTIMIZED     2 /* 'PNG' for opaque pixels, else 'STANDARD' */
    PNG_ALPHA_BROKEN        3 /* the alpha channel is gamma encoded */

PNG_ALPHA_PNG is the default libpng handling of the alpha channel. It is not
pre-multiplied into the color components. In addition the call states
that the output is for a sRGB system and causes all PNG files without gAMA
chunks to be assumed to be encoded using sRGB.

png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_GAMMA_MAC);

In this case the output is assumed to be something like an sRGB conformant
display preceeded by a power-law lookup table of power 1.45.  This is how
early Mac systems behaved.

png_set_alpha_mode(pp, PNG_ALPHA_STANDARD, PNG_GAMMA_LINEAR);

This is the classic Jim Blinn approach and will work in academic
environments where everything is done by the book.  It has the shortcoming
of assuming that input PNG data with no gamma information is linear - this
is unlikely to be correct unless the PNG files were generated locally.
Most of the time the output precision will be so low as to show
significant banding in dark areas of the image.

png_set_expand_16(pp);
    png_set_alpha_mode(pp, PNG_ALPHA_STANDARD, PNG_DEFAULT_sRGB);

This is a somewhat more realistic Jim Blinn inspired approach.  PNG files
are assumed to have the sRGB encoding if not marked with a gamma value and
the output is always 16 bits per component.  This permits accurate scaling
and processing of the data.  If you know that your input PNG files were
generated locally you might need to replace PNG_DEFAULT_sRGB with the
correct value for your system.

png_set_alpha_mode(pp, PNG_ALPHA_OPTIMIZED, PNG_DEFAULT_sRGB);

If you just need to composite the PNG image onto an existing background
and if you control the code that does this you can use the optimization
setting.  In this case you just copy completely opaque pixels to the
output.  For pixels that are not completely transparent (you just skip
those) you do the composition math using png_composite or png_composite_16
below then encode the resultant 8-bit or 16-bit values to match the output
encoding.

Other cases

If neither the PNG nor the standard linear encoding work for you because
of the software or hardware you use then you have a big problem.  The PNG
case will probably result in halos around the image.  The linear encoding
will probably result in a washed out, too bright, image (it's actually too
contrasty.)  Try the ALPHA_OPTIMIZED mode above - this will probably
substantially reduce the halos.  Alternatively try:

png_set_alpha_mode(pp, PNG_ALPHA_BROKEN, PNG_DEFAULT_sRGB);

This option will also reduce the halos, but there will be slight dark
halos round the opaque parts of the image where the background is light.
In the OPTIMIZED mode the halos will be light halos where the background
is dark.  Take your pick - the halos are unavoidable unless you can get
your hardware/software fixed!  (The OPTIMIZED approach is slightly
faster.)

When the default gamma of PNG files doesn't match the output gamma.
If you have PNG files with no gamma information png_set_alpha_mode allows
you to provide a default gamma, but it also sets the ouput gamma to the
matching value.  If you know your PNG files have a gamma that doesn't
match the output you can take advantage of the fact that
png_set_alpha_mode always sets the output gamma but only sets the PNG
default if it is not already set:

png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_DEFAULT_sRGB);
    png_set_alpha_mode(pp, PNG_ALPHA_PNG, PNG_GAMMA_MAC);

The first call sets both the default and the output gamma values, the
second call overrides the output gamma without changing the default.  This
is easier than achieving the same effect with png_set_gamma.  You must use
PNG_ALPHA_PNG for the first call - internal checking in png_set_alpha will
fire if more than one call to png_set_alpha_mode and png_set_background is
made in the same read operation, however multiple calls with PNG_ALPHA_PNG
are ignored.

If you don't need, or can't handle, the alpha channel you can call
png_set_background() to remove it by compositing against a fixed color.  Don't
call png_set_strip_alpha() to do this - it will leave spurious pixel values in
transparent parts of this image.

png_set_background(png_ptr, &background_color,
       PNG_BACKGROUND_GAMMA_SCREEN, 0, 1);

The background_color is an RGB or grayscale value according to the data format
libpng will produce for you.  Because you don't yet know the format of the PNG
file, if you call png_set_background at this point you must arrange for the
format produced by libpng to always have 8-bit or 16-bit components and then
store the color as an 8-bit or 16-bit color as appropriate.  The color contains
separate gray and RGB component values, so you can let libpng produce gray or
RGB output according to the input format, but low bit depth grayscale images
must always be converted to at least 8-bit format.  (Even though low bit depth
grayscale images can't have an alpha channel they can have a transparent
color!)

You set the transforms you need later, either as flags to the high level
interface or libpng API calls for the low level interface.  For reference the
settings and API calls required are:

8-bit values:
   PNG_TRANSFORM_SCALE_16 | PNG_EXPAND
   png_set_expand(png_ptr); png_set_scale_16(png_ptr);

If you must get exactly the same inaccurate results
   produced by default in versions prior to libpng-1.5.4,
   use PNG_TRANSFORM_STRIP_16 and png_set_strip_16(png_ptr)
   instead.

16-bit values:
   PNG_TRANSFORM_EXPAND_16
   png_set_expand_16(png_ptr);

In either case palette image data will be expanded to RGB.  If you just want
color data you can add PNG_TRANSFORM_GRAY_TO_RGB or png_set_gray_to_rgb(png_ptr)
to the list.

Calling png_set_background before the PNG file header is read will not work
prior to libpng-1.5.4.  Because the failure may result in unexpected warnings or
errors it is therefore much safer to call png_set_background after the head has
been read.  Unfortunately this means that prior to libpng-1.5.4 it cannot be
used with the high level interface.

The high-level read interface

At this point there are two ways to proceed; through the high-level
read interface, or through a sequence of low-level read operations.
You can use the high-level interface if (a) you are willing to read
the entire image into memory, and (b) the input transformations
you want to do are limited to the following set:

PNG_TRANSFORM_IDENTITY      No transformation
    PNG_TRANSFORM_SCALE_16      Strip 16-bit samples to
                                8-bit accurately
    PNG_TRANSFORM_STRIP_16      Chop 16-bit samples to
                                8-bit less accurately
    PNG_TRANSFORM_STRIP_ALPHA   Discard the alpha channel
    PNG_TRANSFORM_PACKING       Expand 1, 2 and 4-bit
                                samples to bytes
    PNG_TRANSFORM_PACKSWAP      Change order of packed
                                pixels to LSB first
    PNG_TRANSFORM_EXPAND        Perform set_expand()
    PNG_TRANSFORM_INVERT_MONO   Invert monochrome images
    PNG_TRANSFORM_SHIFT         Normalize pixels to the
                                sBIT depth
    PNG_TRANSFORM_BGR           Flip RGB to BGR, RGBA
                                to BGRA
    PNG_TRANSFORM_SWAP_ALPHA    Flip RGBA to ARGB or GA
                                to AG
    PNG_TRANSFORM_INVERT_ALPHA  Change alpha from opacity
                                to transparency
    PNG_TRANSFORM_SWAP_ENDIAN   Byte-swap 16-bit samples
    PNG_TRANSFORM_GRAY_TO_RGB   Expand grayscale samples
                                to RGB (or GA to RGBA)
    PNG_TRANSFORM_EXPAND_16     Expand samples to 16 bits

(This excludes setting a background color, doing gamma transformation,
quantizing, and setting filler.)  If this is the case, simply do this:

png_read_png(png_ptr, info_ptr, png_transforms, NULL)

where png_transforms is an integer containing the bitwise OR of some
set of transformation flags.  This call is equivalent to png_read_info(),
followed the set of transformations indicated by the transform mask,
then png_read_image(), and finally png_read_end().

(The final parameter of this call is not yet used.  Someday it might point
to transformation parameters required by some future input transform.)

You must use png_transforms and not call any png_set_transform() functions
when you use png_read_png().

After you have called png_read_png(), you can retrieve the image data
with

row_pointers = png_get_rows(png_ptr, info_ptr);

where row_pointers is an array of pointers to the pixel data for each row:

png_bytep row_pointers[height];

If you know your image size and pixel size ahead of time, you can allocate
row_pointers prior to calling png_read_png() with

if (height > PNG_UINT_32_MAX/(sizeof (png_byte)))
      png_error (png_ptr,
          "Image is too tall to process in memory");

if (width > PNG_UINT_32_MAX/pixel_size)
      png_error (png_ptr,
          "Image is too wide to process in memory");

row_pointers = png_malloc(png_ptr,
       height*(sizeof (png_bytep)));

for (int i=0; i<height, i++)
      row_pointers[i]=NULL;  /* security precaution */

for (int i=0; i<height, i++)
      row_pointers[i]=png_malloc(png_ptr,
          width*pixel_size);

png_set_rows(png_ptr, info_ptr, &row_pointers);

Alternatively you could allocate your image in one big block and define
row_pointers[i] to point into the proper places in your block, but first
be sure that your platform is able to allocate such a large buffer:

/* Guard against integer overflow */
   if (height > PNG_SIZE_MAX/(width*pixel_size)) {
        png_error(png_ptr,"image_data buffer would be too large");
   }

png_bytep buffer=png_malloc(png_ptr,height*width*pixel_size);

for (int i=0; i<height, i++)
      row_pointers[i]=buffer+i*width*pixel_size;

png_set_rows(png_ptr, info_ptr, &row_pointers);

If you use png_set_rows(), the application is responsible for freeing
row_pointers (and row_pointers[i], if they were separately allocated).

If you don't allocate row_pointers ahead of time, png_read_png() will
do it, and it'll be free'ed by libpng when you call png_destroy_*().

The low-level read interface

If you are going the low-level route, you are now ready to read all
the file information up to the actual image data.  You do this with a
call to png_read_info().

png_read_info(png_ptr, info_ptr);

This will process all chunks up to but not including the image data.

This also copies some of the data from the PNG file into the decode structure
for use in later transformations.  Important information copied in is:

1) The PNG file gamma from the gAMA chunk.  This overwrites the default value
provided by an earlier call to png_set_gamma or png_set_alpha_mode.

2) Prior to libpng-1.5.4 the background color from a bKGd chunk.  This
damages the information provided by an earlier call to png_set_background
resulting in unexpected behavior.  Libpng-1.5.4 no longer does this.

3) The number of significant bits in each component value.  Libpng uses this to
optimize gamma handling by reducing the internal lookup table sizes.

4) The transparent color information from a tRNS chunk.  This can be modified by
a later call to png_set_tRNS.

Querying the info structure

Functions are used to get the information from the info_ptr once it
has been read.  Note that these fields may not be completely filled
in until png_read_end() has read the chunk data following the image.

png_get_IHDR(png_ptr, info_ptr, &width, &height,
       &bit_depth, &color_type, &interlace_type,
       &compression_type, &filter_method);

width          - holds the width of the image
                     in pixels (up to 2^31).

height         - holds the height of the image
                     in pixels (up to 2^31).

bit_depth      - holds the bit depth of one of the
                     image channels.  (valid values are
                     1, 2, 4, 8, 16 and depend also on
                     the color_type.  See also
                     significant bits (sBIT) below).

color_type     - describes which color/alpha channels
                         are present.
                     PNG_COLOR_TYPE_GRAY
                        (bit depths 1, 2, 4, 8, 16)
                     PNG_COLOR_TYPE_GRAY_ALPHA
                        (bit depths 8, 16)
                     PNG_COLOR_TYPE_PALETTE
                        (bit depths 1, 2, 4, 8)
                     PNG_COLOR_TYPE_RGB
                        (bit_depths 8, 16)
                     PNG_COLOR_TYPE_RGB_ALPHA
                        (bit_depths 8, 16)

PNG_COLOR_MASK_PALETTE
                     PNG_COLOR_MASK_COLOR
                     PNG_COLOR_MASK_ALPHA

interlace_type - (PNG_INTERLACE_NONE or
                     PNG_INTERLACE_ADAM7)

compression_type - (must be PNG_COMPRESSION_TYPE_BASE
                     for PNG 1.0)

filter_method  - (must be PNG_FILTER_TYPE_BASE
                     for PNG 1.0, and can also be
                     PNG_INTRAPIXEL_DIFFERENCING if
                     the PNG datastream is embedded in
                     a MNG-1.0 datastream)

Any of width, height, color_type, bit_depth,
    interlace_type, compression_type, or filter_method can
    be NULL if you are not interested in their values.

Note that png_get_IHDR() returns 32-bit data into
    the application's width and height variables.
    This is an unsafe situation if these are not png_uint_32
    variables.  In such situations, the
    png_get_image_width() and png_get_image_height()
    functions described below are safer.

width            = png_get_image_width(png_ptr,
                         info_ptr);

height           = png_get_image_height(png_ptr,
                         info_ptr);

bit_depth        = png_get_bit_depth(png_ptr,
                         info_ptr);

color_type       = png_get_color_type(png_ptr,
                         info_ptr);

interlace_type   = png_get_interlace_type(png_ptr,
                         info_ptr);

compression_type = png_get_compression_type(png_ptr,
                         info_ptr);

filter_method    = png_get_filter_type(png_ptr,
                         info_ptr);

channels = png_get_channels(png_ptr, info_ptr);

channels       - number of channels of info for the
                     color type (valid values are 1 (GRAY,
                     PALETTE), 2 (GRAY_ALPHA), 3 (RGB),
                     4 (RGB_ALPHA or RGB + filler byte))

rowbytes = png_get_rowbytes(png_ptr, info_ptr);

rowbytes       - number of bytes needed to hold a row
                     This value, the bit_depth, color_type,
                     and the number of channels can change
                     if you use transforms such as
                     png_set_expand(). See
                     png_read_update_info(), below.

signature = png_get_signature(png_ptr, info_ptr);

signature      - holds the signature read from the
                     file (if any).  The data is kept in
                     the same offset it would be if the
                     whole signature were read (i.e. if an
                     application had already read in 4
                     bytes of signature before starting
                     libpng, the remaining 4 bytes would
                     be in signature[4] through signature[7]
                     (see png_set_sig_bytes())).

These are also important, but their validity depends on whether the chunk
has been read.  The png_get_valid(png_ptr, info_ptr, PNG_INFO_<chunk>) and
png_get_<chunk>(png_ptr, info_ptr, ...) functions return non-zero if the
data has been read, or zero if it is missing.  The parameters to the
png_get_<chunk> are set directly if they are simple data types, or a
pointer into the info_ptr is returned for any complex types.

The colorspace data from gAMA, cHRM, sRGB, iCCP, and sBIT chunks
is simply returned to give the application information about how the
image was encoded.  Libpng itself only does transformations using the file
gamma when combining semitransparent pixels with the background color, and,
since libpng-1.6.0, when converting between 8-bit sRGB and 16-bit linear pixels
within the simplified API.  Libpng also uses the file gamma when converting
RGB to gray, beginning with libpng-1.0.5, if the application calls
png_set_rgb_to_gray()).

png_get_PLTE(png_ptr, info_ptr, &palette,
                     &num_palette);

palette        - the palette for the file
                     (array of png_color)

num_palette    - number of entries in the palette

png_get_gAMA(png_ptr, info_ptr, &file_gamma);
    png_get_gAMA_fixed(png_ptr, info_ptr, &int_file_gamma);

file_gamma     - the gamma at which the file is
                     written (PNG_INFO_gAMA)

int_file_gamma - 100,000 times the gamma at which the
                     file is written

png_get_cHRM(png_ptr, info_ptr,  &white_x, &white_y, &red_x,
                     &red_y, &green_x, &green_y, &blue_x, &blue_y)
    png_get_cHRM_XYZ(png_ptr, info_ptr, &red_X, &red_Y, &red_Z,
                     &green_X, &green_Y, &green_Z, &blue_X, &blue_Y,
                     &blue_Z)
    png_get_cHRM_fixed(png_ptr, info_ptr, &int_white_x,
                     &int_white_y, &int_red_x, &int_red_y,
                     &int_green_x, &int_green_y, &int_blue_x,
                     &int_blue_y)
    png_get_cHRM_XYZ_fixed(png_ptr, info_ptr, &int_red_X, &int_red_Y,
                     &int_red_Z, &int_green_X, &int_green_Y,
                     &int_green_Z, &int_blue_X, &int_blue_Y,
                     &int_blue_Z)

{white,red,green,blue}_{x,y}
                     A color space encoding specified using the
                     chromaticities of the end points and the
                     white point. (PNG_INFO_cHRM)

{red,green,blue}_{X,Y,Z}
                     A color space encoding specified using the
                     encoding end points - the CIE tristimulus
                     specification of the intended color of the red,
                     green and blue channels in the PNG RGB data.
                     The white point is simply the sum of the three
                     end points. (PNG_INFO_cHRM)

png_get_sRGB(png_ptr, info_ptr, &srgb_intent);

png_get_iCCP(png_ptr, info_ptr, &name,
       &compression_type, &profile, &proflen);

name             - The profile name.

compression_type - The compression type; always
                       PNG_COMPRESSION_TYPE_BASE for PNG 1.0.
                       You may give NULL to this argument to
                       ignore it.

profile          - International Color Consortium color
                       profile data. May contain NULs.

proflen          - length of profile data in bytes.

png_get_sBIT(png_ptr, info_ptr, &sig_bit);

sig_bit        - the number of significant bits for
                     (PNG_INFO_sBIT) each of the gray,
                     red, green, and blue channels,
                     whichever are appropriate for the
                     given color type (png_color_16)

png_get_tRNS(png_ptr, info_ptr, &trans_alpha,
                     &num_trans, &trans_color);

trans_alpha    - array of alpha (transparency)
                     entries for palette (PNG_INFO_tRNS)

num_trans      - number of transparent entries
                     (PNG_INFO_tRNS)

trans_color    - graylevel or color sample values of
                     the single transparent color for
                     non-paletted images (PNG_INFO_tRNS)

png_get_eXIf_1(png_ptr, info_ptr, &num_exif, &exif);
                     (PNG_INFO_eXIf)

exif           - Exif profile (array of png_byte)

png_get_hIST(png_ptr, info_ptr, &hist);
                     (PNG_INFO_hIST)

hist           - histogram of palette (array of
                     png_uint_16)

png_get_tIME(png_ptr, info_ptr, &mod_time);

mod_time       - time image was last modified
                    (PNG_VALID_tIME)

png_get_bKGD(png_ptr, info_ptr, &background);

background     - background color (of type
                     png_color_16p) (PNG_VALID_bKGD)
                     valid 16-bit red, green and blue
                     values, regardless of color_type

num_comments   = png_get_text(png_ptr, info_ptr,
                     &text_ptr, &num_text);

num_comments   - number of comments

text_ptr       - array of png_text holding image
                     comments

text_ptr[i].key   - keyword for comment.  Must contain
                         1-79 characters.

text_ptr[i].text  - text comments for current
                         keyword.  Can be empty.

text_ptr[i].text_length - length of text string,
                 after decompression, 0 for iTXt

text_ptr[i].itxt_length - length of itxt string,
                 after decompression, 0 for tEXt/zTXt

text_ptr[i].lang  - language of comment (empty
                         string for unknown).

text_ptr[i].lang_key  - keyword in UTF-8
                         (empty string for unknown).

Note that the itxt_length, lang, and lang_key
    members of the text_ptr structure only exist when the
    library is built with iTXt chunk support.  Prior to
    libpng-1.4.0 the library was built by default without
    iTXt support. Also note that when iTXt is supported,
    they contain NULL pointers when the "compression"
    field contains PNG_TEXT_COMPRESSION_NONE or
    PNG_TEXT_COMPRESSION_zTXt.

num_text       - number of comments (same as
                     num_comments; you can put NULL here
                     to avoid the duplication)

Note while png_set_text() will accept text, language,
    and translated keywords that can be NULL pointers, the
    structure returned by png_get_text will always contain
    regular zero-terminated C strings.  They might be
    empty strings but they will never be NULL pointers.

num_spalettes = png_get_sPLT(png_ptr, info_ptr,
       &palette_ptr);

num_spalettes  - number of sPLT chunks read.

palette_ptr    - array of palette structures holding
                     contents of one or more sPLT chunks
                     read.

png_get_oFFs(png_ptr, info_ptr, &offset_x, &offset_y,
       &unit_type);

offset_x       - positive offset from the left edge
                     of the screen (can be negative)

offset_y       - positive offset from the top edge
                     of the screen (can be negative)

unit_type      - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER

png_get_pHYs(png_ptr, info_ptr, &res_x, &res_y,
       &unit_type);

res_x          - pixels/unit physical resolution in
                     x direction

res_y          - pixels/unit physical resolution in
                     x direction

unit_type      - PNG_RESOLUTION_UNKNOWN,
                     PNG_RESOLUTION_METER

png_get_sCAL(png_ptr, info_ptr, &unit, &width,
       &height)

unit        - physical scale units (an integer)

width       - width of a pixel in physical scale units

height      - height of a pixel in physical scale units
                 (width and height are doubles)

png_get_sCAL_s(png_ptr, info_ptr, &unit, &width,
       &height)

unit        - physical scale units (an integer)

width       - width of a pixel in physical scale units
                  (expressed as a string)

height      - height of a pixel in physical scale units
                 (width and height are strings like "2.54")

num_unknown_chunks = png_get_unknown_chunks(png_ptr,
       info_ptr, &unknowns)

unknowns          - array of png_unknown_chunk
                        structures holding unknown chunks

unknowns[i].name  - name of unknown chunk

unknowns[i].data  - data of unknown chunk

unknowns[i].size  - size of unknown chunk's data

unknowns[i].location - position of chunk in file

The value of "i" corresponds to the order in which the
    chunks were read from the PNG file or inserted with the
    png_set_unknown_chunks() function.

The value of "location" is a bitwise "or" of

PNG_HAVE_IHDR  (0x01)
         PNG_HAVE_PLTE  (0x02)
         PNG_AFTER_IDAT (0x08)

The data from the pHYs chunk can be retrieved in several convenient
forms:

res_x = png_get_x_pixels_per_meter(png_ptr,
       info_ptr)

res_y = png_get_y_pixels_per_meter(png_ptr,
       info_ptr)

res_x_and_y = png_get_pixels_per_meter(png_ptr,
       info_ptr)

res_x = png_get_x_pixels_per_inch(png_ptr,
       info_ptr)

res_y = png_get_y_pixels_per_inch(png_ptr,
       info_ptr)

res_x_and_y = png_get_pixels_per_inch(png_ptr,
       info_ptr)

aspect_ratio = png_get_pixel_aspect_ratio(png_ptr,
       info_ptr)

Each of these returns 0 [signifying "unknown"] if
       the data is not present or if res_x is 0;
       res_x_and_y is 0 if res_x != res_y

Note that because of the way the resolutions are
       stored internally, the inch conversions won't
       come out to exactly even number.  For example,
       72 dpi is stored as 0.28346 pixels/meter, and
       when this is retrieved it is 71.9988 dpi, so
       be sure to round the returned value appropriately
       if you want to display a reasonable-looking result.

The data from the oFFs chunk can be retrieved in several convenient
forms:

x_offset = png_get_x_offset_microns(png_ptr, info_ptr);

y_offset = png_get_y_offset_microns(png_ptr, info_ptr);

x_offset = png_get_x_offset_inches(png_ptr, info_ptr);

y_offset = png_get_y_offset_inches(png_ptr, info_ptr);

Each of these returns 0 [signifying "unknown" if both
       x and y are 0] if the data is not present or if the
       chunk is present but the unit is the pixel.  The
       remark about inexact inch conversions applies here
       as well, because a value in inches can't always be
       converted to microns and back without some loss
       of precision.

For more information, see the
PNG specification for chunk contents.  Be careful with trusting
rowbytes, as some of the transformations could increase the space
needed to hold a row (expand, filler, gray_to_rgb, etc.).
See png_read_update_info(), below.

A quick word about text_ptr and num_text.  PNG stores comments in
keyword/text pairs, one pair per chunk, with no limit on the number
of text chunks, and a 2^31 byte limit on their size.  While there are
suggested keywords, there is no requirement to restrict the use to these
strings.  It is strongly suggested that keywords and text be sensible
to humans (that's the point), so don't use abbreviations.  Non-printing
symbols are not allowed.  See the PNG specification for more details.
There is also no requirement to have text after the keyword.

Keywords should be limited to 79 Latin-1 characters without leading or
trailing spaces, but non-consecutive spaces are allowed within the
keyword.  It is possible to have the same keyword any number of times.
The text_ptr is an array of png_text structures, each holding a
pointer to a language string, a pointer to a keyword and a pointer to
a text string.  The text string, language code, and translated
keyword may be empty or NULL pointers.  The keyword/text
pairs are put into the array in the order that they are received.
However, some or all of the text chunks may be after the image, so, to
make sure you have read all the text chunks, don't mess with these
until after you read the stuff after the image.  This will be
mentioned again below in the discussion that goes with png_read_end().

Input transformations

After you've read the header information, you can set up the library
to handle any special transformations of the image data.  The various
ways to transform the data will be described in the order that they
should occur.  This is important, as some of these change the color
type and/or bit depth of the data, and some others only work on
certain color types and bit depths.

Transformations you request are ignored if they don't have any meaning for a
particular input data format.  However some transformations can have an effect
as a result of a previous transformation.  If you specify a contradictory set of
transformations, for example both adding and removing the alpha channel, you
cannot predict the final result.

The color used for the transparency values should be supplied in the same
format/depth as the current image data.  It is stored in the same format/depth
as the image data in a tRNS chunk, so this is what libpng expects for this data.

The color used for the background value depends on the need_expand argument as
described below.

Data will be decoded into the supplied row buffers packed into bytes
unless the library has been told to transform it into another format.
For example, 4 bit/pixel paletted or grayscale data will be returned
2 pixels/byte with the leftmost pixel in the high-order bits of the byte,
unless png_set_packing() is called.  8-bit RGB data will be stored
in RGB RGB RGB format unless png_set_filler() or png_set_add_alpha()
is called to insert filler bytes, either before or after each RGB triplet.

16-bit RGB data will be returned RRGGBB RRGGBB, with the most significant
byte of the color value first, unless png_set_scale_16() is called to
transform it to regular RGB RGB triplets, or png_set_filler() or
png_set_add alpha() is called to insert two filler bytes, either before
or after each RRGGBB triplet.  Similarly, 8-bit or 16-bit grayscale data can
be modified with png_set_filler(), png_set_add_alpha(), png_set_strip_16(),
or png_set_scale_16().

The following code transforms grayscale images of less than 8 to 8 bits,
changes paletted images to RGB, and adds a full alpha channel if there is
transparency information in a tRNS chunk.  This is most useful on
grayscale images with bit depths of 2 or 4 or if there is a multiple-image
viewing application that wishes to treat all images in the same way.

if (color_type == PNG_COLOR_TYPE_PALETTE)
        png_set_palette_to_rgb(png_ptr);

if (png_get_valid(png_ptr, info_ptr,
        PNG_INFO_tRNS)) png_set_tRNS_to_alpha(png_ptr);

if (color_type == PNG_COLOR_TYPE_GRAY &&
        bit_depth < 8) png_set_expand_gray_1_2_4_to_8(png_ptr);

The first two functions are actually aliases for png_set_expand(), added
in libpng version 1.0.4, with the function names expanded to improve code
readability.  In some future version they may actually do different
things.

As of libpng version 1.2.9, png_set_expand_gray_1_2_4_to_8() was
added.  It expands the sample depth without changing tRNS to alpha.

As of libpng version 1.5.2, png_set_expand_16() was added.  It behaves as
png_set_expand(); however, the resultant channels have 16 bits rather than 8.
Use this when the output color or gray channels are made linear to avoid fairly
severe accuracy loss.

if (bit_depth < 16)
      png_set_expand_16(png_ptr);

PNG can have files with 16 bits per channel.  If you only can handle
8 bits per channel, this will strip the pixels down to 8-bit.

if (bit_depth == 16)
#if PNG_LIBPNG_VER >= 10504
       png_set_scale_16(png_ptr);
#else
       png_set_strip_16(png_ptr);
#endif

(The more accurate "png_set_scale_16()" API became available in libpng version
1.5.4).

If you need to process the alpha channel on the image separately from the image
data (for example if you convert it to a bitmap mask) it is possible to have
libpng strip the channel leaving just RGB or gray data:

if (color_type & PNG_COLOR_MASK_ALPHA)
       png_set_strip_alpha(png_ptr);

If you strip the alpha channel you need to find some other way of dealing with
the information.  If, instead, you want to convert the image to an opaque
version with no alpha channel use png_set_background; see below.

As of libpng version 1.5.2, almost all useful expansions are supported, the
major ommissions are conversion of grayscale to indexed images (which can be
done trivially in the application) and conversion of indexed to grayscale (which
can be done by a trivial manipulation of the palette.)

In the following table, the 01 means grayscale with depth<8, 31 means
indexed with depth<8, other numerals represent the color type, "T" means
the tRNS chunk is present, A means an alpha channel is present, and O
means tRNS or alpha is present but all pixels in the image are opaque.

FROM  01  31   0  0T  0O   2  2T  2O   3  3T  3O  4A  4O  6A  6O
   TO
   01    -  [G]  -   -   -   -   -   -   -   -   -   -   -   -   -
   31   [Q]  Q  [Q] [Q] [Q]  Q   Q   Q   Q   Q   Q  [Q] [Q]  Q   Q
    0    1   G   +   .   .   G   G   G   G   G   G   B   B  GB  GB
   0T    lt  Gt  t   +   .   Gt  G   G   Gt  G   G   Bt  Bt GBt GBt
   0O    lt  Gt  t   .   +   Gt  Gt  G   Gt  Gt  G   Bt  Bt GBt GBt
    2    C   P   C   C   C   +   .   .   C   -   -  CB  CB   B   B
   2T    Ct  -   Ct  C   C   t   +   t   -   -   -  CBt CBt  Bt  Bt
   2O    Ct  -   Ct  C   C   t   t   +   -   -   -  CBt CBt  Bt  Bt
    3   [Q]  p  [Q] [Q] [Q]  Q   Q   Q   +   .   .  [Q] [Q]  Q   Q
   3T   [Qt] p  [Qt][Q] [Q]  Qt  Qt  Qt  t   +   t  [Qt][Qt] Qt  Qt
   3O   [Qt] p  [Qt][Q] [Q]  Qt  Qt  Qt  t   t   +  [Qt][Qt] Qt  Qt
   4A    lA  G   A   T   T   GA  GT  GT  GA  GT  GT  +   BA  G  GBA
   4O    lA GBA  A   T   T   GA  GT  GT  GA  GT  GT  BA  +  GBA  G
   6A    CA  PA  CA  C   C   A   T  tT   PA  P   P   C  CBA  +   BA
   6O    CA PBA  CA  C   C   A  tT   T   PA  P   P  CBA  C   BA  +

Within the matrix,
     "+" identifies entries where 'from' and 'to' are the same.
     "-" means the transformation is not supported.
     "." means nothing is necessary (a tRNS chunk can just be ignored).
     "t" means the transformation is obtained by png_set_tRNS.
     "A" means the transformation is obtained by png_set_add_alpha().
     "X" means the transformation is obtained by png_set_expand().
     "1" means the transformation is obtained by
         png_set_expand_gray_1_2_4_to_8() (and by png_set_expand()
         if there is no transparency in the original or the final
         format).
     "C" means the transformation is obtained by png_set_gray_to_rgb().
     "G" means the transformation is obtained by png_set_rgb_to_gray().
     "P" means the transformation is obtained by
         png_set_expand_palette_to_rgb().
     "p" means the transformation is obtained by png_set_packing().
     "Q" means the transformation is obtained by png_set_quantize().
     "T" means the transformation is obtained by
         png_set_tRNS_to_alpha().
     "B" means the transformation is obtained by
         png_set_background(), or png_strip_alpha().

When an entry has multiple transforms listed all are required to cause the
right overall transformation.  When two transforms are separated by a comma
either will do the job.  When transforms are enclosed in [] the transform should
do the job but this is currently unimplemented - a different format will result
if the suggested transformations are used.

In PNG files, the alpha channel in an image
is the level of opacity.  If you need the alpha channel in an image to
be the level of transparency instead of opacity, you can invert the
alpha channel (or the tRNS chunk data) after it's read, so that 0 is
fully opaque and 255 (in 8-bit or paletted images) or 65535 (in 16-bit
images) is fully transparent, with

png_set_invert_alpha(png_ptr);

PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as
they can, resulting in, for example, 8 pixels per byte for 1 bit
files.  This code expands to 1 pixel per byte without changing the
values of the pixels:

if (bit_depth < 8)
       png_set_packing(png_ptr);

PNG files have possible bit depths of 1, 2, 4, 8, and 16.  All pixels
stored in a PNG image have been "scaled" or "shifted" up to the next
higher possible bit depth (e.g. from 5 bits/sample in the range [0,31]
to 8 bits/sample in the range [0, 255]).  However, it is also possible
to convert the PNG pixel data back to the original bit depth of the
image.  This call reduces the pixels back down to the original bit depth:

png_color_8p sig_bit;

if (png_get_sBIT(png_ptr, info_ptr, &sig_bit))
       png_set_shift(png_ptr, sig_bit);

PNG files store 3-color pixels in red, green, blue order.  This code
changes the storage of the pixels to blue, green, red:

if (color_type == PNG_COLOR_TYPE_RGB ||
        color_type == PNG_COLOR_TYPE_RGB_ALPHA)
       png_set_bgr(png_ptr);

PNG files store RGB pixels packed into 3 or 6 bytes. This code expands them
into 4 or 8 bytes for windowing systems that need them in this format:

if (color_type == PNG_COLOR_TYPE_RGB)
       png_set_filler(png_ptr, filler, PNG_FILLER_BEFORE);

where "filler" is the 8-bit or 16-bit number to fill with, and the location
is either PNG_FILLER_BEFORE or PNG_FILLER_AFTER, depending upon whether
you want the filler before the RGB or after. When filling an 8-bit pixel,
the least significant 8 bits of the number are used, if a 16-bit number is
supplied.  This transformation does not affect images that already have full
alpha channels.  To add an opaque alpha channel, use filler=0xffff and
PNG_FILLER_AFTER which will generate RGBA pixels.

Note that png_set_filler() does not change the color type.  If you want
to do that, you can add a true alpha channel with

if (color_type == PNG_COLOR_TYPE_RGB ||
       color_type == PNG_COLOR_TYPE_GRAY)
       png_set_add_alpha(png_ptr, filler, PNG_FILLER_AFTER);

where "filler" contains the alpha value to assign to each pixel.
The png_set_add_alpha() function was added in libpng-1.2.7.

If you are reading an image with an alpha channel, and you need the
data as ARGB instead of the normal PNG format RGBA:

if (color_type == PNG_COLOR_TYPE_RGB_ALPHA)
       png_set_swap_alpha(png_ptr);

For some uses, you may want a grayscale image to be represented as
RGB.  This code will do that conversion:

if (color_type == PNG_COLOR_TYPE_GRAY ||
        color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
       png_set_gray_to_rgb(png_ptr);

Conversely, you can convert an RGB or RGBA image to grayscale or grayscale
with alpha.

if (color_type == PNG_COLOR_TYPE_RGB ||
        color_type == PNG_COLOR_TYPE_RGB_ALPHA)
       png_set_rgb_to_gray(png_ptr, error_action,
          double red_weight, double green_weight);

error_action = 1: silently do the conversion

error_action = 2: issue a warning if the original
                      image has any pixel where
                      red != green or red != blue

error_action = 3: issue an error and abort the
                      conversion if the original
                      image has any pixel where
                      red != green or red != blue

red_weight:       weight of red component

green_weight:     weight of green component
                      If either weight is negative, default
                      weights are used.

In the corresponding fixed point API the red_weight and green_weight values are
simply scaled by 100,000:

png_set_rgb_to_gray(png_ptr, error_action,
       png_fixed_point red_weight,
       png_fixed_point green_weight);

If you have set error_action = 1 or 2, you can
later check whether the image really was gray, after processing
the image rows, with the png_get_rgb_to_gray_status(png_ptr) function.
It will return a png_byte that is zero if the image was gray or
1 if there were any non-gray pixels.  Background and sBIT data
will be silently converted to grayscale, using the green channel
data for sBIT, regardless of the error_action setting.

The default values come from the PNG file cHRM chunk if present; otherwise, the
defaults correspond to the ITU-R recommendation 709, and also the sRGB color
space, as recommended in the Charles Poynton's Colour FAQ,
Copyright (c) 2006-11-28 Charles Poynton, in section 9:

<http://www.poynton.com/notes/colour_and_gamma/ColorFAQ.html#RTFToC9>

Y = 0.2126 * R + 0.7152 * G + 0.0722 * B

Previous versions of this document, 1998 through 2002, recommended a slightly
different formula:

Y = 0.212671 * R + 0.715160 * G + 0.072169 * B

Libpng uses an integer approximation:

Y = (6968 * R + 23434 * G + 2366 * B)/32768

The calculation is done in a linear colorspace, if the image gamma
can be determined.

The png_set_background() function has been described already; it tells libpng to
composite images with alpha or simple transparency against the supplied
background color.  For compatibility with versions of libpng earlier than
libpng-1.5.4 it is recommended that you call the function after reading the file
header, even if you don't want to use the color in a bKGD chunk, if one exists.

If the PNG file contains a bKGD chunk (PNG_INFO_bKGD valid),
you may use this color, or supply another color more suitable for
the current display (e.g., the background color from a web page).  You
need to tell libpng how the color is represented, both the format of the
component values in the color (the number of bits) and the gamma encoding of the
color.  The function takes two arguments, background_gamma_mode and need_expand
to convey this information; however, only two combinations are likely to be
useful:

png_color_16 my_background;
    png_color_16p image_background;

if (png_get_bKGD(png_ptr, info_ptr, &image_background))
       png_set_background(png_ptr, image_background,
           PNG_BACKGROUND_GAMMA_FILE, 1/*needs to be expanded*/, 1);
    else
       png_set_background(png_ptr, &my_background,
           PNG_BACKGROUND_GAMMA_SCREEN, 0/*do not expand*/, 1);

The second call was described above - my_background is in the format of the
final, display, output produced by libpng.  Because you now know the format of
the PNG it is possible to avoid the need to choose either 8-bit or 16-bit
output and to retain palette images (the palette colors will be modified
appropriately and the tRNS chunk removed.)  However, if you are doing this,
take great care not to ask for transformations without checking first that
they apply!

In the first call the background color has the original bit depth and color type
of the PNG file.  So, for palette images the color is supplied as a palette
index and for low bit greyscale images the color is a reduced bit value in
image_background->gray.

If you didn't call png_set_gamma() before reading the file header, for example
if you need your code to remain compatible with older versions of libpng prior
to libpng-1.5.4, this is the place to call it.

Do not call it if you called png_set_alpha_mode(); doing so will damage the
settings put in place by png_set_alpha_mode().  (If png_set_alpha_mode() is
supported then you can certainly do png_set_gamma() before reading the PNG
header.)

This API unconditionally sets the screen and file gamma values, so it will
override the value in the PNG file unless it is called before the PNG file
reading starts.  For this reason you must always call it with the PNG file
value when you call it in this position:

if (png_get_gAMA(png_ptr, info_ptr, &file_gamma))
      png_set_gamma(png_ptr, screen_gamma, file_gamma);

else
      png_set_gamma(png_ptr, screen_gamma, 0.45455);

If you need to reduce an RGB file to a paletted file, or if a paletted
file has more entries than will fit on your screen, png_set_quantize()
will do that.  Note that this is a simple match quantization that merely
finds the closest color available.  This should work fairly well with
optimized palettes, but fairly badly with linear color cubes.  If you
pass a palette that is larger than maximum_colors, the file will
reduce the number of colors in the palette so it will fit into
maximum_colors.  If there is a histogram, libpng will use it to make
more intelligent choices when reducing the palette.  If there is no
histogram, it may not do as good a job.

if (color_type & PNG_COLOR_MASK_COLOR)
   {
      if (png_get_valid(png_ptr, info_ptr,
          PNG_INFO_PLTE))
      {
         png_uint_16p histogram = NULL;

png_get_hIST(png_ptr, info_ptr,
             &histogram);
         png_set_quantize(png_ptr, palette, num_palette,
            max_screen_colors, histogram, 1);
      }

else
      {
         png_color std_color_cube[MAX_SCREEN_COLORS] =
            { ... colors ... };

png_set_quantize(png_ptr, std_color_cube,
            MAX_SCREEN_COLORS, MAX_SCREEN_COLORS,
            NULL,0);
      }
   }

PNG files describe monochrome as black being zero and white being one.
The following code will reverse this (make black be one and white be
zero):

if (bit_depth == 1 && color_type == PNG_COLOR_TYPE_GRAY)
      png_set_invert_mono(png_ptr);

This function can also be used to invert grayscale and gray-alpha images:

if (color_type == PNG_COLOR_TYPE_GRAY ||
       color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
      png_set_invert_mono(png_ptr);

PNG files store 16-bit pixels in network byte order (big-endian,
ie. most significant bits first).  This code changes the storage to the
other way (little-endian, i.e. least significant bits first, the
way PCs store them):

if (bit_depth == 16)
       png_set_swap(png_ptr);

If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you
need to change the order the pixels are packed into bytes, you can use:

if (bit_depth < 8)
       png_set_packswap(png_ptr);

Finally, you can write your own transformation function if none of
the existing ones meets your needs.  This is done by setting a callback
with

png_set_read_user_transform_fn(png_ptr,
        read_transform_fn);

You must supply the function

void read_transform_fn(png_structp png_ptr, png_row_infop
        row_info, png_bytep data)

See pngtest.c for a working example.  Your function will be called
after all of the other transformations have been processed.  Take care with
interlaced images if you do the interlace yourself - the width of the row is the
width in 'row_info', not the overall image width.

If supported, libpng provides two information routines that you can use to find
where you are in processing the image:

png_get_current_pass_number(png_structp png_ptr);
   png_get_current_row_number(png_structp png_ptr);

Don't try using these outside a transform callback - firstly they are only
supported if user transforms are supported, secondly they may well return
unexpected results unless the row is actually being processed at the moment they
are called.

With interlaced
images the value returned is the row in the input sub-image image.  Use
PNG_ROW_FROM_PASS_ROW(row, pass) and PNG_COL_FROM_PASS_COL(col, pass) to
find the output pixel (x,y) given an interlaced sub-image pixel (row,col,pass).

The discussion of interlace handling above contains more information on how to
use these values.

You can also set up a pointer to a user structure for use by your
callback function, and you can inform libpng that your transform
function will change the number of channels or bit depth with the
function

png_set_user_transform_info(png_ptr, user_ptr,
        user_depth, user_channels);

The user's application, not libpng, is responsible for allocating and
freeing any memory required for the user structure.

You can retrieve the pointer via the function
png_get_user_transform_ptr().  For example:

voidp read_user_transform_ptr =
        png_get_user_transform_ptr(png_ptr);

The last thing to handle is interlacing; this is covered in detail below,
but you must call the function here if you want libpng to handle expansion
of the interlaced image.

number_of_passes = png_set_interlace_handling(png_ptr);

After setting the transformations, libpng can update your png_info
structure to reflect any transformations you've requested with this
call.

png_read_update_info(png_ptr, info_ptr);

This is most useful to update the info structure's rowbytes
field so you can use it to allocate your image memory.  This function
will also update your palette with the correct screen_gamma and
background if these have been given with the calls above.  You may
only call png_read_update_info() once with a particular info_ptr.

After you call png_read_update_info(), you can allocate any
memory you need to hold the image.  The row data is simply
raw byte data for all forms of images.  As the actual allocation
varies among applications, no example will be given.  If you
are allocating one large chunk, you will need to build an
array of pointers to each row, as it will be needed for some
of the functions below.

Be sure that your platform can allocate the buffer that you'll need.
libpng internally checks for oversize width, but you'll need to
do your own check for number_of_rows*width*pixel_size if you are using
a multiple-row buffer:

/* Guard against integer overflow */
   if (number_of_rows > PNG_SIZE_MAX/(width*pixel_size)) {
        png_error(png_ptr,"image_data buffer would be too large");
   }

Remember: Before you call png_read_update_info(), the png_get_*()
functions return the values corresponding to the original PNG image.
After you call png_read_update_info the values refer to the image
that libpng will output.  Consequently you must call all the png_set_
functions before you call png_read_update_info().  This is particularly
important for png_set_interlace_handling() - if you are going to call
png_read_update_info() you must call png_set_interlace_handling() before
it unless you want to receive interlaced output.

Reading image data

After you've allocated memory, you can read the image data.
The simplest way to do this is in one function call.  If you are
allocating enough memory to hold the whole image, you can just
call png_read_image() and libpng will read in all the image data
and put it in the memory area supplied.  You will need to pass in
an array of pointers to each row.

This function automatically handles interlacing, so you don't
need to call png_set_interlace_handling() (unless you call
png_read_update_info()) or call this function multiple times, or any
of that other stuff necessary with png_read_rows().

png_read_image(png_ptr, row_pointers);

where row_pointers is:

png_bytep row_pointers[height];

You can point to void or char or whatever you use for pixels.

If you don't want to read in the whole image at once, you can
use png_read_rows() instead.  If there is no interlacing (check
interlace_type == PNG_INTERLACE_NONE), this is simple:

png_read_rows(png_ptr, row_pointers, NULL,
        number_of_rows);

where row_pointers is the same as in the png_read_image() call.

If you are doing this just one row at a time, you can do this with
a single row_pointer instead of an array of row_pointers:

png_bytep row_pointer = row;
    png_read_row(png_ptr, row_pointer, NULL);

If the file is interlaced (interlace_type != 0 in the IHDR chunk), things
get somewhat harder.  The only current (PNG Specification version 1.2)
interlacing type for PNG is (interlace_type == PNG_INTERLACE_ADAM7);
a somewhat complicated 2D interlace scheme, known as Adam7, that
breaks down an image into seven smaller images of varying size, based
on an 8x8 grid.  This number is defined (from libpng 1.5) as
PNG_INTERLACE_ADAM7_PASSES in png.h

libpng can fill out those images or it can give them to you "as is".
It is almost always better to have libpng handle the interlacing for you.
If you want the images filled out, there are two ways to do that.  The one
mentioned in the PNG specification is to expand each pixel to cover
those pixels that have not been read yet (the "rectangle" method).
This results in a blocky image for the first pass, which gradually
smooths out as more pixels are read.  The other method is the "sparkle"
method, where pixels are drawn only in their final locations, with the
rest of the image remaining whatever colors they were initialized to
before the start of the read.  The first method usually looks better,
but tends to be slower, as there are more pixels to put in the rows.

If, as is likely, you want libpng to expand the images, call this before
calling png_start_read_image() or png_read_update_info():

if (interlace_type == PNG_INTERLACE_ADAM7)
       number_of_passes
           = png_set_interlace_handling(png_ptr);

This will return the number of passes needed.  Currently, this is seven,
but may change if another interlace type is added.  This function can be
called even if the file is not interlaced, where it will return one pass.
You then need to read the whole image 'number_of_passes' times.  Each time
will distribute the pixels from the current pass to the correct place in
the output image, so you need to supply the same rows to png_read_rows in
each pass.

If you are not going to display the image after each pass, but are
going to wait until the entire image is read in, use the sparkle
effect.  This effect is faster and the end result of either method
is exactly the same.  If you are planning on displaying the image
after each pass, the "rectangle" effect is generally considered the
better looking one.

If you only want the "sparkle" effect, just call png_read_row() or
png_read_rows() as
normal, with the third parameter NULL.  Make sure you make pass over
the image number_of_passes times, and you don't change the data in the
rows between calls.  You can change the locations of the data, just
not the data.  Each pass only writes the pixels appropriate for that
pass, and assumes the data from previous passes is still valid.

png_read_rows(png_ptr, row_pointers, NULL,
        number_of_rows);
    or
    png_read_row(png_ptr, row_pointers, NULL);

If you only want the first effect (the rectangles), do the same as
before except pass the row buffer in the third parameter, and leave
the second parameter NULL.

png_read_rows(png_ptr, NULL, row_pointers,
        number_of_rows);
    or
    png_read_row(png_ptr, NULL, row_pointers);

If you don't want libpng to handle the interlacing details, just call
png_read_rows() PNG_INTERLACE_ADAM7_PASSES times to read in all the images.
Each of the images is a valid image by itself; however, you will almost
certainly need to distribute the pixels from each sub-image to the
correct place.  This is where everything gets very tricky.

If you want to retrieve the separate images you must pass the correct
number of rows to each successive call of png_read_rows().  The calculation
gets pretty complicated for small images, where some sub-images may
not even exist because either their width or height ends up zero.
libpng provides two macros to help you in 1.5 and later versions:

png_uint_32 width = PNG_PASS_COLS(image_width, pass_number);
   png_uint_32 height = PNG_PASS_ROWS(image_height, pass_number);

Respectively these tell you the width and height of the sub-image
corresponding to the numbered pass.  'pass' is in in the range 0 to 6 -
this can be confusing because the specification refers to the same passes
as 1 to 7!  Be careful, you must check both the width and height before
calling png_read_rows() and not call it for that pass if either is zero.

You can, of course, read each sub-image row by row.  If you want to
produce optimal code to make a pixel-by-pixel transformation of an
interlaced image this is the best approach; read each row of each pass,
transform it, and write it out to a new interlaced image.

If you want to de-interlace the image yourself libpng provides further
macros to help that tell you where to place the pixels in the output image.
Because the interlacing scheme is rectangular - sub-image pixels are always
arranged on a rectangular grid - all you need to know for each pass is the
starting column and row in the output image of the first pixel plus the
spacing between each pixel.  As of libpng 1.5 there are four macros to
retrieve this information:

png_uint_32 x = PNG_PASS_START_COL(pass);
   png_uint_32 y = PNG_PASS_START_ROW(pass);
   png_uint_32 xStep = 1U << PNG_PASS_COL_SHIFT(pass);
   png_uint_32 yStep = 1U << PNG_PASS_ROW_SHIFT(pass);

These allow you to write the obvious loop:

png_uint_32 input_y = 0;
   png_uint_32 output_y = PNG_PASS_START_ROW(pass);

while (output_y < output_image_height)
   {
      png_uint_32 input_x = 0;
      png_uint_32 output_x = PNG_PASS_START_COL(pass);

while (output_x < output_image_width)
      {
         image[output_y][output_x] =
             subimage[pass][input_y][input_x++];

output_x += xStep;
      }

++input_y;
      output_y += yStep;
   }

Notice that the steps between successive output rows and columns are
returned as shifts.  This is possible because the pixels in the subimages
are always a power of 2 apart - 1, 2, 4 or 8 pixels - in the original
image.  In practice you may need to directly calculate the output coordinate
given an input coordinate.  libpng provides two further macros for this
purpose:

png_uint_32 output_x = PNG_COL_FROM_PASS_COL(input_x, pass);
   png_uint_32 output_y = PNG_ROW_FROM_PASS_ROW(input_y, pass);

Finally a pair of macros are provided to tell you if a particular image
row or column appears in a given pass:

int col_in_pass = PNG_COL_IN_INTERLACE_PASS(output_x, pass);
   int row_in_pass = PNG_ROW_IN_INTERLACE_PASS(output_y, pass);

Bear in mind that you will probably also need to check the width and height
of the pass in addition to the above to be sure the pass even exists!

With any luck you are convinced by now that you don't want to do your own
interlace handling.  In reality normally the only good reason for doing this
is if you are processing PNG files on a pixel-by-pixel basis and don't want
to load the whole file into memory when it is interlaced.

libpng includes a test program, pngvalid, that illustrates reading and
writing of interlaced images.  If you can't get interlacing to work in your
code and don't want to leave it to libpng (the recommended approach), see
how pngvalid.c does it.

Finishing a sequential read

After you are finished reading the image through the
low-level interface, you can finish reading the file.

If you want to use a different crc action for handling CRC errors in
chunks after the image data, you can call png_set_crc_action()
again at this point.

If you are interested in comments or time, which may be stored either
before or after the image data, you should pass the separate png_info
struct if you want to keep the comments from before and after the image
separate.

png_infop end_info = png_create_info_struct(png_ptr);

if (!end_info)
    {
       png_destroy_read_struct(&png_ptr, &info_ptr,
           (png_infopp)NULL);
       return (ERROR);
    }

png_read_end(png_ptr, end_info);

If you are not interested, you should still call png_read_end()
but you can pass NULL, avoiding the need to create an end_info structure.
If you do this, libpng will not process any chunks after IDAT other than
skipping over them and perhaps (depending on whether you have called
png_set_crc_action) checking their CRCs while looking for the IEND chunk.

png_read_end(png_ptr, (png_infop)NULL);

If you don't call png_read_end(), then your file pointer will be
left pointing to the first chunk after the last IDAT, which is probably
not what you want if you expect to read something beyond the end of
the PNG datastream.

When you are done, you can free all memory allocated by libpng like this:

png_destroy_read_struct(&png_ptr, &info_ptr,
       &end_info);

or, if you didn't create an end_info structure,

png_destroy_read_struct(&png_ptr, &info_ptr,
       (png_infopp)NULL);

It is also possible to individually free the info_ptr members that
point to libpng-allocated storage with the following function:

png_free_data(png_ptr, info_ptr, mask, seq)

mask - identifies data to be freed, a mask
           containing the bitwise OR of one or
           more of
             PNG_FREE_PLTE, PNG_FREE_TRNS,
             PNG_FREE_HIST, PNG_FREE_ICCP,
             PNG_FREE_PCAL, PNG_FREE_ROWS,
             PNG_FREE_SCAL, PNG_FREE_SPLT,
             PNG_FREE_TEXT, PNG_FREE_UNKN,
           or simply PNG_FREE_ALL

seq  - sequence number of item to be freed
           (-1 for all items)

This function may be safely called when the relevant storage has
already been freed, or has not yet been allocated, or was allocated
by the user and not by libpng,  and will in those cases do nothing.
The "seq" parameter is ignored if only one item of the selected data
type, such as PLTE, is allowed.  If "seq" is not -1, and multiple items
are allowed for the data type identified in the mask, such as text or
sPLT, only the n'th item in the structure is freed, where n is "seq".

The default behavior is only to free data that was allocated internally
by libpng.  This can be changed, so that libpng will not free the data,
or so that it will free data that was allocated by the user with png_malloc()
or png_calloc() and passed in via a png_set_*() function, with

png_data_freer(png_ptr, info_ptr, freer, mask)

freer  - one of
               PNG_DESTROY_WILL_FREE_DATA
               PNG_SET_WILL_FREE_DATA
               PNG_USER_WILL_FREE_DATA

mask   - which data elements are affected
             same choices as in png_free_data()

This function only affects data that has already been allocated.
You can call this function after reading the PNG data but before calling
any png_set_*() functions, to control whether the user or the png_set_*()
function is responsible for freeing any existing data that might be present,
and again after the png_set_*() functions to control whether the user
or png_destroy_*() is supposed to free the data.  When the user assumes
responsibility for libpng-allocated data, the application must use
png_free() to free it, and when the user transfers responsibility to libpng
for data that the user has allocated, the user must have used png_malloc()
or png_calloc() to allocate it.

If you allocated your row_pointers in a single block, as suggested above in
the description of the high level read interface, you must not transfer
responsibility for freeing it to the png_set_rows or png_read_destroy function,
because they would also try to free the individual row_pointers[i].

If you allocated text_ptr.text, text_ptr.lang, and text_ptr.translated_keyword
separately, do not transfer responsibility for freeing text_ptr to libpng,
because when libpng fills a png_text structure it combines these members with
the key member, and png_free_data() will free only text_ptr.key.  Similarly,
if you transfer responsibility for free'ing text_ptr from libpng to your
application, your application must not separately free those members.

The png_free_data() function will turn off the "valid" flag for anything
it frees.  If you need to turn the flag off for a chunk that was freed by
your application instead of by libpng, you can use

png_set_invalid(png_ptr, info_ptr, mask);

mask - identifies the chunks to be made invalid,
           containing the bitwise OR of one or
           more of
             PNG_INFO_gAMA, PNG_INFO_sBIT,
             PNG_INFO_cHRM, PNG_INFO_PLTE,
             PNG_INFO_tRNS, PNG_INFO_bKGD,
             PNG_INFO_eXIf,
             PNG_INFO_hIST, PNG_INFO_pHYs,
             PNG_INFO_oFFs, PNG_INFO_tIME,
             PNG_INFO_pCAL, PNG_INFO_sRGB,
             PNG_INFO_iCCP, PNG_INFO_sPLT,
             PNG_INFO_sCAL, PNG_INFO_IDAT

For a more compact example of reading a PNG image, see the file example.c.

Reading PNG files progressively

The progressive reader is slightly different from the non-progressive
reader.  Instead of calling png_read_info(), png_read_rows(), and
png_read_end(), you make one call to png_process_data(), which calls
callbacks when it has the info, a row, or the end of the image.  You
set up these callbacks with png_set_progressive_read_fn().  You don't
have to worry about the input/output functions of libpng, as you are
giving the library the data directly in png_process_data().  I will
assume that you have read the section on reading PNG files above,
so I will only highlight the differences (although I will show
all of the code).

png_structp png_ptr;
png_infop info_ptr;

/*  An example code fragment of how you would
     initialize the progressive reader in your
     application. */
 int
 initialize_png_reader()
 {
    png_ptr = png_create_read_struct
        (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
         user_error_fn, user_warning_fn);

if (!png_ptr)
        return (ERROR);

info_ptr = png_create_info_struct(png_ptr);

if (!info_ptr)
    {
       png_destroy_read_struct(&png_ptr,
          (png_infopp)NULL, (png_infopp)NULL);
       return (ERROR);
    }

if (setjmp(png_jmpbuf(png_ptr)))
    {
       png_destroy_read_struct(&png_ptr, &info_ptr,
          (png_infopp)NULL);
       return (ERROR);
    }

/* This one's new.  You can provide functions
       to be called when the header info is valid,
       when each row is completed, and when the image
       is finished.  If you aren't using all functions,
       you can specify NULL parameters.  Even when all
       three functions are NULL, you need to call
       png_set_progressive_read_fn().  You can use
       any struct as the user_ptr (cast to a void pointer
       for the function call), and retrieve the pointer
       from inside the callbacks using the function

png_get_progressive_ptr(png_ptr);

which will return a void pointer, which you have
       to cast appropriately.
     */
    png_set_progressive_read_fn(png_ptr, (void *)user_ptr,
        info_callback, row_callback, end_callback);

return 0;
 }

/* A code fragment that you call as you receive blocks
   of data */
 int
 process_data(png_bytep buffer, png_uint_32 length)
 {
    if (setjmp(png_jmpbuf(png_ptr)))
    {
       png_destroy_read_struct(&png_ptr, &info_ptr,
           (png_infopp)NULL);
       return (ERROR);
    }

/* This one's new also.  Simply give it a chunk
       of data from the file stream (in order, of
       course).  On machines with segmented memory
       models machines, don't give it any more than
       64K.  The library seems to run fine with sizes
       of 4K. Although you can give it much less if
       necessary (I assume you can give it chunks of
       1 byte, I haven't tried less than 256 bytes
       yet).  When this function returns, you may
       want to display any rows that were generated
       in the row callback if you don't already do
       so there.
     */
    png_process_data(png_ptr, info_ptr, buffer, length);

/* At this point you can call png_process_data_skip if
       you want to handle data the library will skip yourself;
       it simply returns the number of bytes to skip (and stops
       libpng skipping that number of bytes on the next
       png_process_data call).
    return 0;
 }

/* This function is called (as set by
    png_set_progressive_read_fn() above) when enough data
    has been supplied so all of the header has been
    read.
 */
 void
 info_callback(png_structp png_ptr, png_infop info)
 {
    /* Do any setup here, including setting any of
       the transformations mentioned in the Reading
       PNG files section.  For now, you _must_ call
       either png_start_read_image() or
       png_read_update_info() after all the
       transformations are set (even if you don't set
       any).  You may start getting rows before
       png_process_data() returns, so this is your
       last chance to prepare for that.

This is where you turn on interlace handling,
       assuming you don't want to do it yourself.

If you need to you can stop the processing of
       your original input data at this point by calling
       png_process_data_pause.  This returns the number
       of unprocessed bytes from the last png_process_data
       call - it is up to you to ensure that the next call
       sees these bytes again.  If you don't want to bother
       with this you can get libpng to cache the unread
       bytes by setting the 'save' parameter (see png.h) but
       then libpng will have to copy the data internally.
     */
 }

/* This function is called when each row of image
    data is complete */
 void
 row_callback(png_structp png_ptr, png_bytep new_row,
    png_uint_32 row_num, int pass)
 {
    /* If the image is interlaced, and you turned
       on the interlace handler, this function will
       be called for every row in every pass.  Some
       of these rows will not be changed from the
       previous pass.  When the row is not changed,
       the new_row variable will be NULL.  The rows
       and passes are called in order, so you don't
       really need the row_num and pass, but I'm
       supplying them because it may make your life
       easier.

If you did not turn on interlace handling then
       the callback is called for each row of each
       sub-image when the image is interlaced.  In this
       case 'row_num' is the row in the sub-image, not
       the row in the output image as it is in all other
       cases.

For the non-NULL rows of interlaced images when
       you have switched on libpng interlace handling,
       you must call png_progressive_combine_row()
       passing in the row and the old row.  You can
       call this function for NULL rows (it will just
       return) and for non-interlaced images (it just
       does the memcpy for you) if it will make the
       code easier.  Thus, you can just do this for
       all cases if you switch on interlace handling;
     */

png_progressive_combine_row(png_ptr, old_row,
          new_row);

/* where old_row is what was displayed
       previously for the row.  Note that the first
       pass (pass == 0, really) will completely cover
       the old row, so the rows do not have to be
       initialized.  After the first pass (and only
       for interlaced images), you will have to pass
       the current row, and the function will combine
       the old row and the new row.

You can also call png_process_data_pause in this
       callback - see above.
    */
 }

void
 end_callback(png_structp png_ptr, png_infop info)
 {
    /* This function is called after the whole image
       has been read, including any chunks after the
       image (up to and including the IEND).  You
       will usually have the same info chunk as you
       had in the header, although some data may have
       been added to the comments and time fields.

Most people won't do much here, perhaps setting
       a flag that marks the image as finished.
     */
 }

IV. Writing

Much of this is very similar to reading.  However, everything of
importance is repeated here, so you won't have to constantly look
back up in the reading section to understand writing.

Setup

You will want to do the I/O initialization before you get into libpng,
so if it doesn't work, you don't have anything to undo. If you are not
using the standard I/O functions, you will need to replace them with
custom writing functions.  See the discussion under Customizing libpng.

FILE *fp = fopen(file_name, "wb");

if (!fp)
       return (ERROR);

Next, png_struct and png_info need to be allocated and initialized.
As these can be both relatively large, you may not want to store these
on the stack, unless you have stack space to spare.  Of course, you
will want to check if they return NULL.  If you are also reading,
you won't want to name your read structure and your write structure
both "png_ptr"; you can call them anything you like, such as
"read_ptr" and "write_ptr".  Look at pngtest.c, for example.

png_structp png_ptr = png_create_write_struct
       (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
        user_error_fn, user_warning_fn);

if (!png_ptr)
       return (ERROR);

png_infop info_ptr = png_create_info_struct(png_ptr);
    if (!info_ptr)
    {
       png_destroy_write_struct(&png_ptr,
           (png_infopp)NULL);
       return (ERROR);
    }

If you want to use your own memory allocation routines,
define PNG_USER_MEM_SUPPORTED and use
png_create_write_struct_2() instead of png_create_write_struct():

png_structp png_ptr = png_create_write_struct_2
       (PNG_LIBPNG_VER_STRING, (png_voidp)user_error_ptr,
        user_error_fn, user_warning_fn, (png_voidp)
        user_mem_ptr, user_malloc_fn, user_free_fn);

After you have these structures, you will need to set up the
error handling.  When libpng encounters an error, it expects to
longjmp() back to your routine.  Therefore, you will need to call
setjmp() and pass the png_jmpbuf(png_ptr).  If you
write the file from different routines, you will need to update
the png_jmpbuf(png_ptr) every time you enter a new routine that will
call a png_*() function.  See your documentation of setjmp/longjmp
for your compiler for more information on setjmp/longjmp.  See
the discussion on libpng error handling in the Customizing Libpng
section below for more information on the libpng error handling.

if (setjmp(png_jmpbuf(png_ptr)))
    {
    png_destroy_write_struct(&png_ptr, &info_ptr);
       fclose(fp);
       return (ERROR);
    }
    ...
    return;

If you would rather avoid the complexity of setjmp/longjmp issues,
you can compile libpng with PNG_NO_SETJMP, in which case
errors will result in a call to PNG_ABORT() which defaults to abort().

You can #define PNG_ABORT() to a function that does something
more useful than abort(), as long as your function does not
return.

Checking for invalid palette index on write was added at libpng
1.5.10.  If a pixel contains an invalid (out-of-range) index libpng issues
a benign error.  This is enabled by default because this condition is an
error according to the PNG specification, Clause 11.3.2, but the error can
be ignored in each png_ptr with

png_set_check_for_invalid_index(png_ptr, 0);

If the error is ignored, or if png_benign_error() treats it as a warning,
any invalid pixels are written as-is by the encoder, resulting in an
invalid PNG datastream as output.  In this case the application is
responsible for ensuring that the pixel indexes are in range when it writes
a PLTE chunk with fewer entries than the bit depth would allow.

Now you need to set up the output code.  The default for libpng is to
use the C function fwrite().  If you use this, you will need to pass a
valid FILE * in the function png_init_io().  Be sure that the file is
opened in binary mode.  Again, if you wish to handle writing data in
another way, see the discussion on libpng I/O handling in the Customizing
Libpng section below.

png_init_io(png_ptr, fp);

If you are embedding your PNG into a datastream such as MNG, and don't
want libpng to write the 8-byte signature, or if you have already
written the signature in your application, use

png_set_sig_bytes(png_ptr, 8);

to inform libpng that it should not write a signature.

Write callbacks

At this point, you can set up a callback function that will be
called after each row has been written, which you can use to control
a progress meter or the like.  It's demonstrated in pngtest.c.
You must supply a function

void write_row_callback(png_structp png_ptr, png_uint_32 row,
       int pass);
    {
      /* put your code here */
    }

(You can give it another name that you like instead of "write_row_callback")

To inform libpng about your function, use

png_set_write_status_fn(png_ptr, write_row_callback);

When this function is called the row has already been completely processed and
it has also been written out.  The 'row' and 'pass' refer to the next row to be
handled.  For the
non-interlaced case the row that was just handled is simply one less than the
passed in row number, and pass will always be 0.  For the interlaced case the
same applies unless the row value is 0, in which case the row just handled was
the last one from one of the preceding passes.  Because interlacing may skip a
pass you cannot be sure that the preceding pass is just 'pass-1', if you really
need to know what the last pass is record (row,pass) from the callback and use
the last recorded value each time.

As with the user transform you can find the output row using the
PNG_ROW_FROM_PASS_ROW macro.

You now have the option of modifying how the compression library will
run.  The following functions are mainly for testing, but may be useful
in some cases, like if you need to write PNG files extremely fast and
are willing to give up some compression, or if you want to get the
maximum possible compression at the expense of slower writing.  If you
have no special needs in this area, let the library do what it wants by
not calling this function at all, as it has been tuned to deliver a good
speed/compression ratio. The second parameter to png_set_filter() is
the filter method, for which the only valid values are 0 (as of the
July 1999 PNG specification, version 1.2) or 64 (if you are writing
a PNG datastream that is to be embedded in a MNG datastream).  The third
parameter is a flag that indicates which filter type(s) are to be tested
for each scanline.  See the PNG specification for details on the specific
filter types.

/* turn on or off filtering, and/or choose
       specific filters.  You can use either a single
       PNG_FILTER_VALUE_NAME or the bitwise OR of one
       or more PNG_FILTER_NAME masks.
     */
    png_set_filter(png_ptr, 0,
       PNG_FILTER_NONE  | PNG_FILTER_VALUE_NONE |
       PNG_FILTER_SUB   | PNG_FILTER_VALUE_SUB  |
       PNG_FILTER_UP    | PNG_FILTER_VALUE_UP   |
       PNG_FILTER_AVG   | PNG_FILTER_VALUE_AVG  |
       PNG_FILTER_PAETH | PNG_FILTER_VALUE_PAETH|
       PNG_ALL_FILTERS  | PNG_FAST_FILTERS);

If an application wants to start and stop using particular filters during
compression, it should start out with all of the filters (to ensure that
the previous row of pixels will be stored in case it's needed later),
and then add and remove them after the start of compression.

If you are writing a PNG datastream that is to be embedded in a MNG
datastream, the second parameter can be either 0 or 64.

The png_set_compression_*() functions interface to the zlib compression
library, and should mostly be ignored unless you really know what you are
doing.  The only generally useful call is png_set_compression_level()
which changes how much time zlib spends on trying to compress the image
data.  See the Compression Library (zlib.h and algorithm.txt, distributed
with zlib) for details on the compression levels.

#include zlib.h

/* Set the zlib compression level */
    png_set_compression_level(png_ptr,
        Z_BEST_COMPRESSION);

/* Set other zlib parameters for compressing IDAT */
    png_set_compression_mem_level(png_ptr, 8);
    png_set_compression_strategy(png_ptr,
        Z_DEFAULT_STRATEGY);
    png_set_compression_window_bits(png_ptr, 15);
    png_set_compression_method(png_ptr, 8);
    png_set_compression_buffer_size(png_ptr, 8192)

/* Set zlib parameters for text compression
     * If you don't call these, the parameters
     * fall back on those defined for IDAT chunks
     */
    png_set_text_compression_mem_level(png_ptr, 8);
    png_set_text_compression_strategy(png_ptr,
        Z_DEFAULT_STRATEGY);
    png_set_text_compression_window_bits(png_ptr, 15);
    png_set_text_compression_method(png_ptr, 8);

Setting the contents of info for output

You now need to fill in the png_info structure with all the data you
wish to write before the actual image.  Note that the only thing you
are allowed to write after the image is the text chunks and the time
chunk (as of PNG Specification 1.2, anyway).  See png_write_end() and
the latest PNG specification for more information on that.  If you
wish to write them before the image, fill them in now, and flag that
data as being valid.  If you want to wait until after the data, don't
fill them until png_write_end().  For all the fields in png_info and
their data types, see png.h.  For explanations of what the fields
contain, see the PNG specification.

Some of the more important parts of the png_info are:

png_set_IHDR(png_ptr, info_ptr, width, height,
       bit_depth, color_type, interlace_type,
       compression_type, filter_method)

width          - holds the width of the image
                     in pixels (up to 2^31).

height         - holds the height of the image
                     in pixels (up to 2^31).

bit_depth      - holds the bit depth of one of the
                     image channels.
                     (valid values are 1, 2, 4, 8, 16
                     and depend also on the
                     color_type.  See also significant
                     bits (sBIT) below).

color_type     - describes which color/alpha
                     channels are present.
                     PNG_COLOR_TYPE_GRAY
                        (bit depths 1, 2, 4, 8, 16)
                     PNG_COLOR_TYPE_GRAY_ALPHA
                        (bit depths 8, 16)
                     PNG_COLOR_TYPE_PALETTE
                        (bit depths 1, 2, 4, 8)
                     PNG_COLOR_TYPE_RGB
                        (bit_depths 8, 16)
                     PNG_COLOR_TYPE_RGB_ALPHA
                        (bit_depths 8, 16)

PNG_COLOR_MASK_PALETTE
                     PNG_COLOR_MASK_COLOR
                     PNG_COLOR_MASK_ALPHA

interlace_type - PNG_INTERLACE_NONE or
                     PNG_INTERLACE_ADAM7

compression_type - (must be
                     PNG_COMPRESSION_TYPE_DEFAULT)

filter_method  - (must be PNG_FILTER_TYPE_DEFAULT
                     or, if you are writing a PNG to
                     be embedded in a MNG datastream,
                     can also be
                     PNG_INTRAPIXEL_DIFFERENCING)

If you call png_set_IHDR(), the call must appear before any of the
other png_set_*() functions, because they might require access to some of
the IHDR settings.  The remaining png_set_*() functions can be called
in any order.

If you wish, you can reset the compression_type, interlace_type, or
filter_method later by calling png_set_IHDR() again; if you do this, the
width, height, bit_depth, and color_type must be the same in each call.

png_set_PLTE(png_ptr, info_ptr, palette,
       num_palette);

palette        - the palette for the file
                     (array of png_color)
    num_palette    - number of entries in the palette

png_set_gAMA(png_ptr, info_ptr, file_gamma);
    png_set_gAMA_fixed(png_ptr, info_ptr, int_file_gamma);

file_gamma     - the gamma at which the image was
                     created (PNG_INFO_gAMA)

int_file_gamma - 100,000 times the gamma at which
                     the image was created

png_set_cHRM(png_ptr, info_ptr,  white_x, white_y, red_x, red_y,
                     green_x, green_y, blue_x, blue_y)
    png_set_cHRM_XYZ(png_ptr, info_ptr, red_X, red_Y, red_Z, green_X,
                     green_Y, green_Z, blue_X, blue_Y, blue_Z)
    png_set_cHRM_fixed(png_ptr, info_ptr, int_white_x, int_white_y,
                     int_red_x, int_red_y, int_green_x, int_green_y,
                     int_blue_x, int_blue_y)
    png_set_cHRM_XYZ_fixed(png_ptr, info_ptr, int_red_X, int_red_Y,
                     int_red_Z, int_green_X, int_green_Y, int_green_Z,
                     int_blue_X, int_blue_Y, int_blue_Z)

{white,red,green,blue}_{x,y}
                     A color space encoding specified using the chromaticities
                     of the end points and the white point.

{red,green,blue}_{X,Y,Z}
                     A color space encoding specified using the encoding end
                     points - the CIE tristimulus specification of the intended
                     color of the red, green and blue channels in the PNG RGB
                     data.  The white point is simply the sum of the three end
                     points.

png_set_sRGB(png_ptr, info_ptr, srgb_intent);

srgb_intent    - the rendering intent
                     (PNG_INFO_sRGB) The presence of
                     the sRGB chunk means that the pixel
                     data is in the sRGB color space.
                     This chunk also implies specific
                     values of gAMA and cHRM.  Rendering
                     intent is the CSS-1 property that
                     has been defined by the International
                     Color Consortium
                     (http://www.color.org).
                     It can be one of
                     PNG_sRGB_INTENT_SATURATION,
                     PNG_sRGB_INTENT_PERCEPTUAL,
                     PNG_sRGB_INTENT_ABSOLUTE, or
                     PNG_sRGB_INTENT_RELATIVE.

png_set_sRGB_gAMA_and_cHRM(png_ptr, info_ptr,
       srgb_intent);

srgb_intent    - the rendering intent
                     (PNG_INFO_sRGB) The presence of the
                     sRGB chunk means that the pixel
                     data is in the sRGB color space.
                     This function also causes gAMA and
                     cHRM chunks with the specific values
                     that are consistent with sRGB to be
                     written.

png_set_iCCP(png_ptr, info_ptr, name, compression_type,
                       profile, proflen);

name             - The profile name.

profile          - International Color Consortium color
                       profile data. May contain NULs.

proflen          - length of profile data in bytes.

png_set_sBIT(png_ptr, info_ptr, sig_bit);

sig_bit        - the number of significant bits for
                     (PNG_INFO_sBIT) each of the gray, red,
                     green, and blue channels, whichever are
                     appropriate for the given color type
                     (png_color_16)

png_set_tRNS(png_ptr, info_ptr, trans_alpha,
       num_trans, trans_color);

trans_alpha    - array of alpha (transparency)
                     entries for palette (PNG_INFO_tRNS)

num_trans      - number of transparent entries
                     (PNG_INFO_tRNS)

trans_color    - graylevel or color sample values
                     (in order red, green, blue) of the
                     single transparent color for
                     non-paletted images (PNG_INFO_tRNS)

png_set_eXIf_1(png_ptr, info_ptr, num_exif, exif);

exif           - Exif profile (array of
                     png_byte) (PNG_INFO_eXIf)

png_set_hIST(png_ptr, info_ptr, hist);

hist           - histogram of palette (array of
                     png_uint_16) (PNG_INFO_hIST)

png_set_tIME(png_ptr, info_ptr, mod_time);

mod_time       - time image was last modified
                     (PNG_VALID_tIME)

png_set_bKGD(png_ptr, info_ptr, background);

background     - background color (of type
                     png_color_16p) (PNG_VALID_bKGD)

png_set_text(png_ptr, info_ptr, text_ptr, num_text);

text_ptr       - array of png_text holding image
                     comments

text_ptr[i].compression - type of compression used
                 on "text" PNG_TEXT_COMPRESSION_NONE
                           PNG_TEXT_COMPRESSION_zTXt
                           PNG_ITXT_COMPRESSION_NONE
                           PNG_ITXT_COMPRESSION_zTXt
    text_ptr[i].key   - keyword for comment.  Must contain
                 1-79 characters.
    text_ptr[i].text  - text comments for current
                         keyword.  Can be NULL or empty.
    text_ptr[i].text_length - length of text string,
                 after decompression, 0 for iTXt
    text_ptr[i].itxt_length - length of itxt string,
                 after decompression, 0 for tEXt/zTXt
    text_ptr[i].lang  - language of comment (NULL or
                         empty for unknown).
    text_ptr[i].translated_keyword  - keyword in UTF-8 (NULL
                         or empty for unknown).

num_text       - number of comments

png_set_sPLT(png_ptr, info_ptr, &palette_ptr,
       num_spalettes);

palette_ptr    - array of png_sPLT_struct structures
                     to be added to the list of palettes
                     in the info structure.
    num_spalettes  - number of palette structures to be
                     added.

png_set_oFFs(png_ptr, info_ptr, offset_x, offset_y,
        unit_type);

offset_x  - positive offset from the left
                     edge of the screen

offset_y  - positive offset from the top
                     edge of the screen

unit_type - PNG_OFFSET_PIXEL, PNG_OFFSET_MICROMETER

png_set_pHYs(png_ptr, info_ptr, res_x, res_y,
        unit_type);

res_x       - pixels/unit physical resolution
                  in x direction

res_y       - pixels/unit physical resolution
                  in y direction

unit_type   - PNG_RESOLUTION_UNKNOWN,
                  PNG_RESOLUTION_METER

png_set_sCAL(png_ptr, info_ptr, unit, width, height)

unit        - physical scale units (an integer)

width       - width of a pixel in physical scale units

height      - height of a pixel in physical scale units
                  (width and height are doubles)

png_set_sCAL_s(png_ptr, info_ptr, unit, width, height)

unit        - physical scale units (an integer)

width       - width of a pixel in physical scale units
                  expressed as a string

height      - height of a pixel in physical scale units
                 (width and height are strings like "2.54")

png_set_unknown_chunks(png_ptr, info_ptr, &unknowns,
       num_unknowns)

unknowns          - array of png_unknown_chunk
                        structures holding unknown chunks
    unknowns[i].name  - name of unknown chunk
    unknowns[i].data  - data of unknown chunk
    unknowns[i].size  - size of unknown chunk's data
    unknowns[i].location - position to write chunk in file
                           0: do not write chunk
                           PNG_HAVE_IHDR: before PLTE
                           PNG_HAVE_PLTE: before IDAT
                           PNG_AFTER_IDAT: after IDAT

The "location" member is set automatically according to
what part of the output file has already been written.
You can change its value after calling png_set_unknown_chunks()
as demonstrated in pngtest.c.  Within each of the "locations",
the chunks are sequenced according to their position in the
structure (that is, the value of "i", which is the order in which
the chunk was either read from the input file or defined with
png_set_unknown_chunks).

A quick word about text and num_text.  text is an array of png_text
structures.  num_text is the number of valid structures in the array.
Each png_text structure holds a language code, a keyword, a text value,
and a compression type.

The compression types have the same valid numbers as the compression
types of the image data.  Currently, the only valid number is zero.
However, you can store text either compressed or uncompressed, unlike
images, which always have to be compressed.  So if you don't want the
text compressed, set the compression type to PNG_TEXT_COMPRESSION_NONE.
Because tEXt and zTXt chunks don't have a language field, if you
specify PNG_TEXT_COMPRESSION_NONE or PNG_TEXT_COMPRESSION_zTXt
any language code or translated keyword will not be written out.

Until text gets around a few hundred bytes, it is not worth compressing it.
After the text has been written out to the file, the compression type
is set to PNG_TEXT_COMPRESSION_NONE_WR or PNG_TEXT_COMPRESSION_zTXt_WR,
so that it isn't written out again at the end (in case you are calling
png_write_end() with the same struct).

The keywords that are given in the PNG Specification are:

Title            Short (one line) title or
                     caption for image

Author           Name of image's creator

Description      Description of image (possibly long)

Creation Time    Time of original image creation
                     (usually RFC 1123 format, see below)

Software         Software used to create the image

Disclaimer       Legal disclaimer

Warning          Warning of nature of content

Source           Device used to create the image

Comment          Miscellaneous comment; conversion
                     from other image format

The keyword-text pairs work like this.  Keywords should be short
simple descriptions of what the comment is about.  Some typical
keywords are found in the PNG specification, as is some recommendations
on keywords.  You can repeat keywords in a file.  You can even write
some text before the image and some after.  For example, you may want
to put a description of the image before the image, but leave the
disclaimer until after, so viewers working over modem connections
don't have to wait for the disclaimer to go over the modem before
they start seeing the image.  Finally, keywords should be full
words, not abbreviations.  Keywords and text are in the ISO 8859-1
(Latin-1) character set (a superset of regular ASCII) and can not
contain NUL characters, and should not contain control or other
unprintable characters.  To make the comments widely readable, stick
with basic ASCII, and avoid machine specific character set extensions
like the IBM-PC character set.  The keyword must be present, but
you can leave off the text string on non-compressed pairs.
Compressed pairs must have a text string, as only the text string
is compressed anyway, so the compression would be meaningless.

PNG supports modification time via the png_time structure.  Two
conversion routines are provided, png_convert_from_time_t() for
time_t and png_convert_from_struct_tm() for struct tm.  The
time_t routine uses gmtime().  You don't have to use either of
these, but if you wish to fill in the png_time structure directly,
you should provide the time in universal time (GMT) if possible
instead of your local time.  Note that the year number is the full
year (e.g. 1998, rather than 98 - PNG is year 2000 compliant!), and
that months start with 1.

If you want to store the time of the original image creation, you should
use a plain tEXt chunk with the "Creation Time" keyword.  This is
necessary because the "creation time" of a PNG image is somewhat vague,
depending on whether you mean the PNG file, the time the image was
created in a non-PNG format, a still photo from which the image was
scanned, or possibly the subject matter itself.  In order to facilitate
machine-readable dates, it is recommended that the "Creation Time"
tEXt chunk use RFC 1123 format dates (e.g. "22 May 1997 18:07:10 GMT"),
although this isn't a requirement.  Unlike the tIME chunk, the
"Creation Time" tEXt chunk is not expected to be automatically changed
by the software.  To facilitate the use of RFC 1123 dates, a function
png_convert_to_rfc1123_buffer(buffer, png_timep) is provided to
convert from PNG time to an RFC 1123 format string.  The caller must provide
a writeable buffer of at least 29 bytes.

Writing unknown chunks

You can use the png_set_unknown_chunks function to queue up private chunks
for writing.  You give it a chunk name, location, raw data, and a size.  You
also must use png_set_keep_unknown_chunks() to ensure that libpng will
handle them.  That's all there is to it.  The chunks will be written by the
next following png_write_info_before_PLTE, png_write_info, or png_write_end
function, depending upon the specified location.  Any chunks previously
read into the info structure's unknown-chunk list will also be written out
in a sequence that satisfies the PNG specification's ordering rules.

Here is an example of writing two private chunks, prVt and miNE:

#ifdef PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED
    /* Set unknown chunk data */
    png_unknown_chunk unk_chunk[2];
    strcpy((char *) unk_chunk[0].name, "prVt";
    unk_chunk[0].data = (unsigned char *) "PRIVATE DATA";
    unk_chunk[0].size = strlen(unk_chunk[0].data)+1;
    unk_chunk[0].location = PNG_HAVE_IHDR;
    strcpy((char *) unk_chunk[1].name, "miNE";
    unk_chunk[1].data = (unsigned char *) "MY CHUNK DATA";
    unk_chunk[1].size = strlen(unk_chunk[0].data)+1;
    unk_chunk[1].location = PNG_AFTER_IDAT;
    png_set_unknown_chunks(write_ptr, write_info_ptr,
        unk_chunk, 2);
    /* Needed because miNE is not safe-to-copy */
    png_set_keep_unknown_chunks(png, PNG_HANDLE_CHUNK_ALWAYS,
       (png_bytep) "miNE", 1);
    # if PNG_LIBPNG_VER < 10600
      /* Deal with unknown chunk location bug in 1.5.x and earlier */
      png_set_unknown_chunk_location(png, info, 0, PNG_HAVE_IHDR);
      png_set_unknown_chunk_location(png, info, 1, PNG_AFTER_IDAT);
    # endif
    # if PNG_LIBPNG_VER < 10500
      /* PNG_AFTER_IDAT writes two copies of the chunk prior to libpng-1.5.0,
       * one before IDAT and another after IDAT, so don't use it; only use
       * PNG_HAVE_IHDR location.  This call resets the location previously
       * set by assignment and png_set_unknown_chunk_location() for chunk 1.
       */
      png_set_unknown_chunk_location(png, info, 1, PNG_HAVE_IHDR);
    # endif
    #endif

The high-level write interface

At this point there are two ways to proceed; through the high-level
write interface, or through a sequence of low-level write operations.
You can use the high-level interface if your image data is present
in the info structure.  All defined output
transformations are permitted, enabled by the following masks.

PNG_TRANSFORM_IDENTITY      No transformation
    PNG_TRANSFORM_PACKING       Pack 1, 2 and 4-bit samples
    PNG_TRANSFORM_PACKSWAP      Change order of packed
                                pixels to LSB first
    PNG_TRANSFORM_INVERT_MONO   Invert monochrome images
    PNG_TRANSFORM_SHIFT         Normalize pixels to the
                                sBIT depth
    PNG_TRANSFORM_BGR           Flip RGB to BGR, RGBA
                                to BGRA
    PNG_TRANSFORM_SWAP_ALPHA    Flip RGBA to ARGB or GA
                                to AG
    PNG_TRANSFORM_INVERT_ALPHA  Change alpha from opacity
                                to transparency
    PNG_TRANSFORM_SWAP_ENDIAN   Byte-swap 16-bit samples
    PNG_TRANSFORM_STRIP_FILLER        Strip out filler
                                      bytes (deprecated).
    PNG_TRANSFORM_STRIP_FILLER_BEFORE Strip out leading
                                      filler bytes
    PNG_TRANSFORM_STRIP_FILLER_AFTER  Strip out trailing
                                      filler bytes

If you have valid image data in the info structure (you can use
png_set_rows() to put image data in the info structure), simply do this:

png_write_png(png_ptr, info_ptr, png_transforms, NULL)

where png_transforms is an integer containing the bitwise OR of some set of
transformation flags.  This call is equivalent to png_write_info(),
followed the set of transformations indicated by the transform mask,
then png_write_image(), and finally png_write_end().

(The final parameter of this call is not yet used.  Someday it might point
to transformation parameters required by some future output transform.)

You must use png_transforms and not call any png_set_transform() functions
when you use png_write_png().

The low-level write interface

If you are going the low-level route instead, you are now ready to
write all the file information up to the actual image data.  You do
this with a call to png_write_info().

png_write_info(png_ptr, info_ptr);

Note that there is one transformation you may need to do before
png_write_info().  In PNG files, the alpha channel in an image is the
level of opacity.  If your data is supplied as a level of transparency,
you can invert the alpha channel before you write it, so that 0 is
fully transparent and 255 (in 8-bit or paletted images) or 65535
(in 16-bit images) is fully opaque, with

png_set_invert_alpha(png_ptr);

This must appear before png_write_info() instead of later with the
other transformations because in the case of paletted images the tRNS
chunk data has to be inverted before the tRNS chunk is written.  If
your image is not a paletted image, the tRNS data (which in such cases
represents a single color to be rendered as transparent) won't need to
be changed, and you can safely do this transformation after your
png_write_info() call.

If you need to write a private chunk that you want to appear before
the PLTE chunk when PLTE is present, you can write the PNG info in
two steps, and insert code to write your own chunk between them:

png_write_info_before_PLTE(png_ptr, info_ptr);
    png_set_unknown_chunks(png_ptr, info_ptr, ...);
    png_write_info(png_ptr, info_ptr);

After you've written the file information, you can set up the library
to handle any special transformations of the image data.  The various
ways to transform the data will be described in the order that they
should occur.  This is important, as some of these change the color
type and/or bit depth of the data, and some others only work on
certain color types and bit depths.  Even though each transformation
checks to see if it has data that it can do something with, you should
make sure to only enable a transformation if it will be valid for the
data.  For example, don't swap red and blue on grayscale data.

PNG files store RGB pixels packed into 3 or 6 bytes.  This code tells
the library to strip input data that has 4 or 8 bytes per pixel down
to 3 or 6 bytes (or strip 2 or 4-byte grayscale+filler data to 1 or 2
bytes per pixel).

png_set_filler(png_ptr, 0, PNG_FILLER_BEFORE);

where the 0 is unused, and the location is either PNG_FILLER_BEFORE or
PNG_FILLER_AFTER, depending upon whether the filler byte in the pixel
is stored XRGB or RGBX.

PNG files pack pixels of bit depths 1, 2, and 4 into bytes as small as
they can, resulting in, for example, 8 pixels per byte for 1 bit files.
If the data is supplied at 1 pixel per byte, use this code, which will
correctly pack the pixels into a single byte:

png_set_packing(png_ptr);

PNG files reduce possible bit depths to 1, 2, 4, 8, and 16.  If your
data is of another bit depth, you can write an sBIT chunk into the
file so that decoders can recover the original data if desired.

/* Set the true bit depth of the image data */
    if (color_type & PNG_COLOR_MASK_COLOR)
    {
       sig_bit.red = true_bit_depth;
       sig_bit.green = true_bit_depth;
       sig_bit.blue = true_bit_depth;
    }

else
    {
       sig_bit.gray = true_bit_depth;
    }

if (color_type & PNG_COLOR_MASK_ALPHA)
    {
       sig_bit.alpha = true_bit_depth;
    }

png_set_sBIT(png_ptr, info_ptr, &sig_bit);

If the data is stored in the row buffer in a bit depth other than
one supported by PNG (e.g. 3 bit data in the range 0-7 for a 4-bit PNG),
this will scale the values to appear to be the correct bit depth as
is required by PNG.

png_set_shift(png_ptr, &sig_bit);

PNG files store 16-bit pixels in network byte order (big-endian,
ie. most significant bits first).  This code would be used if they are
supplied the other way (little-endian, i.e. least significant bits
first, the way PCs store them):

if (bit_depth > 8)
       png_set_swap(png_ptr);

If you are using packed-pixel images (1, 2, or 4 bits/pixel), and you
need to change the order the pixels are packed into bytes, you can use:

if (bit_depth < 8)
       png_set_packswap(png_ptr);

PNG files store 3 color pixels in red, green, blue order.  This code
would be used if they are supplied as blue, green, red:

png_set_bgr(png_ptr);

PNG files describe monochrome as black being zero and white being
one. This code would be used if the pixels are supplied with this reversed
(black being one and white being zero):

png_set_invert_mono(png_ptr);

Finally, you can write your own transformation function if none of
the existing ones meets your needs.  This is done by setting a callback
with

png_set_write_user_transform_fn(png_ptr,
       write_transform_fn);

You must supply the function

void write_transform_fn(png_structp png_ptr, png_row_infop
       row_info, png_bytep data)

See pngtest.c for a working example.  Your function will be called
before any of the other transformations are processed.  If supported
libpng also supplies an information routine that may be called from
your callback:

png_get_current_row_number(png_ptr);
   png_get_current_pass_number(png_ptr);

This returns the current row passed to the transform.  With interlaced
images the value returned is the row in the input sub-image image.  Use
PNG_ROW_FROM_PASS_ROW(row, pass) and PNG_COL_FROM_PASS_COL(col, pass) to
find the output pixel (x,y) given an interlaced sub-image pixel (row,col,pass).

The discussion of interlace handling above contains more information on how to
use these values.

You can also set up a pointer to a user structure for use by your
callback function.

png_set_user_transform_info(png_ptr, user_ptr, 0, 0);

The user_channels and user_depth parameters of this function are ignored
when writing; you can set them to zero as shown.

You can retrieve the pointer via the function png_get_user_transform_ptr().
For example:

voidp write_user_transform_ptr =
       png_get_user_transform_ptr(png_ptr);

It is possible to have libpng flush any pending output, either manually,
or automatically after a certain number of lines have been written.  To
flush the output stream a single time call:

png_write_flush(png_ptr);

and to have libpng flush the output stream periodically after a certain
number of scanlines have been written, call:

png_set_flush(png_ptr, nrows);

Note that the distance between rows is from the last time png_write_flush()
was called, or the first row of the image if it has never been called.
So if you write 50 lines, and then png_set_flush 25, it will flush the
output on the next scanline, and every 25 lines thereafter, unless
png_write_flush() is called before 25 more lines have been written.
If nrows is too small (less than about 10 lines for a 640 pixel wide
RGB image) the image compression may decrease noticeably (although this
may be acceptable for real-time applications).  Infrequent flushing will
only degrade the compression performance by a few percent over images
that do not use flushing.

Writing the image data

That's it for the transformations.  Now you can write the image data.
The simplest way to do this is in one function call.  If you have the
whole image in memory, you can just call png_write_image() and libpng
will write the image.  You will need to pass in an array of pointers to
each row.  This function automatically handles interlacing, so you don't
need to call png_set_interlace_handling() or call this function multiple
times, or any of that other stuff necessary with png_write_rows().

png_write_image(png_ptr, row_pointers);

where row_pointers is:

png_byte *row_pointers[height];

You can point to void or char or whatever you use for pixels.

If you don't want to write the whole image at once, you can
use png_write_rows() instead.  If the file is not interlaced,
this is simple:

png_write_rows(png_ptr, row_pointers,
       number_of_rows);

row_pointers is the same as in the png_write_image() call.

If you are just writing one row at a time, you can do this with
a single row_pointer instead of an array of row_pointers:

png_bytep row_pointer = row;

png_write_row(png_ptr, row_pointer);

When the file is interlaced, things can get a good deal more complicated.
The only currently (as of the PNG Specification version 1.2, dated July
1999) defined interlacing scheme for PNG files is the "Adam7" interlace
scheme, that breaks down an image into seven smaller images of varying
size.  libpng will build these images for you, or you can do them
yourself.  If you want to build them yourself, see the PNG specification
for details of which pixels to write when.

If you don't want libpng to handle the interlacing details, just
use png_set_interlace_handling() and call png_write_rows() the
correct number of times to write all the sub-images
(png_set_interlace_handling() returns the number of sub-images.)

If you want libpng to build the sub-images, call this before you start
writing any rows:

number_of_passes = png_set_interlace_handling(png_ptr);

This will return the number of passes needed.  Currently, this is seven,
but may change if another interlace type is added.

Then write the complete image number_of_passes times.

png_write_rows(png_ptr, row_pointers, number_of_rows);

Think carefully before you write an interlaced image.  Typically code that
reads such images reads all the image data into memory, uncompressed, before
doing any processing.  Only code that can display an image on the fly can
take advantage of the interlacing and even then the image has to be exactly
the correct size for the output device, because scaling an image requires
adjacent pixels and these are not available until all the passes have been
read.

If you do write an interlaced image you will hardly ever need to handle
the interlacing yourself.  Call png_set_interlace_handling() and use the
approach described above.

The only time it is conceivable that you will really need to write an
interlaced image pass-by-pass is when you have read one pass by pass and
made some pixel-by-pixel transformation to it, as described in the read
code above.  In this case use the PNG_PASS_ROWS and PNG_PASS_COLS macros
to determine the size of each sub-image in turn and simply write the rows
you obtained from the read code.

Finishing a sequential write

After you are finished writing the image, you should finish writing
the file.  If you are interested in writing comments or time, you should
pass an appropriately filled png_info pointer.  If you are not interested,
you can pass NULL.

png_write_end(png_ptr, info_ptr);

When you are done, you can free all memory used by libpng like this:

png_destroy_write_struct(&png_ptr, &info_ptr);

It is also possible to individually free the info_ptr members that
point to libpng-allocated storage with the following function:

png_free_data(png_ptr, info_ptr, mask, seq)

mask  - identifies data to be freed, a mask
            containing the bitwise OR of one or
            more of
              PNG_FREE_PLTE, PNG_FREE_TRNS,
              PNG_FREE_HIST, PNG_FREE_ICCP,
              PNG_FREE_PCAL, PNG_FREE_ROWS,
              PNG_FREE_SCAL, PNG_FREE_SPLT,
              PNG_FREE_TEXT, PNG_FREE_UNKN,
            or simply PNG_FREE_ALL

seq   - sequence number of item to be freed
            (-1 for all items)

This function may be safely called when the relevant storage has
already been freed, or has not yet been allocated, or was allocated
by the user  and not by libpng,  and will in those cases do nothing.
The "seq" parameter is ignored if only one item of the selected data
type, such as PLTE, is allowed.  If "seq" is not -1, and multiple items
are allowed for the data type identified in the mask, such as text or
sPLT, only the n'th item in the structure is freed, where n is "seq".

If you allocated data such as a palette that you passed in to libpng
with png_set_*, you must not free it until just before the call to
png_destroy_write_struct().

png_data_freer(png_ptr, info_ptr, freer, mask)

freer  - one of
               PNG_DESTROY_WILL_FREE_DATA
               PNG_SET_WILL_FREE_DATA
               PNG_USER_WILL_FREE_DATA

mask   - which data elements are affected
             same choices as in png_free_data()

For example, to transfer responsibility for some data from a read structure
to a write structure, you could use

png_data_freer(read_ptr, read_info_ptr,
       PNG_USER_WILL_FREE_DATA,
       PNG_FREE_PLTE|PNG_FREE_tRNS|PNG_FREE_hIST)

png_data_freer(write_ptr, write_info_ptr,
       PNG_DESTROY_WILL_FREE_DATA,
       PNG_FREE_PLTE|PNG_FREE_tRNS|PNG_FREE_hIST)

thereby briefly reassigning responsibility for freeing to the user but
immediately afterwards reassigning it once more to the write_destroy
function.  Having done this, it would then be safe to destroy the read
structure and continue to use the PLTE, tRNS, and hIST data in the write
structure.

This function only affects data that has already been allocated.
You can call this function before calling after the png_set_*() functions
to control whether the user or png_destroy_*() is supposed to free the data.
When the user assumes responsibility for libpng-allocated data, the
application must use
png_free() to free it, and when the user transfers responsibility to libpng
for data that the user has allocated, the user must have used png_malloc()
or png_calloc() to allocate it.

V. Simplified API

The simplified API, which became available in libpng-1.6.0, hides the details
of both libpng and the PNG file format itself.
It allows PNG files to be read into a very limited number of
in-memory bitmap formats or to be written from the same formats.  If these
formats do not accommodate your needs then you can, and should, use the more
sophisticated APIs above - these support a wide variety of in-memory formats
and a wide variety of sophisticated transformations to those formats as well
as a wide variety of APIs to manipulate ancilliary information.

To read a PNG file using the simplified API:

1) Declare a 'png_image' structure (see below) on the stack, set the
     version field to PNG_IMAGE_VERSION and the 'opaque' pointer to NULL
     (this is REQUIRED, your program may crash if you don't do it.)

2) Call the appropriate png_image_begin_read... function.

3) Set the png_image 'format' member to the required sample format.

4) Allocate a buffer for the image and, if required, the color-map.

5) Call png_image_finish_read to read the image and, if required, the
     color-map into your buffers.

There are no restrictions on the format of the PNG input itself; all valid
color types, bit depths, and interlace methods are acceptable, and the
input image is transformed as necessary to the requested in-memory format
during the png_image_finish_read() step.  The only caveat is that if you
request a color-mapped image from a PNG that is full-color or makes
complex use of an alpha channel the transformation is extremely lossy and the
result may look terrible.

To write a PNG file using the simplified API:

1) Declare a 'png_image' structure on the stack and memset()
     it to all zero.

2) Initialize the members of the structure that describe the
     image, setting the 'format' member to the format of the
     image samples.

3) Call the appropriate png_image_write... function with a
     pointer to the image and, if necessary, the color-map to write
     the PNG data.

png_image is a structure that describes the in-memory format of an image
when it is being read or defines the in-memory format of an image that you
need to write.  The "png_image" structure contains the following members:

png_controlp opaque  Initialize to NULL, free with png_image_free
   png_uint_32  version Set to PNG_IMAGE_VERSION
   png_uint_32  width   Image width in pixels (columns)
   png_uint_32  height  Image height in pixels (rows)
   png_uint_32  format  Image format as defined below
   png_uint_32  flags   A bit mask containing informational flags
   png_uint_32  colormap_entries; Number of entries in the color-map
   png_uint_32  warning_or_error;
   char         message[64];

In the event of an error or warning the "warning_or_error"
field will be set to a non-zero value and the 'message' field will contain
a '\0' terminated string with the libpng error or warning message.  If both
warnings and an error were encountered, only the error is recorded.  If there
are multiple warnings, only the first one is recorded.

The upper 30 bits of the "warning_or_error" value are reserved; the low two
bits contain a two bit code such that a value more than 1 indicates a failure
in the API just called:

0 - no warning or error
   1 - warning
   2 - error
   3 - error preceded by warning

The pixels (samples) of the image have one to four channels whose components
have original values in the range 0 to 1.0:

1: A single gray or luminance channel (G).
  2: A gray/luminance channel and an alpha channel (GA).
  3: Three red, green, blue color channels (RGB).
  4: Three color channels and an alpha channel (RGBA).

The channels are encoded in one of two ways:

a) As a small integer, value 0..255, contained in a single byte.  For the
alpha channel the original value is simply value/255.  For the color or
luminance channels the value is encoded according to the sRGB specification
and matches the 8-bit format expected by typical display devices.

The color/gray channels are not scaled (pre-multiplied) by the alpha
channel and are suitable for passing to color management software.

b) As a value in the range 0..65535, contained in a 2-byte integer, in
the native byte order of the platform on which the application is running.
All channels can be converted to the original value by dividing by 65535; all
channels are linear.  Color channels use the RGB encoding (RGB end-points) of
the sRGB specification.  This encoding is identified by the
PNG_FORMAT_FLAG_LINEAR flag below.

When the simplified API needs to convert between sRGB and linear colorspaces,
the actual sRGB transfer curve defined in the sRGB specification (see the
article at https://en.wikipedia.org/wiki/SRGB) is used, not the gamma=1/2.2
approximation used elsewhere in libpng.

When an alpha channel is present it is expected to denote pixel coverage
of the color or luminance channels and is returned as an associated alpha
channel: the color/gray channels are scaled (pre-multiplied) by the alpha
value.

The samples are either contained directly in the image data, between 1 and 8
bytes per pixel according to the encoding, or are held in a color-map indexed
by bytes in the image data.  In the case of a color-map the color-map entries
are individual samples, encoded as above, and the image data has one byte per
pixel to select the relevant sample from the color-map.

PNG_FORMAT_*

The #defines to be used in png_image::format.  Each #define identifies a
particular layout of channel data and, if present, alpha values.  There are
separate defines for each of the two component encodings.

A format is built up using single bit flag values.  All combinations are
valid.  Formats can be built up from the flag values or you can use one of
the predefined values below.  When testing formats always use the FORMAT_FLAG
macros to test for individual features - future versions of the library may
add new flags.

When reading or writing color-mapped images the format should be set to the
format of the entries in the color-map then png_image_{read,write}_colormap
called to read or write the color-map and set the format correctly for the
image data.  Do not set the PNG_FORMAT_FLAG_COLORMAP bit directly!

NOTE: libpng can be built with particular features disabled. If you see
compiler errors because the definition of one of the following flags has been
compiled out it is because libpng does not have the required support.  It is
possible, however, for the libpng configuration to enable the format on just
read or just write; in that case you may see an error at run time.
You can guard against this by checking for the definition of the
appropriate "_SUPPORTED" macro, one of:

PNG_SIMPLIFIED_{READ,WRITE}_{BGR,AFIRST}_SUPPORTED

PNG_FORMAT_FLAG_ALPHA    format with an alpha channel
   PNG_FORMAT_FLAG_COLOR    color format: otherwise grayscale
   PNG_FORMAT_FLAG_LINEAR   2-byte channels else 1-byte
   PNG_FORMAT_FLAG_COLORMAP image data is color-mapped
   PNG_FORMAT_FLAG_BGR      BGR colors, else order is RGB
   PNG_FORMAT_FLAG_AFIRST   alpha channel comes first

Supported formats are as follows.  Future versions of libpng may support more
formats; for compatibility with older versions simply check if the format
macro is defined using #ifdef.  These defines describe the in-memory layout
of the components of the pixels of the image.

First the single byte (sRGB) formats:

PNG_FORMAT_GRAY
   PNG_FORMAT_GA
   PNG_FORMAT_AG
   PNG_FORMAT_RGB
   PNG_FORMAT_BGR
   PNG_FORMAT_RGBA
   PNG_FORMAT_ARGB
   PNG_FORMAT_BGRA
   PNG_FORMAT_ABGR

Then the linear 2-byte formats.  When naming these "Y" is used to
indicate a luminance (gray) channel.  The component order within the pixel
is always the same - there is no provision for swapping the order of the
components in the linear format.  The components are 16-bit integers in
the native byte order for your platform, and there is no provision for
swapping the bytes to a different endian condition.

PNG_FORMAT_LINEAR_Y
   PNG_FORMAT_LINEAR_Y_ALPHA
   PNG_FORMAT_LINEAR_RGB
   PNG_FORMAT_LINEAR_RGB_ALPHA

With color-mapped formats the image data is one byte for each pixel. The byte
is an index into the color-map which is formatted as above.  To obtain a
color-mapped format it is sufficient just to add the PNG_FOMAT_FLAG_COLORMAP
to one of the above definitions, or you can use one of the definitions below.

PNG_FORMAT_RGB_COLORMAP
   PNG_FORMAT_BGR_COLORMAP
   PNG_FORMAT_RGBA_COLORMAP
   PNG_FORMAT_ARGB_COLORMAP
   PNG_FORMAT_BGRA_COLORMAP
   PNG_FORMAT_ABGR_COLORMAP

PNG_IMAGE macros

These are convenience macros to derive information from a png_image
structure.  The PNG_IMAGE_SAMPLE_ macros return values appropriate to the
actual image sample values - either the entries in the color-map or the
pixels in the image.  The PNG_IMAGE_PIXEL_ macros return corresponding values
for the pixels and will always return 1 for color-mapped formats.  The
remaining macros return information about the rows in the image and the
complete image.

NOTE: All the macros that take a png_image::format parameter are compile time
constants if the format parameter is, itself, a constant.  Therefore these
macros can be used in array declarations and case labels where required.
Similarly the macros are also pre-processor constants (sizeof is not used) so
they can be used in #if tests.

PNG_IMAGE_SAMPLE_CHANNELS(fmt)
    Returns the total number of channels in a given format: 1..4

PNG_IMAGE_SAMPLE_COMPONENT_SIZE(fmt)
    Returns the size in bytes of a single component of a pixel or color-map
    entry (as appropriate) in the image: 1 or 2.

PNG_IMAGE_SAMPLE_SIZE(fmt)
    This is the size of the sample data for one sample.  If the image is
    color-mapped it is the size of one color-map entry (and image pixels are
    one byte in size), otherwise it is the size of one image pixel.

PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(fmt)
    The maximum size of the color-map required by the format expressed in a
    count of components.  This can be used to compile-time allocate a
    color-map:

png_uint_16 colormap[PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(linear_fmt)];

png_byte colormap[PNG_IMAGE_MAXIMUM_COLORMAP_COMPONENTS(sRGB_fmt)];

Alternatively use the PNG_IMAGE_COLORMAP_SIZE macro below to use the
    information from one of the png_image_begin_read_ APIs and dynamically
    allocate the required memory.

PNG_IMAGE_COLORMAP_SIZE(fmt)
   The size of the color-map required by the format; this is the size of the
   color-map buffer passed to the png_image_{read,write}_colormap APIs. It is
   a fixed number determined by the format so can easily be allocated on the
   stack if necessary.

Corresponding information about the pixels

PNG_IMAGE_PIXEL_CHANNELS(fmt)
   The number of separate channels (components) in a pixel; 1 for a
   color-mapped image.

PNG_IMAGE_PIXEL_COMPONENT_SIZE(fmt)\
   The size, in bytes, of each component in a pixel; 1 for a color-mapped
   image.

PNG_IMAGE_PIXEL_SIZE(fmt)
   The size, in bytes, of a complete pixel; 1 for a color-mapped image.

Information about the whole row, or whole image

PNG_IMAGE_ROW_STRIDE(image)
   Returns the total number of components in a single row of the image; this
   is the minimum 'row stride', the minimum count of components between each
   row.  For a color-mapped image this is the minimum number of bytes in a
   row.

If you need the stride measured in bytes, row_stride_bytes is
   PNG_IMAGE_ROW_STRIDE(image) * PNG_IMAGE_PIXEL_COMPONENT_SIZE(fmt)
   plus any padding bytes that your application might need, for example
   to start the next row on a 4-byte boundary.

PNG_IMAGE_BUFFER_SIZE(image, row_stride)
   Return the size, in bytes, of an image buffer given a png_image and a row
   stride - the number of components to leave space for in each row.

PNG_IMAGE_SIZE(image)
   Return the size, in bytes, of the image in memory given just a png_image;
   the row stride is the minimum stride required for the image.

PNG_IMAGE_COLORMAP_SIZE(image)
   Return the size, in bytes, of the color-map of this image.  If the image
   format is not a color-map format this will return a size sufficient for
   256 entries in the given format; check PNG_FORMAT_FLAG_COLORMAP if
   you don't want to allocate a color-map in this case.

PNG_IMAGE_FLAG_*

Flags containing additional information about the image are held in
the 'flags' field of png_image.

PNG_IMAGE_FLAG_COLORSPACE_NOT_sRGB == 0x01
    This indicates the the RGB values of the in-memory bitmap do not
    correspond to the red, green and blue end-points defined by sRGB.

PNG_IMAGE_FLAG_FAST == 0x02
   On write emphasise speed over compression; the resultant PNG file will be
   larger but will be produced significantly faster, particular for large
   images.  Do not use this option for images which will be distributed, only
   used it when producing intermediate files that will be read back in
   repeatedly.  For a typical 24-bit image the option will double the read
   speed at the cost of increasing the image size by 25%, however for many
   more compressible images the PNG file can be 10 times larger with only a
   slight speed gain.

PNG_IMAGE_FLAG_16BIT_sRGB == 0x04
    On read if the image is a 16-bit per component image and there is no gAMA
    or sRGB chunk assume that the components are sRGB encoded.  Notice that
    images output by the simplified API always have gamma information; setting
    this flag only affects the interpretation of 16-bit images from an
    external source.  It is recommended that the application expose this flag
    to the user; the user can normally easily recognize the difference between
    linear and sRGB encoding.  This flag has no effect on write - the data
    passed to the write APIs must have the correct encoding (as defined
    above.)

If the flag is not set (the default) input 16-bit per component data is
    assumed to be linear.

NOTE: the flag can only be set after the png_image_begin_read_ call,
    because that call initializes the 'flags' field.

READ APIs

The png_image passed to the read APIs must have been initialized by setting
   the png_controlp field 'opaque' to NULL (or, better, memset the whole thing.)

int png_image_begin_read_from_file( png_imagep image,
     const char *file_name)

The named file is opened for read and the image header
     is filled in from the PNG header in the file.

int png_image_begin_read_from_stdio (png_imagep image,
     FILE* file)

The PNG header is read from the stdio FILE object.

int png_image_begin_read_from_memory(png_imagep image,
      png_const_voidp memory, png_size_t size)

The PNG header is read from the given memory buffer.

int png_image_finish_read(png_imagep image,
      png_colorp background, void *buffer,
      png_int_32 row_stride, void *colormap));

Finish reading the image into the supplied buffer and
      clean up the png_image structure.

row_stride is the step, in png_byte or png_uint_16 units
      as appropriate, between adjacent rows.  A positive stride
      indicates that the top-most row is first in the buffer -
      the normal top-down arrangement.  A negative stride
      indicates that the bottom-most row is first in the buffer.

background need only be supplied if an alpha channel must
      be removed from a png_byte format and the removal is to be
      done by compositing on a solid color; otherwise it may be
      NULL and any composition will be done directly onto the
      buffer.  The value is an sRGB color to use for the
      background, for grayscale output the green channel is used.

For linear output removing the alpha channel is always done
      by compositing on black.

void png_image_free(png_imagep image)

Free any data allocated by libpng in image->opaque,
      setting the pointer to NULL.  May be called at any time
      after the structure is initialized.

WRITE APIS

For write you must initialize a png_image structure to describe the image to
be written:

version: must be set to PNG_IMAGE_VERSION
   opaque: must be initialized to NULL
   width: image width in pixels
   height: image height in rows
   format: the format of the data you wish to write
   flags: set to 0 unless one of the defined flags applies; set
      PNG_IMAGE_FLAG_COLORSPACE_NOT_sRGB for color format images
      where the RGB values do not correspond to the colors in sRGB.
   colormap_entries: set to the number of entries in the color-map (0 to 256)

int png_image_write_to_file, (png_imagep image,
      const char *file, int convert_to_8bit, const void *buffer,
      png_int_32 row_stride, const void *colormap));

Write the image to the named file.

int png_image_write_to_memory (png_imagep image, void *memory,
      png_alloc_size_t * PNG_RESTRICT memory_bytes,
      int convert_to_8_bit, const void *buffer, ptrdiff_t row_stride,
      const void *colormap));

Write the image to memory.

int png_image_write_to_stdio(png_imagep image, FILE *file,
      int convert_to_8_bit, const void *buffer,
      png_int_32 row_stride, const void *colormap)

Write the image to the given (FILE*).

With all write APIs if image is in one of the linear formats with
(png_uint_16) data then setting convert_to_8_bit will cause the output to be
a (png_byte) PNG gamma encoded according to the sRGB specification, otherwise
a 16-bit linear encoded PNG file is written.

With all APIs row_stride is handled as in the read APIs - it is the spacing
from one row to the next in component sized units (float) and if negative
indicates a bottom-up row layout in the buffer.  If you pass zero, libpng will
calculate the row_stride for you from the width and number of channels.

Note that the write API does not support interlacing, sub-8-bit pixels,
indexed (paletted) images, or most ancillary chunks.

VI. Modifying/Customizing libpng

There are two issues here.  The first is changing how libpng does
standard things like memory allocation, input/output, and error handling.
The second deals with more complicated things like adding new chunks,
adding new transformations, and generally changing how libpng works.
Both of those are compile-time issues; that is, they are generally
determined at the time the code is written, and there is rarely a need
to provide the user with a means of changing them.

Memory allocation, input/output, and error handling

All of the memory allocation, input/output, and error handling in libpng
goes through callbacks that are user-settable.  The default routines are
in pngmem.c, pngrio.c, pngwio.c, and pngerror.c, respectively.  To change
these functions, call the appropriate png_set_*_fn() function.

Memory allocation is done through the functions png_malloc(), png_calloc(),
and png_free().  The png_malloc() and png_free() functions currently just
call the standard C functions and png_calloc() calls png_malloc() and then
clears the newly allocated memory to zero; note that png_calloc(png_ptr, size)
is not the same as the calloc(number, size) function provided by stdlib.h.
There is limited support for certain systems with segmented memory
architectures and the types of pointers declared by png.h match this; you
will have to use appropriate pointers in your application.  If you prefer
to use a different method of allocating and freeing data, you can use
png_create_read_struct_2() or png_create_write_struct_2() to register your
own functions as described above.  These functions also provide a void
pointer that can be retrieved via

mem_ptr=png_get_mem_ptr(png_ptr);

Your replacement memory functions must have prototypes as follows:

png_voidp malloc_fn(png_structp png_ptr,
       png_alloc_size_t size);

void free_fn(png_structp png_ptr, png_voidp ptr);

Your malloc_fn() must return NULL in case of failure.  The png_malloc()
function will normally call png_error() if it receives a NULL from the
system memory allocator or from your replacement malloc_fn().

Your free_fn() will never be called with a NULL ptr, since libpng's
png_free() checks for NULL before calling free_fn().

Input/Output in libpng is done through png_read() and png_write(),
which currently just call fread() and fwrite().  The FILE * is stored in
png_struct and is initialized via png_init_io().  If you wish to change
the method of I/O, the library supplies callbacks that you can set
through the function png_set_read_fn() and png_set_write_fn() at run
time, instead of calling the png_init_io() function.  These functions
also provide a void pointer that can be retrieved via the function
png_get_io_ptr().  For example:

png_set_read_fn(png_structp read_ptr,
        voidp read_io_ptr, png_rw_ptr read_data_fn)

png_set_write_fn(png_structp write_ptr,
        voidp write_io_ptr, png_rw_ptr write_data_fn,
        png_flush_ptr output_flush_fn);

voidp read_io_ptr = png_get_io_ptr(read_ptr);
    voidp write_io_ptr = png_get_io_ptr(write_ptr);

The replacement I/O functions must have prototypes as follows:

void user_read_data(png_structp png_ptr,
        png_bytep data, png_size_t length);

void user_write_data(png_structp png_ptr,
        png_bytep data, png_size_t length);

void user_flush_data(png_structp png_ptr);

The user_read_data() function is responsible for detecting and
handling end-of-data errors.

Supplying NULL for the read, write, or flush functions sets them back
to using the default C stream functions, which expect the io_ptr to
point to a standard *FILE structure.  It is probably a mistake
to use NULL for one of write_data_fn and output_flush_fn but not both
of them, unless you have built libpng with PNG_NO_WRITE_FLUSH defined.
It is an error to read from a write stream, and vice versa.

Error handling in libpng is done through png_error() and png_warning().
Errors handled through png_error() are fatal, meaning that png_error()
should never return to its caller.  Currently, this is handled via
setjmp() and longjmp() (unless you have compiled libpng with
PNG_NO_SETJMP, in which case it is handled via PNG_ABORT()),
but you could change this to do things like exit() if you should wish,
as long as your function does not return.

On non-fatal errors, png_warning() is called
to print a warning message, and then control returns to the calling code.
By default png_error() and png_warning() print a message on stderr via
fprintf() unless the library is compiled with PNG_NO_CONSOLE_IO defined
(because you don't want the messages) or PNG_NO_STDIO defined (because
fprintf() isn't available).  If you wish to change the behavior of the error
functions, you will need to set up your own message callbacks.  These
functions are normally supplied at the time that the png_struct is created.
It is also possible to redirect errors and warnings to your own replacement
functions after png_create_*_struct() has been called by calling:

png_set_error_fn(png_structp png_ptr,
        png_voidp error_ptr, png_error_ptr error_fn,
        png_error_ptr warning_fn);

If NULL is supplied for either error_fn or warning_fn, then the libpng
default function will be used, calling fprintf() and/or longjmp() if a
problem is encountered.  The replacement error functions should have
parameters as follows:

void user_error_fn(png_structp png_ptr,
        png_const_charp error_msg);

void user_warning_fn(png_structp png_ptr,
        png_const_charp warning_msg);

Then, within your user_error_fn or user_warning_fn, you can retrieve
the error_ptr if you need it, by calling

png_voidp error_ptr = png_get_error_ptr(png_ptr);

The motivation behind using setjmp() and longjmp() is the C++ throw and
catch exception handling methods.  This makes the code much easier to write,
as there is no need to check every return code of every function call.
However, there are some uncertainties about the status of local variables
after a longjmp, so the user may want to be careful about doing anything
after setjmp returns non-zero besides returning itself.  Consult your
compiler documentation for more details.  For an alternative approach, you
may wish to use the "cexcept" facility (see https://cexcept.sourceforge.io/),
which is illustrated in pngvalid.c and in contrib/visupng.

Beginning in libpng-1.4.0, the png_set_benign_errors() API became available.
You can use this to handle certain errors (normally handled as errors)
as warnings.

png_set_benign_errors (png_ptr, int allowed);

allowed: 0: treat png_benign_error() as an error.
             1: treat png_benign_error() as a warning.

As of libpng-1.6.0, the default condition is to treat benign errors as
warnings while reading and as errors while writing.

Custom chunks

If you need to read or write custom chunks, you may need to get deeper
into the libpng code.  The library now has mechanisms for storing
and writing chunks of unknown type; you can even declare callbacks
for custom chunks.  However, this may not be good enough if the
library code itself needs to know about interactions between your
chunk and existing `intrinsic' chunks.

If you need to write a new intrinsic chunk, first read the PNG
specification. Acquire a first level of understanding of how it works.
Pay particular attention to the sections that describe chunk names,
and look at how other chunks were designed, so you can do things
similarly.  Second, check out the sections of libpng that read and
write chunks.  Try to find a chunk that is similar to yours and use
it as a template.  More details can be found in the comments inside
the code.  It is best to handle private or unknown chunks in a generic method,
via callback functions, instead of by modifying libpng functions. This
is illustrated in pngtest.c, which uses a callback function to handle a
private "vpAg" chunk and the new "sTER" chunk, which are both unknown to
libpng.

If you wish to write your own transformation for the data, look through
the part of the code that does the transformations, and check out some of
the simpler ones to get an idea of how they work.  Try to find a similar
transformation to the one you want to add and copy off of it.  More details
can be found in the comments inside the code itself.

Configuring for gui/windowing platforms:

You will need to write new error and warning functions that use the GUI
interface, as described previously, and set them to be the error and
warning functions at the time that png_create_*_struct() is called,
in order to have them available during the structure initialization.
They can be changed later via png_set_error_fn().  On some compilers,
you may also have to change the memory allocators (png_malloc, etc.).

Configuring zlib:

There are special functions to configure the compression.  Perhaps the
most useful one changes the compression level, which currently uses
input compression values in the range 0 - 9.  The library normally
uses the default compression level (Z_DEFAULT_COMPRESSION = 6).  Tests
have shown that for a large majority of images, compression values in
the range 3-6 compress nearly as well as higher levels, and do so much
faster.  For online applications it may be desirable to have maximum speed
(Z_BEST_SPEED = 1).  With versions of zlib after v0.99, you can also
specify no compression (Z_NO_COMPRESSION = 0), but this would create
files larger than just storing the raw bitmap.  You can specify the
compression level by calling:

#include zlib.h
    png_set_compression_level(png_ptr, level);

Another useful one is to reduce the memory level used by the library.
The memory level defaults to 8, but it can be lowered if you are
short on memory (running DOS, for example, where you only have 640K).
Note that the memory level does have an effect on compression; among
other things, lower levels will result in sections of incompressible
data being emitted in smaller stored blocks, with a correspondingly
larger relative overhead of up to 15% in the worst case.

#include zlib.h
    png_set_compression_mem_level(png_ptr, level);

The other functions are for configuring zlib.  They are not recommended
for normal use and may result in writing an invalid PNG file.  See
zlib.h for more information on what these mean.

#include zlib.h
    png_set_compression_strategy(png_ptr,
        strategy);

png_set_compression_window_bits(png_ptr,
        window_bits);

png_set_compression_method(png_ptr, method);

This controls the size of the IDAT chunks (default 8192):

png_set_compression_buffer_size(png_ptr, size);

As of libpng version 1.5.4, additional APIs became
available to set these separately for non-IDAT
compressed chunks such as zTXt, iTXt, and iCCP:

#include zlib.h
    #if PNG_LIBPNG_VER >= 10504
    png_set_text_compression_level(png_ptr, level);

png_set_text_compression_mem_level(png_ptr, level);

png_set_text_compression_strategy(png_ptr,
        strategy);

png_set_text_compression_window_bits(png_ptr,
        window_bits);

png_set_text_compression_method(png_ptr, method);
    #endif

Controlling row filtering

If you want to control whether libpng uses filtering or not, which
filters are used, and how it goes about picking row filters, you
can call one of these functions.  The selection and configuration
of row filters can have a significant impact on the size and
encoding speed and a somewhat lesser impact on the decoding speed
of an image.  Filtering is enabled by default for RGB and grayscale
images (with and without alpha), but not for paletted images nor
for any images with bit depths less than 8 bits/pixel.

The 'method' parameter sets the main filtering method, which is
currently only '0' in the PNG 1.2 specification.  The 'filters'
parameter sets which filter(s), if any, should be used for each
scanline.  Possible values are PNG_ALL_FILTERS, PNG_NO_FILTERS,
or PNG_FAST_FILTERS to turn filtering on and off, or to turn on
just the fast-decoding subset of filters, respectively.

Individual filter types are PNG_FILTER_NONE, PNG_FILTER_SUB,
PNG_FILTER_UP, PNG_FILTER_AVG, PNG_FILTER_PAETH, which can be bitwise
ORed together with '|' to specify one or more filters to use.
These filters are described in more detail in the PNG specification.
If you intend to change the filter type during the course of writing
the image, you should start with flags set for all of the filters
you intend to use so that libpng can initialize its internal
structures appropriately for all of the filter types.  (Note that this
means the first row must always be adaptively filtered, because libpng
currently does not allocate the filter buffers until png_write_row()
is called for the first time.)

filters = PNG_NO_FILTERS;
    filters = PNG_ALL_FILTERS;
    filters = PNG_FAST_FILTERS;

filters = PNG_FILTER_NONE | PNG_FILTER_SUB |
              PNG_FILTER_UP | PNG_FILTER_AVG |
              PNG_FILTER_PAETH;

png_set_filter(png_ptr, PNG_FILTER_TYPE_BASE,
       filters);

The second parameter can also be
              PNG_INTRAPIXEL_DIFFERENCING if you are
              writing a PNG to be embedded in a MNG
              datastream.  This parameter must be the
              same as the value of filter_method used
              in png_set_IHDR().

Requesting debug printout

The macro definition PNG_DEBUG can be used to request debugging
printout.  Set it to an integer value in the range 0 to 3.  Higher
numbers result in increasing amounts of debugging information.  The
information is printed to the "stderr" file, unless another file
name is specified in the PNG_DEBUG_FILE macro definition.

When PNG_DEBUG > 0, the following functions (macros) become available:

png_debug(level, message)
   png_debug1(level, message, p1)
   png_debug2(level, message, p1, p2)

in which "level" is compared to PNG_DEBUG to decide whether to print
the message, "message" is the formatted string to be printed,
and p1 and p2 are parameters that are to be embedded in the string
according to printf-style formatting directives.  For example,

png_debug1(2, "foo=%d", foo);

is expanded to

if (PNG_DEBUG > 2)
      fprintf(PNG_DEBUG_FILE, "foo=%d\n", foo);

When PNG_DEBUG is defined but is zero, the macros aren't defined, but you
can still use PNG_DEBUG to control your own debugging:

#ifdef PNG_DEBUG
       fprintf(stderr, ...
   #endif

When PNG_DEBUG = 1, the macros are defined, but only png_debug statements
having level = 0 will be printed.  There aren't any such statements in
this version of libpng, but if you insert some they will be printed.

VII.  MNG support

The MNG specification (available at http://www.libpng.org/pub/mng) allows
certain extensions to PNG for PNG images that are embedded in MNG datastreams.
Libpng can support some of these extensions.  To enable them, use the
png_permit_mng_features() function:

feature_set = png_permit_mng_features(png_ptr, mask)

mask is a png_uint_32 containing the bitwise OR of the
        features you want to enable.  These include
        PNG_FLAG_MNG_EMPTY_PLTE
        PNG_FLAG_MNG_FILTER_64
        PNG_ALL_MNG_FEATURES

feature_set is a png_uint_32 that is the bitwise AND of
      your mask with the set of MNG features that is
      supported by the version of libpng that you are using.

It is an error to use this function when reading or writing a standalone
PNG file with the PNG 8-byte signature.  The PNG datastream must be wrapped
in a MNG datastream.  As a minimum, it must have the MNG 8-byte signature
and the MHDR and MEND chunks.  Libpng does not provide support for these
or any other MNG chunks; your application must provide its own support for
them.  You may wish to consider using libmng (available at
https://www.libmng.com/) instead.

VIII.  Changes to Libpng from version 0.88

It should be noted that versions of libpng later than 0.96 are not
distributed by the original libpng author, Guy Schalnat, nor by
Andreas Dilger, who had taken over from Guy during 1996 and 1997, and
distributed versions 0.89 through 0.96, but rather by another member
of the original PNG Group, Glenn Randers-Pehrson.  Guy and Andreas are
still alive and well, but they have moved on to other things.

The old libpng functions png_read_init(), png_write_init(),
png_info_init(), png_read_destroy(), and png_write_destroy() have been
moved to PNG_INTERNAL in version 0.95 to discourage their use.  These
functions will be removed from libpng version 1.4.0.

The preferred method of creating and initializing the libpng structures is
via the png_create_read_struct(), png_create_write_struct(), and
png_create_info_struct() because they isolate the size of the structures
from the application, allow version error checking, and also allow the
use of custom error handling routines during the initialization, which
the old functions do not.  The functions png_read_destroy() and
png_write_destroy() do not actually free the memory that libpng
allocated for these structs, but just reset the data structures, so they
can be used instead of png_destroy_read_struct() and
png_destroy_write_struct() if you feel there is too much system overhead
allocating and freeing the png_struct for each image read.

Setting the error callbacks via png_set_message_fn() before
png_read_init() as was suggested in libpng-0.88 is no longer supported
because this caused applications that do not use custom error functions
to fail if the png_ptr was not initialized to zero.  It is still possible
to set the error callbacks AFTER png_read_init(), or to change them with
png_set_error_fn(), which is essentially the same function, but with a new
name to force compilation errors with applications that try to use the old
method.

Support for the sCAL, iCCP, iTXt, and sPLT chunks was added at libpng-1.0.6;
however, iTXt support was not enabled by default.

Starting with version 1.0.7, you can find out which version of the library
you are using at run-time:

png_uint_32 libpng_vn = png_access_version_number();

The number libpng_vn is constructed from the major version, minor
version with leading zero, and release number with leading zero,
(e.g., libpng_vn for version 1.0.7 is 10007).

Note that this function does not take a png_ptr, so you can call it
before you've created one.

You can also check which version of png.h you used when compiling your
application:

png_uint_32 application_vn = PNG_LIBPNG_VER;

IX.  Changes to Libpng from version 1.0.x to 1.2.x

Support for user memory management was enabled by default.  To
accomplish this, the functions png_create_read_struct_2(),
png_create_write_struct_2(), png_set_mem_fn(), png_get_mem_ptr(),
png_malloc_default(), and png_free_default() were added.

Support for the iTXt chunk has been enabled by default as of
version 1.2.41.

Support for certain MNG features was enabled.

Support for numbered error messages was added.  However, we never got
around to actually numbering the error messages.  The function
png_set_strip_error_numbers() was added (Note: the prototype for this
function was inadvertently removed from png.h in PNG_NO_ASSEMBLER_CODE
builds of libpng-1.2.15.  It was restored in libpng-1.2.36).

The png_malloc_warn() function was added at libpng-1.2.3.  This issues
a png_warning and returns NULL instead of aborting when it fails to
acquire the requested memory allocation.

Support for setting user limits on image width and height was enabled
by default.  The functions png_set_user_limits(), png_get_user_width_max(),
and png_get_user_height_max() were added at libpng-1.2.6.

The png_set_add_alpha() function was added at libpng-1.2.7.

The function png_set_expand_gray_1_2_4_to_8() was added at libpng-1.2.9.
Unlike png_set_gray_1_2_4_to_8(), the new function does not expand the
tRNS chunk to alpha. The png_set_gray_1_2_4_to_8() function is
deprecated.

A number of macro definitions in support of runtime selection of
assembler code features (especially Intel MMX code support) were
added at libpng-1.2.0:

PNG_ASM_FLAG_MMX_SUPPORT_COMPILED
    PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU
    PNG_ASM_FLAG_MMX_READ_COMBINE_ROW
    PNG_ASM_FLAG_MMX_READ_INTERLACE
    PNG_ASM_FLAG_MMX_READ_FILTER_SUB
    PNG_ASM_FLAG_MMX_READ_FILTER_UP
    PNG_ASM_FLAG_MMX_READ_FILTER_AVG
    PNG_ASM_FLAG_MMX_READ_FILTER_PAETH
    PNG_ASM_FLAGS_INITIALIZED
    PNG_MMX_READ_FLAGS
    PNG_MMX_FLAGS
    PNG_MMX_WRITE_FLAGS
    PNG_MMX_FLAGS

We added the following functions in support of runtime
selection of assembler code features:

png_get_mmx_flagmask()
    png_set_mmx_thresholds()
    png_get_asm_flags()
    png_get_mmx_bitdepth_threshold()
    png_get_mmx_rowbytes_threshold()
    png_set_asm_flags()

We replaced all of these functions with simple stubs in libpng-1.2.20,
when the Intel assembler code was removed due to a licensing issue.

These macros are deprecated:

PNG_READ_TRANSFORMS_NOT_SUPPORTED
    PNG_PROGRESSIVE_READ_NOT_SUPPORTED
    PNG_NO_SEQUENTIAL_READ_SUPPORTED
    PNG_WRITE_TRANSFORMS_NOT_SUPPORTED
    PNG_READ_ANCILLARY_CHUNKS_NOT_SUPPORTED
    PNG_WRITE_ANCILLARY_CHUNKS_NOT_SUPPORTED

They have been replaced, respectively, by:

PNG_NO_READ_TRANSFORMS
    PNG_NO_PROGRESSIVE_READ
    PNG_NO_SEQUENTIAL_READ
    PNG_NO_WRITE_TRANSFORMS
    PNG_NO_READ_ANCILLARY_CHUNKS
    PNG_NO_WRITE_ANCILLARY_CHUNKS

PNG_MAX_UINT was replaced with PNG_UINT_31_MAX.  It has been
deprecated since libpng-1.0.16 and libpng-1.2.6.

The function
    png_check_sig(sig, num)
was replaced with
    !png_sig_cmp(sig, 0, num)
It has been deprecated since libpng-0.90.

The function
    png_set_gray_1_2_4_to_8()
which also expands tRNS to alpha was replaced with
    png_set_expand_gray_1_2_4_to_8()
which does not. It has been deprecated since libpng-1.0.18 and 1.2.9.

X.  Changes to Libpng from version 1.0.x/1.2.x to 1.4.x

Private libpng prototypes and macro definitions were moved from
png.h and pngconf.h into a new pngpriv.h header file.

Functions png_set_benign_errors(), png_benign_error(), and
png_chunk_benign_error() were added.

Support for setting the maximum amount of memory that the application
will allocate for reading chunks was added, as a security measure.
The functions png_set_chunk_cache_max() and png_get_chunk_cache_max()
were added to the library.

We implemented support for I/O states by adding png_ptr member io_state
and functions png_get_io_chunk_name() and png_get_io_state() in pngget.c

We added PNG_TRANSFORM_GRAY_TO_RGB to the available high-level
input transforms.

Checking for and reporting of errors in the IHDR chunk is more thorough.

Support for global arrays was removed, to improve thread safety.

Some obsolete/deprecated macros and functions have been removed.

Typecasted NULL definitions such as
   #define png_voidp_NULL            (png_voidp)NULL
were eliminated.  If you used these in your application, just use
NULL instead.

The png_struct and info_struct members "trans" and "trans_values" were
changed to "trans_alpha" and "trans_color", respectively.

The obsolete, unused pnggccrd.c and pngvcrd.c files and related makefiles
were removed.

The PNG_1_0_X and PNG_1_2_X macros were eliminated.

The PNG_LEGACY_SUPPORTED macro was eliminated.

Many WIN32_WCE #ifdefs were removed.

The functions png_read_init(info_ptr), png_write_init(info_ptr),
png_info_init(info_ptr), png_read_destroy(), and png_write_destroy()
have been removed.  They have been deprecated since libpng-0.95.

The png_permit_empty_plte() was removed. It has been deprecated
since libpng-1.0.9.  Use png_permit_mng_features() instead.

We removed the obsolete stub functions png_get_mmx_flagmask(),
png_set_mmx_thresholds(), png_get_asm_flags(),
png_get_mmx_bitdepth_threshold(), png_get_mmx_rowbytes_threshold(),
png_set_asm_flags(), and png_mmx_supported()

We removed the obsolete png_check_sig(), png_memcpy_check(), and
png_memset_check() functions.  Instead use !png_sig_cmp(), memcpy(),
and memset(), respectively.

The function png_set_gray_1_2_4_to_8() was removed. It has been
deprecated since libpng-1.0.18 and 1.2.9, when it was replaced with
png_set_expand_gray_1_2_4_to_8() because the former function also
expanded any tRNS chunk to an alpha channel.

Macros for png_get_uint_16, png_get_uint_32, and png_get_int_32
were added and are used by default instead of the corresponding
functions. Unfortunately,
from libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
function) incorrectly returned a value of type png_uint_32.

We changed the prototype for png_malloc() from
    png_malloc(png_structp png_ptr, png_uint_32 size)
to
    png_malloc(png_structp png_ptr, png_alloc_size_t size)

This also applies to the prototype for the user replacement malloc_fn().

The png_calloc() function was added and is used in place of
of "png_malloc(); memset();" except in the case in png_read_png()
where the array consists of pointers; in this case a "for" loop is used
after the png_malloc() to set the pointers to NULL, to give robust.
behavior in case the application runs out of memory part-way through
the process.

We changed the prototypes of png_get_compression_buffer_size() and
png_set_compression_buffer_size() to work with png_size_t instead of
png_uint_32.

Support for numbered error messages was removed by default, since we
never got around to actually numbering the error messages. The function
png_set_strip_error_numbers() was removed from the library by default.

The png_zalloc() and png_zfree() functions are no longer exported.
The png_zalloc() function no longer zeroes out the memory that it
allocates.  Applications that called png_zalloc(png_ptr, number, size)
can call png_calloc(png_ptr, number*size) instead, and can call
png_free() instead of png_zfree().

Support for dithering was disabled by default in libpng-1.4.0, because
it has not been well tested and doesn't actually "dither".
The code was not
removed, however, and could be enabled by building libpng with
PNG_READ_DITHER_SUPPORTED defined.  In libpng-1.4.2, this support
was re-enabled, but the function was renamed png_set_quantize() to
reflect more accurately what it actually does.  At the same time,
the PNG_DITHER_[RED,GREEN_BLUE]_BITS macros were also renamed to
PNG_QUANTIZE_[RED,GREEN,BLUE]_BITS, and PNG_READ_DITHER_SUPPORTED
was renamed to PNG_READ_QUANTIZE_SUPPORTED.

We removed the trailing '.' from the warning and error messages.

XI.  Changes to Libpng from version 1.4.x to 1.5.x

From libpng-1.4.0 until 1.4.4, the png_get_uint_16 macro (but not the
function) incorrectly returned a value of type png_uint_32.
The incorrect macro was removed from libpng-1.4.5.

png_set_check_for_invalid_index(png_ptr, allowed);

allowed  - one of
                 0: disable benign error (accept the
                    invalid data without warning).
                 1: enable benign error (treat the
                    invalid data as an error or a
                    warning).

If the error is ignored, or if png_benign_error() treats it as a warning,
any invalid pixels are decoded as opaque black by the decoder and written
as-is by the encoder.

Retrieving the maximum palette index found was added at libpng-1.5.15.
This statement must appear after png_read_png() or png_read_image() while
reading, and after png_write_png() or png_write_image() while writing.

int max_palette = png_get_palette_max(png_ptr, info_ptr);

This will return the maximum palette index found in the image, or "-1" if
the palette was not checked, or "0" if no palette was found.  Note that this
does not account for any palette index used by ancillary chunks such as the
bKGD chunk; you must check those separately to determine the maximum
palette index actually used.

There are no substantial API changes between the non-deprecated parts of
the 1.4.5 API and the 1.5.0 API; however, the ability to directly access
members of the main libpng control structures, png_struct and png_info,
deprecated in earlier versions of libpng, has been completely removed from
libpng 1.5, and new private "pngstruct.h", "pnginfo.h", and "pngdebug.h"
header files were created.

We no longer include zlib.h in png.h.  The include statement has been moved
to pngstruct.h, where it is not accessible by applications. Applications that
need access to information in zlib.h will need to add the '#include "zlib.h"'
directive.  It does not matter whether this is placed prior to or after
the '"#include png.h"' directive.

The png_sprintf(), png_strcpy(), and png_strncpy() macros are no longer used
and were removed.

We moved the png_strlen(), png_memcpy(), png_memset(), and png_memcmp()
macros into a private header file (pngpriv.h) that is not accessible to
applications.

In png_get_iCCP, the type of "profile" was changed from png_charpp
to png_bytepp, and in png_set_iCCP, from png_charp to png_const_bytep.

There are changes of form in png.h, including new and changed macros to
declare parts of the API.  Some API functions with arguments that are
pointers to data not modified within the function have been corrected to
declare these arguments with PNG_CONST.

Much of the internal use of C macros to control the library build has also
changed and some of this is visible in the exported header files, in
particular the use of macros to control data and API elements visible
during application compilation may require significant revision to
application code.  (It is extremely rare for an application to do this.)

Any program that compiled against libpng 1.4 and did not use deprecated
features or access internal library structures should compile and work
against libpng 1.5, except for the change in the prototype for
png_get_iCCP() and png_set_iCCP() API functions mentioned above.

libpng 1.5.0 adds PNG_ PASS macros to help in the reading and writing of
interlaced images.  The macros return the number of rows and columns in
each pass and information that can be used to de-interlace and (if
absolutely necessary) interlace an image.

libpng 1.5.0 adds an API png_longjmp(png_ptr, value).  This API calls
the application-provided png_longjmp_ptr on the internal, but application
initialized, longjmp buffer.  It is provided as a convenience to avoid
the need to use the png_jmpbuf macro, which had the unnecessary side
effect of resetting the internal png_longjmp_ptr value.

libpng 1.5.0 includes a complete fixed point API.  By default this is
present along with the corresponding floating point API.  In general the
fixed point API is faster and smaller than the floating point one because
the PNG file format used fixed point, not floating point.  This applies
even if the library uses floating point in internal calculations.  A new
macro, PNG_FLOATING_ARITHMETIC_SUPPORTED, reveals whether the library
uses floating point arithmetic (the default) or fixed point arithmetic
internally for performance critical calculations such as gamma correction.
In some cases, the gamma calculations may produce slightly different
results.  This has changed the results in png_rgb_to_gray and in alpha
composition (png_set_background for example). This applies even if the
original image was already linear (gamma == 1.0) and, therefore, it is
not necessary to linearize the image.  This is because libpng has *not*
been changed to optimize that case correctly, yet.

Fixed point support for the sCAL chunk comes with an important caveat;
the sCAL specification uses a decimal encoding of floating point values
and the accuracy of PNG fixed point values is insufficient for
representation of these values. Consequently a "string" API
(png_get_sCAL_s and png_set_sCAL_s) is the only reliable way of reading
arbitrary sCAL chunks in the absence of either the floating point API or
internal floating point calculations.  Starting with libpng-1.5.0, both
of these functions are present when PNG_sCAL_SUPPORTED is defined.  Prior
to libpng-1.5.0, their presence also depended upon PNG_FIXED_POINT_SUPPORTED
being defined and PNG_FLOATING_POINT_SUPPORTED not being defined.

Applications no longer need to include the optional distribution header
file pngusr.h or define the corresponding macros during application
build in order to see the correct variant of the libpng API.  From 1.5.0
application code can check for the corresponding _SUPPORTED macro:

#ifdef PNG_INCH_CONVERSIONS_SUPPORTED
   /* code that uses the inch conversion APIs. */
#endif

This macro will only be defined if the inch conversion functions have been
compiled into libpng.  The full set of macros, and whether or not support
has been compiled in, are available in the header file pnglibconf.h.
This header file is specific to the libpng build.  Notice that prior to
1.5.0 the _SUPPORTED macros would always have the default definition unless
reset by pngusr.h or by explicit settings on the compiler command line.
These settings may produce compiler warnings or errors in 1.5.0 because
of macro redefinition.

Applications can now choose whether to use these macros or to call the
corresponding function by defining PNG_USE_READ_MACROS or
PNG_NO_USE_READ_MACROS before including png.h.  Notice that this is
only supported from 1.5.0; defining PNG_NO_USE_READ_MACROS prior to 1.5.0
will lead to a link failure.

Prior to libpng-1.5.4, the zlib compressor used the same set of parameters
when compressing the IDAT data and textual data such as zTXt and iCCP.
In libpng-1.5.4 we reinitialized the zlib stream for each type of data.
We added five png_set_text_*() functions for setting the parameters to
use with textual data.

Prior to libpng-1.5.4, the PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
option was off by default, and slightly inaccurate scaling occurred.
This option can no longer be turned off, and the choice of accurate
or inaccurate 16-to-8 scaling is by using the new png_set_scale_16_to_8()
API for accurate scaling or the old png_set_strip_16_to_8() API for simple
chopping.  In libpng-1.5.4, the PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
macro became PNG_READ_SCALE_16_TO_8_SUPPORTED, and the PNG_READ_16_TO_8
macro became PNG_READ_STRIP_16_TO_8_SUPPORTED, to enable the two
png_set_*_16_to_8() functions separately.

Prior to libpng-1.5.4, the png_set_user_limits() function could only be
used to reduce the width and height limits from the value of
PNG_USER_WIDTH_MAX and PNG_USER_HEIGHT_MAX, although this document said
that it could be used to override them.  Now this function will reduce or
increase the limits.

Starting in libpng-1.5.22, default user limits were established. These
can be overridden by application calls to png_set_user_limits(),
png_set_user_chunk_cache_max(), and/or png_set_user_malloc_max().
The limits are now
                             max possible  default
   png_user_width_max        0x7fffffff    1,000,000
   png_user_height_max       0x7fffffff    1,000,000
   png_user_chunk_cache_max  0 (unlimited) 1000
   png_user_chunk_malloc_max 0 (unlimited) 8,000,000

The png_set_option() function (and the "options" member of the png struct) was
added to libpng-1.5.15, with option PNG_ARM_NEON.

The library now supports a complete fixed point implementation and can
thus be used on systems that have no floating point support or very
limited or slow support.  Previously gamma correction, an essential part
of complete PNG support, required reasonably fast floating point.

As part of this the choice of internal implementation has been made
independent of the choice of fixed versus floating point APIs and all the
missing fixed point APIs have been implemented.

The exact mechanism used to control attributes of API functions has
changed, as described in the INSTALL file.

A new test program, pngvalid, is provided in addition to pngtest.
pngvalid validates the arithmetic accuracy of the gamma correction
calculations and includes a number of validations of the file format.
A subset of the full range of tests is run when "make check" is done
(in the 'configure' build.)  pngvalid also allows total allocated memory
usage to be evaluated and performs additional memory overwrite validation.

Many changes to individual feature macros have been made. The following
are the changes most likely to be noticed by library builders who
configure libpng:

1) All feature macros now have consistent naming:

#define PNG_NO_feature turns the feature off
#define PNG_feature_SUPPORTED turns the feature on

pnglibconf.h contains one line for each feature macro which is either:

#define PNG_feature_SUPPORTED

if the feature is supported or:

/*#undef PNG_feature_SUPPORTED*/

if it is not.  Library code consistently checks for the 'SUPPORTED' macro.
It does not, and libpng applications should not, check for the 'NO' macro
which will not normally be defined even if the feature is not supported.
The 'NO' macros are only used internally for setting or not setting the
corresponding 'SUPPORTED' macros.

Compatibility with the old names is provided as follows:

PNG_INCH_CONVERSIONS turns on PNG_INCH_CONVERSIONS_SUPPORTED

And the following definitions disable the corresponding feature:

PNG_SETJMP_NOT_SUPPORTED disables SETJMP
PNG_READ_TRANSFORMS_NOT_SUPPORTED disables READ_TRANSFORMS
PNG_NO_READ_COMPOSITED_NODIV disables READ_COMPOSITE_NODIV
PNG_WRITE_TRANSFORMS_NOT_SUPPORTED disables WRITE_TRANSFORMS
PNG_READ_ANCILLARY_CHUNKS_NOT_SUPPORTED disables READ_ANCILLARY_CHUNKS
PNG_WRITE_ANCILLARY_CHUNKS_NOT_SUPPORTED disables WRITE_ANCILLARY_CHUNKS

Library builders should remove use of the above, inconsistent, names.

2) Warning and error message formatting was previously conditional on
the STDIO feature. The library has been changed to use the
CONSOLE_IO feature instead. This means that if CONSOLE_IO is disabled
the library no longer uses the printf(3) functions, even though the
default read/write implementations use (FILE) style stdio.h functions.

3) Three feature macros now control the fixed/floating point decisions:

PNG_FLOATING_POINT_SUPPORTED enables the floating point APIs

PNG_FIXED_POINT_SUPPORTED enables the fixed point APIs; however, in
practice these are normally required internally anyway (because the PNG
file format is fixed point), therefore in most cases PNG_NO_FIXED_POINT
merely stops the function from being exported.

PNG_FLOATING_ARITHMETIC_SUPPORTED chooses between the internal floating
point implementation or the fixed point one.  Typically the fixed point
implementation is larger and slower than the floating point implementation
on a system that supports floating point; however, it may be faster on a
system which lacks floating point hardware and therefore uses a software
emulation.

4) Added PNG_{READ,WRITE}_INT_FUNCTIONS_SUPPORTED.  This allows the
functions to read and write ints to be disabled independently of
PNG_USE_READ_MACROS, which allows libpng to be built with the functions
even though the default is to use the macros - this allows applications
to choose at app buildtime whether or not to use macros (previously
impossible because the functions weren't in the default build.)

XII.  Changes to Libpng from version 1.5.x to 1.6.x

A "simplified API" has been added (see documentation in png.h and a simple
example in contrib/examples/pngtopng.c).  The new publicly visible API
includes the following:

macros:
     PNG_FORMAT_*
     PNG_IMAGE_*
   structures:
     png_control
     png_image
   read functions
     png_image_begin_read_from_file()
     png_image_begin_read_from_stdio()
     png_image_begin_read_from_memory()
     png_image_finish_read()
     png_image_free()
   write functions
     png_image_write_to_file()
     png_image_write_to_memory()
     png_image_write_to_stdio()

Starting with libpng-1.6.0, you can configure libpng to prefix all exported
symbols, using the PNG_PREFIX macro.

We no longer include string.h in png.h.  The include statement has been moved
to pngpriv.h, where it is not accessible by applications.  Applications that
need access to information in string.h must add an '#include <string.h>'
directive.  It does not matter whether this is placed prior to or after
the '#include "png.h"' directive.

The following API are now DEPRECATED:
   png_info_init_3()
   png_convert_to_rfc1123() which has been replaced
     with png_convert_to_rfc1123_buffer()
   png_malloc_default()
   png_free_default()
   png_reset_zstream()

The following have been removed:
   png_get_io_chunk_name(), which has been replaced
     with png_get_io_chunk_type().  The new
     function returns a 32-bit integer instead of
     a string.
   The png_sizeof(), png_strlen(), png_memcpy(), png_memcmp(), and
     png_memset() macros are no longer used in the libpng sources and
     have been removed.  These had already been made invisible to applications
     (i.e., defined in the private pngpriv.h header file) since libpng-1.5.0.

The signatures of many exported functions were changed, such that
   png_structp became png_structrp or png_const_structrp
   png_infop became png_inforp or png_const_inforp
where "rp" indicates a "restricted pointer".

Dropped support for 16-bit platforms. The support for FAR/far types has
been eliminated and the definition of png_alloc_size_t is now controlled
by a flag so that 'small size_t' systems can select it if necessary.

Error detection in some chunks has improved; in particular the iCCP chunk
reader now does pretty complete validation of the basic format.  Some bad
profiles that were previously accepted are now accepted with a warning or
rejected, depending upon the png_set_benign_errors() setting, in particular
the very old broken Microsoft/HP 3144-byte sRGB profile.  Starting with
libpng-1.6.11, recognizing and checking sRGB profiles can be avoided by
means of

#if defined(PNG_SKIP_sRGB_CHECK_PROFILE) && \
        defined(PNG_SET_OPTION_SUPPORTED)
       png_set_option(png_ptr, PNG_SKIP_sRGB_CHECK_PROFILE,
           PNG_OPTION_ON);
    #endif

It's not a good idea to do this if you are using the "simplified API",
which needs to be able to recognize sRGB profiles conveyed via the iCCP
chunk.

The PNG spec requirement that only grayscale profiles may appear in images
with color type 0 or 4 and that even if the image only contains gray pixels,
only RGB profiles may appear in images with color type 2, 3, or 6, is now
enforced.  The sRGB chunk is allowed to appear in images with any color type
and is interpreted by libpng to convey a one-tracer-curve gray profile or a
three-tracer-curve RGB profile as appropriate.

Libpng 1.5.x erroneously used /MD for Debug DLL builds; if you used the debug
builds in your app and you changed your app to use /MD you will need to
change it back to /MDd for libpng 1.6.x.

Prior to libpng-1.6.0 a warning would be issued if the iTXt chunk contained
an empty language field or an empty translated keyword.  Both of these
are allowed by the PNG specification, so these warnings are no longer issued.

The library now issues an error if the application attempts to set a
transform after it calls png_read_update_info() or if it attempts to call
both png_read_update_info() and png_start_read_image() or to call either
of them more than once.

The default condition for benign_errors is now to treat benign errors as
warnings while reading and as errors while writing.

The library now issues a warning if both background processing and RGB to
gray are used when gamma correction happens. As with previous versions of
the library the results are numerically very incorrect in this case.

There are some minor arithmetic changes in some transforms such as
png_set_background(), that might be detected by certain regression tests.

Unknown chunk handling has been improved internally, without any API change.
This adds more correct option control of the unknown handling, corrects
a pre-existing bug where the per-chunk 'keep' setting is ignored, and makes
it possible to skip IDAT chunks in the sequential reader.

The machine-generated configure files are no longer included in branches
libpng16 and later of the GIT repository.  They continue to be included
in the tarball releases, however.

Libpng-1.6.0 through 1.6.2 used the CMF bytes at the beginning of the IDAT
stream to set the size of the sliding window for reading instead of using the
default 32-kbyte sliding window size.  It was discovered that there are
hundreds of PNG files in the wild that have incorrect CMF bytes that caused
zlib to issue the "invalid distance too far back" error and reject the file.
Libpng-1.6.3 and later calculate their own safe CMF from the image dimensions,
provide a way to revert to the libpng-1.5.x behavior (ignoring the CMF bytes
and using a 32-kbyte sliding window), by using

png_set_option(png_ptr, PNG_MAXIMUM_INFLATE_WINDOW,
        PNG_OPTION_ON);

and provide a tool (contrib/tools/pngfix) for rewriting a PNG file while
optimizing the CMF bytes in its IDAT chunk correctly.

Libpng-1.6.0 and libpng-1.6.1 wrote uncompressed iTXt chunks with the wrong
length, which resulted in PNG files that cannot be read beyond the bad iTXt
chunk.  This error was fixed in libpng-1.6.3, and a tool (called
contrib/tools/png-fix-itxt) has been added to the libpng distribution.

Starting with libpng-1.6.17, the PNG_SAFE_LIMITS macro was eliminated
and safe limits are used by default (users who need larger limits
can still override them at compile time or run time, as described above).

The new limits are
                                default   spec limit
   png_user_width_max         1,000,000  2,147,483,647
   png_user_height_max        1,000,000  2,147,483,647
   png_user_chunk_cache_max         128  unlimited
   png_user_chunk_malloc_max  8,000,000  unlimited

Starting with libpng-1.6.18, a PNG_RELEASE_BUILD macro was added, which allows
library builders to control compilation for an installed system (a release build).
It can be set for testing debug or beta builds to ensure that they will compile
when the build type is switched to RC or STABLE. In essence this overrides the
PNG_LIBPNG_BUILD_BASE_TYPE definition which is not directly user controllable.

Starting with libpng-1.6.19, attempting to set an over-length PLTE chunk
is an error. Previously this requirement of the PNG specification was not
enforced, and the palette was always limited to 256 entries. An over-length
PLTE chunk found in an input PNG is silently truncated.

Starting with libpng-1.6.31, the eXIf chunk is supported. Libpng does not
attempt to decode the Exif profile; it simply returns a byte array
containing the profile to the calling application which must do its own
decoding.

XIII.  Detecting libpng

The png_get_io_ptr() function has been present since libpng-0.88, has never
changed, and is unaffected by conditional compilation macros.  It is the
best choice for use in configure scripts for detecting the presence of any
libpng version since 0.88.  In an autoconf "configure.in" you could use

AC_CHECK_LIB(png, png_get_io_ptr, ...

XV. Source code repository

Since about February 2009, version 1.2.34, libpng has been under "git" source
control.  The git repository was built from old libpng-x.y.z.tar.gz files
going back to version 0.70.  You can access the git repository (read only)
at

https://github.com/glennrp/libpng or
    https://git.code.sf.net/p/libpng/code.git

or you can browse it with a web browser at

https://github.com/glennrp/libpng or
    https://sourceforge.net/p/libpng/code/ci/libpng16/tree/

Patches can be sent to glennrp at users.sourceforge.net or to
png-mng-implement at lists.sourceforge.net or you can upload them to
the libpng bug tracker at

https://libpng.sourceforge.io/

or as a "pull request" to

https://github.com/glennrp/libpng/pulls

We also accept patches built from the tar or zip distributions, and
simple verbal discriptions of bug fixes, reported either to the
SourceForge bug tracker, to the png-mng-implement at lists.sf.net
mailing list, as github issues, or directly to glennrp.

XV. Coding style

Our coding style is similar to the "Allman" style
(See https://en.wikipedia.org/wiki/Indent_style#Allman_style), with curly
braces on separate lines:

if (condition)
    {
       action;
    }

else if (another condition)
    {
       another action;
    }

The braces can be omitted from simple one-line actions:

if (condition)
       return (0);

We use 3-space indentation, except for continued statements which
are usually indented the same as the first line of the statement
plus four more spaces.

For macro definitions we use 2-space indentation, always leaving the "#"
in the first column.

#ifndef PNG_NO_FEATURE
    #  ifndef PNG_FEATURE_SUPPORTED
    #    define PNG_FEATURE_SUPPORTED
    #  endif
    #endif

Comments appear with the leading "/*" at the same indentation as
the statement that follows the comment:

/* Single-line comment */
    statement;

/* This is a multiple-line
     * comment.
     */
    statement;

Very short comments can be placed after the end of the statement
to which they pertain:

statement;    /* comment */

We don't use C++ style ("//") comments. We have, however,
used them in the past in some now-abandoned MMX assembler
code.

Functions and their curly braces are not indented, and
exported functions are marked with PNGAPI:

/* This is a public function that is visible to
  * application programmers. It does thus-and-so.
  */
 void PNGAPI
 png_exported_function(png_ptr, png_info, foo)
 {
    body;
 }

The return type and decorations are placed on a separate line
ahead of the function name, as illustrated above.

The prototypes for all exported functions appear in png.h,
above the comment that says

/* Maintainer: Put new public prototypes here ... */

We mark all non-exported functions with "/* PRIVATE */"":

void /* PRIVATE */
 png_non_exported_function(png_ptr, png_info, foo)
 {
    body;
 }

The prototypes for non-exported functions (except for those in
pngtest) appear in pngpriv.h above the comment that says

/* Maintainer: Put new private prototypes here ^ */

To avoid polluting the global namespace, the names of all exported
functions and variables begin with "png_", and all publicly visible C
preprocessor macros begin with "PNG".  We request that applications that
use libpng *not* begin any of their own symbols with either of these strings.

We put a space after the "sizeof" operator and we omit the
optional parentheses around its argument when the argument
is an expression, not a type name, and we always enclose the
sizeof operator, with its argument, in parentheses:

(sizeof (png_uint_32))
  (sizeof array)

Prior to libpng-1.6.0 we used a "png_sizeof()" macro, formatted as
though it were a function.

Control keywords if, for, while, and switch are always followed by a space
to distinguish them from function calls, which have no trailing space.

We put a space after each comma and after each semicolon
in "for" statements, and we put spaces before and after each
C binary operator and after "for" or "while", and before
"?".  We don't put a space between a typecast and the expression
being cast, nor do we put one between a function name and the
left parenthesis that follows it:

for (i = 2; i > 0; --i)
       y[i] = a(x) + (int)b;

We prefer #ifdef and #ifndef to #if defined() and #if !defined()
when there is only one macro being tested.  We always use parentheses
with "defined".

We express integer constants that are used as bit masks in hex format,
with an even number of lower-case hex digits, and to make them unsigned
(e.g., 0x00U, 0xffU, 0x0100U) and long if they are greater than 0x7fff
(e.g., 0xffffUL).

We prefer to use underscores rather than camelCase in names, except
for a few type names that we inherit from zlib.h.

We prefer "if (something != 0)" and "if (something == 0)" over
"if (something)" and if "(!something)", respectively, and for pointers
we prefer "if (some_pointer != NULL)" or "if (some_pointer == NULL)".

We do not use the TAB character for indentation in the C sources.

Lines do not exceed 80 characters.

Other rules can be inferred by inspecting the libpng source.

XVI. Y2K Compliance in libpng

Since the PNG Development group is an ad-hoc body, we can't make
an official declaration.

This is your unofficial assurance that libpng from version 0.71 and
upward through 1.6.34 are Y2K compliant.  It is my belief that earlier
versions were also Y2K compliant.

Libpng only has two year fields.  One is a 2-byte unsigned integer
that will hold years up to 65535.  The other, which is deprecated,
holds the date in text format, and will hold years up to 9999.

The integer is
    "png_uint_16 year" in png_time_struct.

The string is
    "char time_buffer[29]" in png_struct.  This is no longer used
in libpng-1.6.x and will be removed from libpng-1.7.0.

There are seven time-related functions:

png_convert_to_rfc_1123_buffer() in png.c
      (formerly png_convert_to_rfc_1152() in error, and
      also formerly png_convert_to_rfc_1123())
    png_convert_from_struct_tm() in pngwrite.c, called
      in pngwrite.c
    png_convert_from_time_t() in pngwrite.c
    png_get_tIME() in pngget.c
    png_handle_tIME() in pngrutil.c, called in pngread.c
    png_set_tIME() in pngset.c
    png_write_tIME() in pngwutil.c, called in pngwrite.c

All appear to handle dates properly in a Y2K environment.  The
png_convert_from_time_t() function calls gmtime() to convert from system
clock time, which returns (year - 1900), which we properly convert to
the full 4-digit year.  There is a possibility that applications using
libpng are not passing 4-digit years into the png_convert_to_rfc_1123()
function, or that they are incorrectly passing only a 2-digit year
instead of "year - 1900" into the png_convert_from_struct_tm() function,
but this is not under our control.  The libpng documentation has always
stated that it works with 4-digit years, and the APIs have been
documented as such.

The tIME chunk itself is also Y2K compliant.  It uses a 2-byte unsigned
integer to hold the year, and can hold years as large as 65535.

zlib, upon which libpng depends, is also Y2K compliant.  It contains
no date-related code.

Glenn Randers-Pehrson
   libpng maintainer
   PNG Development Group
PKJM�\����CHANGESnu�[���PKJM�\V���/�TODOnu�[���PKJM�\��]��	B�example.cnu�[���PKJM�\����S�S��=libpng-manual.txtnu�[���PK%&�

/home/emeraadmin/www/PSCU/../node_modules/jquery/../interpret/../debug/../../4d695/libpng-devel.zip