[libpng17] Imported from libpng-1.7.0beta39.tar

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.
Removed "option WRITE_COMPRESSED_TEXT enables WRITE_TEXT" from pnglibconf.dfa
Only mark text chunks as written after successfully writing them.

ANNOUNCE

@@ -1,5 +1,5 @@
 
-Libpng 1.5.0beta47 - September 11, 2010
+Libpng 1.7.0beta39 - November 1, 2014
 
 This is not intended to be a public release.  It will be replaced
 within a few weeks by a public version or by another test version.
@@ -9,397 +9,621 @@ Files available for download:
 Source files with LF line endings (for Unix/Linux) and with a
 "configure" script
 
-   1.5.0beta47.tar.xz (LZMA-compressed, recommended)
-   1.5.0beta47.tar.gz
-   1.5.0beta47.tar.bz2
+   1.7.0beta39.tar.xz (LZMA-compressed, recommended)
+   1.7.0beta39.tar.gz
 
 Source files with CRLF line endings (for Windows), without the
 "configure" script
 
-   lp150b47.zip
-   lp150b47.7z
+   lp170b39.7z  (LZMA-compressed, recommended)
+   lp170b39.zip
 
 Other information:
 
-   1.5.0beta47-README.txt
-   1.5.0beta47-LICENSE.txt
+   1.7.0beta39-README.txt
+   1.7.0beta39-LICENSE.txt
+   libpng-1.7.0beta39-*.asc (armored detached GPG signatures)
 
-Changes since the last public release (1.4.1):
+Changes since the last public release (1.6.0):
 
-version 1.5.0beta01 [February 7, 2010]
-  Moved declaration of png_struct into private pngstruct.h and png_info
-    into pnginfo.h
+Version 1.7.0alpha01 [December 15, 2012]
+  Started 1.7.0 branch from libpng-1.6.0beta33.
+  Made 8-bit compose and rgb_to_grayscale accuracy improvements.  These
+    changes cause 16-bit arithmetic to be used for 8-bit data in the gamma
+    corrected compose and grayscale operations.  The arithmetic errors have
+    three sources all of which are fixed in this commit:
+    1) 8-bit linear calculations produce massive errors for lower intensity
+       values.
+    2) The old 16-bit "16 to 8" gamma table code erroneously wrote the lowest
+       output value into a table entry which corresponded to multiple output
+       values (so where the value written should have been the closest to the
+       transformed input value.)
+    3) In a number of cases the code to access the 16-bit table did not round;
+       it did a simple shift, which was wrong and made the side effects of (2)
+       even worse.
+  The new gamma code does not have the 16-to-8 problem at the cost of slightly
+    more calculations and the algorithm used to minimize the number of
+    calculations has been extended to all the 16-bit tables; it has advantages
+    for any significant gamma correction.
+  Rearranged png_struct, remove unused members, change png_set_filter handling
+    png_struct members rearranged - partly to reorder to avoid packing, partly
+    to put frequently accessed members at the start and partly to make
+    the grouping more clear. png_set_filter code has been rewritten and the
+    code shared with png_write_start_row moved to a common function.  Comments
+    in png.h have been made more clear. Minor fixes to
+    contrib/libtests/timepng.c and some of the png_*_tRNS logic, including
+    more error detection in png_set_tRNS.
+  Cleaned up USER_LIMITS feature, removing members from png_struct when not
+    required.
+  Rearranged ARM-NEON optimizations to isolate the machine specific code to
+    the hardware subdirectory, and add comments to pngrutil.c so that
+    implementors of other optimizations will know what to do.
+  Fixed cases of unquoted DESTDIR in Makefile.am.
+  Rebuilt Makefile.in, configure, etc., with autoconf-2.69 and automake-1.12.5.
 
-version 1.5.0beta02 [February 7, 2010]
-  Reverted to original png_push_save_buffer() code.
+Version 1.7.0alpha02 [December 17, 2012]
+  Removed functions that were deprecated in libpng-1.6.0:
+    png_reset_zstream(), png_info_init_3(), and png_data_freer() and its
+    associated flags.
+  Removed some duplicated lines from contrib/tools/scale.c and png.c.
+  Changed some instances of png_warning() to png_app_error().
+  Updated some left over "1.6.0beta32" in code sources.
+  Fixed a "png_structp" prototype (should be png_structrp) in arm_init.c
+  Updated the version-number hack in pngvalid.c
 
-version 1.5.0beta03 [February 8, 2010]
-  Return allocated "old_buffer" in png_push_save_buffer() before
-    calling png_error().
-  Updated configure script to use SO number 15.
+Version 1.7.0alpha03 [December 19, 2012]
+  Cleaned up and enhanced the configure option; libpng now correctly
+    links and tests against zlib with a prefix; tests have been clarified; and
+    irrelevant or obsolete things (as defined by the autotools man page) have
+    been removed.
+  Documented new PNG_ABORT behavior in the manual and commentary in pngerror.c
+  Cleaned up whitespace in the synopsis portion of the manpage "libpng.3"
+  Fixed previous support for Z_PREFIX in configure builds, corrected sCAL APIs;
+    some of these use floating point arithmetic so they need to be disabled if
+    floating point arithmetic is switched off.  This is a quiet API change -
+    previously it appeared that the APIs were supported if fixed point
+    arithmetic was used internally, however they required certain APIs (floor,
+    modf, frexp, atof) that are part of C floating point support. Changed
+    png_fixed and the gamma code specific version of the same to avoid floor(),
+    which may be a library function (not an intrinsic). Removed unused #if 0
+    code.
+  Disassembled the version number in scripts/options.awk (necessary for
+    building on SunOs).
 
-version 1.5.0beta04 [February 9, 2010]
-  Removed malformed "incomplete struct declaration" of png_info from png.h
+Version 1.7.0alpha04 [December 23, 2012]
+  Added scripts/makefile.msys contributed by Christopher M. Wheeler
+  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 1.5.0beta05 [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 1.7.0alpha05 [December 24, 2012]
+  Discontinued distributing libpng-1.x.x.tar.bz2.
+  Discontinued distributing libpng-1.7.0-1.6.0-diff.txt and similar.
 
-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 1.7.0alpha06 [January 1, 2013]
+  Fixed 'make distcheck' on SUN OS - libpng.so was not being removed
 
-version 1.5.0beta07 [omitted]
+Version 1.7.0alpha07 [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 1.5.0beta08 [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 1.7.0alpha08 [January 17, 2013]
+  Corrected previous 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 1.5.0beta09 [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 1.7.0alpha09 [January 21, 2013]
+  Pulled changes to multi-chunk handling from libpng-1.6.0beta40.
 
-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 1.7.0alpha10 [February 5, 2013]
+  Make 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.
+  Added png_get_palette_max() function.
+  Changed png_size_t to size_t throughout headers, make sRGB check numbers
+    consistent.
+  Fixed ARM support (Mans Rullgard).
+  Removed stray out-of-order #endif and #ifdef (Mans Rullgard).
+  Ported libpng 1.5 options.awk/dfn file handling to 1.7, fixed one bug.
+  Rebuilt Makefile.in, configure, etc., with autoconf-2.69 and automake-1.13.
+  Minor fixes to assembler changes, verified with a dummy .S file
+  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.
 
-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 1.7.0beta01 [February 15, 2013]
+  Enable parallel tests and rearrange TESTS order to take advantage of
+    the massive speed improvements use a make capable of parallel builds
+    on a multi-CPU machine and pass the right arguments to make (-j10000
+    for GNU make) to get the build to run in parallel.
+  Fixed previous support for Z_PREFIX in configure builds, corrected
+    sCAL APIs; some of these use floating point arithmetic so need to be
+    disabled if floating point arithmetic is switched off.  This is a quiet API
+    change - previously it appeared that the APIs were supported if fixed point
+    arithmetic was used internally, however they required certain APIs (floor,
+    modf, frexp, atof) that are part of C floating point support. Changed
+    png_fixed and the gamma code specific version of the same to avoid floor(),
+    which may be a library function (not an intrinsic). Removed unused #if 0
+    code.
+  Fixed make distcheck to skip over the test result files. They were
+    being included because the whole of the 'tests' directory is included,
+    fixed to just include the actual test files and tests/pngstest.  Also
+    reverted the previous erroneous change to remove the ${srcdir} from the
+    front of contrib, scripts, projects and added an explanation of what is
+    going on to Makefile.am
+  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.
+  Accept "," as a separator in pnglibconf.dfa and allow for
+    continuation lines. This adds to the syntax of "option" and "chunk" lines,
+    allowing the elements to be separated by a "," at the end of an element and
+    interpreting a "," at end of line as a continuation - so the following line
+    is read.  The new syntax is optional and solely provided for readability.
+  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 1.5.0beta12 [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 1.7.0beta02 [February 18, 2013]
+  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).
+  Use parentheses more consistently in "#if defined(MACRO)" tests.
+  Folded long lines.
 
-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 1.7.0beta03 [February 22, 2013]
+  Reenabled code to allow zero length PLTE chunks for MNG.
+  Fixed ALIGNED_MEMORY support.
+  Allow run-time ARM NEON checking to be disabled. 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 1.5.0beta14 [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 1.7.0beta04 [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 1.5.0beta15 [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 1.7.0beta05 [March 5, 2013]
+  Avoid a possible memory leak in contrib/gregbook/readpng.c
+  Changed user chunk callback API to respect global and per chunk defaults.
+    Previously a return of 0 from a user chunk callback would result in the
+    chunk being saved (if this was safe, even if the chunk was unsafe-to-copy);
+    this change respects the defaults set by the application, so unknown chunks
+    can be discarded by default and known-safe ones preserved.
+  Corrected Android builds and corrected libpng.vers with symbol
+    prefixing. This adds an API to set optimization options externally,
+    providing an alternative and general solution for the non-portable
+    run-time tests used by the ARM Neon code.  It also makes those tests
+    compile and link on Android.  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 1.5.0beta16 [April 1, 2010]
-  Made 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 1.7.0beta06 [March 13, 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.
+  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.
+  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 1.5.0beta17 [April 17, 2010]
-  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.
+Version 1.7.0beta07 [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).
 
-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 1.7.0beta08 [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 1.5.0beta19 [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 1.7.0beta09 [April 24, 2013]
+  Exposed PNG chunk types in png.h
+  Modified png_uint_32 macros/code to make as few assumptions as possible
+  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 1.5.0beta20 [April 24, 2010]
-  Relaxed more overly-restrictive permissions of some files.
+Version 1.7.0beta10 [April 24, 2013]
+  Attempt to fix the PNG_ARM_NEON configuration mess.
 
-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 1.7.0beta11 [April 26, 2013]
+  Test for 'arm*', not just 'arm' in the host_cpu configure variable.
 
-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 1.7.0beta12 [April 30, 2013]
+  Added png_app_warning for out-of-range unknown chunk index in
+    png_set_unknown_chunk_location().
+  Expanded manual paragraph about writing private chunks.
+  Avoid dereferencing NULL pointer possibly returned from
+     png_create_write_struct() (Andrew Church).
 
-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 1.7.0beta13 [May 12, 2013]
+  Revised contrib/pngminim/*/makefile to generate pnglibconf.h with the
+    right zlib header files.
+  Separated CPPFLAGS and CFLAGS in contrib/pngminim/*/makefile
+  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.
+  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.
+  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.
+  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 1.5.0beta24 [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 1.7.0beta14 [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 STDIO/#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 1.5.0beta25 [September 11, 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 1.7.0beta15 [June 18, 2013]
+  Revised libpng.3 so that "doclifter" can process it.
 
-version 1.5.0beta26 [June 18, 2010]
-  In pngpread.c: png_push_process_row() add check for too many rows.
+Version 1.7.0beta16 [July 5, 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.
+  Updated documentation to show default behavior of benign errors correctly.
+  Added perfect hash code generation for lists of PNG chunks. This is
+    a work in progress; checked in for use in pngfix.c
+  Ported ARM no-read patch from libpng16.
+  Ported pngfix.c patches from libpng16.
 
-version 1.5.0beta27 [June 18, 2010]
-  Removed the check added in beta25 as it is now redundant.
+Version 1.7.0beta17 [August 21, 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.
+  Added information about png_set_options() to the manual.
+  Delay calling png_init_filter_functions() until a row with nonzero filter
+    is found.
 
-version 1.5.0beta28 [June 19, 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 1.7.0beta18 [September 16, 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 1.5.0beta29 [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 1.7.0beta19 [September 30, 2013]
+  Reverted the change to unknown handling #defines; the change breaks 'NOREAD'
+    builds.
+  Fixed default behavior of ARM_NEON_API. If the ARM NEON API option is
+    compiled without the CHECK option it defaulted to on, not off.
+  Catch up with recent libpng16 changes; unknown handling and spelling
+    corrections
+  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 some grammatical changes and updated the png chunks list.
 
-version 1.5.0beta30 [June 22, 2010]
-  Stop memory leak when reading a malformed sCAL chunk.
+Version 1.7.0beta20 [October 13, 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 1.5.0beta31 [June 26, 2010]
-  Revised pngpread.c patch of beta28 to avoid an endless loop.
-  Removed some trailing blanks.
+Version 1.7.0beta21 [November 2, 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.
+  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 1.5.0beta32 [June 26, 2010]
-  Removed leftover scripts/options.patch and scripts/options.rej
+Version 1.7.0beta22 [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 1.5.0beta33 [July 6, 2010]
-  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 1.7.0beta23 [November 24, 2013]
+  Changed #ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED in pngpread.c to
+    #ifdef PNG_SET_UNKNOWN_CHUNKS_SUPPORTED to be consistent with
+    what is in pngpriv.h.
+  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 1.5.0beta34 [July 12, 2010]
-  Put #ifndef PNG_EXTERN, #endif around the define PNG_EXTERN in pngpriv.h
+Version 1.7.0beta24 [December 15, 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
+  Tidied up pngfix inits and fixed pngtest no-write builds.
+  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.
+  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 and libpng-1.7.0beta03
+    through 1.7.0beta23 are vulnerable.  Libpng-1.6.0 and earlier do not
+    have the bug.
 
-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 1.7.0beta25 [December 26, 2013]
+  Merged files with version 1.6.8 that have not changed since then.
+  Rebuilt configure scripts with automake-1.14.1 and autoconf-2.69
 
-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 1.7.0beta26 [January 1, 2014]
+  Added libpng 1.5 checks to pngvalid.c
+  Merged with 1.5 and 1.6 changes to create a single pngvalid.c
+  Added dSIG to chunk list in png.h
+  Removed #error macro from contrib/tools/pngfix.c (Thomas Klausner).
+  Merged pngrio.c, pngtrans.c, and pngwio.c with libpng-1.6.9
+  Fixed test programs for interlace options. Made pngvalid.c and
+    pngtest.c work correctly when READ_INTERLACING and/or WRITE_INTERLACING
+    are switched off.
+  Changed 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 1.5.0beta37 [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 1.7.0beta27 [January 10, 2014]
+  Removed potentially misleading warning from png_check_IHDR().
 
-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 1.7.0beta28 [January 20, 2014]
+  Updated scripts/makefile.* to use CPPFLAGS (Cosmin).
+  Added clang attribute support (Cosmin).
 
-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 1.7.0beta29 [January 30, 2014]
+  Quiet an uninitialized memory warning from VS2013 in png_get_png().
 
-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 1.7.0beta30 [February 2, 2014]
+  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. This
+    is a work-in-progress; no tests are run automatically at present and
+    the program by virtue of exhaustively testing all the transforms is
+    very slow.
 
-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 1.7.0beta31 [February 6, 2014]
+  Make png_read_png() and png_write_png() prototypes in png.h depend
+    upon PNG_READ_SUPPORTED and PNG_WRITE_SUPPORTED.
+  Removed dependence on !PNG_READ_EXPAND_SUPPORTED for calling
+    png_set_packing() in png_read_png().
+  Completed full working pngimage test program and added two test scripts.
+  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).
+  Restored a line, "c = b;", that was inadvertently deleted from the
+    PAETH filtering code from libpng-1.7.0beta24/pngrutil.c. Deleted
+    the other instance of "c = b;" which is the one that triggered a
+    scan-build warning.
+  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 1.5.0beta42 [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 1.7.0beta32 [February 26, 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).
+  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.
+  Minor editing of contrib/arm-neon/README and contrib/examples/*.c
+  Moved some documentation from png.h to libpng.3 and libpng-manual.txt
 
-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 1.7.0beta33 [February 27, 2014]
+  Fixed typos in the manual and in scripts/pnglibconf.dfa (CFLAGS -> CPPFLAGS
+    and PNG_USR_CONFIG -> PNG_USER_CONFIG).
 
-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 1.7.0beta34 [March 17, 2014]
+  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 "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 1.5.0beta45 [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 1.7.0beta35 [August 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).
+  $(CP) is now used alongside $(RM_F).  Also, use 'copy' instead
+    of 'cp' where applicable, and applied other minor makefile changes.
+  Don't warn about invalid dimensions exceeding user limits.
+  Allow an easy replacement of the default pre-built configuration
+    header with a custom header, via the make PNGLIBCONF_H_PREBUILT
+    macro.
+  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).
+  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).
+  Added opinion that the ECCN (Export Control Classification Number) for
+    libpng is EAR99 to the README file.
 
-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 1.7.0beta36 [September 27, 2014]
+  Updated scripts/makefiles from libpng-1.6.13beta04.
+  Minor updates to the manual from libpng-1.6.14beta04.
+  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 1.5.0beta47 [September 11, 2010]
-  Fixed a number of problems with 64-bit compilation reported by Visual
-    Studio 2010 (John Bowler).
+Version 1.7.0beta37 [September 28, 2014]
+  Merged scripts/*, pngvalid.c and pngfix.c with libpng-1.6.14beta04.
 
-Send comments/corrections/commendations to png-mng-implement at lists.sf.net:
+Version 1.7.0beta38 [October 17, 2014]
+  Added "option READ_iCCP enables READ_COMPRESSED_TEXT" to pnglibconf.dfa
+  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.
+  Removed "option WRITE_COMPRESSED_TEXT enables WRITE_TEXT" from pnglibconf.dfa
+  Only mark text chunks as written after successfully writing them.
+
+Version 1.7.0beta39 [November 1, 2014]
+  Ported cosmetic changes from libpng-1.6.15beta02.
+
+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
+to subscribe)
+or to glennrp at users.sourceforge.net
 
 Glenn R-P

CMakeLists.txt

@@ -1,27 +1,21 @@
-cmake_minimum_required(VERSION 2.4.3)
-set(CMAKE_ALLOW_LOOSE_LOOP_CONSTRUCTS true)
+# CMakeLists.txt
 
-if(UNIX AND NOT DEFINED CMAKE_BUILD_TYPE)
-  set(CMAKE_BUILD_TYPE "RelWithDebInfo" CACHE STRING 
-      "Choose the type of build, options are:
-         None(CMAKE_CXX_FLAGS or CMAKE_C_FLAGS used)
-         Debug
-         Release
-         RelWithDebInfo
-         MinSizeRel.")
-endif()
-
-project(libpng C)
-enable_testing()
-
-# Copyright (C) 2007-2010 Glenn Randers-Pehrson
+# Copyright (C) 2007-2014 Glenn Randers-Pehrson
 
 # This code is released under the libpng license.
 # For conditions of distribution and use, see the disclaimer
 # and license in png.h
 
+cmake_minimum_required(VERSION 2.4.4)
+set(CMAKE_ALLOW_LOOSE_LOOP_CONSTRUCTS true)
+
+set(CMAKE_CONFIGURATION_TYPES "Release;Debug;MinSizeRel;RelWithDebInfo")
+
+project(libpng C)
+enable_testing()
+
 set(PNGLIB_MAJOR 1)
-set(PNGLIB_MINOR 5)
+set(PNGLIB_MINOR 7)
 set(PNGLIB_RELEASE 0)
 set(PNGLIB_NAME libpng${PNGLIB_MAJOR}${PNGLIB_MINOR})
 set(PNGLIB_VERSION ${PNGLIB_MAJOR}.${PNGLIB_MINOR}.${PNGLIB_RELEASE})
@@ -56,54 +50,37 @@ else()
   option(PNG_STATIC "Build static lib" ON)
 endif()
 
-if(MINGW)
-  option(PNG_TESTS  "Build pngtest" NO)
-else()
-  option(PNG_TESTS  "Build pngtest" YES)
-endif()
+option(PNG_TESTS  "Build libpng tests" YES)
 
-option(PNG_NO_CONSOLE_IO "FIXME" YES)
-option(PNG_NO_STDIO      "FIXME" YES)
+# Many more configuration options could be added here
 option(PNG_DEBUG         "Build with debug output" NO)
-option(PNGARG            "FIXME" YES)
-#TODO:
-# PNG_CONSOLE_IO_SUPPORTED
-
-# maybe needs improving, but currently I don't know when we can enable what :)
-set(png_asm_tmp "OFF")
-if(NOT WIN32)
-  find_program(uname_executable NAMES uname PATHS /bin /usr/bin /usr/local/bin)
-  if(uname_executable)
-    exec_program(${uname_executable}
-                 ARGS --machine OUTPUT_VARIABLE uname_output)
-    if("uname_output" MATCHES "^.*i[1-9]86.*$")
-      set(png_asm_tmp "ON")
-    else("uname_output" MATCHES "^.*i[1-9]86.*$")
-      set(png_asm_tmp "OFF")
-    endif("uname_output" MATCHES "^.*i[1-9]86.*$")
-  endif(uname_executable)
-else()
-  # this env var is normally only set on win64
-  set(TEXT "ProgramFiles(x86)")
-  if("$ENV{${TEXT}}" STREQUAL "")
-    set(png_asm_tmp "ON")
-  endif("$ENV{${TEXT}}" STREQUAL "")
-endif()
+option(PNGARG            "Disable ANSI-C prototypes" NO)
 
 # SET LIBNAME
-set(PNG_LIB_NAME libpng${PNGLIB_MAJOR}${PNGLIB_MINOR})
+set(PNG_LIB_NAME png${PNGLIB_MAJOR}${PNGLIB_MINOR})
 
 # to distinguish between debug and release lib
 set(CMAKE_DEBUG_POSTFIX "d")
 
+# Use the prebuilt pnglibconf.h file from the scripts folder
+# TODO: fix this by building with awk; without this no cmake build can be
+# configured directly (to do so indirectly use your local awk to build a
+# pnglibconf.h in the build directory.)
+configure_file(${CMAKE_CURRENT_SOURCE_DIR}/scripts/pnglibconf.h.prebuilt
+               ${CMAKE_CURRENT_BINARY_DIR}/pnglibconf.h)
+include_directories(${CMAKE_CURRENT_BINARY_DIR})
 
 # OUR SOURCES
-set(libpng_sources
+set(libpng_public_hdrs
   png.h
   pngconf.h
-  pngpriv.h
+  ${CMAKE_CURRENT_BINARY_DIR}/pnglibconf.h
+)
+set(libpng_sources
+  ${libpng_public_hdrs}
   pngdebug.h
   pnginfo.h
+  pngpriv.h
   pngstruct.h
   png.c
   pngerror.c
@@ -124,52 +101,34 @@ set(libpng_sources
 set(pngtest_sources
   pngtest.c
 )
+set(pngvalid_sources
+  contrib/libtests/pngvalid.c
+)
+set(pngstest_sources
+  contrib/libtests/pngstest.c
+)
 # SOME NEEDED DEFINITIONS
 
-add_definitions(-DPNG_CONFIGURE_LIBPNG)
-
-if(_AIX)
-  add_definitions(-D_ALL_SOURCE)
-endif(_AIX)
-
 if(MSVC)
-  add_definitions(-DPNG_NO_MODULEDEF -D_CRT_SECURE_NO_DEPRECATE)
+  add_definitions(-D_CRT_SECURE_NO_DEPRECATE)
 endif(MSVC)
 
-if(PNG_SHARED OR  NOT MSVC)
-  #if building msvc static this has NOT to be defined
-  add_definitions(-DZLIB_DLL)
-endif()
-
-
-if(PNG_CONSOLE_IO_SUPPORTED)
-  add_definitions(-DPNG_CONSOLE_IO_SUPPORTED)
-endif()
-
-if(PNG_NO_CONSOLE_IO)
-  add_definitions(-DPNG_NO_CONSOLE_IO)
-endif()
-
-if(PNG_NO_STDIO)
-  add_definitions(-DPNG_NO_STDIO)
-endif()
-
 if(PNG_DEBUG)
   add_definitions(-DPNG_DEBUG)
 endif()
 
-if(NOT M_LIBRARY AND NOT WIN32)
-  add_definitions(-DPNG_NO_FLOATING_POINT_SUPPORTED)
-endif()
-
 # NOW BUILD OUR TARGET
 include_directories(${CMAKE_CURRENT_SOURCE_DIR} ${ZLIB_INCLUDE_DIR})
 
+unset(PNG_LIB_TARGETS)
+
 if(PNG_SHARED)
   add_library(${PNG_LIB_NAME} SHARED ${libpng_sources})
+  set(PNG_LIB_TARGETS ${PNG_LIB_NAME})
   if(MSVC)
     # msvc does not append 'lib' - do it here to have consistent name
     set_target_properties(${PNG_LIB_NAME} PROPERTIES PREFIX "lib")
+    set_target_properties(${PNG_LIB_NAME} PROPERTIES IMPORT_PREFIX "lib")
   endif()
   target_link_libraries(${PNG_LIB_NAME} ${ZLIB_LIBRARY} ${M_LIBRARY})
 endif()
@@ -178,12 +137,19 @@ if(PNG_STATIC)
 # does not work without changing name
   set(PNG_LIB_NAME_STATIC ${PNG_LIB_NAME}_static)
   add_library(${PNG_LIB_NAME_STATIC} STATIC ${libpng_sources})
+  list(APPEND PNG_LIB_TARGETS ${PNG_LIB_NAME_STATIC})
   if(MSVC)
     # msvc does not append 'lib' - do it here to have consistent name
     set_target_properties(${PNG_LIB_NAME_STATIC} PROPERTIES PREFIX "lib")
   endif()
+  target_link_libraries(${PNG_LIB_NAME_STATIC} ${ZLIB_LIBRARY} ${M_LIBRARY})
 endif()
 
+if(NOT PNG_LIB_TARGETS)
+  message(SEND_ERROR
+    "No library variant selected to build. "
+    "Please enable at least one of the following options: PNG_STATIC, PNG_SHARED")
+endif()
 
 if(PNG_SHARED AND WIN32)
   set_target_properties(${PNG_LIB_NAME} PROPERTIES DEFINE_SYMBOL PNG_BUILD_DLL)
@@ -193,115 +159,203 @@ if(PNG_TESTS AND PNG_SHARED)
   # does not work with msvc due to png_lib_ver issue
   add_executable(pngtest ${pngtest_sources})
   target_link_libraries(pngtest ${PNG_LIB_NAME})
-  add_test(pngtest pngtest ${CMAKE_CURRENT_SOURCE_DIR}/pngtest.png)
+  add_test(pngtest ./pngtest ${CMAKE_CURRENT_SOURCE_DIR}/pngtest.png)
+  #
+  add_executable(pngvalid ${pngvalid_sources})
+  target_link_libraries(pngvalid ${PNG_LIB_NAME})
+  add_test(pngvalid ./pngvalid)
+  add_executable(pngstest ${pngstest_sources})
+  target_link_libraries(pngstest ${PNG_LIB_NAME})
+  add_test(pngstest ./pngstest
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn0g01.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn0g02.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn0g04.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn0g08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn0g16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn2c08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn2c16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn3p01.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn3p02.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn3p04.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn4a08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn4a16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn6a08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/basn6a16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbbn0g01.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbbn0g02.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbbn0g04.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbbn2c16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbbn3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbgn2c16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbgn3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbrn2c08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbwn0g16.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbwn3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftbyn3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftp0n0g08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftp0n2c08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftp0n3p08.png
+    ${CMAKE_CURRENT_SOURCE_DIR}/contrib/pngsuite/ftp1n3p08.png
+  )
 endif()
 
+# Ensure the CMAKE_LIBRARY_OUTPUT_DIRECTORY is set
+IF(NOT CMAKE_LIBRARY_OUTPUT_DIRECTORY)
+  SET(CMAKE_LIBRARY_OUTPUT_DIRECTORY "lib")
+ENDIF(NOT CMAKE_LIBRARY_OUTPUT_DIRECTORY)
+
+# Set a variable with CMake code which:
+# Creates a symlink from src to dest (if possible) or alternatively
+# copies if different.
+macro(CREATE_SYMLINK SRC_FILE DEST_FILE)
+  FILE(REMOVE ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/${DEST_FILE})
+  if(WIN32 AND NOT CYGWIN AND NOT MSYS)
+    ADD_CUSTOM_COMMAND(
+        OUTPUT ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/${DEST_FILE}   ${CMAKE_CURRENT_BINARY_DIR}/${DEST_FILE}
+        COMMAND ${CMAKE_COMMAND} -E copy_if_different  "${SRC_FILE}" ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/${DEST_FILE}
+        COMMAND ${CMAKE_COMMAND} -E copy_if_different  "${SRC_FILE}" ${CMAKE_CURRENT_BINARY_DIR}/${DEST_FILE}
+        DEPENDS ${PNG_LIB_TARGETS}
+        )
+    ADD_CUSTOM_TARGET(${DEST_FILE}_COPY ALL DEPENDS ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/${DEST_FILE})
+  else(WIN32 AND NOT CYGWIN AND NOT MSYS)
+    get_filename_component(LINK_TARGET "${SRC_FILE}" NAME)
+    execute_process(COMMAND ${CMAKE_COMMAND} -E make_directory ${CMAKE_CURRENT_BINARY_DIR}/${CMAKE_LIBRARY_OUTPUT_DIRECTORY})
+    execute_process(COMMAND ${CMAKE_COMMAND} -E create_symlink "${LINK_TARGET}" ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/${DEST_FILE} WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
+    execute_process(COMMAND ${CMAKE_COMMAND} -E create_symlink "${LINK_TARGET}" ${DEST_FILE} WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})
+  endif(WIN32 AND NOT CYGWIN AND NOT MSYS)
+endmacro()
+
+# libpng is a library so default to 'lib'
+if(NOT DEFINED CMAKE_INSTALL_LIBDIR)
+  set(CMAKE_INSTALL_LIBDIR lib)
+endif(NOT DEFINED CMAKE_INSTALL_LIBDIR)
 
 # CREATE PKGCONFIG FILES
 # we use the same files like ./configure, so we have to set its vars
-if(NOT DEFINED CMAKE_INSTALL_LIBDIR)
-  set(CMAKE_INSTALL_LIBDIR ${CMAKE_INSTALL_PREFIX}/lib)
-endif(NOT DEFINED CMAKE_INSTALL_LIBDIR)
-set(prefix      ${CMAKE_INSTALL_PREFIX})
-set(exec_prefix ${CMAKE_INSTALL_PREFIX})
-set(libdir      ${CMAKE_INSTALL_LIBDIR})
-set(includedir  ${CMAKE_INSTALL_PREFIX}/include)
-set(LIBS	"-lz -lm")
-
+# Only do this on Windows for Cygwin - the files don't make much sense outside
+# a UNIX look alike
+if(NOT WIN32 OR CYGWIN OR MINGW)
+  set(prefix      ${CMAKE_INSTALL_PREFIX})
+  set(exec_prefix ${CMAKE_INSTALL_PREFIX})
+  set(libdir      ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR})
+  set(includedir  ${CMAKE_INSTALL_PREFIX}/include)
+  set(LIBS        "-lz -lm")
   configure_file(${CMAKE_CURRENT_SOURCE_DIR}/libpng.pc.in
-  ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}.pc @ONLY)
-install(CODE "
-  MESSAGE(STATUS \"Symlinking ${PNGLIB_NAME}.pc to libpng.pc\")
-  execute_process(COMMAND \${CMAKE_COMMAND} -E create_symlink
-    \"${PNGLIB_NAME}.pc\"
-    \"libpng.pc\")
-")
+    ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}.pc @ONLY)
+  CREATE_SYMLINK(${PNGLIB_NAME}.pc libpng.pc)
+
   configure_file(${CMAKE_CURRENT_SOURCE_DIR}/libpng-config.in
-  ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config @ONLY)
-install(CODE "
-  MESSAGE(STATUS \"Symlinking ${PNGLIB_NAME}-config to libpng-config\")
-  execute_process(COMMAND \${CMAKE_COMMAND} -E create_symlink
-    \"${PNGLIB_NAME}-config\"
-    \"libpng-config\")
-")
+    ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config @ONLY)
+  CREATE_SYMLINK(${PNGLIB_NAME}-config libpng-config)
+endif(NOT WIN32 OR CYGWIN OR MINGW)
 
 # SET UP LINKS
 if(PNG_SHARED)
   set_target_properties(${PNG_LIB_NAME} PROPERTIES
-#   VERSION 15.${PNGLIB_RELEASE}.1.5.0beta47
-    VERSION 15.${PNGLIB_RELEASE}.0
-    SOVERSION 15
+#   VERSION 17.${PNGLIB_RELEASE}.1.7.0beta39
+    VERSION 17.${PNGLIB_RELEASE}.0
+    SOVERSION 17
     CLEAN_DIRECT_OUTPUT 1)
 endif()
 if(PNG_STATIC)
-  if(NOT WIN32)
-    # that's uncool on win32 - it overwrites our static import lib...
+  # MSVC doesn't use a different file extension for shared vs. static
+  # libs.  We are able to change OUTPUT_NAME to remove the _static
+  # for all other platforms.
+  if(NOT MSVC)
     set_target_properties(${PNG_LIB_NAME_STATIC} PROPERTIES
       OUTPUT_NAME ${PNG_LIB_NAME}
       CLEAN_DIRECT_OUTPUT 1)
   endif()
 endif()
 
+# If CMake > 2.4.x, we set a variable used below to export
+# targets to an export file.
+# TODO: Use VERSION_GREATER after our cmake_minimum_required >= 2.6.2
+if(CMAKE_MAJOR_VERSION GREATER 1 AND CMAKE_MINOR_VERSION GREATER 4)
+  set(PNG_EXPORT_RULE EXPORT libpng)
+elseif(CMAKE_MAJOR_VERSION GREATER 2) # future proof
+  set(PNG_EXPORT_RULE EXPORT libpng)
+endif()
+
 # INSTALL
 if(NOT SKIP_INSTALL_LIBRARIES AND NOT SKIP_INSTALL_ALL )
+  install(TARGETS ${PNG_LIB_TARGETS}
+      ${PNG_EXPORT_RULE}
+      RUNTIME DESTINATION bin
+      LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
+      ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR})
+
   if(PNG_SHARED)
-    install(TARGETS ${PNG_LIB_NAME}
-            RUNTIME DESTINATION bin
-	    LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
-	    ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR})
-    install(CODE "
-      MESSAGE(STATUS \"Symlinking ${PNGLIB_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX} to libpng${CMAKE_SHARED_LIBRARY_SUFFIX}\")
-      execute_process(COMMAND \${CMAKE_COMMAND} -E create_symlink
-	      \"${PNGLIB_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX}\"
-	      \"libpng${CMAKE_SHARED_LIBRARY_SUFFIX}\")
-    ")
-    install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng${CMAKE_SHARED_LIBRARY_SUFFIX} DESTINATION ${CMAKE_INSTALL_LIBDIR})
-  endif()
+    # Create a symlink for libpng.dll.a => libpng17.dll.a on Cygwin
+    if(CYGWIN OR MINGW)
+       get_target_property(BUILD_TARGET_LOCATION ${PNG_LIB_NAME} LOCATION_${CMAKE_BUILD_TYPE})
+       CREATE_SYMLINK(${BUILD_TARGET_LOCATION} libpng${CMAKE_IMPORT_LIBRARY_SUFFIX})
+       install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng${CMAKE_IMPORT_LIBRARY_SUFFIX}
+         DESTINATION ${CMAKE_INSTALL_LIBDIR})
+    endif(CYGWIN OR MINGW)
+
+    if(NOT WIN32)
+      get_target_property(BUILD_TARGET_LOCATION ${PNG_LIB_NAME} LOCATION_${CMAKE_BUILD_TYPE})
+      CREATE_SYMLINK(${BUILD_TARGET_LOCATION} libpng${CMAKE_SHARED_LIBRARY_SUFFIX})
+      install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng${CMAKE_SHARED_LIBRARY_SUFFIX}
+         DESTINATION ${CMAKE_INSTALL_LIBDIR})
+    endif(NOT WIN32)
+  endif(PNG_SHARED)
+
   if(PNG_STATIC)
-    install(TARGETS ${PNG_LIB_NAME_STATIC}
-	    LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
-	    ARCHIVE DESTINATION ${CMAKE_INSTALL_LIBDIR})
-    install(CODE "
-      MESSAGE(STATUS \"Symlinking ${PNGLIB_NAME}${CMAKE_STATIC_LIBRARY_SUFFIX} to libpng${CMAKE_STATIC_LIBRARY_SUFFIX}\")
-      execute_process(COMMAND \${CMAKE_COMMAND} -E create_symlink
-	      \"${PNGLIB_NAME}${CMAKE_STATIC_LIBRARY_SUFFIX}\"
-	      \"libpng${CMAKE_STATIC_LIBRARY_SUFFIX}\")
-    ")
-    install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng${CMAKE_STATIC_LIBRARY_SUFFIX} DESTINATION ${CMAKE_INSTALL_LIBDIR})
-  endif()
+    if(NOT WIN32 OR CYGWIN OR MINGW)
+      get_target_property(BUILD_TARGET_LOCATION ${PNG_LIB_NAME_STATIC} LOCATION_${CMAKE_BUILD_TYPE})
+      CREATE_SYMLINK(${BUILD_TARGET_LOCATION} libpng${CMAKE_STATIC_LIBRARY_SUFFIX})
+      install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng${CMAKE_STATIC_LIBRARY_SUFFIX}
+         DESTINATION ${CMAKE_INSTALL_LIBDIR})
+    endif(NOT WIN32 OR CYGWIN OR MINGW)
+ endif()
 endif()
 
 if(NOT SKIP_INSTALL_HEADERS AND NOT SKIP_INSTALL_ALL )
-  install(FILES png.h pngconf.h         DESTINATION include)
-  install(FILES png.h pngconf.h         DESTINATION include/${PNGLIB_NAME})
+  install(FILES ${libpng_public_hdrs}   DESTINATION include)
+  install(FILES ${libpng_public_hdrs}   DESTINATION include/${PNGLIB_NAME})
 endif()
 if(NOT SKIP_INSTALL_EXECUTABLES AND NOT SKIP_INSTALL_ALL )
-  install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/libpng-config DESTINATION bin)
-  install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config
-          DESTINATION bin)
+  if(NOT WIN32 OR CYGWIN OR MINGW)
+    install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/libpng-config DESTINATION bin)
+    install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config
+            DESTINATION bin)
+  endif(NOT WIN32 OR CYGWIN OR MINGW)
 endif()
+
 if(NOT SKIP_INSTALL_FILES AND NOT SKIP_INSTALL_ALL )
   # Install man pages
-  install(FILES libpng.3 libpngpf.3             DESTINATION share/man/man3)
-  install(FILES png.5                           DESTINATION share/man/man5)
+  if(NOT PNG_MAN_DIR)
+    set(PNG_MAN_DIR "share/man")
+  endif()
+  install(FILES libpng.3 libpngpf.3      DESTINATION ${PNG_MAN_DIR}/man3)
+  install(FILES png.5                    DESTINATION ${PNG_MAN_DIR}/man5)
   # Install pkg-config files
-  install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng.pc
-	  DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
-  install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/libpng-config
-          DESTINATION bin)
-  install(FILES ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}.pc
-	  DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
-  install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config
-          DESTINATION bin)
+  if(NOT WIN32 OR CYGWIN OR MINGW)
+    install(FILES ${CMAKE_CURRENT_BINARY_DIR}/libpng.pc
+            DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
+    install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/libpng-config
+            DESTINATION bin)
+    install(FILES ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}.pc
+            DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig)
+    install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/${PNGLIB_NAME}-config
+            DESTINATION bin)
+  endif(NOT WIN32 OR CYGWIN OR MINGW)
 endif()
 
-# what's with libpng.txt and all the extra files?
+# On versions of CMake that support it, create an export file CMake
+# users can include() to import our targets
+if(PNG_EXPORT_RULE AND NOT SKIP_INSTALL_EXPORT AND NOT SKIP_INSTALL_ALL )
+  install(EXPORT libpng DESTINATION lib/libpng FILE lib${PNG_LIB_NAME}.cmake)
+endif()
 
+# what's with libpng-manual.txt and all the extra files?
 
 # UNINSTALL
 # do we need this?
 
-
 # DIST
 # do we need this?
 

INSTALL

@@ -1,5 +1,26 @@
 
-Installing libpng version 1.5.0beta47 - September 11, 2010
+Installing libpng
+
+Contents
+
+   I. Simple installation
+  II. Rebuilding the configure scripts
+ III. Using scripts/makefile*
+  IV. Using cmake
+   V. Directory structure
+  VI. Building with project files
+ VII. Building with makefiles
+VIII. Configuring libpng for 16-bit platforms
+  IX. Configuring for DOS
+   X. Configuring for Medium Model
+  XI. Prepending a prefix to exported symbols
+ XII. Configuring for compiler xxx:
+XIII. Removing unwanted object code
+ XIV. Changes to the build and configuration of libpng in libpng-1.5.x
+  XV. Configuring libpng for multiprocessing
+ XVI. Other sources of information about libpng
+
+I. Simple installation
 
 On Unix/Linux and similar systems, you can simply type
 
@@ -7,20 +28,42 @@ On Unix/Linux and similar systems, you can simply type
     make check
     make install
 
-and ignore the rest of this document.
+and ignore the rest of this document.  "/path" is the path to the directory
+where you want to install the libpng "lib", "include", and "bin"
+subdirectories.
 
-If configure does not work on your system and you have a reasonably
-up-to-date set of tools, running ./autogen.sh before running ./configure
-may fix the problem.  You can also run the individual commands in
-autogen.sh with the --force option, if supported by your version of
-the tools.  If you run 'libtoolize --force', though, this will replace
-the distributed, patched, version of ltmain.sh with an unpatched version
-and your shared library builds may fail to produce libraries with the
-correct version numbers.
+If you downloaded a GIT clone, you will need to run ./autogen.sh before
+running ./configure, to create "configure" and "Makefile.in" which are
+not included in the GIT repository.
+
+Note that "configure" is only included in the "*.tar" distributions and not
+in the "*.zip" or "*.7z" distributions. If you downloaded one of those
+distributions, see "Building with project files" or "Building with makefiles",
+below.
+
+II. Rebuilding the configure scripts
+
+If configure does not work on your system, or if you have a need to
+change configure.ac or Makefile.am, and you have a reasonably
+up-to-date set of tools, running ./autogen.sh in a git clone before
+running ./configure may fix the problem.  To be really sure that you
+aren't using any of the included pre-built scripts, you can do this:
+
+    ./configure --enable-maintainer-mode
+    make maintainer-clean
+    ./autogen.sh --maintainer --clean
+    ./autogen.sh --maintainer
+    ./configure [--prefix=/path] [other options]
+    make
+    make install
+    make check
+
+III. Using scripts/makefile*
 
 Instead, you can use one of the custom-built makefiles in the
 "scripts" directory
 
+    cp scripts/pnglibconf.h.prebuilt pnglibconf.h
     cp scripts/makefile.system makefile
     make test
     make install
@@ -32,8 +75,28 @@ Or you can use one of the "projects" in the "projects" directory.
 
 Before installing libpng, you must first install zlib, if it
 is not already on your system.  zlib can usually be found
-wherever you got libpng.  zlib can be placed in another directory,
-at the same level as libpng.
+wherever you got libpng; otherwise go to http://zlib.net.  You can place
+zlib in in the same directory as libpng or in another directory.
+
+If your system already has a preinstalled zlib you will still need
+to have access to the zlib.h and zconf.h include files that
+correspond to the version of zlib that's installed.
+
+If you wish to test with a particular zlib that is not first in the
+standard library search path, put ZLIBLIB, ZLIBINC, CPPFLAGS, LDFLAGS,
+and LD_LIBRARY_PATH in your environment before running "make test"
+or "make distcheck":
+
+ZLIBLIB=/path/to/lib export ZLIBLIB
+ZLIBINC=/path/to/include export ZLIBINC
+CPPFLAGS="-I$ZLIBINC" export CPPFLAGS
+LDFLAGS="-L$ZLIBLIB" export LDFLAGS
+LD_LIBRARY_PATH="$ZLIBLIB:$LD_LIBRARY_PATH" export LD_LIBRARY_PATH
+
+If you are using one of the makefile scripts, put ZLIBLIB and ZLIBINC
+in your environment and type "make ZLIBLIB=$ZLIBLIB ZLIBINC=$ZLIBINC test".
+
+IV. Using cmake
 
 If you want to use "cmake" (see www.cmake.org), type
 
@@ -41,13 +104,15 @@ If you want to use "cmake" (see www.cmake.org), type
    make
    make install
 
-If your system already has a preinstalled zlib you will still need
-to have access to the zlib.h and zconf.h include files that
-correspond to the version of zlib that's installed.
+As when using the simple configure method described above, "/path" points to
+the installation directory where you want to put the libpng "lib", "include",
+and "bin" subdirectories.
+
+V. Directory structure
 
 You can rename the directories that you downloaded (they
-might be called "libpng-1.5.0beta47" or "libpng15" and "zlib-1.2.3"
-or "zlib123") so that you have directories called "zlib" and "libpng".
+might be called "libpng-x.y.z" or "libpngNN" and "zlib-1.2.8"
+or "zlib128") so that you have directories called "zlib" and "libpng".
 
 Your directory structure should look like this:
 
@@ -55,27 +120,18 @@ Your directory structure should look like this:
       libpng  (this directory)
           INSTALL (this file)
           README
-          *.h
-          *.c
+          *.h, *.c  => libpng source files
           CMakeLists.txt    =>  "cmake" script
           configuration files:
              configure.ac, configure, Makefile.am, Makefile.in,
-             autogen.sh, config.guess, ltmain.sh, missing,
-             aclocal.m4, config.h.in, config.sub,
              autogen.sh, config.guess, ltmain.sh, missing, libpng.pc.in,
              libpng-config.in, aclocal.m4, config.h.in, config.sub,
              depcomp, install-sh, mkinstalldirs, test-pngtest.sh
           contrib
-             gregbook
-             pngminim
-             pngminus
-             pngsuite
-             visupng
+             arm-neon, conftest, examples, gregbook, libtests, pngminim,
+             pngminus, pngsuite, tools, visupng
           projects
-             cbuilder5 (Borland)
-             visualc6 (msvc)
-             visualc71
-             xcode
+             cbuilder5, owatcom, visualc71, vstudio, xcode
           scripts
              makefile.*
              *.def (module definition files)
@@ -83,34 +139,36 @@ Your directory structure should look like this:
           pngtest.png
           etc.
       zlib
-          README
-          *.h
-          *.c
-          contrib
-          etc.
+          README, *.h, *.c contrib, etc.
 
 If the line endings in the files look funny, you may wish to get the other
 distribution of libpng.  It is available in both tar.gz (UNIX style line
 endings) and zip (DOS style line endings) formats.
 
+VI. Building with project files
+
 If you are building libpng with MSVC, you can enter the
-libpng projects\visualc6 or visualc71 directory and follow the instructions
+libpng projects\visualc71 or vstudio directory and follow the instructions
 in README.txt.
 
 Otherwise enter the zlib directory and follow the instructions in zlib/README,
 then come back here and run "configure" or choose the appropriate
 makefile.sys in the scripts directory.
 
+VII. Building with makefiles
+
 Copy the file (or files) that you need from the
 scripts directory into this directory, for example
 
    MSDOS example: copy scripts\makefile.msc makefile
-   UNIX example:    cp scripts/makefile.std makefile
+                  copy scripts\pnglibconf.h.prebuilt pnglibconf.h
+   UNIX example:  cp scripts/makefile.std makefile
+                  cp scripts/pnglibconf.h.prebuilt pnglibconf.h
 
 Read the makefile to see if you need to change any source or
 target directories to match your preferences.
 
-Then read pngconf.h to see if you want to make any configuration
+Then read pnglibconf.dfa to see if you want to make any configuration
 changes.
 
 Then just run "make" which will create the libpng library in
@@ -128,19 +186,217 @@ do that, run "make install" in the zlib directory first if necessary).
 Some also allow you to run "make test-installed" after you have
 run "make install".
 
-If you encounter a compiler error message complaining about the
-lines
+VIII. Configuring libpng for 16-bit platforms
 
-      __png.h__ already includes setjmp.h;
-      __dont__ include it again.;
+You will want to look into zconf.h to tell zlib (and thus libpng) that
+it cannot allocate more then 64K at a time.  Even if you can, the memory
+won't be accessible.  So limit zlib and libpng to 64K by defining MAXSEG_64K.
 
-this means you have compiled another module that includes setjmp.h,
-which is hazardous because the two modules might not include exactly
-the same setjmp.h.  If you are sure that you know what you are doing
-and that they are exactly the same, then you can comment out or
-delete the two lines.  Better yet, use the cexcept interface
-instead, as demonstrated in contrib/visupng of the libpng distribution.
+IX. Configuring for DOS
 
-Further information can be found in the README and libpng.txt
+For DOS users who only have access to the lower 640K, you will
+have to limit zlib's memory usage via a png_set_compression_mem_level()
+call.  See zlib.h or zconf.h in the zlib library for more information.
+
+X. Configuring for Medium Model
+
+Libpng's support for medium model has been tested on most of the popular
+compilers.  Make sure MAXSEG_64K gets defined, USE_FAR_KEYWORD gets
+defined, and FAR gets defined to far in pngconf.h, and you should be
+all set.  Everything in the library (except for zlib's structure) is
+expecting far data.  You must use the typedefs with the p or pp on
+the end for pointers (or at least look at them and be careful).  Make
+note that the rows of data are defined as png_bytepp, which is
+an "unsigned char far * far *".
+
+XI. Prepending a prefix to exported symbols
+
+Starting with libpng-1.6.0, you can configure libpng (when using the
+"configure" script) to prefix all exported symbols by means of the
+configuration option "--with-libpng-prefix=FOO_", where FOO_ can be any
+string beginning with a letter and containing only uppercase
+and lowercase letters, digits, and the underscore (i.e., a C language
+identifier).  This creates a set of macros in pnglibconf.h, so this is
+transparent to applications; their function calls get transformed by
+the macros to use the modified names.
+
+XII. Configuring for compiler xxx:
+
+All includes for libpng are in pngconf.h.  If you need to add, change
+or delete an include, this is the place to do it.
+The includes that are not needed outside libpng are placed in pngpriv.h,
+which is only used by the routines inside libpng itself.
+The files in libpng proper only include pngpriv.h and png.h, which
+in turn includes pngconf.h and, as of libpng-1.5.0, pnglibconf.h.
+As of libpng-1.5.0, pngpriv.h also includes three other private header
+files, pngstruct.h, pnginfo.h, and pngdebug.h, which contain material
+that previously appeared in the public headers.
+
+XIII. Removing unwanted object code
+
+There are a bunch of #define's in pngconf.h that control what parts of
+libpng are compiled.  All the defines end in _SUPPORTED.  If you are
+never going to use a capability, you can change the #define to #undef
+before recompiling libpng and save yourself code and data space, or
+you can turn off individual capabilities with defines that begin with
+PNG_NO_.
+
+In libpng-1.5.0 and later, the #define's are in pnglibconf.h instead.
+
+You can also turn all of the transforms and ancillary chunk capabilities
+off en masse with compiler directives that define
+PNG_NO_READ[or WRITE]_TRANSFORMS, or PNG_NO_READ[or WRITE]_ANCILLARY_CHUNKS,
+or all four, along with directives to turn on any of the capabilities that
+you do want.  The PNG_NO_READ[or WRITE]_TRANSFORMS directives disable the
+extra transformations but still leave the library fully capable of reading
+and writing PNG files with all known public chunks. Use of the
+PNG_NO_READ[or WRITE]_ANCILLARY_CHUNKS directive produces a library
+that is incapable of reading or writing ancillary chunks.  If you are
+not using the progressive reading capability, you can turn that off
+with PNG_NO_PROGRESSIVE_READ (don't confuse this with the INTERLACING
+capability, which you'll still have).
+
+All the reading and writing specific code are in separate files, so the
+linker should only grab the files it needs.  However, if you want to
+make sure, or if you are building a stand alone library, all the
+reading files start with "pngr" and all the writing files start with "pngw".
+The files that don't match either (like png.c, pngtrans.c, etc.)
+are used for both reading and writing, and always need to be included.
+The progressive reader is in pngpread.c
+
+If you are creating or distributing a dynamically linked library (a .so
+or DLL file), you should not remove or disable any parts of the library,
+as this will cause applications linked with different versions of the
+library to fail if they call functions not available in your library.
+The size of the library itself should not be an issue, because only
+those sections that are actually used will be loaded into memory.
+
+XIV. Changes to the build and configuration of libpng in libpng-1.5.x
+
+Details of internal changes to the library code can be found in the CHANGES
+file and in the GIT repository logs.  These will be of no concern to the vast
+majority of library users or builders; however, the few who configure libpng
+to a non-default feature set may need to change how this is done.
+
+There should be no need for library builders to alter build scripts if
+these use the distributed build support - configure or the makefiles -
+however, users of the makefiles may care to update their build scripts
+to build pnglibconf.h where the corresponding makefile does not do so.
+
+Building libpng with a non-default configuration has changed completely.
+The old method using pngusr.h should still work correctly even though the
+way pngusr.h is used in the build has been changed; however, library
+builders will probably want to examine the changes to take advantage of
+new capabilities and to simplify their build system.
+
+A. Specific changes to library configuration capabilities
+
+The exact mechanism used to control attributes of API functions has
+changed.  A single set of operating system independent macro definitions
+is used and operating system specific directives are defined in
+pnglibconf.h
+
+As part of this the mechanism used to choose procedure call standards on
+those systems that allow a choice has been changed.  At present this only
+affects certain Microsoft (DOS, Windows) and IBM (OS/2) operating systems
+running on Intel processors.  As before, PNGAPI is defined where required
+to control the exported API functions; however, two new macros, PNGCBAPI
+and PNGCAPI, are used instead for callback functions (PNGCBAPI) and
+(PNGCAPI) for functions that must match a C library prototype (currently
+only png_longjmp_ptr, which must match the C longjmp function.)  The new
+approach is documented in pngconf.h
+
+Despite these changes, libpng 1.5.0 only supports the native C function
+calling standard on those platforms tested so far (__cdecl on Microsoft
+Windows).  This is because the support requirements for alternative
+calling conventions seem to no longer exist.  Developers who find it
+necessary to set PNG_API_RULE to 1 should advise the mailing list
+(png-mng-implement) of this and library builders who use Openwatcom and
+therefore set PNG_API_RULE to 2 should also contact the mailing list.
+
+B. Changes to the configuration mechanism
+
+Prior to libpng-1.5.0 library builders who needed to configure libpng
+had either to modify the exported pngconf.h header file to add system
+specific configuration or had to write feature selection macros into
+pngusr.h and cause this to be included into pngconf.h by defining
+PNG_USER_CONFIG. The latter mechanism had the disadvantage that an
+application built without PNG_USER_CONFIG defined would see the
+unmodified, default, libpng API and thus would probably fail to link.
+
+These mechanisms still work in the configure build and in any makefile
+build that builds pnglibconf.h, although the feature selection macros
+have changed somewhat as described above.  In 1.5.0, however, pngusr.h is
+processed only once, at the time the exported header file pnglibconf.h is
+built.  pngconf.h no longer includes pngusr.h; therefore, pngusr.h is ignored
+after the build of pnglibconf.h and it is never included in an application
+build.
+
+The formerly used alternative of adding a list of feature macros to the
+CPPFLAGS setting in the build also still works; however, the macros will be
+copied to pnglibconf.h and this may produce macro redefinition warnings
+when the individual C files are compiled.
+
+All configuration now only works if pnglibconf.h is built from
+scripts/pnglibconf.dfa.  This requires the program awk.  Brian Kernighan
+(the original author of awk) maintains C source code of that awk and this
+and all known later implementations (often called by subtly different
+names - nawk and gawk for example) are adequate to build pnglibconf.h.
+The Sun Microsystems (now Oracle) program 'awk' is an earlier version
+and does not work; this may also apply to other systems that have a
+functioning awk called 'nawk'.
+
+Configuration options are now documented in scripts/pnglibconf.dfa.  This
+file also includes dependency information that ensures a configuration is
+consistent; that is, if a feature is switched off, dependent features are
+also switched off.  As a recommended alternative to using feature macros in
+pngusr.h a system builder may also define equivalent options in pngusr.dfa
+(or, indeed, any file) and add that to the configuration by setting
+DFA_XTRA to the file name.  The makefiles in contrib/pngminim illustrate
+how to do this, and also illustrate a case where pngusr.h is still required.
+
+After you have built libpng, the definitions that were recorded in
+pnglibconf.h are available to your application (pnglibconf.h is included
+in png.h and gets installed alongside png.h and pngconf.h in your
+$PREFIX/include directory).  Do not edit pnglibconf.h after you have built
+libpng, because than the settings would not accurately reflect the settings
+that were used to build libpng.
+
+XV. Configuring libpng for multiprocessing
+
+Libpng uses setjmp()/longjmp() for error handling.  Unfortunately setjmp()
+is known to be not thread-safe on some platforms and we don't know of
+any platform where it is guaranteed to be thread-safe.  Therefore, if
+your application is going to be using multiple threads, you should
+configure libpng with PNG_NO_SETJMP in your pngusr.dfa file, with
+-DPNG_NO_SETJMP on your compile line, or with
+
+  #undef PNG_SETJMP_SUPPORTED
+
+in your pnglibconf.h or pngusr.h.
+
+XVI. Other sources of information about libpng:
+
+Further information can be found in the README and libpng-manual.txt
 files, in the individual makefiles, in png.h, and the manual pages
 libpng.3 and png.5.
+
+Using the ./configure script -- 16 December 2002.
+=================================================
+
+The ./configure script should work compatibly with what scripts/makefile.*
+did, however there are some options you might need to add to configure
+explicitly, which previously was done semi-automatically (if you didn't edit
+scripts/makefile.* yourself, that is)
+
+CFLAGS="-Wall -O -funroll-loops \
+-malign-loops=2 -malign-functions=2" ./configure --prefix=/usr/include \
+--with-pkgconfigdir=/usr/lib/pkgconfig --includedir=/usr/include
+
+You can alternatively specify --includedir=/usr/include, /usr/local/include,
+/usr/include/libpng17, or whatever.
+
+If you find that the configure script is out-of-date or is not supporting
+your platform properly, try running autogen.sh to regenerate "configure",
+"Makefile.in", and the other configuration files. Then try configure again.
+

LICENSE

@@ -10,8 +10,8 @@ this sentence.
 
 This code is released under the libpng license.
 
-libpng versions 1.2.6, August 15, 2004, through 1.5.0beta47, September 11, 2010, are
-Copyright (c) 2004, 2006-2010 Glenn Randers-Pehrson, and are
+libpng versions 1.2.6, August 15, 2004, through 1.7.0beta39, November 1, 2014, are
+Copyright (c) 2004, 2006-2014 Glenn Randers-Pehrson, and are
 distributed according to the same disclaimer and license as libpng-1.2.5
 with the following individual added to the list of Contributing Authors
 
@@ -108,4 +108,4 @@ certification mark of the Open Source Initiative.
 
 Glenn Randers-Pehrson
 glennrp at users.sourceforge.net
-September 11, 2010
+November 1, 2014

Makefile.am

@@ -4,23 +4,75 @@
 
 PNGLIB_BASENAME= libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@
 
-# libpng does not follow GNU file name conventions
-AUTOMAKE_OPTIONS = foreign color-tests
+ACLOCAL_AMFLAGS = -I scripts
 
 # test programs - run on make check, make distcheck
-check_PROGRAMS= pngtest pngvalid
+check_PROGRAMS= pngtest pngunknown pngstest pngvalid pngimage
+
+# Utilities - installed
+bin_PROGRAMS= pngfix png-fix-itxt
+
+# This ensures that pnglibconf.h gets built at the start of 'make all' or
+# 'make check', but it does not add dependencies to the individual programs,
+# this is done below.
+#
+# IMPORTANT: always add the object modules of new programs to the list below
+# because otherwise the sequence 'configure; make new-program' will *sometimes*
+# result in the installed (system) pnglibconf.h being used and the result is
+# always wrong and always very confusing.
+BUILT_SOURCES = pnglibconf.h
+
 pngtest_SOURCES = pngtest.c
 pngtest_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
-pngvalid_SOURCES = pngvalid.c
+
+pngvalid_SOURCES = contrib/libtests/pngvalid.c
 pngvalid_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
-TESTS = test-pngtest.sh pngvalid
-TESTS_ENVIRONMENT= srcdir=$(srcdir) 
+
+pngstest_SOURCES = contrib/libtests/pngstest.c
+pngstest_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
+
+pngunknown_SOURCES = contrib/libtests/pngunknown.c
+pngunknown_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
+
+pngimage_SOURCES = contrib/libtests/pngimage.c
+pngimage_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
+
+pngfix_SOURCES = contrib/tools/pngfix.c
+pngfix_LDADD = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
+
+png_fix_itxt_SOURCES = contrib/tools/png-fix-itxt.c
+
+# Generally these are single line shell scripts to run a test with a particular
+# set of parameters:
+TESTS =\
+   tests/pngtest\
+   tests/pngvalid-gamma-16-to-8 tests/pngvalid-gamma-alpha-mode\
+   tests/pngvalid-gamma-background tests/pngvalid-gamma-expand16-alpha-mode\
+   tests/pngvalid-gamma-expand16-background\
+   tests/pngvalid-gamma-expand16-transform tests/pngvalid-gamma-sbit\
+   tests/pngvalid-gamma-threshold tests/pngvalid-gamma-transform\
+   tests/pngvalid-progressive-interlace-size\
+   tests/pngvalid-progressive-interlace-standard\
+   tests/pngvalid-progressive-interlace-transform\
+   tests/pngvalid-progressive-standard tests/pngvalid-standard\
+   tests/pngstest-0g01 tests/pngstest-0g02 tests/pngstest-0g04\
+   tests/pngstest-0g08 tests/pngstest-0g16 tests/pngstest-2c08\
+   tests/pngstest-2c16 tests/pngstest-3p01 tests/pngstest-3p02\
+   tests/pngstest-3p04 tests/pngstest-3p08 tests/pngstest-4a08\
+   tests/pngstest-4a16 tests/pngstest-6a08 tests/pngstest-6a16\
+   tests/pngstest-error tests/pngunknown-IDAT\
+   tests/pngunknown-discard tests/pngunknown-if-safe tests/pngunknown-sAPI\
+   tests/pngunknown-sTER tests/pngunknown-save tests/pngunknown-vpAg\
+   tests/pngimage-quick tests/pngimage-full
+
+# These tests are expected, and required, to fail:
+XFAIL_TESTS = tests/pngstest-error
 
 # man pages
 dist_man_MANS= libpng.3 libpngpf.3 png.5
 
 # generate the -config scripts if required
-binconfigs= libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@-config 
+binconfigs= libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@-config
 EXTRA_SCRIPTS= libpng-config libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@-config
 bin_SCRIPTS= @binconfigs@
 
@@ -30,21 +82,29 @@ lib_LTLIBRARIES=libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.la
 libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES = png.c pngerror.c\
 	pngget.c pngmem.c pngpread.c pngread.c pngrio.c pngrtran.c pngrutil.c\
 	pngset.c pngtrans.c pngwio.c pngwrite.c pngwtran.c pngwutil.c\
-	png.h pngconf.h pngdebug.h pnginfo.h pngpriv.h pngstruct.h
+	png.h pngconf.h pngdebug.h pnginfo.h pngpriv.h pngstruct.h pngusr.dfa
+
+if PNG_ARM_NEON
+libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES += arm/arm_init.c\
+	arm/filter_neon.S arm/filter_neon_intrinsics.c
+endif
 
 nodist_libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_SOURCES = pnglibconf.h
 
-libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_CPPFLAGS = @LIBPNG_DEFINES@
-
 libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_LDFLAGS = -no-undefined -export-dynamic \
 	-version-number @PNGLIB_MAJOR@@PNGLIB_MINOR@:@PNGLIB_RELEASE@:0
 
 if HAVE_LD_VERSION_SCRIPT
-  # Versioned symbols and restricted exports
+#   Versioned symbols and restricted exports
+if HAVE_SOLARIS_LD
+  libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_LDFLAGS += -Wl,-M -Wl,libpng.vers
+else
   libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_LDFLAGS += -Wl,--version-script=libpng.vers
+endif
+
   libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_DEPENDENCIES = libpng.vers
 else
-  # Only restricted exports when possible
+#   Only restricted exports when possible
   libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_LDFLAGS += -export-symbols libpng.sym
   libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_DEPENDENCIES = libpng.sym
 endif
@@ -59,24 +119,39 @@ nodist_pkginclude_HEADERS= pnglibconf.h
 pkgconfigdir = @pkgconfigdir@
 pkgconfig_DATA = libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.pc
 
-#extra source distribution files.
+# Extra source distribution files, '${srcdir}' is used below to stop build files
+# from those directories being included.  This only works if the configure is
+# not done in the source directory!
 EXTRA_DIST= \
 	ANNOUNCE CHANGES INSTALL LICENSE README TODO \
 	pngtest.png pngbar.png pngnow.png pngbar.jpg autogen.sh \
 	${srcdir}/contrib ${srcdir}/projects ${srcdir}/scripts \
-	$(TESTS) \
-	CMakeLists.txt example.c libpng-@PNGLIB_VERSION@.txt
+	$(TESTS) $(XFAIL_TESTS) tests/pngstest \
+	CMakeLists.txt example.c libpng-manual.txt
 
 SCRIPT_CLEANFILES=scripts/*.out scripts/*.chk scripts/pnglibconf.dfn
 
-CLEANFILES= dfn.c dfn?.out pngout.png libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.pc \
+CLEANFILES= *.tf? pngout.png libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@.pc \
 	libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@-config libpng.vers libpng.sym \
-	pngwin.def check.new pnglibconf.* symbols.new \
+	check.new pnglibconf.* pngprefix.h symbols.new pngtest-log.txt \
 	$(SCRIPT_CLEANFILES)
 
 MAINTAINERCLEANFILES = Makefile.in aclocal.m4 config.guess config.h.in \
 config.sub configure depcomp install-sh ltmain.sh missing
 
+# PNG_COPTS give extra options for the C compiler to be used on all compilation
+# steps (unless targe_CFLAGS is specified; that will take precedence over
+# AM_CFLAGS)
+PNG_COPTS = @PNG_COPTS@
+AM_CFLAGS = ${PNG_COPTS}
+
+# DFNCPP is normally just "CPP -E" - the C preprocessor - but on Solaris and
+# maybe other operating systems (NeXT?) the C preprocessor selected by configure
+# checks input tokens for validity - effectively it performs part of the ANSI-C
+# parsing - and therefore fails with the .df files.  configure.ac has special
+# checks for this and sets DFNCPP appropriately.
+DFNCPP = @DFNCPP@
+
 SUFFIXES = .chk .dfn .out
 
 $(PNGLIB_BASENAME).pc: libpng.pc
@@ -86,7 +161,11 @@ $(PNGLIB_BASENAME)-config: libpng-config
 	cp libpng-config $@
 
 scripts/sym.out scripts/vers.out: png.h pngconf.h pnglibconf.h
-scripts/symbols.out scripts/pngwin.out: png.h pngconf.h scripts/pnglibconf.h.prebuilt
+scripts/prefix.out: png.h pngconf.h pnglibconf.out
+scripts/symbols.out: png.h pngconf.h $(srcdir)/scripts/pnglibconf.h.prebuilt
+scripts/intprefix.out: pnglibconf.h
+
+pnglibconf.out: config.h
 
 libpng.sym: scripts/sym.out
 	rm -f $@
@@ -94,107 +173,197 @@ libpng.sym: scripts/sym.out
 libpng.vers: scripts/vers.out
 	rm -f $@
 	cp $? $@
-pngwin.def: scripts/pngwin.out
+
+if DO_PNG_PREFIX
+# Rename functions in scripts/prefix.out with a PNG_PREFIX prefix.
+# Rename macros in scripts/macro.lst from PNG_PREFIXpng_ to PNG_ (the actual
+# implementation of the macro).
+pnglibconf.h: pnglibconf.out scripts/prefix.out scripts/macro.lst config.h
 	rm -f $@
-	cp $? $@
+	$(AWK) 's==0 && NR>1{print prev}\
+	   s==0{prev=$$0}\
+	   s==1{print "#define", $$1, "@PNG_PREFIX@" $$1}\
+	   s==2{print "#define @PNG_PREFIX@png_" $$1, "PNG_" $$1}\
+	   END{print prev}' s=0 pnglibconf.out s=1 scripts/prefix.out\
+	   s=2 ${srcdir}/scripts/macro.lst >pnglibconf.tf8
+	mv pnglibconf.tf8 $@
+
+pngprefix.h: scripts/intprefix.out
+	rm -f pngprefix.tf1
+	$(AWK) '{print "#define", $$1, "@PNG_PREFIX@" $$1}' $? >pngprefix.tf1
+	mv pngprefix.tf1 $@
+else
 pnglibconf.h: pnglibconf.out
 	rm -f $@
 	cp $? $@
-scripts/pnglibconf.h.prebuilt:
+
+pngprefix.h: # is empty
+	:>$@
+endif
+
+$(srcdir)/scripts/pnglibconf.h.prebuilt:
+	@echo "Attempting to build $@" >&2
 	@echo "This is a machine generated file, but if you want to make" >&2
-	@echo "a new one simply make 'scripts/pnglibconf.out' and copy that" >&2
+	@echo "a new one simply make 'scripts/pnglibconf.out', copy that" >&2
+	@echo "AND set PNG_ZLIB_VERNUM to 0 (you MUST do this)" >&2
 	@exit 1
 
 # The following is necessary to ensure that the local pnglibconf.h is used, not
 # an installed one (this can happen immediately after on a clean system if
-# 'make test' is the first thing the user does.)
-pngvalid.o pngtest.o: pnglibconf.h
+# 'make test' is the first thing the user does.)  Only files which include
+# one of the png source files (typically png.h or pngpriv.h) need to be listed
+# here:
+pngtest.o: pnglibconf.h
 
+contrib/libtests/makepng.o: pnglibconf.h
+contrib/libtests/pngstest.o: pnglibconf.h
+contrib/libtests/pngunknown.o: pnglibconf.h
+contrib/libtests/pngimage.o: pnglibconf.h
+contrib/libtests/pngvalid.o: pnglibconf.h
+contrib/libtests/readpng.o: pnglibconf.h
+contrib/libtests/tarith.o: pnglibconf.h
+contrib/libtests/timepng.o: pnglibconf.h
+
+contrib/tools/makesRGB.o: pnglibconf.h
+contrib/tools/pngfix.o: pnglibconf.h
+
+# We must use -DPNG_NO_USE_READ_MACROS here even when the library may actually
+# be built with PNG_USE_READ_MACROS; this prevents the read macros from
+# interfering with the symbol file format.
 SYMBOL_CFLAGS = -DPNGLIB_LIBNAME='PNG@PNGLIB_MAJOR@@PNGLIB_MINOR@_0'\
 		-DPNGLIB_VERSION='@PNGLIB_VERSION@'\
-		-DSYMBOL_PREFIX='$(SYMBOL_PREFIX)'
+		-DSYMBOL_PREFIX='$(SYMBOL_PREFIX)'\
+		-DPNG_NO_USE_READ_MACROS -DPNG_BUILDING_SYMBOL_TABLE
+
+if DO_PNG_PREFIX
+SYMBOL_CFLAGS += -DPNG_PREFIX='@PNG_PREFIX@'
+endif
 
 .dfn.out:
-	rm -f $@ dfn.c dfn?.out
-	test -d scripts || mkdir scripts
-	echo '#include "$<"' >dfn.c
-	$(CPP) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES) @LIBPNG_DEFINES@\
-	    $(CPPFLAGS) $(SYMBOL_CFLAGS) dfn.c >dfn1.out
-	$(SED) -n -e 's|^.*PNG_DEFN_MAGIC-\(.*\)-PNG_DEFN_END.*$$|\1|p'\
-	    dfn1.out >dfn2.out
-	$(SED) -e 's| *@@@ *||g' -e 's| *$$||' dfn2.out >dfn3.out
-	rm -f dfn.c dfn[12].out
-	mv dfn3.out $@
+	rm -f $@ $*.c $*.tf[12]
+	test -d scripts || mkdir scripts || test -d scripts
+	echo '#include "$<"' >$*.c
+	$(DFNCPP) $(DEFS) $(DEFAULT_INCLUDES) $(INCLUDES)\
+	    $(CPPFLAGS) $(SYMBOL_CFLAGS) $*.c > $*.tf1
+	$(AWK) -f "${srcdir}/scripts/dfn.awk" out="$*.tf2" $*.tf1 1>&2
+	rm -f $*.c $*.tf1
+	mv $*.tf2 $@
 
 # The .dfn file for pnglibconf.h is machine generated
-pnglibconf.dfn: scripts/pnglibconf.dfa scripts/options.awk
-	rm -f $@ dfn?.out
-	$(AWK) -f ${srcdir}/scripts/options.awk out=dfn1.out\
-	    ${srcdir}/scripts/pnglibconf.dfa $(DFA_XTRA) 1>&2
-	$(AWK) -f ${srcdir}/scripts/options.awk out=dfn2.out dfn1.out 1>&2
-	rm dfn1.out
-	mv dfn2.out $@
+pnglibconf.dfn: scripts/pnglibconf.dfa scripts/options.awk pngconf.h pngusr.dfa $(DFA_XTRA)
+	rm -f $@ $*.tf[45]
+	$(AWK) -f ${srcdir}/scripts/options.awk out=$*.tf4 version=search\
+	    ${srcdir}/pngconf.h ${srcdir}/scripts/pnglibconf.dfa\
+	    ${srcdir}/pngusr.dfa $(DFA_XTRA) 1>&2
+	$(AWK) -f ${srcdir}/scripts/options.awk out=$*.tf5 $*.tf4 1>&2
+	rm $*.tf4
+	mv $*.tf5 $@
 
 # Symbol checks (.def and .out files should match)
-scripts/pngwin.chk: scripts/checksym.awk scripts/pngwin.def scripts/pngwin.out
 scripts/symbols.chk: scripts/checksym.awk scripts/symbols.def scripts/symbols.out
+
 .out.chk:
-	rm -f $@ symbols.new
+	rm -f $@ $*.new
 	$(AWK) -f ${srcdir}/scripts/checksym.awk ${srcdir}/scripts/${*F}.def\
-	    $< >&2
-	mv symbols.new $@
+	    of="$*.new" $< >&2
+	mv $*.new $@
 
 # used on demand to regenerate the standard header, CPPFLAGS should
 # be empty - no non-standard defines
-scripts/pnglibconf.dfn: scripts/pnglibconf.dfa scripts/options.awk
-	rm -f $@ dfn?.out
-	test -z "$(CPPFLAGS)" 
+scripts/pnglibconf.dfn: scripts/pnglibconf.dfa scripts/options.awk pngconf.h
+	rm -f $@ pnglibconf.tf[67]
+	test -z "$(CPPFLAGS)"
 	echo "com @PNGLIB_VERSION@ STANDARD API DEFINITION" |\
-	$(AWK) -f ${srcdir}/scripts/options.awk out=dfn1.out\
-	    logunsupported=1 - ${srcdir}/scripts/pnglibconf.dfa 1>&2
-	$(AWK) -f ${srcdir}/scripts/options.awk out=dfn2.out dfn1.out 1>&2
-	rm dfn1.out
-	mv dfn2.out $@
+	$(AWK) -f ${srcdir}/scripts/options.awk out=pnglibconf.tf6\
+	    logunsupported=1 version=search ${srcdir}/pngconf.h -\
+	    ${srcdir}/scripts/pnglibconf.dfa 1>&2
+	$(AWK) -f ${srcdir}/scripts/options.awk out=pnglibconf.tf7\
+	    pnglibconf.tf6 1>&2
+	rm pnglibconf.tf6
+	mv pnglibconf.tf7 $@
 
 $(libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@_la_OBJECTS): png.h pngconf.h \
-	pnglibconf.h pngpriv.h pngdebug.h pnginfo.h pngstruct.h
+	pnglibconf.h pngpriv.h pngdebug.h pnginfo.h pngstruct.h pngprefix.h
 
 test: check-am
 
 # Extra checks
-check: scripts/symbols.chk scripts/pngwin.chk
+check: scripts/symbols.chk
 
 # Don't distribute the generated script files
 dist-hook:
 	cd '$(top_distdir)'; rm -f $(SCRIPT_CLEANFILES)
 
-# install the .../include headers as links to the new ones
-install-data-hook:
-	cd $(DESTDIR)$(includedir); rm -f png.h pngconf.h pnglibconf.h
-	cd $(DESTDIR)$(includedir); $(LN_S) $(PNGLIB_BASENAME)/png.h png.h
-	cd $(DESTDIR)$(includedir); $(LN_S) $(PNGLIB_BASENAME)/pngconf.h \
-		pngconf.h
-	cd $(DESTDIR)$(includedir); $(LN_S) $(PNGLIB_BASENAME)/pnglibconf.h \
-		pnglibconf.h
-	cd $(DESTDIR)$(pkgconfigdir); rm -f libpng.pc
-	cd $(DESTDIR)$(pkgconfigdir); $(LN_S) $(PNGLIB_BASENAME).pc libpng.pc
+# Make links between installed files with release-specific names and the generic
+# file names.  If this install rule is run the generic names will be deleted and
+# recreated - this has obvious issues for systems with multiple installations.
 
-# do evil things to libpng to cause libpng@PNGLIB_MAJOR@@PNGLIB_MINOR@ to be used
-install-exec-hook:
-	cd $(DESTDIR)$(bindir); rm -f libpng-config
-	cd $(DESTDIR)$(bindir); $(LN_S) $(PNGLIB_BASENAME)-config libpng-config
-	@set -x;\
-	cd $(DESTDIR)$(libdir);\
-	for ext in a la so so.@PNGLIB_MAJOR@@PNGLIB_MINOR@.@PNGLIB_RELEASE@ sl dylib dll.a; do\
-		rm -f libpng.$$ext;\
-                if test -f $(PNGLIB_BASENAME).$$ext; then\
-                       $(LN_S) $(PNGLIB_BASENAME).$$ext libpng.$$ext;\
-                fi;\
+install-header-links:
+	@set -ex; cd '$(DESTDIR)$(includedir)'; for f in $(HEADERS); do \
+	   rm -f "$$f"; $(LN_S) "$(PNGLIB_BASENAME)/$$f" "$$f"; done
+
+uninstall-header-links:
+	cd '$(DESTDIR)$(includedir)'; rm -f $(HEADERS)
+
+install-libpng-pc:
+	@set -ex; cd '$(DESTDIR)$(pkgconfigdir)'; rm -f libpng.pc; \
+	   $(LN_S) '$(PNGLIB_BASENAME).pc' libpng.pc
+
+uninstall-libpng-pc:
+	rm -f '$(DESTDIR)$(pkgconfigdir)/libpng.pc'
+
+# EXT_LIST is a list of the possibly library directory extensions, this exists
+# because we can't find a good way of discovering the file extensions that are
+# actually installed on a given system, so instead we check for every extension
+# we have seen.
+
+EXT_LIST = a dll.a so so.@PNGLIB_MAJOR@@PNGLIB_MINOR@.@PNGLIB_RELEASE@ la sl dylib
+
+install-library-links:
+	@set -x; cd '$(DESTDIR)$(libdir)';\
+	for ext in $(EXT_LIST); do\
+	   rm -f "libpng.$$ext";\
+           if test -f "$(PNGLIB_BASENAME).$$ext"; then\
+              $(LN_S) "$(PNGLIB_BASENAME).$$ext" "libpng.$$ext" || exit 1;\
+           fi;\
 	done
 
-uninstall-hook:
-	cd $(DESTDIR)$(includedir); rm -f png.h pngconf.h pnglibconf.h
-	rm -f $(DESTDIR)$(pkgconfigdir)/libpng.pc
-	rm -f $(DESTDIR)$(bindir)/libpng-config
-	rm -f $(DESTDIR)$(libdir)/libpng.a
-	rm -f $(DESTDIR)$(libdir)/libpng.la
+uninstall-library-links:
+	@set -x; cd '$(DESTDIR)$(libdir)'; for ext in $(EXT_LIST); do\
+	   rm -f "libpng.$$ext"; done
+
+install-libpng-config:
+	@set -ex; cd '$(DESTDIR)$(bindir)'; rm -f libpng-config; \
+	   $(LN_S) '$(PNGLIB_BASENAME)-config' libpng-config
+
+uninstall-libpng-config:
+	rm -f '$(DESTDIR)$(bindir)/libpng-config'
+
+if DO_INSTALL_LINKS
+# If --enable-unversioned-links is specified the header and lib file links
+# will be automatically made on a 'make install':
+
+install-data-hook: install-header-links
+uninstall-hook: uninstall-header-links
+install-exec-hook: install-library-links
+uninstall-hook: uninstall-library-links
+endif
+
+if DO_INSTALL_LIBPNG_PC
+# Likewise, --install-pc causes libpng.pc to be constructed:
+
+install-data-hook: install-libpng-pc
+uninstall-hook: uninstall-libpng-pc
+endif
+
+if DO_INSTALL_LIBPNG_CONFIG
+# And --install-config:
+
+install-exec-hook: install-libpng-config
+uninstall-hook: uninstall-libpng-config
+endif
+
+# The following addition ensures that 'make all' always builds the test programs
+# too.  It used to, but some change either in libpng or configure stopped this
+# working.
+all-am: $(check_PROGRAMS)

README

@@ -1,11 +1,11 @@
-README for libpng version 1.5.0beta47 - September 11, 2010 (shared library 15.0)
+README for libpng version 1.7.0beta39 - November 1, 2014 (shared library 17.0)
 See the note about version numbers near the top of png.h
 
 See INSTALL for instructions on how to install libpng.
 
-Libpng comes in several distribution formats.  Get libpng-*.tar.gz,
-libpng-*.tar.xz or libpng-*.tar.bz2 if you want UNIX-style line endings
-in the text files, or lpng*.zip if you want DOS-style line endings.
+Libpng comes in several distribution formats.  Get libpng-*.tar.gz or
+libpng-*.tar.xz or if you want UNIX-style line endings in the text files,
+or lpng*.7z or lpng*.zip if you want DOS-style line endings.
 
 Version 0.89 was the first official release of libpng.  Don't let the
 fact that it's the first release fool you.  The libpng library has been in
@@ -23,18 +23,25 @@ earlier versions if you are using a shared library.  The type of the
 png_uint_32, which will affect shared-library applications that use
 this function.
 
-To avoid problems with changes to the internals of png_info_struct,
+To avoid problems with changes to the internals of png info_struct,
 new APIs have been made available in 0.95 to avoid direct application
 access to info_ptr.  These functions are the png_set_<chunk> and
 png_get_<chunk> functions.  These functions should be used when
 accessing/storing the info_struct data, rather than manipulating it
 directly, to avoid such problems in the future.
 
-It is important to note that the APIs do not make current programs
+It is important to note that the APIs did not make current programs
 that access the info struct directly incompatible with the new
-library.  However, it is strongly suggested that new programs use
-the new APIs (as shown in example.c and pngtest.c), and older programs
-be converted to the new format, to facilitate upgrades in the future.
+library, through libpng-1.2.x.  In libpng-1.4.x, which was meant to
+be a transitional release, members of the png_struct and the
+info_struct can still be accessed, but the compiler will issue a
+warning about deprecated usage.  Since libpng-1.5.0, direct access
+to these structs is not allowed, and the definitions of the structs
+reside in private pngstruct.h and pnginfo.h header files that are not
+accessible to applications.  It is strongly suggested that new
+programs use the new APIs (as shown in example.c and pngtest.c), and
+older programs be converted to the new format, to facilitate upgrades
+in the future.
 ****
 
 Additions since 0.90 include the ability to compile libpng as a
@@ -58,7 +65,7 @@ on the PNG-implement mailing list and not on material submitted
 privately to Guy, Andreas, or Glenn.  They will forward any good
 suggestions to the list.
 
-For a detailed description on using libpng, read libpng.txt.  For
+For a detailed description on using libpng, read libpng-manual.txt.  For
 examples of libpng in a program, see example.c and pngtest.c.  For usage
 information and restrictions (what little they are) on libpng, see
 png.h.  For a description on using zlib (the compression library used by
@@ -77,17 +84,21 @@ compression library that is useful for more things than just PNG files.
 You can use zlib as a drop-in replacement for fread() and fwrite() if
 you are so inclined.
 
-zlib should be available at the same place that libpng is, or at.
-ftp://ftp.info-zip.org/pub/infozip/zlib
+zlib should be available at the same place that libpng is, or at zlib.net.
 
 You may also want a copy of the PNG specification.  It is available
 as an RFC, a W3C Recommendation, and an ISO/IEC Standard.  You can find
 these at http://www.libpng.org/pub/png/documents/
 
 This code is currently being archived at libpng.sf.net in the
-[DOWNLOAD] area, and on CompuServe, Lib 20 (PNG SUPPORT)
-at GO GRAPHSUP.  If you can't find it in any of those places,
-e-mail me, and I'll help you find it.
+[DOWNLOAD] area, and at ftp://ftp.simplesystems.org.  If you can't find it
+in any of those places, e-mail me, and I'll help you find it.
+
+I am not a lawyer, but I believe that the Export Control Classification
+Number (ECCN) for libpng is EAR99, which means not subject to export
+controls or International Traffic in Arms Regulations (ITAR) because it
+is open source, publicly available software, that does not contain any
+encryption software.  See the EAR, paragraphs 734.3(b)(3) and 734.7(b).
 
 If you have any code changes, requests, problems, etc., please e-mail
 them to me.  Also, I'd appreciate any make files or project files,
@@ -105,7 +116,7 @@ based in a large way on Guy's and Andreas' earlier work), and the PNG
 development group.
 
 Send comments/corrections/commendations to png-mng-implement at
-lists.sourceforge.net (subscription required; visit 
+lists.sourceforge.net (subscription required; visit
 https://lists.sourceforge.net/lists/listinfo/png-mng-implement
 to subscribe) or to glennrp at users.sourceforge.net
 
@@ -114,19 +125,16 @@ given in previous versions of this document.  He and Andreas will
 read mail addressed to the png-implement list, however.
 
 Please do not send general questions about PNG.  Send them to
-the (png-list at ccrc.wustl.edu, subscription required, write to
-majordomo at ccrc.wustl.edu with "subscribe png-list" in your message).
-On the other hand,
-please do not send libpng questions to that address, send them to me
-or to the png-implement list.  I'll
-get them in the end anyway.  If you have a question about something
+png-mng-misc at lists.sf.net (subscription required; visit
+https://lists.sourceforge.net/lists/listinfo/png-mng-misc to
+subscribe).  If you have a question about something
 in the PNG specification that is related to using libpng, send it
 to me.  Send me any questions that start with "I was using libpng,
 and ...".  If in doubt, send questions to me.  I'll bounce them
 to others, if necessary.
 
 Please do not send suggestions on how to change PNG.  We have
-been discussing PNG for nine years now, and it is official and
+been discussing PNG for nineteen years now, and it is official and
 finished.  If you have suggestions for libpng, however, I'll
 gladly listen.  Even if your suggestion is not used immediately,
 it may be used later.
@@ -141,8 +149,8 @@ Files in this distribution:
       TODO          =>  Things not implemented in the current library
       Y2KINFO       =>  Statement of Y2K compliance
       example.c     =>  Example code for using libpng functions
-      libpng.3      =>  manual page for libpng (includes libpng.txt)
-      libpng.txt    =>  Description of libpng and its functions
+      libpng.3      =>  manual page for libpng (includes libpng-manual.txt)
+      libpng-manual.txt  =>  Description of libpng and its functions
       libpngpf.3    =>  manual page for libpng's private functions
       png.5         =>  manual page for the PNG format
       png.c         =>  Basic interface functions common to library
@@ -170,23 +178,25 @@ Files in this distribution:
       pngwrite.c    =>  High-level write functions
       pngwtran.c    =>  Write data transformations
       pngwutil.c    =>  Write utility functions
+      arm           =>  Contains optimized code for the ARM platform
       contrib       =>  Contributions
+       examples         =>  Example programs
        gregbook         =>  source code for PNG reading and writing, from
                             Greg Roelofs' "PNG: The Definitive Guide",
                             O'Reilly, 1999
-       msvctest     =>  Builds and runs pngtest using a MSVC workspace
-       pngminus     =>  Simple pnm2png and png2pnm programs
-       pngsuite     =>  Test images
-       visupng      =>  Contains a MSVC workspace for VisualPng
+       libtests         =>  Test programs
+       pngminim         =>  Minimal decoder, encoder, and progressive decoder
+                            programs demonstrating use of pngusr.dfa
+       pngminus         =>  Simple pnm2png and png2pnm programs
+       pngsuite         =>  Test images
+       tools            =>  Various tools
+       visupng          =>  Contains a MSVC workspace for VisualPng
       projects      =>  Contains project files and workspaces for
                         building a DLL
-       cbuilder5        =>  Contains a Borland workspace for building
-                            libpng and zlib
-       visualc6         =>  Contains a Microsoft Visual C++ (MSVC)
-                            workspace for building libpng and zlib
+       owatcom          =>  Contains a WATCOM project for building libpng
        visualc71        =>  Contains a Microsoft Visual C++ (MSVC)
                             workspace for building libpng and zlib
-       xcode            =>  Contains an Apple xcode
+       vstudio          =>  Contains a Microsoft Visual C++ (MSVC)
                             workspace for building libpng and zlib
       scripts       =>  Directory containing scripts for building libpng:
                             (see scripts/README.txt for the list of scripts)

TODO

@@ -6,10 +6,12 @@ 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.
+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).
+Make profile checking optional via a png_set_something() call.
 Man pages for function calls.
 Better documentation.
 Better filter selection

arm/arm_init.c

@@ -0,0 +1,133 @@
+
+/* arm_init.c - NEON optimised filter functions
+ *
+ * Copyright (c) 2014 Glenn Randers-Pehrson
+ * Written by Mans Rullgard, 2011.
+ * Last changed in libpng 1.6.10 [March 6, 2014]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ */
+/* Below, after checking __linux__, various non-C90 POSIX 1003.1 functions are
+ * called.
+ */
+#define _POSIX_SOURCE 1
+
+#include "../pngpriv.h"
+
+#ifdef PNG_READ_SUPPORTED
+#if PNG_ARM_NEON_OPT > 0
+#ifdef PNG_ARM_NEON_CHECK_SUPPORTED /* Do run-time checks */
+/* WARNING: it is strongly recommended that you do not build libpng with
+ * run-time checks for CPU features if at all possible.  In the case of the ARM
+ * NEON instructions there is no processor-specific way of detecting the
+ * presence of the required support, therefore run-time detection is extremely
+ * OS specific.
+ *
+ * You may set the macro PNG_ARM_NEON_FILE to the file name of file containing
+ * a fragment of C source code which defines the png_have_neon function.  There
+ * are a number of implementations in contrib/arm-neon, but the only one that
+ * has partial support is contrib/arm-neon/linux.c - a generic Linux
+ * implementation which reads /proc/cpufino.
+ */
+#ifndef PNG_ARM_NEON_FILE
+#  ifdef __linux__
+#     define PNG_ARM_NEON_FILE "contrib/arm-neon/linux.c"
+#  endif
+#endif
+
+#ifdef PNG_ARM_NEON_FILE
+
+#include <signal.h> /* for sig_atomic_t */
+static int png_have_neon(png_structp png_ptr);
+#include PNG_ARM_NEON_FILE
+
+#else  /* PNG_ARM_NEON_FILE */
+#  error "PNG_ARM_NEON_FILE undefined: no support for run-time ARM NEON checks"
+#endif /* PNG_ARM_NEON_FILE */
+#endif /* PNG_ARM_NEON_CHECK_SUPPORTED */
+
+#ifndef PNG_ALIGNED_MEMORY_SUPPORTED
+#  error "ALIGNED_MEMORY is required; set: -DPNG_ALIGNED_MEMORY_SUPPORTED"
+#endif
+
+void
+png_init_filter_functions_neon(png_structp pp, unsigned int bpp)
+{
+   /* The switch statement is compiled in for ARM_NEON_API, the call to
+    * png_have_neon is compiled in for ARM_NEON_CHECK.  If both are defined
+    * the check is only performed if the API has not set the NEON option on
+    * or off explicitly.  In this case the check controls what happens.
+    *
+    * If the CHECK is not compiled in and the option is UNSET the behavior prior
+    * to 1.6.7 was to use the NEON code - this was a bug caused by having the
+    * wrong order of the 'ON' and 'default' cases.  UNSET now defaults to OFF,
+    * as documented in png.h
+    */
+#ifdef PNG_ARM_NEON_API_SUPPORTED
+   switch ((pp->options >> PNG_ARM_NEON) & 3)
+   {
+      case PNG_OPTION_UNSET:
+         /* Allow the run-time check to execute if it has been enabled -
+          * thus both API and CHECK can be turned on.  If it isn't supported
+          * this case will fall through to the 'default' below, which just
+          * returns.
+          */
+#endif /* PNG_ARM_NEON_API_SUPPORTED */
+#ifdef PNG_ARM_NEON_CHECK_SUPPORTED
+         {
+            static volatile sig_atomic_t no_neon = -1; /* not checked */
+
+            if (no_neon < 0)
+               no_neon = !png_have_neon(pp);
+
+            if (no_neon)
+               return;
+         }
+#ifdef PNG_ARM_NEON_API_SUPPORTED
+         break;
+#endif
+#endif /* PNG_ARM_NEON_CHECK_SUPPORTED */
+
+#ifdef PNG_ARM_NEON_API_SUPPORTED
+      default: /* OFF or INVALID */
+         return;
+
+      case PNG_OPTION_ON:
+         /* Option turned on */
+         break;
+   }
+#endif
+
+   /* IMPORTANT: any new external functions used here must be declared using
+    * PNG_INTERNAL_FUNCTION in ../pngpriv.h.  This is required so that the
+    * 'prefix' option to configure works:
+    *
+    *    ./configure --with-libpng-prefix=foobar_
+    *
+    * Verify you have got this right by running the above command, doing a build
+    * and examining pngprefix.h; it must contain a #define for every external
+    * function you add.  (Notice that this happens automatically for the
+    * initialization function.)
+    */
+   pp->read_filter[PNG_FILTER_VALUE_UP-1] = png_read_filter_row_up_neon;
+
+   if (bpp == 3)
+   {
+      pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub3_neon;
+      pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg3_neon;
+      pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
+         png_read_filter_row_paeth3_neon;
+   }
+
+   else if (bpp == 4)
+   {
+      pp->read_filter[PNG_FILTER_VALUE_SUB-1] = png_read_filter_row_sub4_neon;
+      pp->read_filter[PNG_FILTER_VALUE_AVG-1] = png_read_filter_row_avg4_neon;
+      pp->read_filter[PNG_FILTER_VALUE_PAETH-1] =
+          png_read_filter_row_paeth4_neon;
+   }
+}
+#endif /* PNG_ARM_NEON_OPT > 0 */
+#endif /* PNG_READ_SUPPORTED */

arm/filter_neon.S

@@ -0,0 +1,246 @@
+
+/* filter_neon.S - NEON optimised filter functions
+ *
+ * Copyright (c) 2013 Glenn Randers-Pehrson
+ * Written by Mans Rullgard, 2011.
+ * Last changed in libpng 1.6.8 [December 19, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ */
+
+/* This is required to get the symbol renames, which are #defines, and also
+ * includes the definition (or not) of PNG_ARM_NEON_OPT and
+ * PNG_ARM_NEON_IMPLEMENTATION.
+ */
+#define PNG_VERSION_INFO_ONLY
+#include "../pngpriv.h"
+
+#if defined(__linux__) && defined(__ELF__)
+.section .note.GNU-stack,"",%progbits /* mark stack as non-executable */
+#endif
+
+/* Assembler NEON support - only works for 32-bit ARM (i.e. it does not work for
+ * ARM64).  The code in arm/filter_neon_intrinsics.c supports ARM64, however it
+ * only works if -mfpu=neon is specified on the GCC command line.  See pngpriv.h
+ * for the logic which sets PNG_USE_ARM_NEON_ASM:
+ */
+#if PNG_ARM_NEON_IMPLEMENTATION == 2 /* hand-coded assembler */
+
+#ifdef PNG_READ_SUPPORTED
+#if PNG_ARM_NEON_OPT > 0
+
+#ifdef __ELF__
+#   define ELF
+#else
+#   define ELF @
+#endif
+
+        .arch armv7-a
+        .fpu  neon
+
+.macro  func    name, export=0
+    .macro endfunc
+ELF     .size   \name, . - \name
+        .endfunc
+        .purgem endfunc
+    .endm
+        .text
+    .if \export
+        .global \name
+    .endif
+ELF     .type   \name, STT_FUNC
+        .func   \name
+\name:
+.endm
+
+func    png_read_filter_row_sub4_neon, export=1
+        ldr             r3,  [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+1:
+        vld4.32         {d4[],d5[],d6[],d7[]},    [r1,:128]
+        vadd.u8         d0,  d3,  d4
+        vadd.u8         d1,  d0,  d5
+        vadd.u8         d2,  d1,  d6
+        vadd.u8         d3,  d2,  d7
+        vst4.32         {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
+        subs            r3,  r3,  #16
+        bgt             1b
+
+        bx              lr
+endfunc
+
+func    png_read_filter_row_sub3_neon, export=1
+        ldr             r3,  [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+        mov             r0,  r1
+        mov             r2,  #3
+        mov             r12, #12
+        vld1.8          {q11},    [r0], r12
+1:
+        vext.8          d5,  d22, d23, #3
+        vadd.u8         d0,  d3,  d22
+        vext.8          d6,  d22, d23, #6
+        vadd.u8         d1,  d0,  d5
+        vext.8          d7,  d23, d23, #1
+        vld1.8          {q11},    [r0], r12
+        vst1.32         {d0[0]},  [r1,:32], r2
+        vadd.u8         d2,  d1,  d6
+        vst1.32         {d1[0]},  [r1], r2
+        vadd.u8         d3,  d2,  d7
+        vst1.32         {d2[0]},  [r1], r2
+        vst1.32         {d3[0]},  [r1], r2
+        subs            r3,  r3,  #12
+        bgt             1b
+
+        bx              lr
+endfunc
+
+func    png_read_filter_row_up_neon, export=1
+        ldr             r3,  [r0, #4]           @ rowbytes
+1:
+        vld1.8          {q0}, [r1,:128]
+        vld1.8          {q1}, [r2,:128]!
+        vadd.u8         q0,  q0,  q1
+        vst1.8          {q0}, [r1,:128]!
+        subs            r3,  r3,  #16
+        bgt             1b
+
+        bx              lr
+endfunc
+
+func    png_read_filter_row_avg4_neon, export=1
+        ldr             r12, [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+1:
+        vld4.32         {d4[],d5[],d6[],d7[]},    [r1,:128]
+        vld4.32         {d16[],d17[],d18[],d19[]},[r2,:128]!
+        vhadd.u8        d0,  d3,  d16
+        vadd.u8         d0,  d0,  d4
+        vhadd.u8        d1,  d0,  d17
+        vadd.u8         d1,  d1,  d5
+        vhadd.u8        d2,  d1,  d18
+        vadd.u8         d2,  d2,  d6
+        vhadd.u8        d3,  d2,  d19
+        vadd.u8         d3,  d3,  d7
+        vst4.32         {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
+        subs            r12, r12, #16
+        bgt             1b
+
+        bx              lr
+endfunc
+
+func    png_read_filter_row_avg3_neon, export=1
+        push            {r4,lr}
+        ldr             r12, [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+        mov             r0,  r1
+        mov             r4,  #3
+        mov             lr,  #12
+        vld1.8          {q11},    [r0], lr
+1:
+        vld1.8          {q10},    [r2], lr
+        vext.8          d5,  d22, d23, #3
+        vhadd.u8        d0,  d3,  d20
+        vext.8          d17, d20, d21, #3
+        vadd.u8         d0,  d0,  d22
+        vext.8          d6,  d22, d23, #6
+        vhadd.u8        d1,  d0,  d17
+        vext.8          d18, d20, d21, #6
+        vadd.u8         d1,  d1,  d5
+        vext.8          d7,  d23, d23, #1
+        vld1.8          {q11},    [r0], lr
+        vst1.32         {d0[0]},  [r1,:32], r4
+        vhadd.u8        d2,  d1,  d18
+        vst1.32         {d1[0]},  [r1], r4
+        vext.8          d19, d21, d21, #1
+        vadd.u8         d2,  d2,  d6
+        vhadd.u8        d3,  d2,  d19
+        vst1.32         {d2[0]},  [r1], r4
+        vadd.u8         d3,  d3,  d7
+        vst1.32         {d3[0]},  [r1], r4
+        subs            r12, r12, #12
+        bgt             1b
+
+        pop             {r4,pc}
+endfunc
+
+.macro  paeth           rx,  ra,  rb,  rc
+        vaddl.u8        q12, \ra, \rb           @ a + b
+        vaddl.u8        q15, \rc, \rc           @ 2*c
+        vabdl.u8        q13, \rb, \rc           @ pa
+        vabdl.u8        q14, \ra, \rc           @ pb
+        vabd.u16        q15, q12, q15           @ pc
+        vcle.u16        q12, q13, q14           @ pa <= pb
+        vcle.u16        q13, q13, q15           @ pa <= pc
+        vcle.u16        q14, q14, q15           @ pb <= pc
+        vand            q12, q12, q13           @ pa <= pb && pa <= pc
+        vmovn.u16       d28, q14
+        vmovn.u16       \rx, q12
+        vbsl            d28, \rb, \rc
+        vbsl            \rx, \ra, d28
+.endm
+
+func    png_read_filter_row_paeth4_neon, export=1
+        ldr             r12, [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+        vmov.i8         d20, #0
+1:
+        vld4.32         {d4[],d5[],d6[],d7[]},    [r1,:128]
+        vld4.32         {d16[],d17[],d18[],d19[]},[r2,:128]!
+        paeth           d0,  d3,  d16, d20
+        vadd.u8         d0,  d0,  d4
+        paeth           d1,  d0,  d17, d16
+        vadd.u8         d1,  d1,  d5
+        paeth           d2,  d1,  d18, d17
+        vadd.u8         d2,  d2,  d6
+        paeth           d3,  d2,  d19, d18
+        vmov            d20, d19
+        vadd.u8         d3,  d3,  d7
+        vst4.32         {d0[0],d1[0],d2[0],d3[0]},[r1,:128]!
+        subs            r12, r12, #16
+        bgt             1b
+
+        bx              lr
+endfunc
+
+func    png_read_filter_row_paeth3_neon, export=1
+        push            {r4,lr}
+        ldr             r12, [r0, #4]           @ rowbytes
+        vmov.i8         d3,  #0
+        vmov.i8         d4,  #0
+        mov             r0,  r1
+        mov             r4,  #3
+        mov             lr,  #12
+        vld1.8          {q11},    [r0], lr
+1:
+        vld1.8          {q10},    [r2], lr
+        paeth           d0,  d3,  d20, d4
+        vext.8          d5,  d22, d23, #3
+        vadd.u8         d0,  d0,  d22
+        vext.8          d17, d20, d21, #3
+        paeth           d1,  d0,  d17, d20
+        vst1.32         {d0[0]},  [r1,:32], r4
+        vext.8          d6,  d22, d23, #6
+        vadd.u8         d1,  d1,  d5
+        vext.8          d18, d20, d21, #6
+        paeth           d2,  d1,  d18, d17
+        vext.8          d7,  d23, d23, #1
+        vld1.8          {q11},    [r0], lr
+        vst1.32         {d1[0]},  [r1], r4
+        vadd.u8         d2,  d2,  d6
+        vext.8          d19, d21, d21, #1
+        paeth           d3,  d2,  d19, d18
+        vst1.32         {d2[0]},  [r1], r4
+        vmov            d4,  d19
+        vadd.u8         d3,  d3,  d7
+        vst1.32         {d3[0]},  [r1], r4
+        subs            r12, r12, #12
+        bgt             1b
+
+        pop             {r4,pc}
+endfunc
+#endif /* PNG_ARM_NEON_OPT > 0 */
+#endif /* PNG_READ_SUPPORTED */
+#endif /* PNG_ARM_NEON_IMPLEMENTATION == 2 (assembler) */

arm/filter_neon_intrinsics.c

@@ -0,0 +1,372 @@
+
+/* filter_neon_intrinsics.c - NEON optimised filter functions
+ *
+ * Copyright (c) 2013 Glenn Randers-Pehrson
+ * Written by James Yu <james.yu at linaro.org>, October 2013.
+ * Based on filter_neon.S, written by Mans Rullgard, 2011.
+ *
+ * Last changed in libpng 1.6.8 [December 19, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ */
+
+#include "../pngpriv.h"
+
+/* This code requires -mfpu=neon on the command line: */
+#if PNG_ARM_NEON_IMPLEMENTATION == 1 /* intrinsics code from pngpriv.h */
+
+#include <arm_neon.h>
+
+/* libpng row pointers are not necessarily aligned to any particular boundary,
+ * however this code will only work with appropriate alignment.  arm/arm_init.c
+ * checks for this (and will not compile unless it is done). This code uses
+ * variants of png_aligncast to avoid compiler warnings.
+ */
+#define png_ptr(type,pointer) png_aligncast(type *,pointer)
+#define png_ptrc(type,pointer) png_aligncastconst(const type *,pointer)
+
+/* The following relies on a variable 'temp_pointer' being declared with type
+ * 'type'.  This is written this way just to hide the GCC strict aliasing
+ * warning; note that the code is safe because there never is an alias between
+ * the input and output pointers.
+ */
+#define png_ldr(type,pointer)\
+   (temp_pointer = png_ptr(type,pointer), *temp_pointer)
+
+#ifdef PNG_READ_SUPPORTED
+#if PNG_ARM_NEON_OPT > 0
+
+void
+png_read_filter_row_up_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+   png_const_bytep pp = prev_row;
+
+   for (; rp < rp_stop; rp += 16, pp += 16)
+   {
+      uint8x16_t qrp, qpp;
+
+      qrp = vld1q_u8(rp);
+      qpp = vld1q_u8(pp);
+      qrp = vaddq_u8(qrp, qpp);
+      vst1q_u8(rp, qrp);
+   }
+}
+
+void
+png_read_filter_row_sub3_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+
+   uint8x16_t vtmp = vld1q_u8(rp);
+   uint8x8x2_t *vrpt = png_ptr(uint8x8x2_t, &vtmp);
+   uint8x8x2_t vrp = *vrpt;
+
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   for (; rp < rp_stop;)
+   {
+      uint8x8_t vtmp1, vtmp2;
+      uint32x2_t *temp_pointer;
+
+      vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
+      vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]);
+      vtmp2 = vext_u8(vrp.val[0], vrp.val[1], 6);
+      vdest.val[1] = vadd_u8(vdest.val[0], vtmp1);
+
+      vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
+      vdest.val[2] = vadd_u8(vdest.val[1], vtmp2);
+      vdest.val[3] = vadd_u8(vdest.val[2], vtmp1);
+
+      vtmp = vld1q_u8(rp + 12);
+      vrpt = png_ptr(uint8x8x2_t, &vtmp);
+      vrp = *vrpt;
+
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
+      rp += 3;
+   }
+
+   PNG_UNUSED(prev_row)
+}
+
+void
+png_read_filter_row_sub4_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   for (; rp < rp_stop; rp += 16)
+   {
+      uint32x2x4_t vtmp = vld4_u32(png_ptr(uint32_t,rp));
+      uint8x8x4_t *vrpt = png_ptr(uint8x8x4_t,&vtmp);
+      uint8x8x4_t vrp = *vrpt;
+      uint32x2x4_t *temp_pointer;
+
+      vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]);
+      vdest.val[1] = vadd_u8(vdest.val[0], vrp.val[1]);
+      vdest.val[2] = vadd_u8(vdest.val[1], vrp.val[2]);
+      vdest.val[3] = vadd_u8(vdest.val[2], vrp.val[3]);
+      vst4_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2x4_t,&vdest), 0);
+   }
+
+   PNG_UNUSED(prev_row)
+}
+
+void
+png_read_filter_row_avg3_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_const_bytep pp = prev_row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+
+   uint8x16_t vtmp;
+   uint8x8x2_t *vrpt;
+   uint8x8x2_t vrp;
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   vtmp = vld1q_u8(rp);
+   vrpt = png_ptr(uint8x8x2_t,&vtmp);
+   vrp = *vrpt;
+
+   for (; rp < rp_stop; pp += 12)
+   {
+      uint8x8_t vtmp1, vtmp2, vtmp3;
+
+      uint8x8x2_t *vppt;
+      uint8x8x2_t vpp;
+
+      uint32x2_t *temp_pointer;
+
+      vtmp = vld1q_u8(pp);
+      vppt = png_ptr(uint8x8x2_t,&vtmp);
+      vpp = *vppt;
+
+      vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
+      vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]);
+      vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
+
+      vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3);
+      vtmp3 = vext_u8(vrp.val[0], vrp.val[1], 6);
+      vdest.val[1] = vhadd_u8(vdest.val[0], vtmp2);
+      vdest.val[1] = vadd_u8(vdest.val[1], vtmp1);
+
+      vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 6);
+      vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
+
+      vtmp = vld1q_u8(rp + 12);
+      vrpt = png_ptr(uint8x8x2_t,&vtmp);
+      vrp = *vrpt;
+
+      vdest.val[2] = vhadd_u8(vdest.val[1], vtmp2);
+      vdest.val[2] = vadd_u8(vdest.val[2], vtmp3);
+
+      vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1);
+
+      vdest.val[3] = vhadd_u8(vdest.val[2], vtmp2);
+      vdest.val[3] = vadd_u8(vdest.val[3], vtmp1);
+
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
+      rp += 3;
+   }
+}
+
+void
+png_read_filter_row_avg4_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+   png_const_bytep pp = prev_row;
+
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   for (; rp < rp_stop; rp += 16, pp += 16)
+   {
+      uint32x2x4_t vtmp;
+      uint8x8x4_t *vrpt, *vppt;
+      uint8x8x4_t vrp, vpp;
+      uint32x2x4_t *temp_pointer;
+
+      vtmp = vld4_u32(png_ptr(uint32_t,rp));
+      vrpt = png_ptr(uint8x8x4_t,&vtmp);
+      vrp = *vrpt;
+      vtmp = vld4_u32(png_ptrc(uint32_t,pp));
+      vppt = png_ptr(uint8x8x4_t,&vtmp);
+      vpp = *vppt;
+
+      vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]);
+      vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
+      vdest.val[1] = vhadd_u8(vdest.val[0], vpp.val[1]);
+      vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]);
+      vdest.val[2] = vhadd_u8(vdest.val[1], vpp.val[2]);
+      vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]);
+      vdest.val[3] = vhadd_u8(vdest.val[2], vpp.val[3]);
+      vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]);
+
+      vst4_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2x4_t,&vdest), 0);
+   }
+}
+
+static uint8x8_t
+paeth(uint8x8_t a, uint8x8_t b, uint8x8_t c)
+{
+   uint8x8_t d, e;
+   uint16x8_t p1, pa, pb, pc;
+
+   p1 = vaddl_u8(a, b); /* a + b */
+   pc = vaddl_u8(c, c); /* c * 2 */
+   pa = vabdl_u8(b, c); /* pa */
+   pb = vabdl_u8(a, c); /* pb */
+   pc = vabdq_u16(p1, pc); /* pc */
+
+   p1 = vcleq_u16(pa, pb); /* pa <= pb */
+   pa = vcleq_u16(pa, pc); /* pa <= pc */
+   pb = vcleq_u16(pb, pc); /* pb <= pc */
+
+   p1 = vandq_u16(p1, pa); /* pa <= pb && pa <= pc */
+
+   d = vmovn_u16(pb);
+   e = vmovn_u16(p1);
+
+   d = vbsl_u8(d, b, c);
+   e = vbsl_u8(e, a, d);
+
+   return e;
+}
+
+void
+png_read_filter_row_paeth3_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_const_bytep pp = prev_row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+
+   uint8x16_t vtmp;
+   uint8x8x2_t *vrpt;
+   uint8x8x2_t vrp;
+   uint8x8_t vlast = vdup_n_u8(0);
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   vtmp = vld1q_u8(rp);
+   vrpt = png_ptr(uint8x8x2_t,&vtmp);
+   vrp = *vrpt;
+
+   for (; rp < rp_stop; pp += 12)
+   {
+      uint8x8x2_t *vppt;
+      uint8x8x2_t vpp;
+      uint8x8_t vtmp1, vtmp2, vtmp3;
+      uint32x2_t *temp_pointer;
+
+      vtmp = vld1q_u8(pp);
+      vppt = png_ptr(uint8x8x2_t,&vtmp);
+      vpp = *vppt;
+
+      vdest.val[0] = paeth(vdest.val[3], vpp.val[0], vlast);
+      vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
+
+      vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3);
+      vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3);
+      vdest.val[1] = paeth(vdest.val[0], vtmp2, vpp.val[0]);
+      vdest.val[1] = vadd_u8(vdest.val[1], vtmp1);
+
+      vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 6);
+      vtmp3 = vext_u8(vpp.val[0], vpp.val[1], 6);
+      vdest.val[2] = paeth(vdest.val[1], vtmp3, vtmp2);
+      vdest.val[2] = vadd_u8(vdest.val[2], vtmp1);
+
+      vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1);
+      vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1);
+
+      vtmp = vld1q_u8(rp + 12);
+      vrpt = png_ptr(uint8x8x2_t,&vtmp);
+      vrp = *vrpt;
+
+      vdest.val[3] = paeth(vdest.val[2], vtmp2, vtmp3);
+      vdest.val[3] = vadd_u8(vdest.val[3], vtmp1);
+
+      vlast = vtmp2;
+
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0);
+      rp += 3;
+      vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0);
+      rp += 3;
+   }
+}
+
+void
+png_read_filter_row_paeth4_neon(png_row_infop row_info, png_bytep row,
+   png_const_bytep prev_row)
+{
+   png_bytep rp = row;
+   png_bytep rp_stop = row + row_info->rowbytes;
+   png_const_bytep pp = prev_row;
+
+   uint8x8_t vlast = vdup_n_u8(0);
+   uint8x8x4_t vdest;
+   vdest.val[3] = vdup_n_u8(0);
+
+   for (; rp < rp_stop; rp += 16, pp += 16)
+   {
+      uint32x2x4_t vtmp;
+      uint8x8x4_t *vrpt, *vppt;
+      uint8x8x4_t vrp, vpp;
+      uint32x2x4_t *temp_pointer;
+
+      vtmp = vld4_u32(png_ptr(uint32_t,rp));
+      vrpt = png_ptr(uint8x8x4_t,&vtmp);
+      vrp = *vrpt;
+      vtmp = vld4_u32(png_ptrc(uint32_t,pp));
+      vppt = png_ptr(uint8x8x4_t,&vtmp);
+      vpp = *vppt;
+
+      vdest.val[0] = paeth(vdest.val[3], vpp.val[0], vlast);
+      vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]);
+      vdest.val[1] = paeth(vdest.val[0], vpp.val[1], vpp.val[0]);
+      vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]);
+      vdest.val[2] = paeth(vdest.val[1], vpp.val[2], vpp.val[1]);
+      vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]);
+      vdest.val[3] = paeth(vdest.val[2], vpp.val[3], vpp.val[2]);
+      vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]);
+
+      vlast = vpp.val[3];
+
+      vst4_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2x4_t,&vdest), 0);
+   }
+}
+
+#endif /* PNG_ARM_NEON_OPT > 0 */
+#endif /* PNG_READ_SUPPORTED */
+#endif /* PNG_ARM_NEON_IMPLEMENTATION == 1 (intrinsics) */

autogen.sh

@@ -1,25 +1,224 @@
 #! /bin/sh
-# a quick hack script to generate necessary files from 
-# auto* tools.
 #
-# WARNING: if you run this you will change the versions
-# of the tools which are used and, maybe, required!
-touch Makefile.am configure.ac
-{
-   echo "running libtoolize" >&2
-   libtoolize --force --copy --automake
-} && {
-   echo "running aclocal" >&2
-   aclocal
-} && {
-   echo "running autoheader [ignore the warnings]" >&2
-   autoheader
-} && {
-   echo "running automake" >&2
-   automake --force-missing --foreign -a -c
-} && {
-   echo "running autoconf" >&2
-   autoconf
-} &&
-   echo "autogen complete" >&2 ||
-   echo "ERROR: autogen.sh failed, autogen is incomplete" >&2
+# Run 'autoreconf' to build 'configure', 'Makefile.in' and other configure
+# control files.
+#
+# The first time this is run on a GIT checkout the only files that exist are
+# configure.ac and Makefile.am; all of the autotools support scripts are
+# missing.  They are instantiated with autoreconf --force --install.
+#
+# For regular ("tarball") distributions all the files should exist.  We do not
+# want them to be updated *under any circumstances*.  It should never be
+# necessary to run autogen.sh because ./configure --enable-maintainer-mode says
+# what to do if Makefile.am or configure.ac are changed.
+#
+# It is *probably* OK to update the files on a GIT checkout, because they have
+# come from the local tools, but leave that to the user who is assumed to know
+# whether it is ok or required.
+#
+# This script is intended to work without arguments, there are, however, hidden
+# arguments (a) for use while testing the script and (b) to fix up systems that
+# have been broken.  If (b) is required the script prompts for the correct
+# options.  For this reason the options are *NOT* documented in the help; this
+# is deliberate; UTSL.
+#
+clean=
+maintainer=
+while test $# -gt 0
+do
+   case "$1" in
+      --maintainer)
+         maintainer=1;;
+
+      --clean)
+         clean=1;;
+
+      *)
+         exec >&2
+         echo "$0: usage: ./autogen.sh"
+         if test -d .git
+         then
+            echo "  ./autogen.sh generates the configure script and"
+            echo "  Makefile.in, or refreshes them after changes to Makefile.am"
+            echo "  or configure.ac.  You may prefer to just run autoreconf."
+         elif test -z "$maintainer"
+         then
+            echo "  DO NOT RUN THIS SCRIPT."
+            echo "  If you need to change Makefile.am or configure.ac then you"
+            echo "  also need to run ./configure --enable-maintainer-mode and"
+            echo "  use the appropriate autotools, *NOT* this script, to update"
+            echo "  everything, please check the documentation of autoreconf."
+            echo "  WARNING: libpng is intentionally generated with a known,"
+            echo "  fixed, set of autotools.  It is known *NOT* to work with"
+            echo "  the collection of autotools distributed on highly reputable"
+            echo "  operating systems."
+            echo "  Remember: autotools is GNU software, you are expected to"
+            echo "  pay for support."
+         else
+            echo "  You have run autogen.sh with --maintainer enabled and you"
+            echo "  are not using a GIT distribution, then you have given an"
+            echo "  unrecognized argument.  This is not good. --maintainer"
+            echo "  switches off any assumptions that you might not know what"
+            echo "  you are doing."
+         fi
+         exit 1;;
+   esac
+
+   shift
+done
+#
+# First check for a set of the autotools files; if absent then this is assumed
+# to be a GIT version and the local autotools must be used.  If present this
+# is a tarball distribution and the script should not be used.  If partially
+# present bad things are happening.
+#
+# The autotools generated files:
+libpng_autotools_files="Makefile.in aclocal.m4 config.guess config.h.in\
+   config.sub configure depcomp install-sh ltmain.sh missing test-driver"
+#
+# Files generated by versions of configue >2.68 or automake >1.13 (i.e. later
+# versions than those required by configure.ac):
+libpng_autotools_extra="compile"
+#
+# These are separate because 'maintainer-clean' does not remove them.
+libpng_libtool_files="scripts/libtool.m4 scripts/ltoptions.m4\
+   scripts/ltsugar.m4 scripts/ltversion.m4 scripts/lt~obsolete.m4"
+
+libpng_autotools_dirs="autom4te.cache" # not required
+#
+# The configure generated files:
+libpng_configure_files="Makefile config.h config.log config.status\
+   libpng-config libpng.pc libtool stamp-h1"
+
+libpng_configure_dirs=".deps"
+#
+# We must remove the configure generated files as well as the autotools
+# generated files if autotools are regenerated because otherwise if configure
+# has been run without "--enable-maintainer-mode" make can do a partial update
+# of Makefile.  These functions do the two bits of cleaning.
+clean_autotools(){
+   rm -rf $libpng_autotools_files $libpng_libtool_files $libpng_autotools_dirs
+   rm -rf $libpng_autotools_extra
+}
+
+clean_configure(){
+   rm -rf $libpng_configure_files $libpng_configure_dirs
+}
+#
+# Clean: remove everything (this is to help with testing)
+if test -n "$clean"
+then
+   clean_configure
+   if test -n "$maintainer"
+   then
+      clean_autotools
+   fi
+
+   exit 0
+fi
+#
+# Validate the distribution.
+libpng_autotools_file_found=
+libpng_autotools_file_missing=
+for file in $libpng_autotools_files
+do
+   if test -f  "$file"
+   then
+      libpng_autotools_file_found=1
+   else
+      libpng_autotools_file_missing=1
+   fi
+done
+#
+# Presence of one of these does not *invalidate* missing, but absence
+# invalidates found.
+for file in $libpng_libtool_files
+do
+   if test ! -f "$file"
+   then
+      libpng_autotools_file_missing=1
+   fi
+done
+#
+# The cache directory doesn't matter - it will be regenerated and does not exist
+# anyway in a tarball.
+#
+# Either everything is missing or everything is there, the --maintainer option
+# just changes this so that the mode is set to generate all the files.
+mode=
+if test -z "$libpng_autotools_file_found" -o -n "$maintainer"
+then
+   mode="autoreconf"
+else
+   if test -n "$libpng_autotools_file_missing"
+   then
+      mode="broken"
+   else
+      mode="configure"
+   fi
+fi
+#
+# So:
+case "$mode" in
+   autoreconf)
+      # Clean in case configure files exist
+      clean_configure
+      clean_autotools
+      # Everything must be initialized, so use --force
+      if autoreconf --warnings=all --force --install
+      then
+         missing=
+         for file in $libpng_autotools_files
+         do
+            test -f "$file" || missing=1
+         done
+         # ignore the cache directory
+         test -z "$missing" || {
+            exec >&2
+            echo "autoreconf was run, but did not produce all the expected"
+            echo "files.  It is likely that your autotools installation is"
+            echo "not compatible with that expected by libpng."
+            exit 1
+         }
+      else
+         exec >&2
+         echo "autoreconf failed: your version of autotools is incompatible"
+         echo "with this libpng version.  Please use a distributed archive"
+         echo "(which includes the autotools generated files) and run configure"
+         echo "instead."
+         exit 1
+      fi;;
+
+   configure)
+      if test -d .git
+      then
+         exec >&2
+         echo "ERROR: running autoreconf on an initialized sytem"
+         echo "  This is not necessary; it is only necessary to remake the"
+         echo "  autotools generated files if Makefile.am or configure.ac"
+         echo "  change and make does the right thing with:"
+         echo
+         echo "     ./configure --enable-maintainer-mode."
+         echo
+         echo "  You can run autoreconf yourself if you don't like maintainer"
+         echo "  mode and you can also just run autoreconf -f -i to initialize"
+         echo "  everything in the first place; this script is only for"
+         echo "  compatibility with prior releases."
+         exit 1
+      else
+         exec >&2
+         echo "autogen.sh is intended only to generate 'configure' on systems"
+         echo "that do not have it.  You have a complete 'configure', if you"
+         echo "need to change Makefile.am or configure.ac you also need to"
+         echo "run configure with the --enable-maintainer-mode option."
+         exit 1
+      fi;;
+
+   broken)
+      exec >&2
+      echo "Your system has a partial set of autotools generated files."
+      echo "autogen.sh is unable to proceed.  The full set of files is"
+      echo "contained in the libpng 'tar' distribution archive and you do"
+      echo "not need to run autogen.sh if you use it."
+      exit 1;;
+esac

config.h.in

@@ -1,86 +0,0 @@
-/* config.h.in.  Generated from configure.ac by autoheader.  */
-
-/* Define to 1 if you have the <dlfcn.h> header file. */
-#undef HAVE_DLFCN_H
-
-/* Define to 1 if you have the <inttypes.h> header file. */
-#undef HAVE_INTTYPES_H
-
-/* Define to 1 if you have the `m' library (-lm). */
-#undef HAVE_LIBM
-
-/* Define to 1 if you have the `z' library (-lz). */
-#undef HAVE_LIBZ
-
-/* Define to 1 if you have the <malloc.h> header file. */
-#undef HAVE_MALLOC_H
-
-/* Define to 1 if you have the <memory.h> header file. */
-#undef HAVE_MEMORY_H
-
-/* Define to 1 if you have the `memset' function. */
-#undef HAVE_MEMSET
-
-/* Define to 1 if you have the `pow' function. */
-#undef HAVE_POW
-
-/* Define to 1 if you have the <stdint.h> header file. */
-#undef HAVE_STDINT_H
-
-/* Define to 1 if you have the <stdlib.h> header file. */
-#undef HAVE_STDLIB_H
-
-/* Define to 1 if you have the <strings.h> header file. */
-#undef HAVE_STRINGS_H
-
-/* Define to 1 if you have the <string.h> header file. */
-#undef HAVE_STRING_H
-
-/* Define to 1 if you have the <sys/stat.h> header file. */
-#undef HAVE_SYS_STAT_H
-
-/* Define to 1 if you have the <sys/types.h> header file. */
-#undef HAVE_SYS_TYPES_H
-
-/* Define to 1 if you have the <unistd.h> header file. */
-#undef HAVE_UNISTD_H
-
-/* Define to the sub-directory in which libtool stores uninstalled libraries.
-   */
-#undef LT_OBJDIR
-
-/* Name of package */
-#undef PACKAGE
-
-/* Define to the address where bug reports for this package should be sent. */
-#undef PACKAGE_BUGREPORT
-
-/* Define to the full name of this package. */
-#undef PACKAGE_NAME
-
-/* Define to the full name and version of this package. */
-#undef PACKAGE_STRING
-
-/* Define to the one symbol short name of this package. */
-#undef PACKAGE_TARNAME
-
-/* Define to the home page for this package. */
-#undef PACKAGE_URL
-
-/* Define to the version of this package. */
-#undef PACKAGE_VERSION
-
-/* Define to 1 if you have the ANSI C header files. */
-#undef STDC_HEADERS
-
-/* Define to 1 if your <sys/time.h> declares `struct tm'. */
-#undef TM_IN_SYS_TIME
-
-/* Version number of package */
-#undef VERSION
-
-/* Define to empty if `const' does not conform to ANSI C. */
-#undef const
-
-/* Define to `unsigned int' if <sys/types.h> does not define. */
-#undef size_t

configure.ac

@@ -14,66 +14,270 @@ dnl Makefile.am to upgrade the package name.
 dnl This is here to prevent earlier autoconf from being used, it
 dnl should not be necessary to regenerate configure if the time
 dnl stamps are correct
-AC_PREREQ(2.59)
+AC_PREREQ([2.68])
 
 dnl Version number stuff here:
 
-AC_INIT([libpng], [1.5.0beta47], [png-mng-implement@lists.sourceforge.net])
-AM_INIT_AUTOMAKE
-dnl stop configure from automagically running automake
+AC_INIT([libpng],[1.7.0beta39],[png-mng-implement@lists.sourceforge.net])
+AC_CONFIG_MACRO_DIR([scripts])
+
+# libpng does not follow GNU file name conventions (hence 'foreign')
+# color-tests requires automake 1.11 or later
+# silent-rules requires automake 1.11 or later
+# dist-xz requires automake 1.11 or later
+# 1.12.2 fixes a security issue in 1.11.2 and 1.12.1
+# 1.13 is required for parallel tests
+AM_INIT_AUTOMAKE([1.13 foreign dist-xz color-tests silent-rules subdir-objects])
+# The following line causes --disable-maintainer-mode to be the default to
+# configure, this is necessary because libpng distributions cannot rely on the
+# time stamps of the autotools generated files being correct
 AM_MAINTAINER_MODE
 
-PNGLIB_VERSION=1.5.0beta47
+dnl configure.ac and Makefile.am expect automake 1.11.2 or a compatible later
+dnl version; aclocal.m4 will generate a failure if you use a prior version of
+dnl automake, so the following is not necessary (and is not defined anyway):
+dnl AM_PREREQ([1.11.2])
+dnl stop configure from automagically running automake
+
+PNGLIB_VERSION=1.7.0beta39
 PNGLIB_MAJOR=1
-PNGLIB_MINOR=5
+PNGLIB_MINOR=7
 PNGLIB_RELEASE=0
 
 dnl End of version number stuff
 
 AC_CONFIG_SRCDIR([pngget.c])
-AM_CONFIG_HEADER(config.h)
+AC_CONFIG_HEADERS([config.h])
 
 # Checks for programs.
+AC_LANG([C])
 AC_PROG_CC
-AC_PROG_LD
+AM_PROG_AS
+LT_PATH_LD
 AC_PROG_CPP
-AC_CHECK_TOOL(SED, sed, :)
-AC_CHECK_TOOL(AWK, awk, :)
+AC_PROG_AWK
 AC_PROG_INSTALL
 AC_PROG_LN_S
 AC_PROG_MAKE_SET
+
+dnl libtool/libtoolize; version 2.4.2 is the tested version, this or any
+dnl compatible later version may be used
 LT_INIT([win32-dll])
+LT_PREREQ([2.4.2])
+
+# Some awks crash when confronted with pnglibconf.dfa, do a test run now
+# to make sure this doesn't happen
+AC_MSG_CHECKING([that AWK works])
+if ${AWK} -f ${srcdir}/scripts/options.awk out="/dev/null" version=search\
+   ${srcdir}/pngconf.h ${srcdir}/scripts/pnglibconf.dfa\
+   ${srcdir}/pngusr.dfa 1>&2
+then
+   AC_MSG_RESULT([ok])
+else
+   AC_MSG_FAILURE([failed], 1)
+fi
+
+# This is a remnant of the old cc -E validation, where it may have been
+# necessary to use a different preprocessor for .dfn files
+DFNCPP="$CPP -E"
+AC_SUBST(DFNCPP)
+
+# -Werror cannot be passed to GCC in CFLAGS because configure will fail (it
+# checks the compiler with a program that generates a warning), add the
+# following option to deal with this
+AC_ARG_VAR(PNG_COPTS,
+   [additional flags for the C compiler, use this for options that would]
+   [cause configure itself to fail])
+AC_ARG_ENABLE(werror,
+   AS_HELP_STRING([[[--enable-werror[=OPT]]]],
+      [Pass -Werror or the given argument to the compiler if it is supported]),
+   [test "$enable_werror" = "yes" && enable_werror="-Werror"
+    if test "$enable_werror" != "no"; then
+      sav_CFLAGS="$CFLAGS"
+      CFLAGS="$enable_werror $CFLAGS"
+      AC_MSG_CHECKING([if the compiler allows $enable_werror])
+      AC_COMPILE_IFELSE(
+         [AC_LANG_SOURCE([
+            [int main(int argc, char **argv){]
+            [return argv[argc-1][0];]
+            [}]])],
+         AC_MSG_RESULT(yes)
+         PNG_COPTS="$PNG_COPTS $enable_werror",
+         AC_MSG_RESULT(no))
+      CFLAGS="$sav_CFLAGS"
+    fi],)
 
 # Checks for header files.
 AC_HEADER_STDC
-AC_CHECK_HEADERS([malloc.h stdlib.h string.h strings.h])
+# Headers used in pngpriv.h:
+AC_CHECK_HEADERS([assert.h stdlib.h string.h float.h fp.h math.h m68881.h])
+# Additional headers used in pngconf.h:
+AC_CHECK_HEADERS([limits.h stddef.h stdio.h setjmp.h time.h])
+# Used by pngread.c (for the simplified API):
+AC_CHECK_HEADERS([errno.h])
+# The following are not required for the standard 'make' build of libpng:
+# Additional requirements of contrib/libtests/makepng.c:
+AC_CHECK_HEADERS([ctype.h])
+# Additional requirements of contrib/libtests/pngvalid.c:
+AC_CHECK_HEADERS([fenv.h])
 
 # Checks for typedefs, structures, and compiler characteristics.
-AC_C_CONST
-AC_TYPE_SIZE_T
-AC_STRUCT_TM
+# Most of these have been removed from 1.7.0 because they checked for non-ANSI-C
+# behaviors and the result were, anyway, not used by earlier versions of libpng.
+AC_C_RESTRICT
 
-# Checks for library functions.
-AC_FUNC_STRTOD
-AC_CHECK_FUNCS([memset], , AC_ERROR([memset not found in libc]))
-AC_CHECK_FUNCS([pow], , AC_CHECK_LIB(m, pow, , AC_ERROR([cannot find pow])) )
-AC_CHECK_LIB(z, zlibVersion, , AC_ERROR([zlib not installed]))
+# Checks for library functions; these need to be in order, least demanding to
+# most because the -l directives are added to the head of the list at each
+# check.
+# ANSI-C functions: all are required but they don't have to be implemented in
+# libc, it is perfectly ok to configure with an extra, wrapper, library preset
+# in LIBS.  NOTE: this list may not be complete
+AC_CHECK_FUNCS([memset memcpy strncpy],,
+               AC_MSG_FAILURE([missing ANSI-C functions]))
 
-LIBPNG_DEFINES=-DPNG_CONFIGURE_LIBPNG
-LIBPNG_DEFINES=$LIBPNG_DEFINES
-AC_SUBST(LIBPNG_DEFINES)
+# floor and ceil are required for FLOATING_POINT, but they can be turned off on
+# the configure command line (and they are independent.)  Default is to 'on'
+AC_ARG_ENABLE([floating-point],
+   AS_HELP_STRING([[[--enabled-floating-point]]],
+      [Enable floating point APIs (default: yes)]),,
+   [enable_floating_point=yes])
+
+AC_ARG_ENABLE([fixed-point],
+   AS_HELP_STRING([[[--enable-fixed-point]]],
+      [Enable fixed point APIs (default: yes)]),,
+   [enable_fixed_point=yes])
+
+AC_ARG_ENABLE([floating-arithmetic],
+   AS_HELP_STRING([[[--enable-floating-arithmetic]]],
+      [Use floating point arithmetic internally (default: yes)]),,
+   [enable_floating_arithmetic=yes])
+
+# Check for a requirement for math library support
+if test "$enable_floating_arithmetic" = yes
+then
+   AC_SEARCH_LIBS([floor],[m],,
+                  [AC_MSG_WARN([not found])
+                   enable_floating_arithmetic=no])
+   AC_SEARCH_LIBS([frexp],[m],,
+                  [AC_MSG_WARN([not found])
+                   enable_floating_arithmetic=no])
+   AC_SEARCH_LIBS([modf],[m],,
+                  [AC_MSG_WARN([not found])
+                   enable_floating_arithmetic=no])
+   AC_SEARCH_LIBS([atof],[m],,
+                  [AC_MSG_WARN([not found])
+                   enable_floating_arithmetic=no])
+   AC_SEARCH_LIBS([pow],[m],,
+                  [AC_MSG_WARN([not found])
+                   enable_floating_arithmetic=no])
+fi
+
+# Set the relevant defines if required (only if required), this will override
+# anything on the command line.
+if test "$enable_floating_point" != yes
+then
+   AC_MSG_NOTICE([Floating point APIs disabled])
+   AC_DEFINE([PNG_NO_FLOATING_POINT],[1],
+             [No support for floating point APIs])
+fi
+if test "$enable_fixed_point" != yes
+then
+   AC_MSG_NOTICE([Fixed point APIs disabled])
+   AC_DEFINE([PNG_NO_FIXED_POINT],[1],
+             [No support for fixed point APIs])
+fi
+if test "$enable_floating_arithmetic" != yes
+then
+   AC_MSG_NOTICE([Fixed point arithmetic will be used])
+   AC_DEFINE([PNG_NO_FLOATING_ARITHMETIC],[1],
+             [No support for floating point arithmetic])
+fi
+
+# The following is for pngvalid, to ensure it catches FP errors even on
+# platforms that don't enable FP exceptions, the function appears in the math
+# library (typically), it's not an error if it is not found.
+AC_SEARCH_LIBS([feenableexcept],[m],AC_DEFINE([HAVE_FEENABLEEXCEPT],[1],
+      [Define to 1 if the feenableexcept function is available]))
+# Required by pngvalid:
+AC_SEARCH_LIBS([ceil],[m])
+
+# zlib checks and arguments;
+#  package zlib: argument is the name of the library
+#  zlib-header: the name of the header file
+#  zlib-prefix: the prefix for library functions, if any
+AC_ARG_WITH([zlib],
+   AS_HELP_STRING([[[--with-zlib=<name>]]],
+                  [base name of the zlib implementation (e.g. 'z')]),,
+   [with_zlib='z'])
+
+z_prefix=
+AC_ARG_WITH([zlib-prefix],
+   AS_HELP_STRING([[[--with-zlib-prefix=<prefix>]]],
+                  [prefix for zlib API functions]),
+   [if test "$withval" = "" -o "$withval" = yes; then
+      with_zlib_prefix='z_'
+      AC_DEFINE([PNG_ZLIB_PREFIX],[z_],[Standard zlib prefix])
+   elif test "${withval}" != "no"; then
+      AC_DEFINE_UNQUOTED([PNG_ZLIB_PREFIX],[${withval}],
+         [User defined zlib prefix])
+   else
+      with_zlib_prefix=
+   fi
+   if test "$with_zlib_prefix" != ""; then
+      AC_DEFINE([Z_PREFIX],[1],[Tell zlib.h to prefix function names])
+      z_prefix="#define Z_PREFIX 1"
+   fi])
+
+AC_ARG_WITH([zlib-header],
+   AS_HELP_STRING([[[--with-zlib-header='<file>' or '"file"']]],
+      [C include argument to locate zlib header file, e.g. <zlib.h>, take
+       care to put single quotes round this on the command line]),
+   [AC_DEFINE_UNQUOTED([PNG_ZLIB_HEADER],[${withval}],[Zlib header file])],
+   with_zlib_header='<zlib.h>')
+
+# AC_CHECK_LIB isn't quite enough for this because we need the special header
+# file too.
+LIBS="-l${with_zlib} $LIBS"
+AC_MSG_CHECKING(
+   [[Library '${with_zlib}' and #include ${with_zlib_header} for symbol '${with_zlib_prefix}zlibVersion']])
+AC_LINK_IFELSE(
+   [AC_LANG_PROGRAM(
+      [[
+${z_prefix}
+#include ${with_zlib_header}
+      ]],
+      [[(void)${with_zlib_prefix}zlibVersion();]])],
+   AC_MSG_RESULT([ok]),
+   AC_MSG_FAILURE([not found]))
+
+AC_MSG_CHECKING([if using Solaris linker])
+SLD=`$LD --version 2>&1 | grep Solaris`
+if test "$SLD"; then
+    have_solaris_ld=yes
+    AC_MSG_RESULT(yes)
+else
+    have_solaris_ld=no
+    AC_MSG_RESULT(no)
+fi
+AM_CONDITIONAL(HAVE_SOLARIS_LD, test "$have_solaris_ld" = "yes")
 
 AC_MSG_CHECKING([if libraries can be versioned])
 # Special case for PE/COFF platforms: ld reports
 # support for version-script, but doesn't actually
 # DO anything with it.
 case $host in
-*cygwin* | *mingw32* )
+*cygwin* | *mingw32* | *interix* )
     have_ld_version_script=no
     AC_MSG_RESULT(no)
 ;;
 * )
-GLD=`$LD --help < /dev/null 2>/dev/null | grep version-script`
+
+if test "$have_solaris_ld" = "yes"; then
+    GLD=`$LD --help < /dev/null 2>&1 | grep 'M mapfile'`
+else
+    GLD=`$LD --help < /dev/null 2>/dev/null | grep version-script`
+fi
+
 if test "$GLD"; then
     have_ld_version_script=yes
     AC_MSG_RESULT(yes)
@@ -90,9 +294,9 @@ AM_CONDITIONAL(HAVE_LD_VERSION_SCRIPT, test "$have_ld_version_script" = "yes")
 if test "$have_ld_version_script" = "yes"; then
     AC_MSG_CHECKING([for symbol prefix])
     SYMBOL_PREFIX=`echo "PREFIX=__USER_LABEL_PREFIX__" \
-                   | ${CPP-${CC-gcc} -E} - 2>&1 \
-                   | ${EGREP-grep} "^PREFIX=" \
-                   | ${SED-sed} "s:^PREFIX=::"`
+                  | ${DFNCPP-${CC-gcc -E}} - 2>&1 \
+                  | ${EGREP-grep} "^PREFIX=" \
+                  | ${SED-sed} -e "s:^PREFIX=::" -e "s:__USER_LABEL_PREFIX__::"`
     AC_SUBST(SYMBOL_PREFIX)
     AC_MSG_RESULT($SYMBOL_PREFIX)
 fi
@@ -106,31 +310,130 @@ AC_SUBST(PNGLIB_RELEASE)
 # Additional arguments (and substitutions)
 # Allow the pkg-config directory to be set
 AC_ARG_WITH(pkgconfigdir,
-	AC_HELP_STRING([--with-pkgconfigdir],
-	[Use the specified pkgconfig dir (default is libdir/pkgconfig)]),
-	[pkgconfigdir=${withval}],
-	[pkgconfigdir='${libdir}/pkgconfig'])
+   AS_HELP_STRING([[[--with-pkgconfigdir]]],
+      [Use the specified pkgconfig dir (default is libdir/pkgconfig)]),
+   [pkgconfigdir=${withval}],
+   [pkgconfigdir='${libdir}/pkgconfig'])
 
 AC_SUBST([pkgconfigdir])
-AC_MSG_NOTICE([pkgconfig directory is ${pkgconfigdir}])
+AC_MSG_NOTICE([[pkgconfig directory is ${pkgconfigdir}]])
 
 # Make the *-config binary config scripts optional
 AC_ARG_WITH(binconfigs,
-	AC_HELP_STRING([--with-binconfigs],
-		[Generate shell libpng-config scripts as well as pkg-config data]
-		[@<:@default=yes@:>@]),
-	[if test "${withval}" = no; then
-		binconfigs=
-		AC_MSG_NOTICE([libpng-config scripts will not be built])
-	else
-		binconfigs='${binconfigs}'
-	fi],
-	[binconfigs='${binconfigs}'])
+   AS_HELP_STRING([[[--with-binconfigs]]],
+      [Generate shell libpng-config scripts as well as pkg-config data]
+      [@<:@default=yes@:>@]),
+   [if test "${withval}" = no; then
+      binconfigs=
+      AC_MSG_NOTICE([[libpng-config scripts will not be built]])
+    else
+      binconfigs='${binconfigs}'
+    fi],
+   [binconfigs='${binconfigs}'])
 AC_SUBST([binconfigs])
 
+# Support for prefixes to the API function names; this will generate defines
+# at the start of the build to rename exported library functions
+AC_ARG_WITH(libpng-prefix,
+   AS_HELP_STRING([[[--with-libpng-prefix]]],
+      [prefix libpng exported function (API) names with the given value]),
+   [if test "${withval:-no}" != "no"; then
+      AC_SUBST([PNG_PREFIX], [${withval}])
+    fi])
+AM_CONDITIONAL([DO_PNG_PREFIX], [test "${with_libpng_prefix:-no}" != "no"])
+
+# Control over what links are made for installed files.  Versioned files are
+# always installed, when the following options are turned on corresponding
+# unversioned links are also created (normally as symbolic links):
+AC_ARG_ENABLE([unversioned-links],
+   AS_HELP_STRING([[[--enable-unversioned-links]]],
+      [Installed libpng header files are placed in a versioned subdirectory]
+      [and installed libpng library (including DLL) files are versioned.]
+      [If this option is enabled unversioned links will be created pointing to]
+      [the corresponding installed files.  If you use libpng.pc or]
+      [libpng-config for all builds you do not need these links, but if you]
+      [compile programs directly they will typically #include <png.h> and]
+      [link with -lpng; in that case you need the links.]
+      [The links can be installed manually using 'make install-header-links']
+      [and 'make install-library-links' and can be removed using the]
+      [corresponding uninstall- targets.  If you do enable this option every]
+      [libpng 'make install' will recreate the links to point to the just]
+      [installed version of libpng.  The default is to create the links;]
+      [use --disable-unversioned-links to change this]))
+
+# The AM_CONDITIONAL test is written so that the default is enabled;
+# --disable-unversioned-links must be given to turn the option off.
+AM_CONDITIONAL([DO_INSTALL_LINKS],[test "$enable_unversioned_links" != "no"])
+
+AC_ARG_ENABLE([unversioned-libpng-pc],
+   AS_HELP_STRING([[[--enable-unversioned-libpng-pc]]],
+      [Install the configuration file 'libpng.pc' as a link to the versioned]
+      [version.  This is done by default - use --disable-unversioned-libpng-pc]
+      [to change this.]))
+AM_CONDITIONAL([DO_INSTALL_LIBPNG_PC],
+   [test "$enable_unversioned_libpng_pc" != "no"])
+
+AC_ARG_ENABLE([unversioned-libpng-config],
+   AS_HELP_STRING([[[--enable-unversioned-libpng-config]]],
+      [Install the configuration file 'libpng-config' as a link to the]
+      [versioned version.  This is done by default - use]
+      [--disable-unversioned-libpng-config to change this.]))
+AM_CONDITIONAL([DO_INSTALL_LIBPNG_CONFIG],
+   [test "$enable_unversioned_libpng_config" != "no"])
+
+# HOST SPECIFIC OPTIONS
+# =====================
+#
+# ARM
+# ===
+#
+# ARM NEON (SIMD) support.
+
+AC_ARG_ENABLE([arm-neon],
+   AS_HELP_STRING([[[--enable-arm-neon]]],
+      [Enable ARM NEON optimizations: =no/off, check, api, yes/on:]
+      [no/off: disable the optimizations; check: use internal checking code]
+      [(deprecated and poorly supported); api: disable by default, enable by]
+      [a call to png_set_option; yes/on: turn on unconditionally.]
+      [If not specified: determined by the compiler.]),
+   [case "$enableval" in
+      no|off)
+         # disable the default enabling on __ARM_NEON__ systems:
+         AC_DEFINE([PNG_ARM_NEON_OPT], [0],
+                   [Disable ARM Neon optimizations])
+         # Prevent inclusion of the assembler files below:
+         enable_arm_neon=no;;
+      check)
+         AC_DEFINE([PNG_ARM_NEON_CHECK_SUPPORTED], [],
+                   [Check for ARM Neon support at run-time]);;
+      api)
+         AC_DEFINE([PNG_ARM_NEON_API_SUPPORTED], [],
+                   [Turn on ARM Neon optimizations at run-time]);;
+      yes|on)
+         AC_DEFINE([PNG_ARM_NEON_OPT], [2],
+                   [Enable ARM Neon optimizations])
+         AC_MSG_WARN([--enable-arm-neon: please specify 'check' or 'api', if]
+            [you want the optimizations unconditionally pass -mfpu=neon]
+            [to the compiler.]);;
+      *)
+         AC_MSG_ERROR([--enable-arm-neon=${enable_arm_neon}: invalid value])
+   esac])
+
+# Add ARM specific files to all builds where the host_cpu is arm ('arm*') or
+# where ARM optimizations were explicitly requested (this allows a fallback if a
+# future host CPU does not match 'arm*')
+
+AM_CONDITIONAL([PNG_ARM_NEON],
+   [test "$enable_arm_neon" != 'no' &&
+    case "$host_cpu" in
+      arm*) :;;
+      *)    test "$enable_arm_neon" != '';;
+    esac])
+
+AC_MSG_NOTICE([[Extra options for compiler: $PNG_COPTS]])
+
 # Config files, substituting as above
-AC_CONFIG_FILES([Makefile libpng.pc:libpng.pc.in])
-AC_CONFIG_FILES([libpng-config:libpng-config.in],
-		[chmod +x libpng-config])
+AC_CONFIG_FILES([Makefile libpng.pc])
+AC_CONFIG_FILES([libpng-config], [chmod +x libpng-config])
 
 AC_OUTPUT

contrib/README.txt

@@ -0,0 +1,4 @@
+
+This "contrib" directory contains contributions which are not necessarily under
+the libpng license, although all are open source.  They are not part of
+libpng proper and are not used for building the library.

contrib/arm-neon/README

@@ -0,0 +1,83 @@
+OPERATING SYSTEM SPECIFIC ARM NEON DETECTION
+--------------------------------------------
+
+Detection of the ability to exexcute ARM NEON on an ARM processor requires
+operating system support.  (The information is not available in user mode.)
+
+HOW TO USE THIS
+---------------
+
+This directory contains C code fragments that can be included in arm/arm_init.c
+by setting the macro PNG_ARM_NEON_FILE to the file name in "" or <> at build
+time.  This setting is not recorded in pnglibconf.h and can be changed simply by
+rebuilding arm/arm_init.o with the required macro definition.
+
+For any of this code to be used the ARM NEON code must be enabled and run time
+checks must be supported.  I.e.:
+
+#if PNG_ARM_NEON_OPT > 0
+#ifdef PNG_ARM_NEON_CHECK_SUPPORTED
+
+This is done in a 'configure' build by passing configure the argument:
+
+   --enable-arm-neon=check
+
+Apart from the basic Linux implementation in contrib/arm-neon/linux.c this code
+is unsupported.  That means that it is not even compiled on a regular basis and
+may be broken in any given minor release.
+
+FILE FORMAT
+-----------
+
+Each file documents its testing status as of the last time it was tested (which
+may have been a long time ago):
+
+STATUS: one of:
+   SUPPORTED: This indicates that the file is included in the regularly
+         performed test builds and bugs are fixed when discovered.
+   COMPILED: This indicates that the code did compile at least once.  See the
+         more detailed description for the extent to which the result was
+         successful.
+   TESTED: This means the code was fully compiled into the libpng test programs
+         and these were run at least once.
+
+BUG REPORTS: an email address to which to send reports of problems
+
+The file is a fragment of C code. It should not define any 'extern' symbols;
+everything should be static.  It must define the function:
+
+static int png_have_neon(png_structp png_ptr);
+
+That function must return 1 if ARM NEON instructions are supported, 0 if not.
+It must not execute png_error unless it detects a bug.  A png_error will prevent
+the reading of the PNG and in the future, writing too.
+
+BUG REPORTS
+-----------
+
+If you mail a bug report for any file that is not SUPPORTED there may only be
+limited response.  Consider fixing it and sending a patch to fix the problem -
+this is more likely to result in action.
+
+CONTRIBUTIONS
+-------------
+
+You may send contributions of new implementations to
+png-mng-implement@sourceforge.net.  Please write code in strict C90 C where
+possible.  Obviously OS dependencies are to be expected.  If you submit code you
+must have the authors permission and it must have a license that is acceptable
+to the current maintainer; in particular that license must permit modification
+and redistribution.
+
+Please try to make the contribution a single file and give the file a clear and
+unambiguous name that identifies the target OS.  If multiple files really are
+required put them all in a sub-directory.
+
+You must also be prepared to handle bug reports from users of the code, either
+by joining the png-mng-implement mailing list or by providing an email for the
+"BUG REPORTS" entry or both.  Please make sure that the header of the file
+contains the STATUS and BUG REPORTS fields as above.
+
+Please list the OS requirements as precisely as possible.  Ideally you should
+also list the environment in which the code has been tested and certainly list
+any environments where you suspect it might not work.

contrib/arm-neon/android-ndk.c

@@ -0,0 +1,39 @@
+/* contrib/arm-neon/android-ndk.c
+ *
+ * Copyright (c) 2014 Glenn Randers-Pehrson
+ * Written by John Bowler, 2014.
+ * Last changed in libpng 1.6.10 [March 6, 2014]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * SEE contrib/arm-neon/README before reporting bugs
+ *
+ * STATUS: COMPILED, UNTESTED
+ * BUG REPORTS: png-mng-implement@sourceforge.net
+ *
+ * png_have_neon implemented for the Android NDK, see:
+ *
+ * Documentation:
+ *    http://www.kandroid.org/ndk/docs/CPU-ARM-NEON.html
+ *    http://code.google.com/p/android/issues/detail?id=49065
+ *
+ * NOTE: this requires that libpng is built against the Android NDK and linked
+ * with an implementation of the Android ARM 'cpu-features' library.  The code
+ * has been compiled only, not linked: no version of the library has been found,
+ * only the header files exist in the NDK.
+ */
+#include <cpu-features.h>
+
+static int
+png_have_neon(png_structp png_ptr)
+{
+   /* This is a whole lot easier than the linux code, however it is probably
+    * implemented as below, therefore it is better to cache the result (these
+    * function calls may be slow!)
+    */
+   PNG_UNUSED(png_ptr)
+   return android_getCpuFamily() == ANDROID_CPU_FAMILY_ARM &&
+      (android_getCpuFeatures() & ANDROID_CPU_ARM_FEATURE_NEON) != 0;
+}

contrib/arm-neon/linux-auxv.c

@@ -0,0 +1,120 @@
+/* contrib/arm-neon/linux-auxv.c
+ *
+ * Copyright (c) 2014 Glenn Randers-Pehrson
+ * Written by Mans Rullgard, 2011.
+ * Last changed in libpng 1.6.10 [March 6, 2014]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ * 
+ * SEE contrib/arm-neon/README before reporting bugs
+ *
+ * STATUS: COMPILED, TESTED
+ * BUG REPORTS: png-mng-implement@sourceforge.net
+ *
+ * png_have_neon implemented for Linux versions which allow access to
+ * /proc/self/auxv.  This is probably faster, cleaner and safer than the code to
+ * read /proc/cpuinfo in contrib/arm-neon/linux, however it is yet another piece
+ * of potentially untested code and has more complex dependencies than the code
+ * to read cpuinfo.
+ *
+ * This generic __linux__ implementation requires reading /proc/self/auxv and
+ * looking at each element for one that records NEON capabilities.
+ */
+#include <unistd.h> /* for POSIX 1003.1 */
+#include <errno.h>  /* for EINTR */
+
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <elf.h>
+#include <asm/hwcap.h>
+
+/* A read call may be interrupted, in which case it returns -1 and sets errno to
+ * EINTR if nothing was done, otherwise (if something was done) a partial read
+ * may result.
+ */
+static size_t
+safe_read(png_structp png_ptr, int fd, void *buffer_in, size_t nbytes)
+{
+   size_t ntotal = 0;
+   char *buffer = png_voidcast(char*, buffer_in);
+
+   while (nbytes > 0)
+   {
+      unsigned int nread;
+      int iread;
+
+      /* Passing nread > INT_MAX to read is implementation defined in POSIX
+       * 1003.1, therefore despite the unsigned argument portable code must
+       * limit the value to INT_MAX!
+       */
+      if (nbytes > INT_MAX)
+         nread = INT_MAX;
+
+      else
+         nread = (unsigned int)/*SAFE*/nbytes;
+
+      iread = read(fd, buffer, nread);
+
+      if (iread == -1)
+      {
+         /* This is the devil in the details, a read can terminate early with 0
+          * bytes read because of EINTR, yet it still returns -1 otherwise end
+          * of file cannot be distinguished.
+          */
+         if (errno != EINTR)
+         {
+            png_warning(png_ptr, "/proc read failed");
+            return 0; /* I.e., a permanent failure */
+         }
+      }
+
+      else if (iread < 0)
+      {
+         /* Not a valid 'read' result: */
+         png_warning(png_ptr, "OS /proc read bug");
+         return 0;
+      }
+
+      else if (iread > 0)
+      {
+         /* Continue reading until a permanent failure, or EOF */
+         buffer += iread;
+         nbytes -= (unsigned int)/*SAFE*/iread;
+         ntotal += (unsigned int)/*SAFE*/iread;
+      }
+
+      else
+         return ntotal;
+   }
+
+   return ntotal; /* nbytes == 0 */
+}
+
+static int
+png_have_neon(png_structp png_ptr)
+{
+   int fd = open("/proc/self/auxv", O_RDONLY);
+   Elf32_auxv_t aux;
+
+   /* Failsafe: failure to open means no NEON */
+   if (fd == -1)
+   {
+      png_warning(png_ptr, "/proc/self/auxv open failed");
+      return 0;
+   }
+
+   while (safe_read(png_ptr, fd, &aux, sizeof aux) == sizeof aux)
+   {
+      if (aux.a_type == AT_HWCAP && (aux.a_un.a_val & HWCAP_NEON) != 0)
+      {
+         close(fd);
+         return 1;
+      }
+   }
+
+   close(fd);
+   return 0;
+}

contrib/arm-neon/linux.c

@@ -0,0 +1,159 @@
+/* contrib/arm-neon/linux.c
+ *
+ * Copyright (c) 2014 Glenn Randers-Pehrson
+ * Written by John Bowler, 2014.
+ * Last changed in libpng 1.6.10 [March 6, 2014]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * SEE contrib/arm-neon/README before reporting bugs
+ *
+ * STATUS: SUPPORTED
+ * BUG REPORTS: png-mng-implement@sourceforge.net
+ *
+ * png_have_neon implemented for Linux by reading the widely available
+ * pseudo-file /proc/cpuinfo.
+ *
+ * This code is strict ANSI-C and is probably moderately portable, it does
+ * however use <stdio.h> and assumes that /proc/cpuinfo is never localized.
+ */
+#include <stdio.h>
+
+static int
+png_have_neon(png_structp png_ptr)
+{
+   FILE *f = fopen("/proc/cpuinfo", "rb");
+
+   if (f != NULL)
+   {
+      /* This is a simple state machine which reads the input byte-by-byte until
+       * it gets a match on the 'neon' feature or reaches the end of the stream.
+       */
+      static const char ch_feature[] = { 70, 69, 65, 84, 85, 82, 69, 83 };
+      static const char ch_neon[] = { 78, 69, 79, 78 };
+
+      enum
+      {
+         StartLine, Feature, Colon, StartTag, Neon, HaveNeon, SkipTag, SkipLine
+      }  state;
+      int counter;
+
+      for (state=StartLine, counter=0;;)
+      {
+         int ch = fgetc(f);
+
+         if (ch == EOF)
+         {
+            /* EOF means error or end-of-file, return false; neon at EOF is
+             * assumed to be a mistake.
+             */
+            fclose(f);
+            return 0;
+         }
+
+         switch (state)
+         {
+            case StartLine:
+               /* Match spaces at the start of line */
+               if (ch <= 32) /* skip control characters and space */
+                  break;
+
+               counter=0;
+               state = Feature;
+               /* FALL THROUGH */
+
+            case Feature:
+               /* Match 'FEATURE', ASCII case insensitive. */
+               if ((ch & ~0x20) == ch_feature[counter])
+               {
+                  if (++counter == (sizeof ch_feature))
+                     state = Colon;
+                  break;
+               }
+               
+               /* did not match 'feature' */
+               state = SkipLine;
+               /* FALL THROUGH */
+
+            case SkipLine:
+            skipLine:
+               /* Skip everything until we see linefeed or carriage return */
+               if (ch != 10 && ch != 13)
+                  break;
+
+               state = StartLine;
+               break;
+
+            case Colon:
+               /* Match any number of space or tab followed by ':' */
+               if (ch == 32 || ch == 9)
+                  break;
+
+               if (ch == 58) /* i.e. ':' */
+               {
+                  state = StartTag;
+                  break;
+               }
+
+               /* Either a bad line format or a 'feature' prefix followed by
+                * other characters.
+                */
+               state = SkipLine;
+               goto skipLine;
+
+            case StartTag:
+               /* Skip space characters before a tag */
+               if (ch == 32 || ch == 9)
+                  break;
+
+               state = Neon;
+               counter = 0;
+               /* FALL THROUGH */
+
+            case Neon:
+               /* Look for 'neon' tag */
+               if ((ch & ~0x20) == ch_neon[counter])
+               {
+                  if (++counter == (sizeof ch_neon))
+                     state = HaveNeon;
+                  break;
+               }
+
+               state = SkipTag;
+               /* FALL THROUGH */
+
+            case SkipTag:
+               /* Skip non-space characters */
+               if (ch == 10 || ch == 13)
+                  state = StartLine;
+
+               else if (ch == 32 || ch == 9)
+                  state = StartTag;
+               break;
+
+            case HaveNeon:
+               /* Have seen a 'neon' prefix, but there must be a space or new
+                * line character to terminate it.
+                */
+               if (ch == 10 || ch == 13 || ch == 32 || ch == 9)
+               {
+                  fclose(f);
+                  return 1;
+               }
+
+               state = SkipTag;
+               break;
+
+            default:
+               png_error(png_ptr, "png_have_neon: internal error (bug)");
+         }
+      }
+   }
+
+   else
+      png_warning(png_ptr, "/proc/cpuinfo open failed");
+
+   return 0;
+}

contrib/conftest/README

@@ -0,0 +1,49 @@
+This directory contains test configuration files, currently always '.dfa' files
+intended to be used in the build by setting the make macro DFA_XTRA to the name
+of the file.
+
+These files are used in release validation of the 'configure' builds of libpng
+by building 'make check', or 'make all-am' for cross-builds, with each .dfa
+file.
+
+The files in this directory may change between minor releases, however
+contributions describing specific builds of libpng are welcomed.  There is no
+guarantee that libpng will continue to build with such configurations; support
+for given configurations can be, and has been, dropped between successive minor
+releases.  However if a .dfa file describing a configuration is not in this
+directory it is very unlikely that it will be tested before a minor release!
+
+You can use these .dfa files as the basis of new configurations.  Files in this
+directory should not have any use restrictions or restrictive licenses.
+
+This directory is not included in the .zip and .7z distributions, which do
+not contain 'configure' scripts.
+
+DOCUMENTATION
+=============
+
+Examples:
+   ${srcdir}/pngusr.dfa
+   ${srcdir}/contrib/pngminim/*/pngusr.dfa
+
+Documentation of the options:
+   ${srcdir}/scripts/pnglibconf.dfa
+
+Documentation of the file format:
+   ${srcdir}/scripts/options.awk
+
+FILE NAMING
+===========
+
+File names in this directory may NOT contain any of the five characters:
+
+   - , + * ?
+
+Neither may they contain any space character.
+
+While other characters may be used it is strongly suggested that file names be
+limited to lower case Latiin alphabetic characters (a-z), digits (0-9) and, if
+necessary the underscore (_) character.  File names should be about 8 characters
+long (excluding the .dfa extension).  Submitted .dfa files should have names
+between 7 and 16 characters long, shorter names (6 characters or less) are
+reserved for standard tests.

contrib/conftest/read.dfa

@@ -0,0 +1,58 @@
+# read.dfa
+#  Build time configuration of libpng
+#
+# Author: John Bowler
+# Copyright: (c) John Bowler, 2013
+# Usage rights:
+#  To the extent possible under law, the author has waived all copyright and
+#  related or neighboring rights to this work.  This work is published from:
+#  United States.
+#
+# Build libpng with basic read support.  This enables the lowest level libpng
+# read API - the one where the calling code has to use a loop to read each row.
+# At present this is the API used by most programs.
+#
+# Support is enabled only for those chunks and transformations that are
+# typically required - others can be added easily.
+#
+
+everything = off
+
+# The sequential read code is enabled here; the progressive code can be used
+# instead but there is no point enabling both.
+
+option SEQUENTIAL_READ on
+
+# Likewise it is pointless enabling both fixed and floating point APIs.  Choose
+# one or the other for both the API and the internal math.
+
+#Fixed point:
+#option FIXED_POINT on
+#option FLOATING_ARITHMETIC off
+
+#Floating point:
+option FLOATING_POINT on
+option FLOATING_ARITHMETIC on
+
+# Basic error handling, IO and user memory support.  The latter allows the
+# application program to provide its own implementations of 'malloc' and 'free'.
+option SETJMP on
+option STDIO on
+option USER_MEM on
+
+# To read the full set of PNG images correctly interlace, transparency and
+# 16-bit support is required.  The application can implement interlace itself,
+# but very few do and it's no longer possible to disable it when READ is
+# enabled.
+option READ_tRNS on
+option READ_16BIT on
+
+# Everything else is application dependent.  This file assumes the app handles
+# all the native PNG bit layouts, so it doesn't need any of layout change
+# transforms, but needs libpng to perform gamma correction.  It doesn't do any
+# colorspace stuff and ignores the 'significant bit' information.
+#
+# If your app always expands the image to a limited set of bit layouts you
+# probably want to consider using the simplified API instead of the low level
+# one - see png.h and s_read.dfa.
+option READ_GAMMA on

contrib/conftest/s_read.dfa

@@ -0,0 +1,35 @@
+# s_read.dfa
+#  Build time configuration of libpng
+#
+# Author: John Bowler
+# Copyright: (c) John Bowler, 2013
+# Usage rights:
+#  To the extent possible under law, the author has waived all copyright and
+#  related or neighboring rights to this work.  This work is published from:
+#  United States.
+#
+# Build libpng with simplified read support (only).  This builds a minimal
+# libpng able to read all PNG formats and convert them into a small number of
+# well understood memory formats.
+#
+
+everything = off
+
+option SIMPLIFIED_READ on
+
+# It isn't necessary to chose fixed or floating point for the APIs because the
+# simplified API doesn't need fixed or floating point numbers.  It is necessary
+# to chose an internal math implementation.  The default (because of 'everything
+# = off') is fixed point - turn the floating point implementation on if you have
+# hardware floating point or prefer your software floating point implementation.
+option FLOATING_ARITHMETIC on
+
+# This is not strictly necessary, but without it the message strings in the API
+# will not be filled in
+option ERROR_TEXT on
+
+# Switching these options on enables the 'AFIRST' and 'BGR' formats - you don't
+# need this if you don't use them, they just allow the in-memory layout to be
+# changed to match common hardware formats.
+option SIMPLIFIED_READ_AFIRST on
+option SIMPLIFIED_READ_BGR on

contrib/conftest/s_write.dfa

@@ -0,0 +1,33 @@
+# s_write.dfa
+#  Build time configuration of libpng
+#
+# Author: John Bowler
+# Copyright: (c) John Bowler, 2013
+# Usage rights:
+#  To the extent possible under law, the author has waived all copyright and
+#  related or neighboring rights to this work.  This work is published from:
+#  United States.
+#
+# Build libpng with (just) simplified write support
+#
+
+everything = off
+
+option SIMPLIFIED_WRITE on
+
+# It isn't necessary to chose fixed or floating point for the APIs because the
+# simplified API doesn't need fixed or floating point numbers.  It is necessary
+# to chose an internal math implementation.  The default (because of 'everything
+# = off') is fixed point - turn the floating point implementation on if you have
+# hardware floating point or prefer your software floating point implementation.
+option FLOATING_ARITHMETIC on
+
+# This is not strictly necessary, but without it the message strings in the API
+# will not be filled in
+option ERROR_TEXT on
+
+# Switching these options on enables the 'AFIRST' and 'BGR' formats - you don't
+# need this if you don't use them, they just allow the in-memory layout to be
+# changed to match common hardware formats.
+option SIMPLIFIED_WRITE_AFIRST on
+option SIMPLIFIED_WRITE_BGR on

contrib/conftest/simple.dfa

@@ -0,0 +1,36 @@
+# simple.dfa
+#  Build time configuration of libpng
+#
+# Author: John Bowler
+# Copyright: (c) John Bowler, 2013
+# Usage rights:
+#  To the extent possible under law, the author has waived all copyright and
+#  related or neighboring rights to this work.  This work is published from:
+#  United States.
+#
+# Build libpng with just the simplified APIs (read and write).
+#
+
+everything = off
+
+option SIMPLIFIED_WRITE on
+option SIMPLIFIED_READ on
+
+# It isn't necessary to chose fixed or floating point for the APIs because the
+# simplified API doesn't need fixed or floating point numbers.  It is necessary
+# to chose an internal math implementation.  The default (because of 'everything
+# = off') is fixed point - turn the floating point implementation on if you have
+# hardware floating point or prefer your software floating point implementation.
+option FLOATING_ARITHMETIC on
+
+# This is not strictly necessary, but without it the message strings in the API
+# will not be filled in
+option ERROR_TEXT on
+
+# Switching these options on enables the 'AFIRST' and 'BGR' formats - you don't
+# need this if you don't use them, they just allow the in-memory layout to be
+# changed to match common hardware formats.
+option SIMPLIFIED_READ_AFIRST on
+option SIMPLIFIED_READ_BGR on
+option SIMPLIFIED_WRITE_AFIRST on
+option SIMPLIFIED_WRITE_BGR on

contrib/conftest/write.dfa

@@ -0,0 +1,45 @@
+# write.dfa
+#  Build time configuration of libpng
+#
+# Author: John Bowler
+# Copyright: (c) John Bowler, 2013
+# Usage rights:
+#  To the extent possible under law, the author has waived all copyright and
+#  related or neighboring rights to this work.  This work is published from:
+#  United States.
+#
+# Build libpng with no read support and minimal write support.
+#
+
+everything = off
+
+# Switch on the write code - this makes a minimalist encoder
+
+option WRITE on
+
+# Choose fixed or floating point APIs and arithmetic.  The choices are
+# independent but normally they will match.  It is typically better to use the
+# floating point if you have floating point hardware.  If you don't know, or
+# (perhaps) to make libpng smaller used fixed point throughout.
+
+#Fixed point:
+#option FIXED_POINT on
+#option FLOATING_ARITHMETIC off
+
+#Floating point:
+option FLOATING_POINT on
+option FLOATING_ARITHMETIC on
+
+# Basic error handling, IO and user memory support.  The latter allows the
+# application program to provide its own implementations of 'malloc' and 'free'.
+option SETJMP on
+option STDIO on
+option USER_MEM on
+
+# Everything else is optional.  Unlike the read code in libpng the write code
+# does not need to deal with arbitrary formats, so only add support for things
+# you really do write!  For example you might only write sRGB images, sometimes
+# with transparency and never write 16 bit images, so:
+option WRITE_sRGB on
+option WRITE_tRNS on
+#option WRITE_16BIT off (this is the default with 'everything = off')

contrib/examples/README.txt

@@ -0,0 +1,24 @@
+
+This directory (contrib/examples) contains examples of libpng usage.
+
+NO COPYRIGHT RIGHTS ARE CLAIMED TO ANY OF THE FILES IN THIS DIRECTORY.
+
+To the extent possible under law, the authors have waived all copyright and
+related or neighboring rights to this work.  This work is published from:
+United States.
+
+The files may be used freely in any way.  The intention is that appropriate
+parts of the files be used in other libpng-using programs without any need for
+the authors of the using code to seek copyright or license from the original
+authors.
+
+The source code and comments in this directory are the original work of the
+people named below.  No other person or organization has made contributions to
+the work in this directory.
+
+ORIGINAL AUTHORS
+    The following people have contributed to the code in this directory.  None
+    of the people below claim any rights with regard to the contents of this
+    directory.
+
+    John Bowler <jbowler@acm.org>

contrib/examples/iccfrompng.c

@@ -0,0 +1,180 @@
+/*- iccfrompng
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2011.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Extract any icc profiles found in the given PNG files.  This is a simple
+ * example of a program that extracts information from the header of a PNG file
+ * without processing the image.  Notice that some header information may occur
+ * after the image data. Textual data and comments are an example; the approach
+ * in this file won't work reliably for such data because it only looks for the
+ * information in the section of the file that preceeds the image data.
+ *
+ * Compile and link against libpng and zlib, plus anything else required on the
+ * system you use.
+ *
+ * To use supply a list of PNG files containing iCCP chunks, the chunks will be
+ * extracted to a similarly named file with the extension replaced by 'icc',
+ * which will be overwritten without warning.
+ */
+#include <stdlib.h>
+#include <setjmp.h>
+#include <string.h>
+#include <stdio.h>
+
+#include <png.h>
+
+static int verbose = 1;
+static png_byte no_profile[] = "no profile";
+
+static png_bytep
+extract(FILE *fp, png_uint_32 *proflen)
+{
+   png_structp png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,0,0,0);
+   png_infop info_ptr = NULL;
+   png_bytep result = NULL;
+
+   /* Initialize for error or no profile: */
+   *proflen = 0;
+
+   if (png_ptr == NULL)
+   {
+      fprintf(stderr, "iccfrompng: version library mismatch?\n");
+      return 0;
+   }
+
+   if (setjmp(png_jmpbuf(png_ptr)))
+   {
+      png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+      return 0;
+   }
+
+   png_init_io(png_ptr, fp);
+
+   info_ptr = png_create_info_struct(png_ptr);
+   if (info_ptr == NULL)
+      png_error(png_ptr, "OOM allocating info structure");
+
+   png_read_info(png_ptr, info_ptr);
+
+   {
+      png_charp name;
+      int compression_type;
+      png_bytep profile;
+
+      if (png_get_iCCP(png_ptr, info_ptr, &name, &compression_type, &profile,
+         proflen) & PNG_INFO_iCCP)
+      {
+         result = malloc(*proflen);
+         if (result != NULL)
+            memcpy(result, profile, *proflen);
+
+         else
+            png_error(png_ptr, "OOM allocating profile buffer");
+      }
+
+      else
+	result = no_profile;
+   }
+
+   png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+   return result;
+}
+
+static int
+extract_one_file(const char *filename)
+{
+   int result = 0;
+   FILE *fp = fopen(filename, "rb");
+
+   if (fp != NULL)
+   {
+      png_uint_32 proflen = 0;
+      png_bytep profile = extract(fp, &proflen);
+
+      if (profile != NULL && profile != no_profile)
+      {
+         size_t len;
+         char *output;
+
+         {
+            const char *ep = strrchr(filename, '.');
+
+            if (ep != NULL)
+               len = ep-filename;
+
+            else
+               len = strlen(filename);
+         }
+
+         output = malloc(len + 5);
+         if (output != NULL)
+         {
+            FILE *of;
+
+            memcpy(output, filename, len);
+            strcpy(output+len, ".icc");
+
+            of = fopen(output, "wb");
+            if (of != NULL)
+            {
+               if (fwrite(profile, proflen, 1, of) == 1 &&
+                  fflush(of) == 0 &&
+                  fclose(of) == 0)
+               {
+                  if (verbose)
+                     printf("%s -> %s\n", filename, output);
+                  /* Success return */
+                  result = 1;
+               }
+
+               else
+               {
+                  fprintf(stderr, "%s: error writing profile\n", output);
+                  if (remove(output))
+                     fprintf(stderr, "%s: could not remove file\n", output);
+               }
+            }
+
+            else
+               fprintf(stderr, "%s: failed to open output file\n", output);
+
+            free(output);
+         }
+
+         else
+            fprintf(stderr, "%s: OOM allocating string!\n", filename);
+
+         free(profile);
+      }
+
+      else if (verbose && profile == no_profile)
+	printf("%s has no profile\n", filename);
+   }
+
+   else
+      fprintf(stderr, "%s: could not open file\n", filename);
+
+   return result;
+}
+
+int
+main(int argc, char **argv)
+{
+   int i;
+   int extracted = 0;
+
+   for (i=1; i<argc; ++i)
+   {
+      if (strcmp(argv[i], "-q") == 0)
+         verbose = 0;
+
+      else if (extract_one_file(argv[i]))
+         extracted = 1;
+   }
+
+   /* Exit code is true if any extract succeeds */
+   return extracted == 0;
+}

contrib/examples/pngpixel.c

@@ -0,0 +1,368 @@
+/*- pngpixel
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2011.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Read a single pixel value from a PNG file.
+ *
+ * This code illustrates basic 'by-row' reading of a PNG file using libpng.
+ * Rows are read until a particular pixel is found; the value of this pixel is
+ * then printed on stdout.
+ *
+ * The code illustrates how to do this on interlaced as well as non-interlaced
+ * images.  Normally you would call png_set_interlace_handling() to have libpng
+ * deal with the interlace for you, but that obliges you to buffer half of the
+ * image to assemble the interlaced rows.  In this code
+ * png_set_interlace_handling() is not called and, instead, the code handles the
+ * interlace passes directly looking for the required pixel.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <setjmp.h> /* required for error handling */
+
+/* Normally use <png.h> here to get the installed libpng, but this is done to
+ * ensure the code picks up the local libpng implementation:
+ */
+#include "../../png.h"
+
+/* Return component 'c' of pixel 'x' from the given row. */
+static unsigned int
+component(png_const_bytep row, png_uint_32 x, unsigned int c,
+   unsigned int bit_depth, unsigned int channels)
+{
+   /* PNG images can be up to 2^31 pixels wide, but this means they can be up to
+    * 2^37 bits wide (for a 64-bit pixel - the largest possible) and hence 2^34
+    * bytes wide.  Since the row fitted into memory, however, the following must
+    * work:
+    */
+   png_uint_32 bit_offset_hi = bit_depth * ((x >> 6) * channels);
+   png_uint_32 bit_offset_lo = bit_depth * ((x & 0x3f) * channels + c);
+
+   row = (png_const_bytep)(((PNG_CONST png_byte (*)[8])row) + bit_offset_hi);
+   row += bit_offset_lo >> 3;
+   bit_offset_lo &= 0x07;
+
+   /* PNG pixels are packed into bytes to put the first pixel in the highest
+    * bits of the byte and into two bytes for 16-bit values with the high 8 bits
+    * first, so:
+    */
+   switch (bit_depth)
+   {
+      case 1: return (row[0] >> (7-bit_offset_lo)) & 0x01;
+      case 2: return (row[0] >> (6-bit_offset_lo)) & 0x03;
+      case 4: return (row[0] >> (4-bit_offset_lo)) & 0x0f;
+      case 8: return row[0];
+      case 16: return (row[0] << 8) + row[1];
+      default:
+         /* This should never happen; it indicates a bug in this program or in
+          * libpng itself:
+          */
+         fprintf(stderr, "pngpixel: invalid bit depth %u\n", bit_depth);
+         exit(1);
+   }
+}
+
+/* Print a pixel from a row returned by libpng; determine the row format, find
+ * the pixel, and print the relevant information to stdout.
+ */
+static void
+print_pixel(png_structp png_ptr, png_infop info_ptr, png_const_bytep row,
+   png_uint_32 x)
+{
+   PNG_CONST unsigned int bit_depth = png_get_bit_depth(png_ptr, info_ptr);
+
+   switch (png_get_color_type(png_ptr, info_ptr))
+   {
+      case PNG_COLOR_TYPE_GRAY:
+         printf("GRAY %u\n", component(row, x, 0, bit_depth, 1));
+         return;
+
+      /* The palette case is slightly more difficult - the palette and, if
+       * present, the tRNS ('transparency', though the values are really
+       * opacity) data must be read to give the full picture:
+       */
+      case PNG_COLOR_TYPE_PALETTE:
+         {
+            PNG_CONST unsigned int index = component(row, x, 0, bit_depth, 1);
+            png_colorp palette = NULL;
+            int num_palette = 0;
+
+            if ((png_get_PLTE(png_ptr, info_ptr, &palette, &num_palette) &
+               PNG_INFO_PLTE) && num_palette > 0 && palette != NULL)
+            {
+               png_bytep trans_alpha = NULL;
+               int num_trans = 0;
+               if ((png_get_tRNS(png_ptr, info_ptr, &trans_alpha, &num_trans,
+                  NULL) & PNG_INFO_tRNS) && num_trans > 0 &&
+                  trans_alpha != NULL)
+                  printf("INDEXED %u = %d %d %d %d\n", index,
+                     palette[index].red, palette[index].green,
+                     palette[index].blue,
+                     index < num_trans ? trans_alpha[index] : 255);
+
+               else /* no transparency */
+                  printf("INDEXED %u = %d %d %d\n", index,
+                     palette[index].red, palette[index].green,
+                     palette[index].blue);
+            }
+
+            else
+               printf("INDEXED %u = invalid index\n", index);
+         }
+         return;
+
+      case PNG_COLOR_TYPE_RGB:
+         printf("RGB %u %u %u\n", component(row, x, 0, bit_depth, 3),
+            component(row, x, 1, bit_depth, 3),
+            component(row, x, 2, bit_depth, 3));
+         return;
+
+      case PNG_COLOR_TYPE_GRAY_ALPHA:
+         printf("GRAY+ALPHA %u %u\n", component(row, x, 0, bit_depth, 2),
+            component(row, x, 1, bit_depth, 2));
+         return;
+
+      case PNG_COLOR_TYPE_RGB_ALPHA:
+         printf("RGBA %u %u %u %u\n", component(row, x, 0, bit_depth, 4),
+            component(row, x, 1, bit_depth, 4),
+            component(row, x, 2, bit_depth, 4),
+            component(row, x, 3, bit_depth, 4));
+         return;
+
+      default:
+         png_error(png_ptr, "pngpixel: invalid color type");
+   }
+}
+
+int main(int argc, const char **argv)
+{
+   /* This program uses the default, <setjmp.h> based, libpng error handling
+    * mechanism, therefore any local variable that exists before the call to
+    * setjmp and is changed after the call to setjmp returns successfully must
+    * be declared with 'volatile' to ensure that their values don't get
+    * destroyed by longjmp:
+    */
+   volatile int result = 1/*fail*/;
+
+   if (argc == 4)
+   {
+      long x = atol(argv[1]);
+      long y = atol(argv[2]);
+      FILE *f = fopen(argv[3], "rb");
+      volatile png_bytep row = NULL;
+
+      if (f != NULL)
+      {
+         /* libpng requires a callback function for handling errors; this
+          * callback must not return.  The default callback function uses a
+          * stored <setjmp.h> style jmp_buf which is held in a png_struct and
+          * writes error messages to stderr.  Creating the png_struct is a
+          * little tricky; just copy the following code.
+          */
+         png_structp png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,
+            NULL, NULL, NULL);
+
+         if (png_ptr != NULL)
+         {
+            png_infop info_ptr = png_create_info_struct(png_ptr);
+
+            if (info_ptr != NULL)
+            {
+               /* Declare stack variables to hold pointers to locally allocated
+                * data.
+                */
+
+               /* Initialize the error control buffer: */
+               if (setjmp(png_jmpbuf(png_ptr)) == 0)
+               {
+                  png_uint_32 width, height;
+                  int bit_depth, color_type, interlace_method,
+                     compression_method, filter_method;
+                  png_bytep row_tmp;
+
+                  /* Now associate the recently opened (FILE*) with the default
+                   * libpng initialization functions.  Sometimes libpng is
+                   * compiled without stdio support (it can be difficult to do
+                   * in some environments); in that case you will have to write
+                   * your own read callback to read data from the (FILE*).
+                   */
+                  png_init_io(png_ptr, f);
+
+                  /* And read the first part of the PNG file - the header and
+                   * all the information up to the first pixel.
+                   */
+                  png_read_info(png_ptr, info_ptr);
+
+                  /* This fills in enough information to tell us the width of
+                   * each row in bytes, allocate the appropriate amount of
+                   * space.  In this case png_malloc is used - it will not
+                   * return if memory isn't available.
+                   */
+                  row = png_malloc(png_ptr, png_get_rowbytes(png_ptr,
+                     info_ptr));
+
+                  /* To avoid the overhead of using a volatile auto copy row_tmp
+                   * to a local here - just use row for the png_free below.
+                   */
+                  row_tmp = row;
+
+                  /* All the information we need is in the header is returned by
+                   * png_get_IHDR, if this fails we can now use 'png_error' to
+                   * signal the error and return control to the setjmp above.
+                   */
+                  if (png_get_IHDR(png_ptr, info_ptr, &width, &height,
+                     &bit_depth, &color_type, &interlace_method,
+                     &compression_method, &filter_method))
+                  {
+                     int passes, pass;
+
+                     /* png_set_interlace_handling returns the number of
+                      * passes required as well as turning on libpng's
+                      * handling, but since we do it ourselves this is
+                      * necessary:
+                      */
+                     switch (interlace_method)
+                     {
+                        case PNG_INTERLACE_NONE:
+                           passes = 1;
+                           break;
+
+                        case PNG_INTERLACE_ADAM7:
+                           passes = PNG_INTERLACE_ADAM7_PASSES;
+                           break;
+
+                        default:
+                           png_error(png_ptr, "pngpixel: unknown interlace");
+                     }
+
+                     /* Now read the pixels, pass-by-pass, row-by-row: */
+                     png_start_read_image(png_ptr);
+
+                     for (pass=0; pass<passes; ++pass)
+                     {
+                        png_uint_32 ystart, xstart, ystep, xstep;
+                        png_uint_32 py;
+
+                        if (interlace_method == PNG_INTERLACE_ADAM7)
+                        {
+                           /* Sometimes the whole pass is empty because the
+                            * image is too narrow or too short.  libpng
+                            * expects to be called for each row that is
+                            * present in the pass, so it may be necessary to
+                            * skip the loop below (over py) if the image is
+                            * too narrow.
+                            */
+                           if (PNG_PASS_COLS(width, pass) == 0)
+                              continue;
+
+                           /* We need the starting pixel and the offset
+                            * between each pixel in this pass; use the macros
+                            * in png.h:
+                            */
+                           xstart = PNG_PASS_START_COL(pass);
+                           ystart = PNG_PASS_START_ROW(pass);
+                           xstep = PNG_PASS_COL_OFFSET(pass);
+                           ystep = PNG_PASS_ROW_OFFSET(pass);
+                        }
+
+                        else
+                        {
+                           ystart = xstart = 0;
+                           ystep = xstep = 1;
+                        }
+
+                        /* To find the pixel, loop over 'py' for each pass
+                         * reading a row and then checking to see if it
+                         * contains the pixel.
+                         */
+                        for (py = ystart; py < height; py += ystep)
+                        {
+                           png_uint_32 px, ppx;
+
+                           /* png_read_row takes two pointers.  When libpng
+                            * handles the interlace the first is filled in
+                            * pixel-by-pixel, and the second receives the same
+                            * pixels but they are replicated across the
+                            * unwritten pixels so far for each pass.  When we
+                            * do the interlace, however, they just contain
+                            * the pixels from the interlace pass - giving
+                            * both is wasteful and pointless, so we pass a
+                            * NULL pointer.
+                            */
+                           png_read_row(png_ptr, row_tmp, NULL);
+
+                           /* Now find the pixel if it is in this row; there
+                            * are, of course, much better ways of doing this
+                            * than using a for loop:
+                            */
+                           if (y == py) for (px = xstart, ppx = 0;
+                              px < width; px += xstep, ++ppx) if (x == px)
+                           {
+                              /* 'ppx' is the index of the pixel in the row
+                               * buffer.
+                               */
+                              print_pixel(png_ptr, info_ptr, row_tmp, ppx);
+
+                              /* Now terminate the loops early - we have
+                               * found and handled the required data.
+                               */
+                              goto pass_loop_end;
+                           } /* x loop */
+                        } /* y loop */
+                     } /* pass loop */
+
+                     /* Finally free the temporary buffer: */
+                  pass_loop_end:
+                     row = NULL;
+                     png_free(png_ptr, row_tmp);
+                  }
+
+                  else
+                     png_error(png_ptr, "pngpixel: png_get_IHDR failed");
+
+               }
+
+               else
+               {
+                  /* Else libpng has raised an error.  An error message has
+                   * already been output, so it is only necessary to clean up
+                   * locally allocated data:
+                   */
+                  if (row != NULL)
+                  {
+                     /* The default implementation of png_free never errors out
+                      * (it just crashes if something goes wrong), but the safe
+                      * way of using it is still to clear 'row' before calling
+                      * png_free:
+                      */
+                     png_bytep row_tmp = row;
+                     row = NULL;
+                     png_free(png_ptr, row_tmp);
+                  }
+               }
+
+               png_destroy_info_struct(png_ptr, &info_ptr);
+            }
+
+            else
+               fprintf(stderr, "pngpixel: out of memory allocating png_info\n");
+
+            png_destroy_read_struct(&png_ptr, NULL, NULL);
+         }
+
+         else
+            fprintf(stderr, "pngpixel: out of memory allocating png_struct\n");
+      }
+
+      else
+         fprintf(stderr, "pngpixel: %s: could not open file\n", argv[3]);
+   }
+
+   else
+      /* Wrong number of arguments */
+      fprintf(stderr, "pngpixel: usage: pngpixel x y png-file\n");
+
+   return result;
+}

contrib/examples/pngtopng.c

@@ -0,0 +1,92 @@
+/*- pngtopng
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2011.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Read a PNG and write it out in a fixed format, using the 'simplified API'
+ * that was introduced in libpng-1.6.0.
+ *
+ * This sample code is just the code from the top of 'example.c' with some error
+ * handling added.  See example.c for more comments.
+ */
+#include <stddef.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+
+/* Normally use <png.h> here to get the installed libpng, but this is done to
+ * ensure the code picks up the local libpng implementation:
+ */
+#include "../../png.h"
+
+int main(int argc, const char **argv)
+{
+   int result = 1;
+
+   if (argc == 3)
+   {
+      png_image image;
+
+      /* Only the image structure version number needs to be set. */
+      memset(&image, 0, sizeof image);
+      image.version = PNG_IMAGE_VERSION;
+
+      if (png_image_begin_read_from_file(&image, argv[1]))
+      {
+         png_bytep buffer;
+
+         /* Change this to try different formats!  If you set a colormap format
+          * then you must also supply a colormap below.
+          */
+         image.format = PNG_FORMAT_RGBA;
+
+         buffer = malloc(PNG_IMAGE_SIZE(image));
+
+         if (buffer != NULL)
+         {
+            if (png_image_finish_read(&image, NULL/*background*/, buffer,
+               0/*row_stride*/, NULL/*colormap for PNG_FORMAT_FLAG_COLORMAP */))
+            {
+               if (png_image_write_to_file(&image, argv[2],
+                  0/*convert_to_8bit*/, buffer, 0/*row_stride*/,
+                  NULL/*colormap*/))
+                  result = 0;
+
+               else
+                  fprintf(stderr, "pngtopng: write %s: %s\n", argv[2],
+                      image.message);
+
+               free(buffer);
+            }
+
+            else
+            {
+               fprintf(stderr, "pngtopng: read %s: %s\n", argv[1],
+                   image.message);
+
+               /* This is the only place where a 'free' is required; libpng does
+                * the cleanup on error and success, but in this case we couldn't
+                * complete the read because of running out of memory.
+                */
+               png_image_free(&image);
+            }
+         }
+
+         else
+            fprintf(stderr, "pngtopng: out of memory: %lu bytes\n",
+               (unsigned long)PNG_IMAGE_SIZE(image));
+      }
+
+      else
+         /* Failed to read the first argument: */
+         fprintf(stderr, "pngtopng: %s: %s\n", argv[1], image.message);
+   }
+
+   else
+      /* Wrong number of arguments */
+      fprintf(stderr, "pngtopng: usage: pngtopng input-file output-file\n");
+
+   return result;
+}

contrib/gregbook/Makefile.mingw32

@@ -47,7 +47,8 @@ CC = gcc
 #CC = i386-mingw32msvc-gcc #	e.g., Linux -> Win32 cross-compilation
 LD = $(CC)
 RM = rm -f
-CFLAGS = -O -Wall $(INCS) $(MINGW_CCFLAGS)
+CPPFLAGS = $(INCS)
+CFLAGS = -O -Wall $(MINGW_CCFLAGS)
 # [note that -Wall is a gcc-specific compilation flag ("most warnings on")]
 # [-ansi, -pedantic and -W can also be used]
 LDFLAGS = $(MINGW_LDFLAGS)
@@ -85,10 +86,10 @@ EXES = $(STATIC_EXES) $(DYNAMIC_EXES)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) -c $(CPPFLAGS) $(CFLAGS) $<
 
 %.pic$(O): %.c
-	$(CC) -c $(CFLAGS) -DPNG_BUILD_DLL -o $@ $<
+	$(CC) -c $(CPPFLAGS) $(CFLAGS) -DPNG_BUILD_DLL -o $@ $<
 
 
 # dependencies --------------------------------------------------------------

contrib/gregbook/Makefile.sgi

@@ -23,9 +23,9 @@
 
 # macros --------------------------------------------------------------------
 
-PNGINC = -I/usr/local/include/libpng15
-PNGLIB = -L/usr/local/lib -lpng15	  # dynamically linked against libpng
-#PNGLIB = /usr/local/lib/libpng15.a # statically linked against libpng
+PNGINC = -I/usr/local/include/libpng17
+PNGLIB = -L/usr/local/lib -lpng17	  # dynamically linked against libpng
+#PNGLIB = /usr/local/lib/libpng17.a # statically linked against libpng
 # or:
 #PNGINC = -I../..
 #PNGLIB = -L../.. -lpng
@@ -53,7 +53,8 @@ CC = cc
 LD = cc
 RM = rm -f
 # ABI must be the same as that used to build libpng.
-ABI=
+ABI =
+CPPFLAGS =
 CFLAGS = $(ABI) -O -fullwarn $(INCS)
 LDFLAGS = $(ABI)
 O = .o
@@ -73,7 +74,7 @@ EXES = $(RPNG)$(E) $(RPNG2)$(E) $(WPNG)$(E)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) -c $(CPPFLAGS) $(CFLAGS) $<
 
 
 # dependencies --------------------------------------------------------------

contrib/gregbook/Makefile.unx

@@ -26,14 +26,14 @@
 # macros --------------------------------------------------------------------
 
 #PNGDIR = /usr/local/lib
-#PNGINC = -I/usr/local/include/libpng15
-#PNGLIBd = -L$(PNGDIR) -lpng15 # dynamically linked, installed libpng
-#PNGLIBs = $(PNGDIR)/libpng15.a # statically linked, installed libpng
+#PNGINC = -I/usr/local/include/libpng17
+#PNGLIBd = -L$(PNGDIR) -lpng17 # dynamically linked, installed libpng
+#PNGLIBs = $(PNGDIR)/libpng17.a # statically linked, installed libpng
 # or:
 PNGDIR = ../..#	this one is for libpng-x.y.z/contrib/gregbook builds
 #PNGDIR = ../libpng
 PNGINC = -I$(PNGDIR)
-PNGLIBd = -Wl,-rpath,$(PNGDIR) -L$(PNGDIR) -lpng15	# dynamically linked
+PNGLIBd = -Wl,-rpath,$(PNGDIR) -L$(PNGDIR) -lpng17	# dynamically linked
 PNGLIBs = $(PNGDIR)/libpng.a		# statically linked, local libpng
 
 ZDIR = /usr/local/lib
@@ -64,7 +64,8 @@ WLIBSs = $(PNGLIBs) $(ZLIBs)
 CC = gcc
 LD = gcc
 RM = rm -f
-CFLAGS = -O -Wall $(INCS) -DFEATURE_LOOP
+CPPFLAGS = $(INCS) -DFEATURE_LOOP
+CFLAGS = -O -Wall
 # [note that -Wall is a gcc-specific compilation flag ("most warnings on")]
 # [-ansi, -pedantic and -W can also be used]
 LDFLAGS =
@@ -92,7 +93,7 @@ EXES = $(STATIC_EXES) $(DYNAMIC_EXES)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) -c $(CPPFLAGS) $(CFLAGS) $<
 
 
 # dependencies --------------------------------------------------------------

contrib/gregbook/Makefile.w32

@@ -18,7 +18,7 @@
 # "c:\windows\command.com /e:4096" as the program command line and set the
 # working directory to this directory.  Then double-click to open the new
 # DOS-prompt window with a bigger environment and retry the commands above.
-# 
+#
 # This makefile assumes libpng and zlib have already been built or downloaded
 # and are in subdirectories at the same level as the current subdirectory
 # (as indicated by the PNGPATH and ZPATH macros below).  Edit as appropriate.
@@ -53,7 +53,8 @@ WLIBS = $(PNGLIB) $(ZLIB)
 CC = cl
 LD = link
 RM = del
-CFLAGS = -nologo -O -W3 $(INCS) $(cvars)
+CPPFLAGS = $(INCS)
+CFLAGS = -nologo -O -W3 $(cvars)
 # [note that -W3 is an MSVC-specific compilation flag ("all warnings on")]
 # [see %devstudio%\vc\include\win32.mak for cvars macro definition]
 O = .obj
@@ -76,7 +77,7 @@ EXES = $(RPNG)$(E) $(RPNG2)$(E) $(WPNG)$(E)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) -c $(CPPFLAGS) $(CFLAGS) $<
 
 
 # dependencies --------------------------------------------------------------

contrib/gregbook/README

@@ -104,7 +104,7 @@ BUILD INSTRUCTIONS
      unpacked the source code.
 
      For MSVC, set up the necessary environment variables by invoking
- 
+
         %devstudio%\vc\bin\vcvars32.bat
 
      where where %devstudio% is the installation directory for MSVC /

contrib/gregbook/makevms.com

@@ -20,13 +20,13 @@ $ if zpath .eqs. ""
 $ then
 $   write sys$output "zlib include not found. Exiting..."
 $   exit 2
-$ endif 
+$ endif
 $!
 $ if pngpath .eqs. ""
 $ then
 $   write sys$output "libpng include not found. Exiting..."
 $   exit 2
-$ endif 
+$ endif
 $!
 $!    Look for the compiler used.
 $!

contrib/gregbook/readpng.c

@@ -55,8 +55,9 @@
 
 #include <stdio.h>
 #include <stdlib.h>
+#include <zlib.h>
 
-#include "png.h"        /* libpng header; includes zlib.h */
+#include "png.h"        /* libpng header */
 #include "readpng.h"    /* typedefs, common macros, public prototypes */
 
 /* future versions of libpng will provide this macro: */
@@ -214,6 +215,10 @@ uch *readpng_get_image(double display_exponent, int *pChannels, ulg *pRowbytes)
      * libpng function */
 
     if (setjmp(png_jmpbuf(png_ptr))) {
+        free(image_data);
+        image_data = NULL;
+        free(row_pointers);
+        row_pointers = NULL;
         png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
         return NULL;
     }
@@ -229,8 +234,14 @@ uch *readpng_get_image(double display_exponent, int *pChannels, ulg *pRowbytes)
         png_set_expand(png_ptr);
     if (png_get_valid(png_ptr, info_ptr, PNG_INFO_tRNS))
         png_set_expand(png_ptr);
+#ifdef PNG_READ_16_TO_8_SUPPORTED
     if (bit_depth == 16)
+#  ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
+        png_set_scale_16(png_ptr);
+#  else
         png_set_strip_16(png_ptr);
+#  endif
+#endif
     if (color_type == PNG_COLOR_TYPE_GRAY ||
         color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
         png_set_gray_to_rgb(png_ptr);

contrib/gregbook/readpng2.c

@@ -69,6 +69,7 @@ static void readpng2_row_callback(png_structp png_ptr, png_bytep new_row,
                                  png_uint_32 row_num, int pass);
 static void readpng2_end_callback(png_structp png_ptr, png_infop info_ptr);
 static void readpng2_error_handler(png_structp png_ptr, png_const_charp msg);
+static void readpng2_warning_handler(png_structp png_ptr, png_const_charp msg);
 
 
 
@@ -104,7 +105,7 @@ int readpng2_init(mainprog_info *mainprog_ptr)
     /* could also replace libpng warning-handler (final NULL), but no need: */
 
     png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING, mainprog_ptr,
-      readpng2_error_handler, NULL);
+      readpng2_error_handler, readpng2_warning_handler);
     if (!png_ptr)
         return 4;   /* out of memory */
 
@@ -136,29 +137,23 @@ int readpng2_init(mainprog_info *mainprog_ptr)
      * used, i.e., all chunks recognized by libpng except for IHDR, PLTE, IDAT,
      * IEND, tRNS, bKGD, gAMA, and sRGB (small performance improvement) */
     {
-        /* These byte strings were copied from png.h.  If a future libpng
-         * version recognizes more chunks, add them to this list.  If a
-         * future version of readpng2.c recognizes more chunks, delete them
-         * from this list. */
-        static /* const */ png_byte chunks_to_ignore[] = {
-             99,  72,  82,  77, '\0',  /* cHRM */
-            104,  73,  83,  84, '\0',  /* hIST */
-            105,  67,  67,  80, '\0',  /* iCCP */
-            105,  84,  88, 116, '\0',  /* iTXt */
-            111,  70,  70, 115, '\0',  /* oFFs */
-            112,  67,  65,  76, '\0',  /* pCAL */
-            112,  72,  89, 115, '\0',  /* pHYs */
-            115,  66,  73,  84, '\0',  /* sBIT */
-            115,  67,  65,  76, '\0',  /* sCAL */
-            115,  80,  76,  84, '\0',  /* sPLT */
-            115,  84,  69,  82, '\0',  /* sTER */
-            116,  69,  88, 116, '\0',  /* tEXt */
-            116,  73,  77,  69, '\0',  /* tIME */
-            122,  84,  88, 116, '\0'   /* zTXt */
-        };
+        /* These byte strings were copied from png.h.  If a future version
+         * of readpng2.c recognizes more chunks, add them to this list.
+         */
+        static PNG_CONST png_byte chunks_to_process[] = {
+            98,  75,  71,  68, '\0',  /* bKGD */
+           103,  65,  77,  65, '\0',  /* gAMA */
+           115,  82,  71,  66, '\0',  /* sRGB */
+           };
 
-        png_set_keep_unknown_chunks(png_ptr, 1 /* PNG_HANDLE_CHUNK_NEVER */,
-          chunks_to_ignore, sizeof(chunks_to_ignore)/5);
+       /* Ignore all chunks except for IHDR, PLTE, tRNS, IDAT, and IEND */
+       png_set_keep_unknown_chunks(png_ptr, -1 /* PNG_HANDLE_CHUNK_NEVER */,
+          NULL, -1);
+
+       /* But do not ignore chunks in the "chunks_to_process" list */
+       png_set_keep_unknown_chunks(png_ptr,
+          0 /* PNG_HANDLE_CHUNK_AS_DEFAULT */, chunks_to_process,
+          sizeof(chunks_to_process)/5);
     }
 #endif /* PNG_HANDLE_AS_UNKNOWN_SUPPORTED */
 
@@ -311,8 +306,14 @@ static void readpng2_info_callback(png_structp png_ptr, png_infop info_ptr)
         png_set_expand(png_ptr);
     if (png_get_valid(png_ptr, info_ptr, PNG_INFO_tRNS))
         png_set_expand(png_ptr);
+#ifdef PNG_READ_16_TO_8_SUPPORTED
     if (bit_depth == 16)
+#  ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
+        png_set_scale_16(png_ptr);
+#  else
         png_set_strip_16(png_ptr);
+#  endif
+#endif
     if (color_type == PNG_COLOR_TYPE_GRAY ||
         color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
         png_set_gray_to_rgb(png_ptr);
@@ -447,6 +448,8 @@ static void readpng2_end_callback(png_structp png_ptr, png_infop info_ptr)
 
     /* all done */
 
+    (void)info_ptr; /* Unused */
+
     return;
 }
 
@@ -467,7 +470,12 @@ void readpng2_cleanup(mainprog_info *mainprog_ptr)
 }
 
 
-
+static void readpng2_warning_handler(png_structp png_ptr, png_const_charp msg)
+{
+    fprintf(stderr, "readpng2 libpng warning: %s\n", msg);
+    fflush(stderr);
+    (void)png_ptr; /* Unused */
+}
 
 
 static void readpng2_error_handler(png_structp png_ptr, png_const_charp msg)

contrib/gregbook/readppm.c

@@ -163,8 +163,12 @@ uch *readpng_get_image(double display_exponent, int *pChannels, ulg *pRowbytes)
 
     /* now we can go ahead and just read the whole image */
 
-    fread(image_data, 1L, rowbytes*height, saved_infile);
-
+    if (fread(image_data, 1L, rowbytes*height, saved_infile) <
+       rowbytes*height) {
+        free (image_data);
+        image_data = NULL;
+        return NULL;
+    }
 
     return image_data;
 }

contrib/gregbook/rpng-x.c

@@ -252,8 +252,8 @@ int main(int argc, char **argv)
             else {
                 bgstr = *argv;
                 if (strlen(bgstr) != 7 || bgstr[0] != '#')
-                    ++error; 
-                else 
+                    ++error;
+                else
                     have_bg = TRUE;
             }
         } else {

contrib/gregbook/rpng2-win.c

@@ -590,7 +590,7 @@ int WINAPI WinMain(HINSTANCE hInst, HINSTANCE hPrevInst, PSTR cmd, int showmode)
                   "(unexpectedly) while reading PNG image file\n");
                 exit(3);
             } else /* if (error) */ {
-                // will print error message below
+                /* will print error message below */
             }
             break;
         }

contrib/gregbook/rpng2-x.c

@@ -591,7 +591,7 @@ int main(int argc, char **argv)
                   "(unexpectedly) while reading PNG image file\n");
                 exit(3);
             } else /* if (error) */ {
-                // will print error message below
+                /* will print error message below */
             }
             break;
         }

contrib/gregbook/writepng.c

@@ -55,8 +55,9 @@
 
 
 #include <stdlib.h>     /* for exit() prototype */
+#include <zlib.h>
 
-#include "png.h"        /* libpng header; includes zlib.h and setjmp.h */
+#include "png.h"        /* libpng header, includes setjmp.h */
 #include "writepng.h"   /* typedefs, common macros, public prototypes */
 
 

contrib/libtests/fakepng.c

@@ -0,0 +1,57 @@
+/* Fake a PNG - just write it out directly. */
+#include <stdio.h>
+#include <zlib.h> /* for crc32 */
+
+void
+put_uLong(uLong val)
+{
+   putchar(val >> 24);
+   putchar(val >> 16);
+   putchar(val >>  8);
+   putchar(val >>  0);
+}
+
+void
+put_chunk(const unsigned char *chunk, uInt length)
+{
+   uLong crc;
+
+   put_uLong(length-4); /* Exclude the tag */
+
+   fwrite(chunk, length, 1, stdout);
+
+   crc = crc32(0, Z_NULL, 0);
+   put_uLong(crc32(crc, chunk, length));
+}
+
+const unsigned char signature[] =
+{
+   137, 80, 78, 71, 13, 10, 26, 10
+};
+
+const unsigned char IHDR[] =
+{
+   73, 72, 68, 82, /* IHDR */
+   0, 0, 0, 1, /* width */
+   0, 0, 0, 1, /* height */
+   1, /* bit depth */
+   0, /* color type: greyscale */
+   0, /* compression method */
+   0, /* filter method */
+   0  /* interlace method: none */
+};
+
+const unsigned char unknown[] =
+{
+   'u', 'n', 'K', 'n' /* "unKn" - private safe to copy */
+};
+
+int
+main(void)
+{
+   fwrite(signature, sizeof signature, 1, stdout);
+   put_chunk(IHDR, sizeof IHDR);
+
+   for(;;)
+      put_chunk(unknown, sizeof unknown);
+}

contrib/libtests/gentests.sh

@@ -0,0 +1,102 @@
+#!/bin/sh
+#
+# Copyright (c) 2013 John Cunningham Bowler
+#
+# Last changed in libpng 1.6.0 [February 14, 2013]
+#
+# This code is released under the libpng license.
+# For conditions of distribution and use, see the disclaimer
+# and license in png.h
+#
+# Generate a set of PNG test images.  The images are generated in a
+# sub-directory called 'tests' by default, however a command line argument will
+# change that name.  The generation requires a built version of makepng in the
+# current directory.
+#
+usage(){
+   exec >&2
+   echo "$0 [<directory>]"
+   echo '  Generate a set of PNG test files in "directory" ("tests" by default)'
+   exit 1
+}
+
+mp="$PWD/makepng"
+test -x "$mp" || {
+   exec >&2
+   echo "$0: the 'makepng' program must exist"
+   echo "  in the directory within which this program:"
+   echo "    $mp"
+   echo "  is executed"
+   usage
+}
+
+# Just one argument: the directory
+testdir="tests"
+test $# -gt 1 && {
+   testdir="$1"
+   shift
+}
+test $# -eq 0 || usage
+
+# Take care not to clobber something
+if test -e "$testdir"
+then
+   test -d "$testdir" || usage
+else
+   # mkdir -p isn't portable, so do the following
+   mkdir "$testdir" 2>/dev/null || mkdir -p "$testdir" || usage
+fi
+
+# This fails in a very satisfactory way if it's not accessible
+cd "$testdir"
+:>"test$$.png" || {
+   exec >&2
+   echo "$testdir: directory not writable"
+   usage
+}
+rm "test$$.png" || {
+   exec >&2
+   echo "$testdir: you have create but not write privileges here."
+   echo "  This is unexpected.  You have a spurion; "'"'"test$$.png"'"'"."
+   echo "  You need to remove this yourself.  Try a different directory."
+   exit 1
+}
+
+# Now call makepng ($mp) to create every file we can think of with a
+# reasonable name
+doit(){
+   for gamma in "" --sRGB --linear --1.8
+   do
+      case "$gamma" in
+         "")
+            gname=;;
+         --sRGB)
+            gname="-srgb";;
+         --linear)
+            gname="-lin";;
+         --1.8)
+            gname="-18";;
+         *)
+            gname="-$gamma";;
+      esac
+      "$mp" $gamma "$1" "$2" "test-$1-$2$gname.png"
+   done
+}
+#
+for ct in gray palette
+do
+   for bd in 1 2 4 8
+   do
+      doit "$ct" "$bd"
+   done
+done
+#
+doit "gray" "16"
+#
+for ct in gray-alpha rgb rgb-alpha
+do
+   for bd in 8 16
+   do
+      doit "$ct" "$bd"
+   done
+done

contrib/libtests/readpng.c

@@ -0,0 +1,104 @@
+/* readpng.c
+ *
+ * Copyright (c) 2013 John Cunningham Bowler
+ *
+ * Last changed in libpng 1.6.1 [March 28, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * Load an arbitrary number of PNG files (from the command line, or, if there
+ * are no arguments on the command line, from stdin) then run a time test by
+ * reading each file by row.  The test does nothing with the read result and
+ * does no transforms.  The only output is a time as a floating point number of
+ * seconds with 9 decimal digits.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+
+#if defined(HAVE_CONFIG_H) && !defined(PNG_NO_CONFIG_H)
+#  include <config.h>
+#endif
+
+/* Define the following to use this test against your installed libpng, rather
+ * than the one being built here:
+ */
+#ifdef PNG_FREESTANDING_TESTS
+#  include <png.h>
+#else
+#  include "../../png.h"
+#endif
+
+static int
+read_png(FILE *fp)
+{
+   png_structp png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,0,0,0);
+   png_infop info_ptr = NULL;
+   png_bytep row = NULL, display = NULL;
+
+   if (png_ptr == NULL)
+      return 0;
+
+   if (setjmp(png_jmpbuf(png_ptr)))
+   {
+      png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+      if (row != NULL) free(row);
+      if (display != NULL) free(display);
+      return 0;
+   }
+
+   png_init_io(png_ptr, fp);
+
+   info_ptr = png_create_info_struct(png_ptr);
+   if (info_ptr == NULL)
+      png_error(png_ptr, "OOM allocating info structure");
+
+   png_set_keep_unknown_chunks(png_ptr, PNG_HANDLE_CHUNK_ALWAYS, NULL, 0);
+
+   png_read_info(png_ptr, info_ptr);
+
+   {
+      png_size_t rowbytes = png_get_rowbytes(png_ptr, info_ptr);
+
+      row = malloc(rowbytes);
+      display = malloc(rowbytes);
+
+      if (row == NULL || display == NULL)
+         png_error(png_ptr, "OOM allocating row buffers");
+
+      {
+         png_uint_32 height = png_get_image_height(png_ptr, info_ptr);
+         int passes = png_set_interlace_handling(png_ptr);
+         int pass;
+
+         png_start_read_image(png_ptr);
+
+         for (pass = 0; pass < passes; ++pass)
+         {
+            png_uint_32 y = height;
+
+            /* NOTE: this trashes the row each time; interlace handling won't
+             * work, but this avoids memory thrashing for speed testing.
+             */
+            while (y-- > 0)
+               png_read_row(png_ptr, row, display);
+         }
+      }
+   }
+
+   /* Make sure to read to the end of the file: */
+   png_read_end(png_ptr, info_ptr);
+   png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+   free(row);
+   free(display);
+   return 1;
+}
+
+int
+main(void)
+{
+   /* Exit code 0 on success. */
+   return !read_png(stdin);
+}

contrib/libtests/tarith.c

@@ -0,0 +1,999 @@
+
+/* tarith.c
+ *
+ * Copyright (c) 2011-2013 John Cunningham Bowler
+ *
+ * Last changed in libpng 1.6.0 [February 14, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * Test internal arithmetic functions of libpng.
+ *
+ * This code must be linked against a math library (-lm), but does not require
+ * libpng or zlib to work.  Because it includes the complete source of 'png.c'
+ * it tests the code with whatever compiler options are used to build it.
+ * Changing these options can substantially change the errors in the
+ * calculations that the compiler chooses!
+ */
+#define _POSIX_SOURCE 1
+#define _ISOC99_SOURCE 1
+
+/* Obtain a copy of the code to be tested (plus other things), disabling
+ * stuff that is not required.
+ */
+#include <math.h>
+#include <stdlib.h>
+#include <ctype.h>
+#include <string.h>
+#include <assert.h>
+
+#include "../../pngpriv.h"
+
+#define png_error png_warning
+
+void png_warning(png_const_structrp png_ptr, png_const_charp msg)
+{
+   fprintf(stderr, "validation: %s\n", msg);
+}
+
+#define png_fixed_error png_fixed_warning
+
+void png_fixed_warning(png_const_structrp png_ptr, png_const_charp msg)
+{
+   fprintf(stderr, "overflow in: %s\n", msg);
+}
+
+#define png_set_error_fn(pp, ep, efp, wfp) ((void)0)
+#define png_malloc(pp, s) malloc(s)
+#define png_malloc_warn(pp, s) malloc(s)
+#define png_malloc_base(pp, s) malloc(s)
+#define png_calloc(pp, s) calloc(1, (s))
+#define png_free(pp, s) free(s)
+
+#define png_safecat(b, sb, pos, str) (pos)
+#define png_format_number(start, end, format, number) (start)
+
+#define crc32(crc, pp, s) (crc)
+#define inflateReset(zs) Z_OK
+
+#define png_create_struct(type) (0)
+#define png_destroy_struct(pp) ((void)0)
+#define png_create_struct_2(type, m, mm) (0)
+#define png_destroy_struct_2(pp, f, mm) ((void)0)
+
+#undef PNG_SIMPLIFIED_READ_SUPPORTED
+#undef PNG_SIMPLIFIED_WRITE_SUPPORTED
+#undef PNG_USER_MEM_SUPPORTED
+
+#include "../../png.c"
+
+/* Validate ASCII to fp routines. */
+static int verbose = 0;
+
+int validation_ascii_to_fp(int count, int argc, char **argv)
+{
+   int    showall = 0;
+   double max_error=2;      /* As a percentage error-in-last-digit/.5 */
+   double max_error_abs=17; /* Used when precision is DBL_DIG */
+   double max = 0;
+   double max_abs = 0;
+   double test = 0; /* Important to test this. */
+   int    precision = 5;
+   int    nonfinite = 0;
+   int    finite = 0;
+   int    ok = 0;
+   int    failcount = 0;
+   int    minorarith = 0;
+
+   while (--argc > 0)
+      if (strcmp(*++argv, "-a") == 0)
+         showall = 1;
+      else if (strcmp(*argv, "-e") == 0 && argc > 0)
+      {
+         --argc;
+         max_error = atof(*++argv);
+      }
+      else if (strcmp(*argv, "-E") == 0 && argc > 0)
+      {
+         --argc;
+         max_error_abs = atof(*++argv);
+      }
+      else
+      {
+         fprintf(stderr, "unknown argument %s\n", *argv);
+         return 1;
+      }
+
+   do
+   {
+      png_size_t index;
+      int state, failed = 0;
+      char buffer[64];
+
+      if (isfinite(test))
+         ++finite;
+      else
+         ++nonfinite;
+
+      if (verbose)
+         fprintf(stderr, "%.*g %d\n", DBL_DIG, test, precision);
+
+      /* Check for overflow in the buffer by setting a marker. */
+      memset(buffer, 71, sizeof buffer);
+
+      png_ascii_from_fp(0, buffer, precision+10, test, precision);
+
+      /* Allow for a three digit exponent, this stuff will fail if
+       * the exponent is bigger than this!
+       */
+      if (buffer[precision+7] != 71)
+      {
+         fprintf(stderr, "%g[%d] -> '%s'[%lu] buffer overflow\n", test,
+            precision, buffer, (unsigned long)strlen(buffer));
+         failed = 1;
+      }
+
+      /* Following are used for the number parser below and must be
+       * initialized to zero.
+       */
+      state = 0;
+      index = 0;
+      if (!isfinite(test))
+      {
+         /* Expect 'inf' */
+         if (test >= 0 && strcmp(buffer, "inf") ||
+             test <  0 && strcmp(buffer, "-inf"))
+         {
+            fprintf(stderr, "%g[%d] -> '%s' but expected 'inf'\n", test,
+               precision, buffer);
+            failed = 1;
+         }
+      }
+      else if (!png_check_fp_number(buffer, precision+10, &state, &index) ||
+          buffer[index] != 0)
+      {
+         fprintf(stderr, "%g[%d] -> '%s' but has bad format ('%c')\n", test,
+         precision, buffer, buffer[index]);
+         failed = 1;
+      }
+      else if (PNG_FP_IS_NEGATIVE(state) && !(test < 0))
+      {
+         fprintf(stderr, "%g[%d] -> '%s' but negative value not so reported\n",
+            test, precision, buffer);
+         failed = 1;
+         assert(!PNG_FP_IS_ZERO(state));
+         assert(!PNG_FP_IS_POSITIVE(state));
+      }
+      else if (PNG_FP_IS_ZERO(state) && !(test == 0))
+      {
+         fprintf(stderr, "%g[%d] -> '%s' but zero value not so reported\n",
+            test, precision, buffer);
+         failed = 1;
+         assert(!PNG_FP_IS_NEGATIVE(state));
+         assert(!PNG_FP_IS_POSITIVE(state));
+      }
+      else if (PNG_FP_IS_POSITIVE(state) && !(test > 0))
+      {
+         fprintf(stderr, "%g[%d] -> '%s' but postive value not so reported\n",
+            test, precision, buffer);
+         failed = 1;
+         assert(!PNG_FP_IS_NEGATIVE(state));
+         assert(!PNG_FP_IS_ZERO(state));
+      }
+      else
+      {
+         /* Check the result against the original. */
+         double out = atof(buffer);
+         double change = fabs((out - test)/test);
+         double allow = .5/pow(10,
+            (precision >= DBL_DIG) ? DBL_DIG-1 : precision-1);
+
+         /* NOTE: if you hit this error case are you compiling with gcc
+          * and -O0?  Try -O2 - the errors can accumulate if the FP
+          * code above is not optimized and may drift outside the .5 in
+          * DBL_DIG allowed.  In any case a small number of errors may
+          * occur (very small ones - 1 or 2%) because of rounding in the
+          * calculations, either in the conversion API or in atof.
+          */
+         if (change >= allow && (isfinite(out) ||
+             fabs(test/DBL_MAX) <= 1-allow))
+         {
+            double percent = (precision >= DBL_DIG) ? max_error_abs : max_error;
+            double allowp = (change-allow)*100/allow;
+
+            if (precision >= DBL_DIG)
+            {
+               if (max_abs < allowp) max_abs = allowp;
+            }
+
+            else
+            {
+               if (max < allowp) max = allowp;
+            }
+
+            if (showall || allowp >= percent)
+            {
+               fprintf(stderr,
+                  "%.*g[%d] -> '%s' -> %.*g number changed (%g > %g (%d%%))\n",
+                  DBL_DIG, test, precision, buffer, DBL_DIG, out, change, allow,
+                  (int)round(allowp));
+               failed = 1;
+            }
+            else
+               ++minorarith;
+         }
+      }
+
+      if (failed)
+         ++failcount;
+      else
+         ++ok;
+
+skip:
+      /* Generate a new number and precision. */
+      precision = rand();
+      if (precision & 1) test = -test;
+      precision >>= 1;
+
+      /* Generate random numbers. */
+      if (test == 0 || !isfinite(test))
+         test = precision+1;
+      else
+      {
+         /* Derive the exponent from the previous rand() value. */
+         int exponent = precision % (DBL_MAX_EXP - DBL_MIN_EXP) + DBL_MIN_EXP;
+         int tmp;
+         test = frexp(test * rand(), &tmp);
+         test = ldexp(test, exponent);
+         precision >>= 8; /* arbitrary */
+      }
+
+      /* This limits the precision to 32 digits, enough for standard
+       * IEEE implementations which have at most 15 digits.
+       */
+      precision = (precision & 0x1f) + 1;
+   }
+   while (--count);
+
+   printf("Tested %d finite values, %d non-finite, %d OK (%d failed) %d minor "
+      "arithmetic errors\n", finite, nonfinite, ok, failcount, minorarith);
+   printf(" Error with >=%d digit precision %.2f%%\n", DBL_DIG, max_abs);
+   printf(" Error with < %d digit precision %.2f%%\n", DBL_DIG, max);
+
+   return 0;
+}
+
+/* Observe that valid FP numbers have the forms listed in the PNG extensions
+ * specification:
+ *
+ * [+,-]{integer,integer.fraction,.fraction}[{e,E}[+,-]integer]
+ *
+ * Test each of these in turn, including invalid cases.
+ */
+typedef enum checkfp_state
+{
+   start, fraction, exponent, states
+} checkfp_state;
+
+/* The characters (other than digits) that characterize the states: */
+static const char none[] = "";
+static const char hexdigits[16] = "0123456789ABCDEF";
+
+static const struct
+{
+   const char *start; /* Characters valid at the start */
+   const char *end;   /* Valid characters that end the state */
+   const char *tests; /* Characters to test after 2 digits seen */
+}
+state_characters[states] =
+{
+   /* start:    */ { "+-.", ".eE", "+-.e*0369" },
+   /* fraction: */ { none, "eE",  "+-.E#0147" },
+   /* exponent: */ { "+-", none,  "+-.eE^0258" }
+};
+
+typedef struct
+{
+   char number[1024];  /* Buffer for number being tested */
+   int  limit;         /* Command line limit */
+   int  verbose;       /* Shadows global variable */
+   int  ctimes;        /* Number of numbers tested */
+   int  cmillions;     /* Count of millions of numbers */
+   int  cinvalid;      /* Invalid strings checked */
+   int  cnoaccept;     /* Characters not accepted */
+}
+checkfp_command;
+
+typedef struct
+{
+   int           cnumber;          /* Index into number string */
+   checkfp_state check_state;      /* Current number state */
+   int           at_start;         /* At start (first character) of state */
+   int           cdigits_in_state; /* Digits seen in that state */
+   int           limit;            /* Limit on same for checking all chars */
+   int           state;            /* Current parser state */
+   int           is_negative;      /* Number is negative */
+   int           is_zero;          /* Number is (still) zero */
+   int           number_was_valid; /* Previous character validity */
+}
+checkfp_control;
+
+static int check_all_characters(checkfp_command *co, checkfp_control c);
+
+static int check_some_characters(checkfp_command *co, checkfp_control c,
+   const char *tests);
+
+static int check_one_character(checkfp_command *co, checkfp_control c, int ch)
+{
+   /* Test this character (ch) to ensure the parser does the correct thing.
+    */
+   png_size_t index = 0;
+   const char test = (char)ch;
+   const int number_is_valid = png_check_fp_number(&test, 1, &c.state, &index);
+   const int character_accepted = (index == 1);
+
+   if (c.check_state != exponent && isdigit(ch) && ch != '0')
+      c.is_zero = 0;
+
+   if (c.check_state == start && c.at_start && ch == '-')
+      c.is_negative = 1;
+
+   if (isprint(ch))
+      co->number[c.cnumber++] = (char)ch;
+   else
+   {
+      co->number[c.cnumber++] = '<';
+      co->number[c.cnumber++] = hexdigits[(ch >> 4) & 0xf];
+      co->number[c.cnumber++] = hexdigits[ch & 0xf];
+      co->number[c.cnumber++] = '>';
+   }
+   co->number[c.cnumber] = 0;
+
+   if (co->verbose > 1)
+      fprintf(stderr, "%s\n", co->number);
+
+   if (++(co->ctimes) == 1000000)
+   {
+      if (co->verbose == 1)
+         fputc('.', stderr);
+      co->ctimes = 0;
+      ++(co->cmillions);
+   }
+
+   if (!number_is_valid)
+      ++(co->cinvalid);
+
+   if (!character_accepted)
+      ++(co->cnoaccept);
+
+   /* This should never fail (it's a serious bug if it does): */
+   if (index != 0 && index != 1)
+   {
+      fprintf(stderr, "%s: read beyond end of string (%lu)\n", co->number,
+         (unsigned long)index);
+      return 0;
+   }
+
+   /* Validate the new state, note that the PNG_FP_IS_ macros all return
+    * false unless the number is valid.
+    */
+   if (PNG_FP_IS_NEGATIVE(c.state) !=
+      (number_is_valid && !c.is_zero && c.is_negative))
+   {
+      fprintf(stderr, "%s: negative when it is not\n", co->number);
+      return 0;
+   }
+
+   if (PNG_FP_IS_ZERO(c.state) != (number_is_valid && c.is_zero))
+   {
+      fprintf(stderr, "%s: zero when it is not\n", co->number);
+      return 0;
+   }
+
+   if (PNG_FP_IS_POSITIVE(c.state) !=
+      (number_is_valid && !c.is_zero && !c.is_negative))
+   {
+      fprintf(stderr, "%s: positive when it is not\n", co->number);
+      return 0;
+   }
+
+   /* Testing a digit */
+   if (isdigit(ch))
+   {
+      if (!character_accepted)
+      {
+         fprintf(stderr, "%s: digit '%c' not accepted\n", co->number, ch);
+         return 0;
+      }
+
+      if (!number_is_valid)
+      {
+         fprintf(stderr, "%s: saw a digit (%c) but number not valid\n",
+            co->number, ch);
+         return 0;
+      }
+
+      ++c.cdigits_in_state;
+      c.at_start = 0;
+      c.number_was_valid = 1;
+
+      /* Continue testing characters in this state.  Either test all of
+       * them or, if we have already seen one digit in this state, just test a
+       * limited set.
+       */
+      if (c.cdigits_in_state < 1)
+         return check_all_characters(co, c);
+
+      else
+         return check_some_characters(co, c,
+            state_characters[c.check_state].tests);
+   }
+
+   /* A non-digit; is it allowed here? */
+   else if (((ch == '+' || ch == '-') && c.check_state != fraction &&
+               c.at_start) ||
+            (ch == '.' && c.check_state == start) ||
+            ((ch == 'e' || ch == 'E') && c.number_was_valid &&
+               c.check_state != exponent))
+   {
+      if (!character_accepted)
+      {
+         fprintf(stderr, "%s: character '%c' not accepted\n", co->number, ch);
+         return 0;
+      }
+
+      /* The number remains valid after start of fraction but nowhere else. */
+      if (number_is_valid && (c.check_state != start || ch != '.'))
+      {
+         fprintf(stderr, "%s: saw a non-digit (%c) but number valid\n",
+            co->number, ch);
+         return 0;
+      }
+
+      c.number_was_valid = number_is_valid;
+
+      /* Check for a state change.  When changing to 'fraction' if the number
+       * is valid at this point set the at_start to false to allow an exponent
+       * 'e' to come next.
+       */
+      if (c.check_state == start && ch == '.')
+      {
+         c.check_state = fraction;
+         c.at_start = !number_is_valid;
+         c.cdigits_in_state = 0;
+         c.limit = co->limit;
+         return check_all_characters(co, c);
+      }
+
+      else if (c.check_state < exponent && (ch == 'e' || ch == 'E'))
+      {
+         c.check_state = exponent;
+         c.at_start = 1;
+         c.cdigits_in_state = 0;
+         c.limit = co->limit;
+         return check_all_characters(co, c);
+      }
+
+      /* Else it was a sign, and the state doesn't change. */
+      else
+      {
+         if (ch != '-' && ch != '+')
+         {
+            fprintf(stderr, "checkfp: internal error (1)\n");
+            return 0;
+         }
+
+         c.at_start = 0;
+         return check_all_characters(co, c);
+      }
+   }
+
+   /* Testing an invalid character */
+   else
+   {
+      if (character_accepted)
+      {
+         fprintf(stderr, "%s: character '%c' [0x%.2x] accepted\n", co->number,
+            ch, ch);
+         return 0;
+      }
+
+      if (number_is_valid != c.number_was_valid)
+      {
+         fprintf(stderr,
+            "%s: character '%c' [0x%.2x] changed number validity\n", co->number,
+            ch, ch);
+         return 0;
+      }
+
+      /* Do nothing - the parser has stuck; return success and keep going with
+       * the next character.
+       */
+   }
+
+   /* Successful return (the caller will try the next character.) */
+   return 1;
+}
+
+static int check_all_characters(checkfp_command *co, checkfp_control c)
+{
+   int ch;
+
+   if (c.cnumber+4 < sizeof co->number) for (ch=0; ch<256; ++ch)
+   {
+      if (!check_one_character(co, c, ch))
+         return 0;
+   }
+
+   return 1;
+}
+
+static int check_some_characters(checkfp_command *co, checkfp_control c,
+   const char *tests)
+{
+   int i;
+
+   --(c.limit);
+
+   if (c.cnumber+4 < sizeof co->number && c.limit >= 0)
+   {
+      if (c.limit > 0) for (i=0; tests[i]; ++i)
+      {
+         if (!check_one_character(co, c, tests[i]))
+               return 0;
+      }
+
+      /* At the end check all the characters. */
+      else
+         return check_all_characters(co, c);
+   }
+
+   return 1;
+}
+
+int validation_checkfp(int count, int argc, char **argv)
+{
+   int result;
+   checkfp_command command;
+   checkfp_control control;
+
+   command.number[0] = 0;
+   command.limit = 3;
+   command.verbose = verbose;
+   command.ctimes = 0;
+   command.cmillions = 0;
+   command.cinvalid = 0;
+   command.cnoaccept = 0;
+
+   while (--argc > 0)
+   {
+      ++argv;
+      if (argc > 1 && strcmp(*argv, "-l") == 0)
+      {
+         --argc;
+         command.limit = atoi(*++argv);
+      }
+
+      else
+      {
+         fprintf(stderr, "unknown argument %s\n", *argv);
+         return 1;
+      }
+   }
+
+   control.cnumber = 0;
+   control.check_state = start;
+   control.at_start = 1;
+   control.cdigits_in_state = 0;
+   control.limit = command.limit;
+   control.state = 0;
+   control.is_negative = 0;
+   control.is_zero = 1;
+   control.number_was_valid = 0;
+
+   result = check_all_characters(&command, control);
+
+   printf("checkfp: %s: checked %d,%.3d,%.3d,%.3d strings (%d invalid)\n",
+      result ? "pass" : "FAIL", command.cmillions / 1000,
+      command.cmillions % 1000, command.ctimes / 1000, command.ctimes % 1000,
+      command.cinvalid);
+
+   return result;
+}
+
+int validation_muldiv(int count, int argc, char **argv)
+{
+   int tested = 0;
+   int overflow = 0;
+   int error = 0;
+   int error64 = 0;
+   int passed = 0;
+   int randbits = 0;
+   png_uint_32 randbuffer;
+   png_fixed_point a;
+   png_int_32 times, div;
+
+   while (--argc > 0)
+      {
+         fprintf(stderr, "unknown argument %s\n", *++argv);
+         return 1;
+      }
+
+   /* Find out about the random number generator. */
+   randbuffer = RAND_MAX;
+   while (randbuffer != 0) ++randbits, randbuffer >>= 1;
+   printf("Using random number generator that makes %d bits\n", randbits);
+   for (div=0; div<32; div += randbits)
+      randbuffer = (randbuffer << randbits) ^ rand();
+
+   a = 0;
+   times = div = 0;
+   do
+   {
+      png_fixed_point result;
+      /* NOTE: your mileage may vary, a type is required below that can
+       * hold 64 bits or more, if floating point is used a 64 bit or
+       * better mantissa is required.
+       */
+      long long int fp, fpround;
+      unsigned long hi, lo;
+      int ok;
+
+      /* Check the values, png_64bit_product can only handle positive
+       * numbers, so correct for that here.
+       */
+      {
+         long u1, u2;
+         int n = 0;
+         if (a < 0) u1 = -a, n = 1; else u1 = a;
+         if (times < 0) u2 = -times, n = !n; else u2 = times;
+         png_64bit_product(u1, u2, &hi, &lo);
+         if (n)
+         {
+            /* -x = ~x+1 */
+            lo = ((~lo) + 1) & 0xffffffff;
+            hi = ~hi;
+            if (lo == 0) ++hi;
+         }
+      }
+
+      fp = a;
+      fp *= times;
+      if ((fp & 0xffffffff) != lo || ((fp >> 32) & 0xffffffff) != hi)
+      {
+         fprintf(stderr, "png_64bit_product %d * %d -> %lx|%.8lx not %llx\n",
+            a, times, hi, lo, fp);
+         ++error64;
+      }
+
+      if (div != 0)
+      {
+         /* Round - this is C round to zero. */
+         if ((fp < 0) != (div < 0))
+           fp -= div/2;
+         else
+           fp += div/2;
+
+         fp /= div;
+         fpround = fp;
+         /* Assume 2's complement here: */
+         ok = fpround <= PNG_UINT_31_MAX &&
+              fpround >= -1-(long long int)PNG_UINT_31_MAX;
+         if (!ok) ++overflow;
+      }
+      else
+        ok = 0, ++overflow, fpround = fp/*misleading*/;
+
+      if (verbose)
+         fprintf(stderr, "TEST %d * %d / %d -> %lld (%s)\n", a, times, div,
+            fp, ok ? "ok" : "overflow");
+
+      ++tested;
+      if (png_muldiv(&result, a, times, div) != ok)
+      {
+         ++error;
+         if (ok)
+             fprintf(stderr, "%d * %d / %d -> overflow (expected %lld)\n", a,
+                times, div, fp);
+         else
+             fprintf(stderr, "%d * %d / %d -> %d (expected overflow %lld)\n", a,
+                times, div, result, fp);
+      }
+      else if (ok && result != fpround)
+      {
+         ++error;
+         fprintf(stderr, "%d * %d / %d -> %d not %lld\n", a, times, div, result,
+            fp);
+      }
+      else
+         ++passed;
+
+      /* Generate three new values, this uses rand() and rand() only returns
+       * up to RAND_MAX.
+       */
+      /* CRUDE */
+      a += times;
+      times += div;
+      div = randbuffer;
+      randbuffer = (randbuffer << randbits) ^ rand();
+   }
+   while (--count > 0);
+
+   printf("%d tests including %d overflows, %d passed, %d failed (%d 64 bit "
+      "errors)\n", tested, overflow, passed, error, error64);
+   return 0;
+}
+
+/* When FP is on this just becomes a speed test - compile without FP to get real
+ * validation.
+ */
+#ifdef PNG_FLOATING_ARITHMETIC_SUPPORTED
+#define LN2 .000010576586617430806112933839 /* log(2)/65536 */
+#define L2INV 94548.46219969910586572651    /* 65536/log(2) */
+
+/* For speed testing, need the internal functions too: */
+static png_uint_32 png_log8bit(unsigned x)
+{
+   if (x > 0)
+      return (png_uint_32)floor(.5-log(x/255.)*L2INV);
+
+   return 0xffffffff;
+}
+
+static png_uint_32 png_log16bit(png_uint_32 x)
+{
+   if (x > 0)
+      return (png_uint_32)floor(.5-log(x/65535.)*L2INV);
+
+   return 0xffffffff;
+}
+
+static png_uint_32 png_exp(png_uint_32 x)
+{
+   return (png_uint_32)floor(.5 + exp(x * -LN2) * 0xffffffffU);
+}
+
+static png_byte png_exp8bit(png_uint_32 log)
+{
+   return (png_byte)floor(.5 + exp(log * -LN2) * 255);
+}
+
+static png_uint_16 png_exp16bit(png_uint_32 log)
+{
+   return (png_uint_16)floor(.5 + exp(log * -LN2) * 65535);
+}
+#endif /* FLOATING_ARITHMETIC */
+
+int validation_gamma(int argc, char **argv)
+{
+   double gamma[9] = { 2.2, 1.8, 1.52, 1.45, 1., 1/1.45, 1/1.52, 1/1.8, 1/2.2 };
+   double maxerr;
+   int i, silent=0, onlygamma=0;
+
+   /* Silence the output with -s, just test the gamma functions with -g: */
+   while (--argc > 0)
+      if (strcmp(*++argv, "-s") == 0)
+         silent = 1;
+      else if (strcmp(*argv, "-g") == 0)
+         onlygamma = 1;
+      else
+      {
+         fprintf(stderr, "unknown argument %s\n", *argv);
+         return 1;
+      }
+
+   if (!onlygamma)
+   {
+      /* First validate the log functions: */
+      maxerr = 0;
+      for (i=0; i<256; ++i)
+      {
+         double correct = -log(i/255.)/log(2.)*65536;
+         double error = png_log8bit(i) - correct;
+
+         if (i != 0 && fabs(error) > maxerr)
+            maxerr = fabs(error);
+
+         if (i == 0 && png_log8bit(i) != 0xffffffff ||
+             i != 0 && png_log8bit(i) != floor(correct+.5))
+         {
+            fprintf(stderr, "8 bit log error: %d: got %u, expected %f\n",
+               i, png_log8bit(i), correct);
+         }
+      }
+
+      if (!silent)
+         printf("maximum 8 bit log error = %f\n", maxerr);
+
+      maxerr = 0;
+      for (i=0; i<65536; ++i)
+      {
+         double correct = -log(i/65535.)/log(2.)*65536;
+         double error = png_log16bit(i) - correct;
+
+         if (i != 0 && fabs(error) > maxerr)
+            maxerr = fabs(error);
+
+         if (i == 0 && png_log16bit(i) != 0xffffffff ||
+             i != 0 && png_log16bit(i) != floor(correct+.5))
+         {
+            if (error > .68) /* By experiment error is less than .68 */
+            {
+               fprintf(stderr, "16 bit log error: %d: got %u, expected %f"
+                  " error: %f\n", i, png_log16bit(i), correct, error);
+            }
+         }
+      }
+
+      if (!silent)
+         printf("maximum 16 bit log error = %f\n", maxerr);
+
+      /* Now exponentiations. */
+      maxerr = 0;
+      for (i=0; i<=0xfffff; ++i)
+      {
+         double correct = exp(-i/65536. * log(2.)) * (65536. * 65536);
+         double error = png_exp(i) - correct;
+
+         if (fabs(error) > maxerr)
+            maxerr = fabs(error);
+         if (fabs(error) > 1883) /* By experiment. */
+         {
+            fprintf(stderr, "32 bit exp error: %d: got %u, expected %f"
+                  " error: %f\n", i, png_exp(i), correct, error);
+         }
+      }
+
+      if (!silent)
+         printf("maximum 32 bit exp error = %f\n", maxerr);
+
+      maxerr = 0;
+      for (i=0; i<=0xfffff; ++i)
+      {
+         double correct = exp(-i/65536. * log(2.)) * 255;
+         double error = png_exp8bit(i) - correct;
+
+         if (fabs(error) > maxerr)
+            maxerr = fabs(error);
+         if (fabs(error) > .50002) /* By experiment */
+         {
+            fprintf(stderr, "8 bit exp error: %d: got %u, expected %f"
+                  " error: %f\n", i, png_exp8bit(i), correct, error);
+         }
+      }
+
+      if (!silent)
+         printf("maximum 8 bit exp error = %f\n", maxerr);
+
+      maxerr = 0;
+      for (i=0; i<=0xfffff; ++i)
+      {
+         double correct = exp(-i/65536. * log(2.)) * 65535;
+         double error = png_exp16bit(i) - correct;
+
+         if (fabs(error) > maxerr)
+            maxerr = fabs(error);
+         if (fabs(error) > .524) /* By experiment */
+         {
+            fprintf(stderr, "16 bit exp error: %d: got %u, expected %f"
+                  " error: %f\n", i, png_exp16bit(i), correct, error);
+         }
+      }
+
+      if (!silent)
+         printf("maximum 16 bit exp error = %f\n", maxerr);
+   } /* !onlygamma */
+
+   /* Test the overall gamma correction. */
+   for (i=0; i<9; ++i)
+   {
+      unsigned j;
+      double g = gamma[i];
+      png_fixed_point gfp = floor(g * PNG_FP_1 + .5);
+
+      if (!silent)
+         printf("Test gamma %f\n", g);
+
+      maxerr = 0;
+      for (j=0; j<256; ++j)
+      {
+         double correct = pow(j/255., g) * 255;
+         png_byte out = png_gamma_8bit_correct(j, gfp);
+         double error = out - correct;
+
+         if (fabs(error) > maxerr)
+            maxerr = fabs(error);
+         if (out != floor(correct+.5))
+         {
+            fprintf(stderr, "8bit %d ^ %f: got %d expected %f error %f\n",
+               j, g, out, correct, error);
+         }
+      }
+
+      if (!silent)
+         printf("gamma %f: maximum 8 bit error %f\n", g, maxerr);
+
+      maxerr = 0;
+      for (j=0; j<65536; ++j)
+      {
+         double correct = pow(j/65535., g) * 65535;
+         png_uint_16 out = png_gamma_16bit_correct(j, gfp);
+         double error = out - correct;
+
+         if (fabs(error) > maxerr)
+            maxerr = fabs(error);
+         if (fabs(error) > 1.62)
+         {
+            fprintf(stderr, "16bit %d ^ %f: got %d expected %f error %f\n",
+               j, g, out, correct, error);
+         }
+      }
+
+      if (!silent)
+         printf("gamma %f: maximum 16 bit error %f\n", g, maxerr);
+   }
+
+   return 0;
+}
+
+/**************************** VALIDATION TESTS ********************************/
+/* Various validation routines are included herein, they require some
+ * definition for png_warning and png_error, seetings of VALIDATION:
+ *
+ * 1: validates the ASCII to floating point conversions
+ * 2: validates png_muldiv
+ * 3: accuracy test of fixed point gamma tables
+ */
+
+/* The following COUNT (10^8) takes about 1 hour on a 1GHz Pentium IV
+ * processor.
+ */
+#define COUNT 1000000000
+
+int main(int argc, char **argv)
+{
+   int count = COUNT;
+
+   while (argc > 1)
+   {
+      if (argc > 2 && strcmp(argv[1], "-c") == 0)
+      {
+         count = atoi(argv[2]);
+         argc -= 2;
+         argv += 2;
+      }
+
+      else if (strcmp(argv[1], "-v") == 0)
+      {
+         ++verbose;
+         --argc;
+         ++argv;
+      }
+
+      else
+         break;
+   }
+
+   if (count > 0 && argc > 1)
+   {
+      if (strcmp(argv[1], "ascii") == 0)
+         return validation_ascii_to_fp(count, argc-1, argv+1);
+      else if (strcmp(argv[1], "checkfp") == 0)
+         return validation_checkfp(count, argc-1, argv+1);
+      else if (strcmp(argv[1], "muldiv") == 0)
+         return validation_muldiv(count, argc-1, argv+1);
+      else if (strcmp(argv[1], "gamma") == 0)
+         return validation_gamma(argc-1, argv+1);
+   }
+
+   /* Bad argument: */
+   fprintf(stderr,
+      "usage: tarith [-v] [-c count] {ascii,muldiv,gamma} [args]\n");
+   fprintf(stderr, " arguments: ascii [-a (all results)] [-e error%%]\n");
+   fprintf(stderr, "            checkfp [-l max-number-chars]\n");
+   fprintf(stderr, "            muldiv\n");
+   fprintf(stderr, "            gamma -s (silent) -g (only gamma; no log)\n");
+   return 1;
+}

contrib/libtests/timepng.c

@@ -0,0 +1,303 @@
+/* timepng.c
+ *
+ * Copyright (c) 2013 John Cunningham Bowler
+ *
+ * Last changed in libpng 1.6.1 [March 28, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * Load an arbitrary number of PNG files (from the command line, or, if there
+ * are no arguments on the command line, from stdin) then run a time test by
+ * reading each file by row.  The test does nothing with the read result and
+ * does no transforms.  The only output is a time as a floating point number of
+ * seconds with 9 decimal digits.
+ */
+#define _POSIX_C_SOURCE 199309L /* for clock_gettime */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+
+#include <time.h>
+
+#if defined(HAVE_CONFIG_H) && !defined(PNG_NO_CONFIG_H)
+#  include <config.h>
+#endif
+
+/* Define the following to use this test against your installed libpng, rather
+ * than the one being built here:
+ */
+#ifdef PNG_FREESTANDING_TESTS
+#  include <png.h>
+#else
+#  include "../../png.h"
+#endif
+
+static int read_png(FILE *fp)
+{
+   png_structp png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING,0,0,0);
+   png_infop info_ptr = NULL;
+   png_bytep row = NULL, display = NULL;
+
+   if (png_ptr == NULL)
+      return 0;
+
+   if (setjmp(png_jmpbuf(png_ptr)))
+   {
+      png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+      if (row != NULL) free(row);
+      if (display != NULL) free(display);
+      return 0;
+   }
+
+   png_init_io(png_ptr, fp);
+
+   info_ptr = png_create_info_struct(png_ptr);
+   if (info_ptr == NULL)
+      png_error(png_ptr, "OOM allocating info structure");
+
+   png_read_info(png_ptr, info_ptr);
+
+   {
+      png_size_t rowbytes = png_get_rowbytes(png_ptr, info_ptr);
+
+      row = malloc(rowbytes);
+      display = malloc(rowbytes);
+
+      if (row == NULL || display == NULL)
+         png_error(png_ptr, "OOM allocating row buffers");
+
+      {
+         png_uint_32 height = png_get_image_height(png_ptr, info_ptr);
+         int passes = png_set_interlace_handling(png_ptr);
+         int pass;
+
+         png_start_read_image(png_ptr);
+
+         for (pass = 0; pass < passes; ++pass)
+         {
+            png_uint_32 y = height;
+
+            /* NOTE: this trashes the row each time; interlace handling won't
+             * work, but this avoids memory thrashing for speed testing.
+             */
+            while (y-- > 0)
+               png_read_row(png_ptr, row, display);
+         }
+      }
+   }
+
+   /* Make sure to read to the end of the file: */
+   png_read_end(png_ptr, info_ptr);
+   png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
+   free(row);
+   free(display);
+   return 1;
+}
+
+static int mytime(struct timespec *t)
+{
+   /* Do the timing using clock_gettime and the per-process timer. */
+   if (!clock_gettime(CLOCK_PROCESS_CPUTIME_ID, t))
+      return 1;
+
+   perror("CLOCK_PROCESS_CPUTIME_ID");
+   fprintf(stderr, "timepng: could not get the time\n");
+   return 0;
+}
+
+static int perform_one_test(FILE *fp, int nfiles)
+{
+   int i;
+   struct timespec before, after;
+
+   /* Clear out all errors: */
+   rewind(fp);
+
+   if (mytime(&before))
+   {
+      for (i=0; i<nfiles; ++i)
+      {
+         if (read_png(fp))
+         {
+            if (ferror(fp))
+            {
+               perror("temporary file");
+               fprintf(stderr, "file %d: error reading PNG data\n", i);
+               return 0;
+            }
+         }
+
+         else
+         {
+            perror("temporary file");
+            fprintf(stderr, "file %d: error from libpng\n", i);
+            return 0;
+         }
+      }
+   }
+
+   else
+      return 0;
+
+   if (mytime(&after))
+   {
+      /* Work out the time difference and print it - this is the only output,
+       * so flush it immediately.
+       */
+      unsigned long s = after.tv_sec - before.tv_sec;
+      long ns = after.tv_nsec - before.tv_nsec;
+
+      if (ns < 0)
+      {
+         --s;
+         ns += 1000000000;
+
+         if (ns < 0)
+         {
+            fprintf(stderr, "timepng: bad clock from kernel\n");
+            return 0;
+         }
+      }
+
+      printf("%lu.%.9ld\n", s, ns);
+      fflush(stdout);
+      if (ferror(stdout))
+      {
+         fprintf(stderr, "timepng: error writing output\n");
+         return 0;
+      }
+
+      /* Successful return */
+      return 1;
+   }
+
+   else
+      return 0;
+}
+
+static int add_one_file(FILE *fp, char *name)
+{
+   FILE *ip = fopen(name, "rb");
+
+   if (ip != NULL)
+   {
+      int ch;
+      for (;;)
+      {
+         ch = getc(ip);
+         if (ch == EOF) break;
+         putc(ch, fp);
+      }
+
+      if (ferror(ip))
+      {
+         perror(name);
+         fprintf(stderr, "%s: read error\n", name);
+         return 0;
+      }
+
+      (void)fclose(ip);
+
+      if (ferror(fp))
+      {
+         perror("temporary file");
+         fprintf(stderr, "temporary file write error\n");
+         return 0;
+      }
+   }
+
+   else
+   {
+      perror(name);
+      fprintf(stderr, "%s: open failed\n", name);
+      return 0;
+   }
+
+   return 1;
+}
+
+int main(int argc, char **argv)
+{
+   int ok = 0;
+   FILE *fp = tmpfile();
+
+   if (fp != NULL)
+   {
+      int err = 0;
+      int nfiles = 0;
+
+      if (argc > 1)
+      {
+         int i;
+
+         for (i=1; i<argc; ++i)
+         {
+            if (add_one_file(fp, argv[i]))
+               ++nfiles;
+
+            else
+            {
+               err = 1;
+               break;
+            }
+         }
+      }
+
+      else
+      {
+         char filename[FILENAME_MAX+1];
+
+         while (fgets(filename, FILENAME_MAX+1, stdin))
+         {
+            size_t len = strlen(filename);
+
+            if (filename[len-1] == '\n')
+            {
+               filename[len-1] = 0;
+               if (add_one_file(fp, filename))
+                  ++nfiles;
+
+               else
+               {
+                  err = 1;
+                  break;
+               }
+            }
+
+            else
+            {
+               fprintf(stderr, "timepng: truncated file name ...%s\n",
+                  filename+len-32);
+               err = 1;
+               break;
+            }
+         }
+
+         if (ferror(stdin))
+         {
+            fprintf(stderr, "timepng: stdin: read error\n");
+            err = 1;
+         }
+      }
+
+      if (!err)
+      {
+         if (nfiles > 0)
+            ok = perform_one_test(fp, nfiles);
+
+         else
+            fprintf(stderr, "usage: timepng {files} or ls files | timepng\n");
+      }
+
+      (void)fclose(fp);
+   }
+
+   else
+      fprintf(stderr, "timepng: could not open temporary file\n");
+
+   /* Exit code 0 on success. */
+   return ok == 0;
+}

contrib/pngminim/README

@@ -1,3 +1,4 @@
+
 This demonstrates the use of PNG_USER_CONFIG, pngusr.h and pngusr.dfa
 to build minimal decoder, encoder, and progressive reader applications.
 

contrib/pngminim/decoder/README

@@ -7,4 +7,4 @@ Specify the location of the zlib source (1.2.1 or later) as ZLIBSRC
 on the make command line.
 
 If you prefer to use the shared libraries, go to contrib/pngminus
-and build the png2pnm application there. 
+and build the png2pnm application there.

contrib/pngminim/decoder/makefile

@@ -4,15 +4,18 @@
 #CC=cc
 CC=gcc
 LD=$(CC)
-STRIP=strip
 
 # If awk fails try
 # make AWK=nawk
 
+# If cpp fails try
+# make CPP=/lib/cpp
+
 RM=rm -f
 COPY=cp
 
-CFLAGS=-DPNG_USER_CONFIG -DNO_GZCOMPRESS -DNO_GZIP -I. -O1
+CPPFLAGS=-I. -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
+CFLAGS=-O1 -Wall
 
 C=.c
 O=.o
@@ -79,7 +82,7 @@ OBJS    = $(PROGOBJS) $(PNGOBJS) $(ZOBJS)
 
 # note: dependencies do not work on implicit rule lines
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) $(CPPFLAGS) -c $(CFLAGS) $<
 
 # dependencies
 
@@ -87,23 +90,22 @@ all: pngm2pnm$(E)
 
 pngm2pnm$(E): $(OBJS)
 	$(LD) -o pngm2pnm$(E) $(OBJS)
-	$(STRIP) pngm2pnm$(E)
 
 # The DFA_XTRA setting turns all libpng options off then
 # turns on those required for this minimal build.
 # The CPP_FLAGS setting causes pngusr.h to be included in
 # both the build of pnglibconf.h and, subsequently, when
 # building libpng itself.
-$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak\
+$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak $(ZH)\
 	$(PNGSRC)/scripts/pnglibconf.dfa \
 	$(PNGSRC)/scripts/options.awk pngusr.h pngusr.dfa
 	$(RM) pnglibconf.h pnglibconf.dfn
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
-	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG"\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
+	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG -I."\
 	    DFA_XTRA="pngusr.dfa" $@
 
 clean:
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
 	    srcdir=$(PNGSRC) clean
 	$(RM) pngm2pnm$(O)
 	$(RM) pngm2pnm$(E)

contrib/pngminim/decoder/pngusr.dfa

@@ -1,6 +1,6 @@
 # pngminim/decoder/pngusr.dfa
 #
-# Copyright (c) 2010 Glenn Randers-Pehrson
+# Copyright (c) 2010-2013 Glenn Randers-Pehrson
 #
 # This code is released under the libpng license.
 # For conditions of distribution and use, see the disclaimer
@@ -36,4 +36,5 @@ option FLOATING_ARITHMETIC off
 option SETJMP on
 option STDIO on
 option READ_EXPAND on
-option READ_16_TO_8 on
+option READ_STRIP_16_TO_8 on
+option USER_LIMITS on

contrib/pngminim/decoder/pngusr.h

@@ -1,6 +1,6 @@
 /* minrdpngconf.h: headers to make a minimal png-read-only library
  *
- * Copyright (c) 2007, 2010 Glenn Randers-Pehrson
+ * Copyright (c) 2007, 2010-2013 Glenn Randers-Pehrson
  *
  * This code is released under the libpng license.
  * For conditions of distribution and use, see the disclaimer
@@ -18,6 +18,6 @@
  * affect the API (so are not recorded in pnglibconf.h)
  */
 
-#define PNG_NO_WARNINGS
+#define PNG_ALIGN_TYPE PNG_ALIGN_NONE
 
 #endif /* MINRDPNGCONF_H */

contrib/pngminim/encoder/README

@@ -1,10 +1,10 @@
 This demonstrates the use of PNG_USER_CONFIG and pngusr.h
 
-The makefile builds a minimal write-only decoder with embedded libpng
+The makefile builds a minimal write-only encoder with embedded libpng
 and zlib.
 
 Specify the location of the zlib source (1.2.1 or later) as ZLIBSRC
 on the make command line.
 
 If you prefer to use the shared libraries, go to contrib/pngminus
-and build the pnm2png application there. 
+and build the pnm2png application there.

contrib/pngminim/encoder/makefile

@@ -4,15 +4,18 @@
 #CC=cc
 CC=gcc
 LD=$(CC)
-STRIP=strip
 
 # If awk fails try
 # make AWK=nawk
 
+# If cpp fails try
+# make CPP=/lib/cpp
+
 RM=rm -f
 COPY=cp
 
-CFLAGS=-DPNG_USER_CONFIG -DNO_GZIP -I. -O1
+CPPFLAGS=-I. -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
+CFLAGS=-O1 -Wall
 
 C=.c
 O=.o
@@ -78,7 +81,7 @@ OBJS    = $(PROGOBJS) $(PNGOBJS) $(ZOBJS)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) $(CPPFLAGS) -c $(CFLAGS) $<
 
 # dependencies
 
@@ -86,23 +89,22 @@ all: pnm2pngm$(E)
 
 pnm2pngm$(E): $(OBJS)
 	$(LD) -o pnm2pngm$(E) $(OBJS)
-	$(STRIP) pnm2pngm$(E)
 
 # The DFA_XTRA setting turns all libpng options off then
 # turns on those required for this minimal build.
 # The CPP_FLAGS setting causes pngusr.h to be included in
 # both the build of pnglibconf.h and, subsequently, when
 # building libpng itself.
-$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak\
+$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak $(ZH)\
 	$(PNGSRC)/scripts/pnglibconf.dfa \
 	$(PNGSRC)/scripts/options.awk pngusr.h pngusr.dfa
 	$(RM) pnglibconf.h pnglibconf.dfn
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
-	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG"\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
+	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG -I."\
 	    DFA_XTRA="pngusr.dfa" $@
 
 clean:
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
 	    srcdir=$(PNGSRC) clean
 	$(RM) pnm2pngm$(O)
 	$(RM) pnm2pngm$(E)

contrib/pngminim/encoder/pngusr.dfa

@@ -1,6 +1,6 @@
 # pngminim/encoder/pngusr.dfa
 #
-# Copyright (c) 2010 Glenn Randers-Pehrson
+# Copyright (c) 2010-2013 Glenn Randers-Pehrson
 #
 # This code is released under the libpng license.
 # For conditions of distribution and use, see the disclaimer
@@ -14,6 +14,10 @@ everything = off
 
 option WRITE on
 
+# These 2 options are required if you need to read PBM (P1 or P4) files.
+option WRITE_INVERT on
+option WRITE_PACK on
+
 # You must choose fixed or floating point arithmetic:
 # option FLOATING_POINT on
 

contrib/pngminim/encoder/pngusr.h

@@ -1,6 +1,6 @@
 /* minwrpngconf.h: headers to make a minimal png-write-only library
  *
- * Copyright (c) 2007, 2010 Glenn Randers-Pehrson
+ * Copyright (c) 2007, 2010-2013 Glenn Randers-Pehrson
  *
  * This code is released under the libpng license.
  * For conditions of distribution and use, see the disclaimer
@@ -18,6 +18,6 @@
  * affect the API (so are not recorded in pnglibconf.h)
  */
 
-#define PNG_NO_WARNINGS
+#define PNG_ALIGN_TYPE PNG_ALIGN_NONE
 
 #endif /* MINWRPNGCONF_H */

contrib/pngminim/preader/README

@@ -8,8 +8,8 @@ on the make command line.
 
 Edit makefile if required, to find your X library and include files,
 then
-    
+
     make ZLIBSRC=directory
- 
+
 If you prefer to use the shared libraries, go to contrib/gregbook
-and build the rpng2-x application there. 
+and build the rpng2-x application there.

contrib/pngminim/preader/makefile

@@ -4,11 +4,13 @@
 #CC=cc
 CC=gcc
 LD=$(CC)
-STRIP=strip
 
 # If awk fails try
 # make AWK=nawk
 
+# If cpp fails try
+# make CPP=/lib/cpp
+
 RM=rm -f
 COPY=cp
 
@@ -28,7 +30,8 @@ XLIB = -L/usr/X11R6/lib -lX11
 #LIBS = $(XLIB)
 LIBS = $(XLIB) -lm                      #platforms that need libm
 
-CFLAGS=-DPNG_USER_CONFIG -DNO_GZCOMPRESS -DNO_GZIP -I. $(XINC) -O1
+CPPFLAGS=-I. $(XINC) -DPNG_USER_CONFIG -DNO_GZCOMPRESS -DZ_SOLO -DNO_GZIP
+CFLAGS=-O1 -Wall
 
 C=.c
 O=.o
@@ -94,7 +97,7 @@ OBJS    = $(PROGOBJS) $(PNGOBJS) $(ZOBJS)
 # implicit make rules -------------------------------------------------------
 
 .c$(O):
-	$(CC) -c $(CFLAGS) $<
+	$(CC) $(CPPFLAGS) -c $(CFLAGS) $<
 
 # dependencies
 
@@ -102,23 +105,22 @@ all: $(PROGDOCS) rpng2-x$(E)
 
 rpng2-x$(E): $(OBJS)
 	$(LD) -o rpng2-x$(E) $(OBJS) $(LIBS)
-	$(STRIP) rpng2-x$(E)
 
 # The DFA_XTRA setting turns all libpng options off then
 # turns on those required for this minimal build.
 # The CPP_FLAGS setting causes pngusr.h to be included in
 # both the build of pnglibconf.h and, subsequently, when
 # building libpng itself.
-$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak\
+$(PNGCONF): $(PNGSRC)/scripts/pnglibconf.mak $(ZH)\
 	$(PNGSRC)/scripts/pnglibconf.dfa \
 	$(PNGSRC)/scripts/options.awk pngusr.h pngusr.dfa
 	$(RM) pnglibconf.h pnglibconf.dfn
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
-	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG"\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
+	    srcdir=$(PNGSRC) CPPFLAGS="-DPNG_USER_CONFIG -I."\
 	    DFA_XTRA="pngusr.dfa" $@
 
 clean:
-	$(MAKE) $(MAKEFLAGS) -f $(PNGSRC)/scripts/pnglibconf.mak\
+	$(MAKE) -f $(PNGSRC)/scripts/pnglibconf.mak $(MAKEFLAGS)\
 	    srcdir=$(PNGSRC) clean
 	$(RM) rpng2-x$(O)
 	$(RM) rpng2-x$(E)

contrib/pngminim/preader/pngusr.dfa

@@ -1,6 +1,6 @@
 # pngminim/preader/pngusr.dfa
 #
-# Copyright (c) 2010 Glenn Randers-Pehrson
+# Copyright (c) 2010-2013 Glenn Randers-Pehrson
 #
 # This code is released under the libpng license.
 # For conditions of distribution and use, see the disclaimer
@@ -36,5 +36,5 @@ option STDIO on
 option READ_bKGD on
 option READ_GAMMA on
 option READ_EXPAND on
-option READ_16_TO_8 on
+option READ_STRIP_16_TO_8 on
 option READ_GRAY_TO_RGB on

contrib/pngminim/preader/pngusr.h

@@ -1,6 +1,6 @@
 /* minrdpngconf.h: headers to make a minimal png-read-only library
  *
- * Copyright (c) 2009, 2010 Glenn Randers-Pehrson
+ * Copyright (c) 2009, 2010-2013 Glenn Randers-Pehrson
  *
  * This code is released under the libpng license.
  * For conditions of distribution and use, see the disclaimer
@@ -18,6 +18,6 @@
  * affect the API (so are not recorded in pnglibconf.h)
  */
 
-#define PNG_NO_WARNINGS
+#define PNG_ALIGN_TYPE PNG_ALIGN_NONE
 
 #endif /* MINPRDPNGCONF_H */

contrib/pngminus/makefile.std

@@ -8,9 +8,9 @@ LD=$(CC)
 RM=rm -f
 
 #PNGPATH = /usr/local
-#PNGINC = -I$(PNGPATH)/include/libpng12
-#PNGLIB = -L$(PNGPATH)/lib -lpng12
-#PNGLIBS = $(PNGPATH)/lib/libpng12.a
+#PNGINC = -I$(PNGPATH)/include/libpng17
+#PNGLIB = -L$(PNGPATH)/lib -lpng17
+#PNGLIBS = $(PNGPATH)/lib/libpng17.a
 PNGINC = -I../..
 PNGLIB = -L../.. -lpng
 PNGLIBS = ../../libpng.a
@@ -23,9 +23,9 @@ ZINC = -I../../../zlib
 ZLIB = -L../../../zlib -lz
 ZLIBS = ../../../zlib/libz.a
 
-CFLAGS=-O3 $(PNGINC) $(ZINC)
-LDFLAGS=$(PNGLIB) $(ZLIB)
-LDFLAGSS=$(PNGLIBS) $(ZLIBS)
+CFLAGS=$(PNGINC) $(ZINC)
+LDLIBS=$(PNGLIB) $(ZLIB)
+LDLIBSS=$(PNGLIBS) $(ZLIBS)
 C=.c
 O=.o
 L=.a
@@ -40,19 +40,19 @@ png2pnm$(O): png2pnm$(C)
 	$(CC) -c $(CFLAGS) png2pnm$(C)
 
 png2pnm$(E): png2pnm$(O)
-	$(LD) -o png2pnm$(E) png2pnm$(O) $(LDFLAGS) -lm
+	$(LD) $(LDFLAGS) -o png2pnm$(E) png2pnm$(O) $(LDLIBS) -lm
 
 png2pnm-static$(E): png2pnm$(O)
-	$(LD) -o png2pnm-static$(E) png2pnm$(O) $(LDFLAGSS) -lm
+	$(LD) $(LDFLAGS) -o png2pnm-static$(E) png2pnm$(O) $(LDLIBSS) -lm
 
 pnm2png$(O): pnm2png$(C)
 	$(CC) -c $(CFLAGS) pnm2png$(C)
 
 pnm2png$(E): pnm2png$(O)
-	$(LD) -o pnm2png$(E) pnm2png$(O) $(LDFLAGS) -lm
+	$(LD) $(LDFLAGS) -o pnm2png$(E) pnm2png$(O) $(LDLIBS) -lm
 
 pnm2png-static$(E): pnm2png$(O)
-	$(LD) -o pnm2png-static$(E) pnm2png$(O) $(LDFLAGSS) -lm
+	$(LD) $(LDFLAGS) -o pnm2png-static$(E) pnm2png$(O) $(LDLIBSS) -lm
 
 clean:
 	$(RM) png2pnm$(O)

contrib/pngminus/png2pnm.c

@@ -18,6 +18,7 @@
 #include <mem.h>
 #include <fcntl.h>
 #endif
+#include <zlib.h>
 
 #ifndef BOOL
 #define BOOL unsigned char
@@ -144,7 +145,7 @@ int main(int argc, char *argv[])
   if (png2pnm (fp_rd, fp_wr, fp_al, raw, alpha) == FALSE)
   {
     fprintf (stderr, "PNG2PNM\n");
-    fprintf (stderr, "Error:  unsuccessful convertion of PNG-image\n");
+    fprintf (stderr, "Error:  unsuccessful conversion of PNG-image\n");
     exit(1);
   }
 
@@ -403,7 +404,7 @@ BOOL png2pnm (FILE *png_file, FILE *pnm_file, FILE *alpha_file, BOOL raw, BOOL a
             if (bit_depth == 16){
               dep_16 = (long) *pix_ptr++;
               fprintf (alpha_file, "%ld ", (dep_16 << 8) + (long) *pix_ptr++);
-            }  
+            }
             else
               fprintf (alpha_file, "%ld ", (long) *pix_ptr++);
         }

contrib/pngminus/pnm2png.c

@@ -18,6 +18,7 @@
 #include <mem.h>
 #include <fcntl.h>
 #endif
+#include <zlib.h>
 
 #ifndef BOOL
 #define BOOL unsigned char
@@ -197,6 +198,9 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
   char          height_token[16];
   char          maxval_token[16];
   int           color_type;
+  unsigned long   ul_width=0, ul_alpha_width=0;
+  unsigned long   ul_height=0, ul_alpha_height=0;
+  unsigned long   ul_maxval=0;
   png_uint_32   width, alpha_width;
   png_uint_32   height, alpha_height;
   png_uint_32   maxval;
@@ -206,6 +210,9 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
   int           alpha_present;
   int           row, col;
   BOOL          raw, alpha_raw = FALSE;
+#if defined(PNG_WRITE_INVERT_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+  BOOL          packed_bitmap = FALSE;
+#endif
   png_uint_32   tmp16;
   int           i;
 
@@ -218,20 +225,36 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
   }
   else if ((type_token[1] == '1') || (type_token[1] == '4'))
   {
+#if defined(PNG_WRITE_INVERT_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
     raw = (type_token[1] == '4');
     color_type = PNG_COLOR_TYPE_GRAY;
+    get_token(pnm_file, width_token);
+    sscanf (width_token, "%lu", &ul_width);
+    width = (png_uint_32) ul_width;
+    get_token(pnm_file, height_token);
+    sscanf (height_token, "%lu", &ul_height);
+    height = (png_uint_32) ul_height;
     bit_depth = 1;
+    packed_bitmap = TRUE;
+#else
+    fprintf (stderr, "PNM2PNG built without PNG_WRITE_INVERT_SUPPORTED and \n");
+    fprintf (stderr, "PNG_WRITE_PACK_SUPPORTED can't read PBM (P1,P4) files\n");
+#endif
   }
   else if ((type_token[1] == '2') || (type_token[1] == '5'))
   {
     raw = (type_token[1] == '5');
     color_type = PNG_COLOR_TYPE_GRAY;
     get_token(pnm_file, width_token);
-    sscanf (width_token, "%lu", &width);
+    sscanf (width_token, "%lu", &ul_width);
+    width = (png_uint_32) ul_width;
     get_token(pnm_file, height_token);
-    sscanf (height_token, "%lu", &height);
+    sscanf (height_token, "%lu", &ul_height);
+    height = (png_uint_32) ul_height;
     get_token(pnm_file, maxval_token);
-    sscanf (maxval_token, "%lu", &maxval);
+    sscanf (maxval_token, "%lu", &ul_maxval);
+    maxval = (png_uint_32) ul_maxval;
+
     if (maxval <= 1)
       bit_depth = 1;
     else if (maxval <= 3)
@@ -248,11 +271,14 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
     raw = (type_token[1] == '6');
     color_type = PNG_COLOR_TYPE_RGB;
     get_token(pnm_file, width_token);
-    sscanf (width_token, "%lu", &width);
+    sscanf (width_token, "%lu", &ul_width);
+    width = (png_uint_32) ul_width;
     get_token(pnm_file, height_token);
-    sscanf (height_token, "%lu", &height);
+    sscanf (height_token, "%lu", &ul_height);
+    height = (png_uint_32) ul_height;
     get_token(pnm_file, maxval_token);
-    sscanf (maxval_token, "%lu", &maxval);
+    sscanf (maxval_token, "%lu", &ul_maxval);
+    maxval = (png_uint_32) ul_maxval;
     if (maxval <= 1)
       bit_depth = 1;
     else if (maxval <= 3)
@@ -287,15 +313,18 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
     {
       alpha_raw = (type_token[1] == '5');
       get_token(alpha_file, width_token);
-      sscanf (width_token, "%lu", &alpha_width);
+      sscanf (width_token, "%lu", &ul_alpha_width);
+      alpha_width=(png_uint_32) ul_alpha_width;
       if (alpha_width != width)
         return FALSE;
       get_token(alpha_file, height_token);
-      sscanf (height_token, "%lu", &alpha_height);
+      sscanf (height_token, "%lu", &ul_alpha_height);
+      alpha_height = (png_uint_32) ul_alpha_height;
       if (alpha_height != height)
         return FALSE;
       get_token(alpha_file, maxval_token);
-      sscanf (maxval_token, "%lu", &maxval);
+      sscanf (maxval_token, "%lu", &ul_maxval);
+      maxval = (png_uint_32) ul_maxval;
       if (maxval <= 1)
         alpha_depth = 1;
       else if (maxval <= 3)
@@ -329,8 +358,14 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
 
   alpha_present = (channels - 1) % 2;
 
-  /* row_bytes is the width x number of channels x (bit-depth / 8) */
-  row_bytes = width * channels * ((bit_depth <= 8) ? 1 : 2);
+#if defined(PNG_WRITE_INVERT_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+  if (packed_bitmap)
+    /* row data is as many bytes as can fit width x channels x bit_depth */
+    row_bytes = (width * channels * bit_depth + 7) / 8;
+  else
+#endif
+    /* row_bytes is the width x number of channels x (bit-depth / 8) */
+    row_bytes = width * channels * ((bit_depth <= 8) ? 1 : 2);
 
   if ((png_pixels = (png_byte *) malloc (row_bytes * height * sizeof (png_byte))) == NULL)
     return FALSE;
@@ -340,40 +375,48 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
 
   for (row = 0; row < height; row++)
   {
-    for (col = 0; col < width; col++)
+#if defined(PNG_WRITE_INVERT_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+    if (packed_bitmap) {
+      for (i = 0; i < row_bytes; i++)
+        /* png supports this format natively so no conversion is needed */
+        *pix_ptr++ = get_data (pnm_file, 8);
+    } else
+#endif
     {
-      for (i = 0; i < (channels - alpha_present); i++)
+      for (col = 0; col < width; col++)
       {
-        if (raw)
-          *pix_ptr++ = get_data (pnm_file, bit_depth);
-        else
-          if (bit_depth <= 8)
-            *pix_ptr++ = get_value (pnm_file, bit_depth);
+        for (i = 0; i < (channels - alpha_present); i++)
+        {
+          if (raw)
+            *pix_ptr++ = get_data (pnm_file, bit_depth);
           else
-          {
-            tmp16 = get_value (pnm_file, bit_depth);
-            *pix_ptr = (png_byte) ((tmp16 >> 8) & 0xFF);
-            pix_ptr++;
-            *pix_ptr = (png_byte) (tmp16 & 0xFF);
-            pix_ptr++;
-          }
-      }
+            if (bit_depth <= 8)
+              *pix_ptr++ = get_value (pnm_file, bit_depth);
+            else
+            {
+              tmp16 = get_value (pnm_file, bit_depth);
+              *pix_ptr = (png_byte) ((tmp16 >> 8) & 0xFF);
+              pix_ptr++;
+              *pix_ptr = (png_byte) (tmp16 & 0xFF);
+              pix_ptr++;
+            }
+        }
 
-      if (alpha) /* read alpha-channel from pgm file */
-      {
-        if (alpha_raw)
-          *pix_ptr++ = get_data (alpha_file, alpha_depth);
-        else
-          if (alpha_depth <= 8)
-            *pix_ptr++ = get_value (alpha_file, bit_depth);
+        if (alpha) /* read alpha-channel from pgm file */
+        {
+          if (alpha_raw)
+            *pix_ptr++ = get_data (alpha_file, alpha_depth);
           else
-          {
-            tmp16 = get_value (alpha_file, bit_depth);
-            *pix_ptr++ = (png_byte) ((tmp16 >> 8) & 0xFF);
-            *pix_ptr++ = (png_byte) (tmp16 & 0xFF);
-          }
-      } /* if alpha */
-
+            if (alpha_depth <= 8)
+              *pix_ptr++ = get_value (alpha_file, bit_depth);
+            else
+            {
+              tmp16 = get_value (alpha_file, bit_depth);
+              *pix_ptr++ = (png_byte) ((tmp16 >> 8) & 0xFF);
+              *pix_ptr++ = (png_byte) (tmp16 & 0xFF);
+            }
+        } /* if alpha */
+      } /* if packed_bitmap */
     } /* end for col */
   } /* end for row */
 
@@ -390,6 +433,14 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
     return FALSE;
   }
 
+#if defined(PNG_WRITE_INVERT_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+  if (packed_bitmap == TRUE)
+  {
+    png_set_packing (png_ptr);
+    png_set_invert_mono (png_ptr);
+  }
+#endif
+
   /* setjmp() must be called in every function that calls a PNG-reading libpng function */
   if (setjmp (png_jmpbuf(png_ptr)))
   {
@@ -446,19 +497,32 @@ BOOL pnm2png (FILE *pnm_file, FILE *png_file, FILE *alpha_file, BOOL interlace,
 void get_token(FILE *pnm_file, char *token)
 {
   int i = 0;
+  int ret;
 
-  /* remove white-space */
+  /* remove white-space and comment lines */
   do
   {
-    token[i] = (unsigned char) fgetc (pnm_file);
+    ret = fgetc(pnm_file);
+    if (ret == '#') {
+      /* the rest of this line is a comment */
+      do
+      {
+        ret = fgetc(pnm_file);
+      }
+      while ((ret != '\n') && (ret != '\r') && (ret != EOF));
+    }
+    if (ret == EOF) break;
+    token[i] = (unsigned char) ret;
   }
   while ((token[i] == '\n') || (token[i] == '\r') || (token[i] == ' '));
 
   /* read string */
   do
   {
+    ret = fgetc(pnm_file);
+    if (ret == EOF) break;
     i++;
-    token[i] = (unsigned char) fgetc (pnm_file);
+    token[i] = (unsigned char) ret;
   }
   while ((token[i] != '\n') && (token[i] != '\r') && (token[i] != ' '));
 
@@ -510,6 +574,7 @@ png_uint_32 get_value (FILE *pnm_file, int depth)
 {
   static png_uint_32 mask = 0;
   png_byte token[16];
+  unsigned long ul_ret_value;
   png_uint_32 ret_value;
   int i = 0;
 
@@ -518,7 +583,8 @@ png_uint_32 get_value (FILE *pnm_file, int depth)
       mask = (mask << 1) | 0x01;
 
   get_token (pnm_file, (char *) token);
-  sscanf ((const char *) token, "%lu", &ret_value);
+  sscanf ((const char *) token, "%lu", &ul_ret_value);
+  ret_value = (png_uint_32) ul_ret_value;
 
   ret_value &= mask;
 

contrib/pngsuite/README

@@ -1,14 +1,15 @@
 
 pngsuite
 --------
-(c) Willem van Schaik, 1999
+(c) Willem van Schaik, 1999, 2011, 2012
+Two images are by Glenn Randers-Pehrson, 2012
 
-Permission to use, copy, and distribute these images for any purpose and
-without fee is hereby granted.
+Permission to use, copy, modify, and distribute these images for any
+purpose and without fee is hereby granted.
 
-The 15 "bas*.png" images are part of the much larger PngSuite test-set of 
-images, available for developers of PNG supporting software. The 
-complete set, available at http:/www.schaik.com/pngsuite/, contains 
+The 15 "bas*.png" images are part of the much larger PngSuite test-set of
+images, available for developers of PNG supporting software. The
+complete set, available at http:/www.schaik.com/pngsuite/, contains
 a variety of images to test interlacing, gamma settings, ancillary
 chunks, etc.
 
@@ -38,6 +39,8 @@ can use them to test the proper functioning of PNG software.
     basn6a08.png   8-bit RGBA
     basn6a16.png  16-bit RGBA
 
+    ftbbn0g01.png  1-bit grayscale, black bKGD
+    ftbbn0g02.png  2-bit grayscale, black bKGD
     ftbbn0g04.png  4-bit grayscale, black bKGD
     ftbbn2c16.png 16-bit truecolor, black bKGD
     ftbbn3p08.png  8-bit paletted, black bKGD
@@ -48,11 +51,11 @@ can use them to test the proper functioning of PNG software.
     ftbwn3p08.png  8-bit paletted, white bKGD
     ftbyn3p08.png  8-bit paletted, yellow bKGD
     ftp0n0g08.png  8-bit grayscale, opaque
-    ftp0n2c08.png  8-bit truecolor, opaque 
+    ftp0n2c08.png  8-bit truecolor, opaque
     ftp0n3p08.png  8-bit paletted, opaque
     ftp1n3p08.png  8-bit paletted, no bKGD
 
-Here is the correct result of typing "pngtest -m *.png" in
+Here is the correct result of typing "pngtest -m bas*.png" in
 this directory:
 
 Testing basn0g01.png: PASS (524 zero samples)

contrib/tools/README.txt

@@ -0,0 +1,26 @@
+This directory (contrib/tools) contains tools used by the authors of libpng.
+
+Code and data placed in this directory is not required to build libpng,
+however the code in this directory has been used to generate data or code in
+the body of the libpng source.  The source code identifies where this has
+been done.  Code in this directory may not compile on all operating systems
+that libpng supports.
+
+NO COPYRIGHT RIGHTS ARE CLAIMED TO ANY OF THE FILES IN THIS DIRECTORY.
+
+To the extent possible under law, the authors have waived all copyright and
+related or neighboring rights to this work.  This work is published from:
+United States.
+
+The files may be used freely in any way.
+
+The source code and comments in this directory are the original work of the
+people named below.  No other person or organization has made contributions to
+the work in this directory.
+
+ORIGINAL AUTHORS
+    The following people have contributed to the code in this directory.  None
+    of the people below claim any rights with regard to the contents of this
+    directory.
+
+    John Bowler <jbowler@acm.org>

contrib/tools/checksum-icc.c

@@ -0,0 +1,102 @@
+/* checksum-icc.c
+ *
+ * Copyright (c) 2013 John Cunningham Bowler
+ *
+ * Last changed in libpng 1.6.0 [February 14, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * Generate crc32 and adler32 checksums of the given input files, used to
+ * generate check-codes for use when matching ICC profiles within libpng.
+ */
+#include <stdio.h>
+
+#include <zlib.h>
+
+static int
+read_one_file(FILE *ip, const char *name)
+{
+   uLong length = 0;
+   uLong a32 = adler32(0, NULL, 0);
+   uLong c32 = crc32(0, NULL, 0);
+   Byte header[132];
+
+   for (;;)
+   {
+      int ch = getc(ip);
+      Byte b;
+
+      if (ch == EOF) break;
+
+      b = (Byte)ch;
+
+      if (length < sizeof header)
+         header[length] = b;
+
+      ++length;
+      a32 = adler32(a32, &b, 1);
+      c32 = crc32(c32, &b, 1);
+   }
+
+   if (ferror(ip))
+      return 0;
+
+   /* Success */
+   printf("PNG_ICC_CHECKSUM(0x%8.8lx, 0x%8.8lx,\n   PNG_MD5("
+      "0x%2.2x%2.2x%2.2x%2.2x, 0x%2.2x%2.2x%2.2x%2.2x, 0x%2.2x%2.2x%2.2x%2.2x,"
+      " 0x%2.2x%2.2x%2.2x%2.2x), %d,\n"
+      "   \"%4.4d/%2.2d/%2.2d %2.2d:%2.2d:%2.2d\", %lu, \"%s\")\n",
+      (unsigned long)a32, (unsigned long)c32,
+      header[84], header[85], header[86], header[87],
+      header[88], header[89], header[90], header[91],
+      header[92], header[93], header[94], header[95],
+      header[96], header[97], header[98], header[99],
+#     define u16(x) (header[x] * 256 + header[x+1])
+#     define u32(x) (u16(x) * 65536 + u16(x+2))
+      u32(64), u16(24), u16(26), u16(28), u16(30), u16(32), u16(34),
+      (unsigned long)length, name);
+
+   return 1;
+}
+
+int main(int argc, char **argv)
+{
+   int err = 0;
+
+   printf("/* adler32, crc32, MD5[16], intent, date, length, file-name */\n");
+
+   if (argc > 1)
+   {
+      int i;
+
+      for (i=1; i<argc; ++i)
+      {
+         FILE *ip = fopen(argv[i], "rb");
+
+         if (ip == NULL || !read_one_file(ip, argv[i]))
+         {
+            err = 1;
+            perror(argv[i]);
+            fprintf(stderr, "%s: read error\n", argv[i]);
+            printf("/* ERROR: %s */\n", argv[i]);
+         }
+
+         (void)fclose(ip);
+      }
+   }
+
+   else
+   {
+      if (!read_one_file(stdin, "-"))
+      {
+         err = 1;
+         perror("stdin");
+         fprintf(stderr, "stdin: read error\n");
+         printf("/* ERROR: stdin */\n");
+      }
+   }
+
+   return err;
+}

contrib/tools/chunkhash.c

@@ -0,0 +1,437 @@
+/* chunkhash.c -- build a perfect hash code for the chunk names on stdin.
+ *
+ * Last changed in libpng 1.7.0 [(PENDING RELEASE)]
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2014.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Feed this program all the known chunks that libpng must recognize.  It will
+ * generate an appropriate hash key for the macro in pngpriv.h  To generate the
+ * list of chunks currently defined in png.h do this:
+ *
+ *    sed -n -e 's/^#define png_\(....\) PNG_U32.*$/\1/p' png.h
+ *
+ * An alternative to using this program is to pipe the output of the above
+ * through gperf, however the code generated by perf is somewhat more verbose
+ * and it doesn't generate as good a result (however it's a heck of a lot faster
+ * than this program!)
+ */
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+#include <png.h> /* for the libpng types for this platform */
+
+/* The machine generated file is checked in, so this works and it obtains the
+ * definition of the PNG_CHUNK_HASH macro.  To change this definition only ever
+ * alter the strings below!
+ */
+#include "chunkhash.h"
+
+/* The best_ variables are local variables defined in the functions below.  When
+ * used in libpng code, however, they are just constants.
+ */
+#define PNG_CHUNK_HASH_MASK best_mask
+#define PNG_CHUNK_HASH_C0 best_c[0]
+#define PNG_CHUNK_HASH_C1 best_c[1]
+#define PNG_CHUNK_HASH_C2 best_c[2]
+#define PNG_CHUNK_HASH_C3 best_c[3]
+
+/* These strings contain the C text to copy into the new version of
+ * contrib/tools/chunkhash.c
+ */
+static const char *strings[] = {
+"/* chunkhash.h -- a perfect hash code for the chunk names in png.h",
+" *",
+" * Last changed in libpng 1.7.0 [(PENDING RELEASE)]",
+" *",
+" * THIS IS A MACHINE GENERATED FILE.  See contrib/tools/chunkhash.c for",
+" * copyright and other information.",
+" *",
+" * USAGE:  To include the PNG_CHUNK_HASH macro and associated definitions:",
+" *",
+" * #define PNG_CHUNKHASH_DEFS",
+" * #include \"contrib/tools/chunkhash.h\"",
+" *",
+" * To define the png_chunk_hash array used by the macro:",
+" *",
+" * #define PNG_CHUNKHASH_CODE",
+" * #include \"contrib/tools/chunkhash.h\"",
+" *",
+" * One or both of the defines must be given except when building chunkhash",
+" * itself.",
+" */",
+"#ifdef PNG_CHUNKHASH_DEFS",
+"#ifndef PNG_CHUNKHASH_H",
+"#define PNG_CHUNKHASH_H",
+"/* A perfect hash code - returns a value 0..(PNG_KNOWN_CHUNK_COUNT-1) and is",
+" * generated by the ridiculously simple program in contrib/tools/chunkhash.c",
+" *",
+" * The hash code used here multiplies each byte by a different constant to",
+" * return a single number:",
+" *",
+" *    b0 * c[0] + b1 * [c1] + b2 * c[2] + b3 * c[3]",
+" *",
+" * The values of the constants are found by search using the a table of",
+" * primes, including 0, and may be (in fact are at present) 0 for some of the",
+" * bytes, the compiler is expected to optimize multiply by zero, or one!",
+" *",
+" * The lookup table reduces the sparse result of the hash calculation to the",
+" * correct index values.  The chunks are indexed in the string order of the",
+" * png_uint_32 chunk names.",
+" */",
+0, /* HEADER definitions go here */
+"",
+"extern const png_byte png_chunk_hash[PNG_CHUNK_HASH_MASK+1];",
+"#endif /* !PNG_CHUNKHASH_H */",
+"#endif /* PNG_CHUNKHASH_DEFS */",
+"",
+"#ifndef PNG_CHUNK_HASH",
+"#define PNG_CHUNK_HASH(chunk) (png_chunk_hash[PNG_CHUNK_HASH_MASK & (\\",
+"      ((chunk) >> 24) * PNG_CHUNK_HASH_C0 +\\",
+"      ((chunk) >> 16) * PNG_CHUNK_HASH_C1 +\\",
+"      ((chunk) >>  8) * PNG_CHUNK_HASH_C2 +\\",
+"      ((chunk)      ) * PNG_CHUNK_HASH_C3)])",
+"#endif",
+"",
+"#ifdef PNG_CHUNKHASH_CODE",
+"#ifndef PNG_CHUNKHASH_C",
+"#define PNG_CHUNKHASH_C",
+"const png_byte png_chunk_hash[PNG_CHUNK_HASH_MASK+1] = {",
+0, /* png.c definitions go here */
+"};",
+"#endif /* !PNG_CHUNKHASH_C */",
+"#endif /* PNG_CHUNKHASH_CODE */",
+0 /* end of file */
+};
+
+#define CHUNK_COUNT_MAX 32
+   /* If necessary this is easy to increase; just don't use a png_uint_32
+    * bitmask below, however it doesn't seem likely that the limit will be hit
+    * any time soon.
+    */
+
+static const png_byte
+chunk_hash[256] =
+{
+   0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21,
+   22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40,
+   41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59,
+   60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78,
+   79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97,
+   98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113,
+   114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128,
+   129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143,
+   144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158,
+   159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173,
+   174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188,
+   189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203,
+   204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218,
+   219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233,
+   234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248,
+   249, 250, 251, 252, 253, 254, 255
+};
+
+static void
+error(const char *string)
+{
+   fprintf(stderr, "chunkhash: %s\n", string);
+   exit(1);
+}
+
+/* Return a hash code for the given chunk; this is the same as the pngpriv.h
+ * macro, but parameterized.  The names of the parameters have to be chosen to
+ * match the above #defines.
+ */
+static png_uint_32
+test_hash(png_uint_32 chunk, const unsigned int *best_c, png_uint_32 best_mask)
+{
+#  define png_chunk_hash chunk_hash /* prevents lookup */
+      return PNG_CHUNK_HASH(chunk);
+#  undef png_chunk_hash
+}
+
+static void
+print_chunks(const png_uint_32 *chunks, unsigned int nchunks, png_uint_32 first,
+   png_uint_32 last, const unsigned int *best_c, png_uint_32 best_mask,
+   const png_byte *lut)
+{
+   /* 'last' is the last code to print plus 1, this is used to detect duplicates
+    * (which should not happen - if there are any it's an internal error!)
+    */
+   if (nchunks > 0)
+   {
+      /* Recursive algorithm; find the hash code of the next chunk, then print
+       * all remaining chunks with codes in the range first..code (including
+       * duplicates) and after print all remaining chunks with codes in the
+       * range (code+1)..last
+       */
+      const png_uint_32 chunk = *chunks++;
+#     define png_chunk_hash lut
+         const png_uint_32 code = PNG_CHUNK_HASH(chunk);
+#     undef png_chunk_hash
+
+      --nchunks;
+
+      /* The code might be out of the print range */
+      if (code >= first && code <= last)
+      {
+         if (code >= first) /* not time yet */
+            print_chunks(chunks, nchunks, first, code, best_c, best_mask, lut);
+
+         printf("#define PNG_CHUNK_%c%c%c%c_TAG %lu%s\n", (char)(chunk>>24),
+            (char)(chunk>>16), (char)(chunk>>8), (char)chunk,
+            (unsigned long)code, code == last ? " /* DUPLICATE */" : "");
+
+         if (code < last) /* still some to go */
+            print_chunks(chunks, nchunks, code+1, last, best_c, best_mask, lut);
+      }
+
+      else
+         print_chunks(chunks, nchunks, first, last, best_c, best_mask, lut);
+   }
+}
+
+static unsigned int primes[] =
+{
+   0, 1, 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67,
+   71, 73, 79, 83, 89, 97, 101, 103, 107, 109, 113, 127, 131, 137, 139, 149,
+   151, 157, 163, 167, 173, 179, 181, 191, 193, 197, 199, 211, 223, 227, 229,
+   233, 239, 241, 251
+};
+
+#define NPRIMES ((sizeof primes)/(sizeof primes[0]))
+
+int
+main(void)
+{
+   /* Stdin should just contain chunk names, one per line */
+   png_uint_32 best_mask;
+   unsigned int best_c[4];
+   png_uint_32 chunks[CHUNK_COUNT_MAX];
+   png_byte known_chunk_count;
+
+   /* Not required; stop GCC whining: */
+   memset(best_c, 0xab, sizeof best_c);
+   best_mask = 0x12345678;
+
+   {
+      png_uint_32 this_chunk = 0;
+      png_byte n_chunks = 0;
+
+      for (;;)
+      {
+         int ch = getchar();
+
+         if ((ch >= 'a' && ch <= 'z') || (ch >= 'A' && ch <= 'Z'))
+         {
+            if (this_chunk > 0xffffffU)
+               error("chunk name too long");
+
+            this_chunk = (this_chunk << 8) + (unsigned)ch;
+         }
+
+         else
+         {
+            if (this_chunk > 0)
+            {
+               if (this_chunk <= 0xffffffU)
+                  error("chunk name too short");
+
+               if (n_chunks >= CHUNK_COUNT_MAX)
+                  error("too many chunks (check CHUNK_COUNT_MAX)");
+
+               chunks[n_chunks++] = this_chunk;
+               this_chunk = 0;
+            }
+         }
+
+         if (ch < 0)
+            break;
+      }
+
+      /* 22 is the number of chunks currently defined (excluding fRAc), at any
+       * time it should also be in PNG_KNOWN_CHUNK_COUNT, but this run of
+       * chunkhash may be trying to add a chunk so allow bigger numbers.
+       */
+#     ifdef PNG_KNOWN_CHUNK_COUNT
+#        define CHUNK_COUNT_MIN PNG_KNOWN_CHUNK_COUNT
+#     else
+#        define CHUNK_COUNT_MIN 26
+#     endif
+      if (n_chunks < CHUNK_COUNT_MIN)
+         error("too few chunks (expecting at least 26)");
+
+      known_chunk_count = n_chunks;
+   }
+
+   /* Exhaustive search of the hash parameters - in fact this isn't very slow at
+    * all.
+    */
+   {
+      unsigned int i1, c_zero = 0, c_one = 0;
+
+      for (i1=0; i1<NPRIMES; ++i1)
+      {
+         unsigned int i2;
+         unsigned int c[4];
+
+         fprintf(stderr, "TEST: %u\n", primes[i1]);
+         c[0] = primes[i1];
+
+         for (i2=0; i2<NPRIMES; ++i2)
+         {
+            unsigned int i3;
+
+            c[1] = primes[i2];
+
+            for (i3=0; i3<NPRIMES; ++i3)
+            {
+               unsigned int i4;
+
+               c[2] = primes[i3];
+
+               for (i4=0; i4<NPRIMES; ++i4)
+               {
+                  unsigned int i;
+                  png_uint_32 hash_mask = 0xffU;
+                  png_uint_32 hashes[CHUNK_COUNT_MAX];
+
+                  c[3] = primes[i4];
+
+                  while (hash_mask > 0xfU)
+                  {
+                     for (i=0; i<known_chunk_count; ++i)
+                     {
+                        unsigned int j;
+                        png_uint_32 hash = test_hash(chunks[i], c, hash_mask);
+
+                        for (j=0; j<i; ++j) if (hashes[j] == hash) goto next_i4;
+
+                        hashes[i] = hash;
+                     }
+
+                     hash_mask >>= 1;
+                  }
+
+               next_i4:
+                  if (hash_mask < 0xffU)
+                  {
+                     /* This worked */
+                     unsigned int best, c0 = 0, c1 = 0;
+
+                     hash_mask <<= 1;
+                     hash_mask += 1;
+
+                     for (i=0; i<4; ++i)
+                     {
+                        if (c[i] == 0)
+                           ++c0;
+
+                        else if (c[i] == 1)
+                           ++c1;
+                     }
+
+                     if (hash_mask == best_mask)
+                        best = (c0 > c_zero) || (c0 == c_zero && c1 > c_one);
+
+                     else
+                        best = (hash_mask < best_mask);
+
+                     if (best)
+                     {
+                        fprintf(stderr, "{%u,%u,%u,%u} & 0x%lx\n",
+                           c[0], c[1], c[2], c[3],
+                           (unsigned long)hash_mask);
+
+                        c_zero = c0;
+                        c_one = c1;
+                        best_mask = hash_mask;
+                        memcpy(best_c, c, sizeof best_c);
+                     }
+                  }
+               } /* i4 */
+            } /* i3 */
+         } /* i2 */
+      } /* i1 */
+   }
+
+   /* Calculate the LUT (png_chunk_hash) */
+   {
+      png_byte b;
+      png_byte lut[256];
+
+      /* A failure returns PNG_KNOWN_CHUNK_COUNT: */
+      memset(lut, known_chunk_count, sizeof lut);
+
+      for (b=0; b<known_chunk_count; ++b)
+         lut[test_hash(chunks[b], best_c, best_mask)] = b;
+
+      /* Validate the pngpriv.h hash function. */
+#     define png_chunk_hash lut
+         {
+            unsigned int i;
+
+            for (i=0; i<known_chunk_count; ++i)
+            {
+               png_uint_32 chunk = chunks[i];
+
+               if (PNG_CHUNK_HASH(chunk) != i)
+                  error("internal error: hash didn't work!");
+            }
+         }
+#     undef lut
+
+      /* Print all the results, first the stuff for pngpriv.h */
+      {
+         unsigned int i = 0;
+
+         while (strings[i] != 0)
+            puts(strings[i++]);
+
+         printf("#define PNG_CHUNK_HASH_MASK 0x%lxU\n",
+            (unsigned long)best_mask);
+         printf("#define PNG_CHUNK_HASH_C0 %u\n", best_c[0]);
+         printf("#define PNG_CHUNK_HASH_C1 %u\n", best_c[1]);
+         printf("#define PNG_CHUNK_HASH_C2 %u\n", best_c[2]);
+         printf("#define PNG_CHUNK_HASH_C3 %u\n", best_c[3]);
+
+         /* Print the hash codes of all the chunks */
+         putchar('\n');
+         print_chunks(chunks, known_chunk_count, 0, 0xffffffff, best_c,
+            best_mask, lut);
+         putchar('\n');
+         printf("#define PNG_KNOWN_CHUNK_COUNT %u\n", known_chunk_count);
+
+         while (strings[++i] != 0)
+            puts(strings[i]);
+
+         /* Now print the LUT */
+         fputs("   ", stdout);
+         {
+            unsigned int j;
+
+            for (j=0; j<=best_mask; ++j)
+            {
+               printf("%d", lut[j]);
+
+               if ((j % 16) == 15 && j < best_mask)
+                  printf(",\n   ");
+
+               else if (j < best_mask)
+                  printf(", ");
+
+               else
+                  printf("\n");
+            }
+         }
+
+         /* And the trailing text */
+         while (strings[++i] != 0)
+            puts(strings[i]);
+      }
+   }
+
+   exit(0);
+}

contrib/tools/chunkhash.h

@@ -0,0 +1,97 @@
+/* chunkhash.h -- a perfect hash code for the chunk names in png.h
+ *
+ * Last changed in libpng 1.7.0 [(PENDING RELEASE)]
+ *
+ * THIS IS A MACHINE GENERATED FILE.  See contrib/tools/chunkhash.c for
+ * copyright and other information.
+ *
+ * USAGE:  To include the PNG_CHUNK_HASH macro and associated definitions:
+ *
+ * #define PNG_CHUNKHASH_DEFS
+ * #include "contrib/tools/chunkhash.h"
+ *
+ * To define the png_chunk_hash array used by the macro:
+ *
+ * #define PNG_CHUNKHASH_CODE
+ * #include "contrib/tools/chunkhash.h"
+ *
+ * One or both of the defines must be given except when building chunkhash
+ * itself.
+ */
+#ifdef PNG_CHUNKHASH_DEFS
+#ifndef PNG_CHUNKHASH_H
+#define PNG_CHUNKHASH_H
+/* A perfect hash code - returns a value 0..(PNG_KNOWN_CHUNK_COUNT-1) and is
+ * generated by the ridiculously simple program in contrib/tools/chunkhash.c
+ *
+ * The hash code used here multiplies each byte by a different constant to
+ * return a single number:
+ *
+ *    b0 * c[0] + b1 * [c1] + b2 * c[2] + b3 * c[3]
+ *
+ * The values of the constants are found by search using the a table of
+ * primes, including 0, and may be (in fact are at present) 0 for some of the
+ * bytes, the compiler is expected to optimize multiply by zero, or one!
+ *
+ * The lookup table reduces the sparse result of the hash calculation to the
+ * correct index values.  The chunks are indexed in the string order of the
+ * png_uint_32 chunk names.
+ */
+#define PNG_CHUNK_HASH_MASK 0x3fU
+#define PNG_CHUNK_HASH_C0 103
+#define PNG_CHUNK_HASH_C1 1
+#define PNG_CHUNK_HASH_C2 0
+#define PNG_CHUNK_HASH_C3 1
+
+#define PNG_CHUNK_IDAT_TAG 0
+#define PNG_CHUNK_IEND_TAG 1
+#define PNG_CHUNK_IHDR_TAG 2
+#define PNG_CHUNK_PLTE_TAG 3
+#define PNG_CHUNK_bKGD_TAG 4
+#define PNG_CHUNK_cHRM_TAG 5
+#define PNG_CHUNK_fRAc_TAG 6
+#define PNG_CHUNK_gAMA_TAG 7
+#define PNG_CHUNK_gIFg_TAG 8
+#define PNG_CHUNK_gIFt_TAG 9
+#define PNG_CHUNK_gIFx_TAG 10
+#define PNG_CHUNK_hIST_TAG 11
+#define PNG_CHUNK_iCCP_TAG 12
+#define PNG_CHUNK_iTXt_TAG 13
+#define PNG_CHUNK_oFFs_TAG 14
+#define PNG_CHUNK_pCAL_TAG 15
+#define PNG_CHUNK_pHYs_TAG 16
+#define PNG_CHUNK_sBIT_TAG 17
+#define PNG_CHUNK_sCAL_TAG 18
+#define PNG_CHUNK_sPLT_TAG 19
+#define PNG_CHUNK_sRGB_TAG 20
+#define PNG_CHUNK_sTER_TAG 21
+#define PNG_CHUNK_tEXt_TAG 22
+#define PNG_CHUNK_tIME_TAG 23
+#define PNG_CHUNK_tRNS_TAG 24
+#define PNG_CHUNK_zTXt_TAG 25
+
+#define PNG_KNOWN_CHUNK_COUNT 26
+
+extern const png_byte png_chunk_hash[PNG_CHUNK_HASH_MASK+1];
+#endif /* !PNG_CHUNKHASH_H */
+#endif /* PNG_CHUNKHASH_DEFS */
+
+#ifndef PNG_CHUNK_HASH
+#define PNG_CHUNK_HASH(chunk) (png_chunk_hash[PNG_CHUNK_HASH_MASK & (\
+      ((chunk) >> 24) * PNG_CHUNK_HASH_C0 +\
+      ((chunk) >> 16) * PNG_CHUNK_HASH_C1 +\
+      ((chunk) >>  8) * PNG_CHUNK_HASH_C2 +\
+      ((chunk)      ) * PNG_CHUNK_HASH_C3)])
+#endif
+
+#ifdef PNG_CHUNKHASH_CODE
+#ifndef PNG_CHUNKHASH_C
+#define PNG_CHUNKHASH_C
+const png_byte png_chunk_hash[PNG_CHUNK_HASH_MASK+1] = {
+   26, 3, 26, 26, 26, 26, 26, 13, 26, 26, 26, 16, 26, 26, 26, 26,
+   26, 24, 12, 26, 18, 26, 26, 26, 26, 20, 26, 17, 26, 26, 25, 15,
+   26, 8, 14, 26, 26, 22, 26, 26, 1, 19, 5, 21, 26, 26, 9, 26,
+   26, 26, 10, 7, 26, 11, 26, 0, 26, 2, 23, 26, 26, 4, 26, 6
+};
+#endif /* !PNG_CHUNKHASH_C */
+#endif /* PNG_CHUNKHASH_CODE */

contrib/tools/cvtcolor.c

@@ -0,0 +1,188 @@
+/*-
+ * convert.c
+ *
+ * Last changed in libpng 1.6.0 [February 14, 2013]
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2013.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Convert 8-bit sRGB or 16-bit linear values to another format.
+ */
+#define _ISOC99_SOURCE 1
+
+#include <stdlib.h>
+#include <string.h>
+#include <math.h>
+#include <stdio.h>
+
+#include <fenv.h>
+
+#include "sRGB.h"
+
+static void
+usage(const char *prog)
+{
+   fprintf(stderr,
+      "%s: usage: %s [-linear|-sRGB] [-gray|-color] component{1,4}\n",
+      prog, prog);
+   exit(1);
+}
+
+unsigned long
+component(const char *prog, const char *arg, int issRGB)
+{
+   char *ep;
+   unsigned long c = strtoul(arg, &ep, 0);
+
+   if (ep <= arg || *ep || c > 65535 || (issRGB && c > 255))
+   {
+      fprintf(stderr, "%s: %s: invalid component value (%lu)\n", prog, arg, c);
+      usage(prog);
+   }
+
+   return c;
+}
+
+int
+main(int argc, const char **argv)
+{
+   const char *prog = *argv++;
+   int to_linear = 0, to_gray = 0, to_color = 0;
+   int channels = 0;
+   double c[4];
+
+   /* FE_TONEAREST is the IEEE754 round to nearest, preferring even, mode; i.e.
+    * everything rounds to the nearest value except that '.5' rounds to the
+    * nearest even value.
+    */
+   fesetround(FE_TONEAREST);
+
+   c[3] = c[2] = c[1] = c[0] = 0;
+
+   while (--argc > 0 && **argv == '-')
+   {
+      const char *arg = 1+*argv++;
+
+      if (strcmp(arg, "sRGB") == 0)
+         to_linear = 0;
+
+      else if (strcmp(arg, "linear") == 0)
+         to_linear = 1;
+
+      else if (strcmp(arg, "gray") == 0)
+         to_gray = 1, to_color = 0;
+
+      else if (strcmp(arg, "color") == 0)
+         to_gray = 0, to_color = 1;
+
+      else
+         usage(prog);
+   }
+
+   switch (argc)
+   {
+      default:
+         usage(prog);
+         break;
+
+      case 4:
+         c[3] = component(prog, argv[3], to_linear);
+         ++channels;
+      case 3:
+         c[2] = component(prog, argv[2], to_linear);
+         ++channels;
+      case 2:
+         c[1] = component(prog, argv[1], to_linear);
+         ++channels;
+      case 1:
+         c[0] = component(prog, argv[0], to_linear);
+         ++channels;
+         break;
+      }
+
+   if (to_linear)
+   {
+      int i;
+      int components = channels;
+
+      if ((components & 1) == 0)
+         --components;
+
+      for (i=0; i<components; ++i) c[i] = linear_from_sRGB(c[i] / 255);
+      if (components < channels)
+         c[components] = c[components] / 255;
+   }
+
+   else
+   {
+      int i;
+      for (i=0; i<4; ++i) c[i] /= 65535;
+
+      if ((channels & 1) == 0)
+      {
+         double alpha = c[channels-1];
+
+         if (alpha > 0)
+            for (i=0; i<channels-1; ++i) c[i] /= alpha;
+         else
+            for (i=0; i<channels-1; ++i) c[i] = 1;
+      }
+   }
+
+   if (to_gray)
+   {
+      if (channels < 3)
+      {
+         fprintf(stderr, "%s: too few channels (%d) for -gray\n",
+            prog, channels);
+         usage(prog);
+      }
+
+      c[0] = YfromRGB(c[0], c[1], c[2]);
+      channels -= 2;
+   }
+
+   if (to_color)
+   {
+      if (channels > 2)
+      {
+         fprintf(stderr, "%s: too many channels (%d) for -color\n",
+            prog, channels);
+         usage(prog);
+      }
+
+      c[3] = c[1]; /* alpha, if present */
+      c[2] = c[1] = c[0];
+   }
+
+   if (to_linear)
+   {
+      int i;
+      if ((channels & 1) == 0)
+      {
+         double alpha = c[channels-1];
+         for (i=0; i<channels-1; ++i) c[i] *= alpha;
+      }
+
+      for (i=0; i<channels; ++i) c[i] = nearbyint(c[i] * 65535);
+   }
+
+   else /* to sRGB */
+   {
+      int i = (channels+1)&~1;
+      while (--i >= 0)
+         c[i] = sRGB_from_linear(c[i]);
+
+      for (i=0; i<channels; ++i) c[i] = nearbyint(c[i] * 255);
+   }
+
+   {
+      int i;
+      for (i=0; i<channels; ++i) printf(" %g", c[i]);
+   }
+   printf("\n");
+
+   return 0;
+}

contrib/tools/intgamma.sh

@@ -0,0 +1,110 @@
+#!/bin/sh
+#
+# intgamma.sh
+#
+# Last changed in libpng 1.6.0 [February 14, 2013]
+#
+# COPYRIGHT: Written by John Cunningham Bowler, 2013.
+# To the extent possible under law, the author has waived all copyright and
+# related or neighboring rights to this work.  This work is published from:
+# United States.
+#
+# Shell script to generate png.c 8-bit and 16-bit log tables (see the code in
+# png.c for details).
+#
+# This script uses the "bc" arbitrary precision calculator to calculate 32-bit
+# fixed point values of logarithms appropriate to finding the log of an 8-bit
+# (0..255) value and a similar table for the exponent calculation.
+#
+# "bc" must be on the path when the script is executed, and the math library
+# (-lm) must be available
+#
+# function to print out a list of numbers as integers; the function truncates
+# the integers which must be one-per-line
+function print(){
+   awk 'BEGIN{
+      str = ""
+   }
+   {
+      sub("\\.[0-9]*$", "")
+      if ($0 == "")
+         $0 = "0"
+
+      if (str == "")
+         t = "   " $0 "U"
+      else
+         t = str ", " $0 "U"
+
+      if (length(t) >= 80) {
+         print str ","
+         str = "   " $0 "U"
+      } else
+         str = t
+   }
+   END{
+      print str
+   }'
+}
+#
+# The logarithm table.
+cat <<END
+/* 8-bit log table: png_8bit_l2[128]
+ * This is a table of -log(value/255)/log(2) for 'value' in the range 128 to
+ * 255, so it's the base 2 logarithm of a normalized 8-bit floating point
+ * mantissa.  The numbers are 32-bit fractions.
+ */
+static const png_uint_32
+png_8bit_l2[128] =
+{
+END
+#
+bc -lqws <<END | print
+f=65536*65536/l(2)
+for (i=128;i<256;++i) { .5 - l(i/255)*f; }
+END
+echo '};'
+echo
+#
+# The exponent table.
+cat <<END
+/* The 'exp()' case must invert the above, taking a 20-bit fixed point
+ * logarithmic value and returning a 16 or 8-bit number as appropriate.  In
+ * each case only the low 16 bits are relevant - the fraction - since the
+ * integer bits (the top 4) simply determine a shift.
+ *
+ * The worst case is the 16-bit distinction between 65535 and 65534; this
+ * requires perhaps spurious accuracy in the decoding of the logarithm to
+ * distinguish log2(65535/65534.5) - 10^-5 or 17 bits.  There is little chance
+ * of getting this accuracy in practice.
+ *
+ * To deal with this the following exp() function works out the exponent of the
+ * frational part of the logarithm by using an accurate 32-bit value from the
+ * top four fractional bits then multiplying in the remaining bits.
+ */
+static const png_uint_32
+png_32bit_exp[16] =
+{
+END
+#
+bc -lqws <<END | print
+f=l(2)/16
+for (i=0;i<16;++i) {
+   x = .5 + e(-i*f)*2^32;
+   if (x >= 2^32) x = 2^32-1;
+   x;
+}
+END
+echo '};'
+echo
+#
+# And the table of adjustment values.
+cat <<END
+/* Adjustment table; provided to explain the numbers in the code below. */
+#if 0
+END
+bc -lqws <<END | awk '{ printf "%5d %s\n", 12-NR, $0 }'
+for (i=11;i>=0;--i){
+   (1 - e(-(2^i)/65536*l(2))) * 2^(32-i)
+}
+END
+echo '#endif'

contrib/tools/makesRGB.c

@@ -0,0 +1,430 @@
+/* makesRGB.c -- build sRGB-to-linear and linear-to-sRGB conversion tables
+ *
+ * Last changed in libpng 1.6.0 [February 14, 2013]
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2013.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Make a table to convert 8-bit sRGB encoding values into the closest 16-bit
+ * linear value.
+ *
+ * Make two tables to take a linear value scaled to 255*65535 and return an
+ * approximation to the 8-bit sRGB encoded value.  Calculate the error in these
+ * tables and display it.
+ */
+#define _C99_SOURCE 1
+#include <stdio.h>
+#include <math.h>
+#include <stdlib.h>
+
+/* pngpriv.h includes the definition of 'PNG_sRGB_FROM_LINEAR' which is required
+ * to verify the actual code.
+ */
+#include "../../pngpriv.h"
+
+#include "sRGB.h"
+
+/* The tables are declared 'const' in pngpriv.h, so this redefines the tables to
+ * be used.
+ */
+#define png_sRGB_table sRGB_table
+#define png_sRGB_base sRGB_base
+#define png_sRGB_delta sRGB_delta
+
+static png_uint_16 png_sRGB_table[256];
+static png_uint_16 png_sRGB_base[512];
+static png_byte png_sRGB_delta[512];
+
+static const unsigned int max_input = 255*65535;
+
+double
+fsRGB(double l)
+{
+   return sRGB_from_linear(l/max_input);
+}
+
+double
+sRGB(unsigned int i)
+{
+   return fsRGB(i);
+}
+
+double
+finvsRGB(unsigned int i)
+{
+   return 65535 * linear_from_sRGB(i/255.);
+}
+
+png_uint_16
+invsRGB(unsigned int i)
+{
+   unsigned int x = nearbyint(finvsRGB(i));
+
+   if (x > 65535)
+   {
+      fprintf(stderr, "invsRGB(%u) overflows to %u\n", i, x);
+      exit(1);
+   }
+
+   return (png_uint_16)x;
+}
+
+int
+main(int argc, char **argv)
+{
+   unsigned int i, i16, ibase;
+   double min_error = 0;
+   double max_error = 0;
+   double min_error16 = 0;
+   double max_error16 = 0;
+   double adjust;
+   double adjust_lo = 0.4, adjust_hi = 0.6, adjust_mid = 0.5;
+   unsigned int ec_lo = 0, ec_hi = 0, ec_mid = 0;
+   unsigned int error_count = 0;
+   unsigned int error_count16 = 0;
+   int test_only = 0;
+
+   if (argc > 1)
+      test_only = strcmp("--test", argv[1]) == 0;
+
+   /* Initialize the encoding table first. */
+   for (i=0; i<256; ++i)
+   {
+      png_sRGB_table[i] = invsRGB(i);
+   }
+
+   /* Now work out the decoding tables (this is where the error comes in because
+    * there are 512 set points and 512 straight lines between them.)
+    */
+   for (;;)
+   {
+      if (ec_lo == 0)
+         adjust = adjust_lo;
+
+      else if (ec_hi == 0)
+         adjust = adjust_hi;
+
+      else if (ec_mid == 0)
+         adjust = adjust_mid;
+
+      else if (ec_mid < ec_hi)
+         adjust = (adjust_mid + adjust_hi)/2;
+
+      else if (ec_mid < ec_lo)
+         adjust = (adjust_mid + adjust_lo)/2;
+
+      else
+      {
+         fprintf(stderr, "not reached: %u .. %u .. %u\n", ec_lo, ec_mid, ec_hi);
+         exit(1);
+      }
+
+      /* Calculate the table using the current 'adjust' */
+      for (i=0; i<=511; ++i)
+      {
+         double lo = 255 * sRGB(i << 15);
+         double hi = 255 * sRGB((i+1) << 15);
+         unsigned int calc;
+
+         calc = nearbyint((lo+adjust) * 256);
+         if (calc > 65535)
+         {
+            fprintf(stderr, "table[%d][0]: overflow %08x (%d)\n", i, calc,
+               calc);
+            exit(1);
+         }
+         png_sRGB_base[i] = calc;
+
+         calc = nearbyint((hi-lo) * 32);
+         if (calc > 255)
+         {
+            fprintf(stderr, "table[%d][1]: overflow %08x (%d)\n", i, calc,
+               calc);
+            exit(1);
+         }
+         png_sRGB_delta[i] = calc;
+      }
+
+      /* Check the 16-bit linear values alone: */
+      error_count16 = 0;
+      for (i16=0; i16 <= 65535; ++i16)
+      {
+         unsigned int i = 255*i16;
+         unsigned int iexact = nearbyint(255*sRGB(i));
+         unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
+
+         if (icalc != iexact)
+            ++error_count16;
+      }
+
+      /* Now try changing the adjustment. */
+      if (ec_lo == 0)
+         ec_lo = error_count16;
+
+      else if (ec_hi == 0)
+         ec_hi = error_count16;
+
+      else if (ec_mid == 0)
+      {
+         ec_mid = error_count16;
+         printf("/* initial error counts: %u .. %u .. %u */\n", ec_lo, ec_mid,
+            ec_hi);
+      }
+
+      else if (error_count16 < ec_mid)
+      {
+         printf("/* adjust (mid ): %f: %u -> %u */\n", adjust, ec_mid,
+            error_count16);
+         ec_mid = error_count16;
+         adjust_mid = adjust;
+      }
+
+      else if (adjust < adjust_mid && error_count16 < ec_lo)
+      {
+         printf("/* adjust (low ): %f: %u -> %u */\n", adjust, ec_lo,
+            error_count16);
+         ec_lo = error_count16;
+         adjust_lo = adjust;
+      }
+
+      else if (adjust > adjust_mid && error_count16 < ec_hi)
+      {
+         printf("/* adjust (high): %f: %u -> %u */\n", adjust, ec_hi,
+            error_count16);
+         ec_hi = error_count16;
+         adjust_hi = adjust;
+      }
+
+      else
+      {
+         adjust = adjust_mid;
+         printf("/* adjust: %f: %u */\n", adjust, ec_mid);
+         break;
+      }
+   }
+
+   /* For each entry in the table try to adjust it to minimize the error count
+    * in that entry.  Each entry corresponds to 128 input values.
+    */
+   for (ibase=0; ibase<65536; ibase+=128)
+   {
+      png_uint_16 base = png_sRGB_base[ibase >> 7], trybase = base, ob=base;
+      png_byte delta = png_sRGB_delta[ibase >> 7], trydelta = delta, od=delta;
+      unsigned int ecbase = 0, eco;
+
+      for (;;)
+      {
+         png_sRGB_base[ibase >> 7] = trybase;
+         png_sRGB_delta[ibase >> 7] = trydelta;
+
+         /* Check the 16-bit linear values alone: */
+         error_count16 = 0;
+         for (i16=ibase; i16 < ibase+128; ++i16)
+         {
+            unsigned int i = 255*i16;
+            unsigned int iexact = nearbyint(255*sRGB(i));
+            unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
+
+            if (icalc != iexact)
+               ++error_count16;
+         }
+
+         if (error_count16 == 0)
+            break;
+
+         if (ecbase == 0)
+         {
+            eco = ecbase = error_count16;
+            ++trybase; /* First test */
+         }
+
+         else if (error_count16 < ecbase)
+         {
+            if (trybase > base)
+            {
+               base = trybase;
+               ++trybase;
+            }
+            else if (trybase < base)
+            {
+               base = trybase;
+               --trybase;
+            }
+            else if (trydelta > delta)
+            {
+               delta = trydelta;
+               ++trydelta;
+            }
+            else if (trydelta < delta)
+            {
+               delta = trydelta;
+               --trydelta;
+            }
+            else
+            {
+               fprintf(stderr, "makesRGB: impossible\n");
+               exit(1);
+            }
+            ecbase = error_count16;
+         }
+
+         else
+         {
+            if (trybase > base)
+               trybase = base-1;
+            else if (trybase < base)
+            {
+               trybase = base;
+               ++trydelta;
+            }
+            else if (trydelta > delta)
+               trydelta = delta-1;
+            else if (trydelta < delta)
+               break; /* end of tests */
+         }
+      }
+
+      png_sRGB_base[ibase >> 7] = base;
+      png_sRGB_delta[ibase >> 7] = delta;
+      if (base != ob || delta != od)
+      {
+         printf("/* table[%u]={%u,%u} -> {%u,%u} %u -> %u errors */\n",
+            ibase>>7, ob, od, base, delta, eco, ecbase);
+      }
+      else if (0)
+         printf("/* table[%u]={%u,%u} %u errors */\n", ibase>>7, ob, od,
+            ecbase);
+   }
+
+   /* Only do the full (slow) test at the end: */
+   min_error = -.4999;
+   max_error = .4999;
+   error_count = 0;
+
+   for (i=0; i <= max_input; ++i)
+   {
+      unsigned int iexact = nearbyint(255*sRGB(i));
+      unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
+
+      if (icalc != iexact)
+      {
+         double err = 255*sRGB(i) - icalc;
+
+         if (err > (max_error+.001) || err < (min_error-.001))
+         {
+            printf(
+               "/* 0x%08x: exact: %3d, got: %3d [tables: %08x, %08x] (%f) */\n",
+               i, iexact, icalc, png_sRGB_base[i>>15],
+               png_sRGB_delta[i>>15], err);
+         }
+
+         ++error_count;
+         if (err > max_error)
+            max_error = err;
+         else if (err < min_error)
+            min_error = err;
+      }
+   }
+
+   /* Re-check the 16-bit cases too, including the warning if there is an error
+    * bigger than 1.
+    */
+   error_count16 = 0;
+   max_error16 = 0;
+   min_error16 = 0;
+   for (i16=0; i16 <= 65535; ++i16)
+   {
+      unsigned int i = 255*i16;
+      unsigned int iexact = nearbyint(255*sRGB(i));
+      unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
+
+      if (icalc != iexact)
+      {
+         double err = 255*sRGB(i) - icalc;
+
+         ++error_count16;
+         if (err > max_error16)
+            max_error16 = err;
+         else if (err < min_error16)
+            min_error16 = err;
+
+         if (abs(icalc - iexact) > 1)
+            printf(
+               "/* 0x%04x: exact: %3d, got: %3d [tables: %08x, %08x] (%f) */\n",
+               i16, iexact, icalc, png_sRGB_base[i>>15],
+               png_sRGB_delta[i>>15], err);
+      }
+   }
+
+   /* Check the round trip for each 8-bit sRGB value. */
+   for (i16=0; i16 <= 255; ++i16)
+   {
+      unsigned int i = 255 * png_sRGB_table[i16];
+      unsigned int iexact = nearbyint(255*sRGB(i));
+      unsigned int icalc = PNG_sRGB_FROM_LINEAR(i);
+
+      if (i16 != iexact)
+      {
+         fprintf(stderr, "8-bit rounding error: %d -> %d\n", i16, iexact);
+         exit(1);
+      }
+
+      if (icalc != i16)
+      {
+         double finv = finvsRGB(i16);
+
+         printf("/* 8-bit roundtrip error: %d -> %f -> %d(%f) */\n",
+            i16, finv, icalc, fsRGB(255*finv));
+      }
+   }
+
+
+   printf("/* error: %g - %g, %u (%g%%) of readings inexact */\n",
+      min_error, max_error, error_count, (100.*error_count)/max_input);
+   printf("/* 16-bit error: %g - %g, %u (%g%%) of readings inexact */\n",
+      min_error16, max_error16, error_count16, (100.*error_count16)/65535);
+
+   if (!test_only)
+   {
+      printf("PNG_CONST png_uint_16 png_sRGB_table[256] =\n{\n   ");
+      for (i=0; i<255; )
+      {
+         do
+         {
+            printf("%d,", png_sRGB_table[i++]);
+         }
+         while ((i & 0x7) != 0 && i<255);
+         if (i<255) printf("\n   ");
+      }
+      printf("%d\n};\n\n", png_sRGB_table[i]);
+
+
+      printf("PNG_CONST png_uint_16 png_sRGB_base[512] =\n{\n   ");
+      for (i=0; i<511; )
+      {
+         do
+         {
+            printf("%d,", png_sRGB_base[i++]);
+         }
+         while ((i & 0x7) != 0 && i<511);
+         if (i<511) printf("\n   ");
+      }
+      printf("%d\n};\n\n", png_sRGB_base[i]);
+
+      printf("PNG_CONST png_byte png_sRGB_delta[512] =\n{\n   ");
+      for (i=0; i<511; )
+      {
+         do
+         {
+            printf("%d,", png_sRGB_delta[i++]);
+         }
+         while ((i & 0xf) != 0 && i<511);
+         if (i<511) printf("\n   ");
+      }
+      printf("%d\n};\n\n", png_sRGB_delta[i]);
+   }
+
+   return 0;
+}

contrib/tools/png-fix-itxt.c

@@ -0,0 +1,153 @@
+
+/* png-fix-itxt version 1.0.0
+ *
+ * Copyright 2013 Glenn Randers-Pehrson
+ * Last changed in libpng 1.6.3 [July 18, 2013]
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ *
+ * Usage:            
+ *
+ *     png-fix-itxt.exe < bad.png > good.png
+ *
+ * Fixes a PNG file written with libpng-1.6.0 or 1.6.1 that has one or more
+ * uncompressed iTXt chunks.  Assumes that the actual length is greater
+ * than or equal to the value in the length byte, and that the CRC is
+ * correct for the actual length.  This program hunts for the CRC and
+ * adjusts the length byte accordingly.  It is not an error to process a
+ * PNG file that has no iTXt chunks or one that has valid iTXt chunks;
+ * such files will simply be copied.
+ *
+ * Requires zlib (for crc32 and Z_NULL); build with
+ *
+ *     gcc -O -o png-fix-itxt png-fix-itxt.c -lz
+ *
+ * If you need to handle iTXt chunks larger than 500000 kbytes you must
+ * rebuild png-fix-itxt with a larger values of MAX_LENGTH (or a smaller value
+ * if you know you will never encounter such huge iTXt chunks).
+ */
+
+#include <stdio.h>
+#include <zlib.h>
+
+#define MAX_LENGTH 500000
+
+#define GETBREAK ((unsigned char)(inchar=getchar())); if (inchar == EOF) break
+
+int
+main(void)
+{
+   unsigned int i;
+   unsigned char buf[MAX_LENGTH];
+   unsigned long crc;
+   unsigned char c;
+   int inchar;
+
+/* Skip 8-byte signature */
+   for (i=8; i; i--)
+   {
+      c=GETBREAK;
+      putchar(c);
+   }
+
+if (inchar != EOF)
+for (;;)
+ {
+   /* Read the length */
+   unsigned long length; /* must be 32 bits! */
+   c=GETBREAK; buf[0] = c; length  = c; length <<= 8;
+   c=GETBREAK; buf[1] = c; length += c; length <<= 8;
+   c=GETBREAK; buf[2] = c; length += c; length <<= 8;
+   c=GETBREAK; buf[3] = c; length += c;
+
+   /* Read the chunkname */
+   c=GETBREAK; buf[4] = c;
+   c=GETBREAK; buf[5] = c;
+   c=GETBREAK; buf[6] = c;
+   c=GETBREAK; buf[7] = c;
+
+
+   /* The iTXt chunk type expressed as integers is (105, 84, 88, 116) */
+   if (buf[4] == 105 && buf[5] == 84 && buf[6] == 88 && buf[7] == 116)
+   {
+      if (length >= MAX_LENGTH-12)
+         break;  /* To do: handle this more gracefully */
+
+      /* Initialize the CRC */
+      crc = crc32(0, Z_NULL, 0);
+
+      /* Copy the data bytes */
+      for (i=8; i < length + 12; i++)
+      {
+         c=GETBREAK; buf[i] = c;
+      }
+
+      /* Calculate the CRC */
+      crc = crc32(crc, buf+4, (uInt)length+4);
+
+      for (;;)
+      {
+        /* Check the CRC */
+        if (((crc >> 24) & 0xff) == buf[length+8] &&
+            ((crc >> 16) & 0xff) == buf[length+9] &&
+            ((crc >>  8) & 0xff) == buf[length+10] &&
+            ((crc      ) & 0xff) == buf[length+11])
+           break;
+
+        length++;
+
+        if (length >= MAX_LENGTH-12)
+           break;
+
+        c=GETBREAK;
+        buf[length+11]=c;
+
+        /* Update the CRC */
+        crc = crc32(crc, buf+7+length, 1);
+      }
+
+      /* Update length bytes */
+      buf[0] = (unsigned char)((length << 24) & 0xff);
+      buf[1] = (unsigned char)((length << 16) & 0xff);
+      buf[2] = (unsigned char)((length <<  8) & 0xff);
+      buf[3] = (unsigned char)((length      ) & 0xff);
+
+      /* Write the fixed iTXt chunk (length, name, data, crc) */
+      for (i=0; i<length+12; i++)
+         putchar(buf[i]);
+   }
+
+   else
+   {
+      /* Copy bytes that were already read (length and chunk name) */
+      for (i=0; i<8; i++)
+         putchar(buf[i]);
+
+      /* Copy data bytes and CRC */
+      for (i=8; i< length+12; i++)
+      {
+         c=GETBREAK;
+         putchar(c);
+      }
+
+      if (inchar == EOF)
+      {
+         break;
+      }
+
+   /* The IEND chunk type expressed as integers is (73, 69, 78, 68) */
+      if (buf[4] == 73 && buf[5] == 69 && buf[6] == 78 && buf[7] == 68)
+         break;
+   }
+
+   if (inchar == EOF)
+      break;
+
+   if (buf[4] == 73 && buf[5] == 69 && buf[6] == 78 && buf[7] == 68)
+     break;
+ }
+
+ return 0;
+}

contrib/tools/sRGB.h

@@ -0,0 +1,48 @@
+/*-
+ * sRGB.h
+ *
+ * Last changed in libpng 1.6.0 [February 14, 2013]
+ *
+ * COPYRIGHT: Written by John Cunningham Bowler, 2013.
+ * To the extent possible under law, the author has waived all copyright and
+ * related or neighboring rights to this work.  This work is published from:
+ * United States.
+ *
+ * Utility file; not actually a header, this contains definitions of sRGB
+ * calculation functions for inclusion in those test programs that need them.
+ *
+ * All routines take and return a floating point value in the range
+ * 0 to 1.0, doing a calculation according to the sRGB specification
+ * (in fact the source of the numbers is the wikipedia article at
+ * http://en.wikipedia.org/wiki/SRGB).
+ */
+static double
+sRGB_from_linear(double l)
+{
+   if (l <= 0.0031308)
+      l *= 12.92;
+
+   else
+      l = 1.055 * pow(l, 1/2.4) - 0.055;
+
+   return l;
+}
+
+static double
+linear_from_sRGB(double s)
+{
+   if (s <= 0.04045)
+      return s / 12.92;
+
+   else
+      return pow((s+0.055)/1.055, 2.4);
+}
+
+static double
+YfromRGB(double r, double g, double b)
+{
+   /* Use the sRGB (rounded) coefficients for Rlinear, Glinear, Blinear to get
+    * the CIE Y value (also linear).
+    */
+   return 0.2126 * r + 0.7152 * g + 0.0722 * b;
+}

contrib/tools/scale.c

@@ -0,0 +1,238 @@
+/* Given a target range and a source range work out an expression to scale from
+ * the source to the target of the form:
+ *
+ *    (number * mult + add)>>16
+ *
+ * The command arguments are:
+ *
+ *    scale target source
+ *
+ * and the program works out a pair of numbers, mult and add, that evaluate:
+ *
+ *          number * target
+ *   round( --------------- )
+ *              source
+ *
+ * exactly for number in the range 0..source
+ */
+#define _ISOC99_SOURCE 1
+
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+#include <math.h>
+
+static double minerr;
+static unsigned long minmult, minadd, minshift;
+static long mindelta;
+
+static int
+test(unsigned long target, unsigned long source, unsigned long mult,
+   long add, unsigned long shift, long delta)
+{
+   unsigned long i;
+   double maxerr = 0;
+   double rs = (double)target/source;
+
+   for (i=0; i<=source; ++i)
+   {
+      unsigned long t = i*mult+add;
+      double err = fabs((t >> shift) - i*rs);
+
+      if (err > minerr)
+         return 0;
+
+      if (err > maxerr)
+         maxerr = err;
+   }
+
+   if (maxerr < minerr)
+   {
+      minerr = maxerr;
+      minmult = mult;
+      minadd = add;
+      minshift = shift;
+      mindelta = delta;
+   }
+
+   return maxerr < .5;
+}
+
+static int
+dotest(unsigned long target, unsigned long source, unsigned long mult,
+   long add, unsigned long shift, long delta, int print)
+{
+   if (test(target, source, mult, add, shift, delta))
+   {
+      if (print & 4)
+         printf("      {%11lu,%6ld /* >>%lu */ }, /* %lu/%lu */\n",
+            mult, add, shift, target, source);
+
+      else if (print & 2)
+         printf("      {%11lu,%6ld,%3lu }, /* %lu/%lu */\n",
+            mult, add, shift, target, source);
+
+      else if (print)
+         printf("number * %lu/%lu = (number * %lu + %ld) >> %lu [delta %ld]\n",
+            target, source, mult, add, shift, delta);
+
+      return 1;
+   }
+
+   return 0;
+}
+
+static int
+find(unsigned long target, unsigned long source, int print, int fixshift)
+{
+   unsigned long shift = 0;
+   unsigned long shiftlim = 0;
+
+   /* In the final math the sum is at most (source*mult+add) >> shift, so:
+    *
+    *    source*mult+add < 1<<32
+    *    mult < (1<<32)/source
+    *
+    * but:
+    *
+    *    mult = (target<<shift)/source
+    *
+    * so:
+    *
+    *    (target<<shift) < (1<<32)
+    */
+   if (fixshift < 0)
+      while ((target<<shiftlim) < 0x80000000U) ++shiftlim;
+
+   else
+      shift = shiftlim = (unsigned long)fixshift;
+
+   minerr = 1E8;
+
+   for (; shift<=shiftlim; ++shift)
+   {
+      unsigned long mult = ((target<<shift) + (source>>1)) / source;
+      long delta;
+      long limit = 1; /* seems to be sufficient */
+      long add, start, end;
+
+      end = 1<<shift;
+      start = -end;
+
+      for (add=start; add<=end; ++add)
+         if (dotest(target,source,mult,add,shift,0,print))
+            return 1;
+
+      for (delta=1; delta<=limit; ++delta)
+      {
+#        if 0
+            fprintf(stderr, "%lu/%lu: shift %lu, delta %lu\n", target, source,
+               shift, delta);
+#        endif
+
+         for (add=start; add<=end; ++add)
+         {
+            if (dotest(target, source, mult-delta, add, shift, -delta, print))
+               return 1;
+
+            if (dotest(target, source, mult+delta, add, shift, delta, print))
+               return 1;
+         }
+      }
+   }
+
+   if (print & 4)
+      printf("      {%11lu,%6ld /* >>%lu */ }, /* %lu/%lu ERROR: .5+%g*/\n",
+         minmult, minadd, minshift, target, source, minerr-.5);
+
+   else if (print & 2)
+      printf("      {%11lu,%6ld,%3lu }, /* %lu/%lu ERROR: .5+%g*/\n",
+         minmult, minadd, minshift, target, source, minerr-.5);
+
+   else if (print)
+      printf(
+         "number * %lu/%lu ~= (number * %lu + %ld) >> %lu +/-.5+%g [delta %ld]\n",
+         target, source, minmult, minadd, minshift, minerr-.5, mindelta);
+
+   return 0;
+}
+
+static void
+usage(const char *prog)
+{
+   fprintf(stderr,
+      "usage: %s {--denominator|--maxshift|--code} target {source}\n"
+      " For each 'source' prints 'mult' and 'add' such that:\n\n"
+      "   (number * mult + add) >> 16 = round(number*target/source)\n\n"
+      " for all integer values of number in the range 0..source.\n\n"
+      " --denominator: swap target and source (specify a single source first\n"
+      "                and follow with multiple targets.)\n"
+      "    --maxshift: find the lowest shift value that works for all the\n"
+      "                repeated 'source' values\n"
+      "        --code: output C code for array/structure initialization\n",
+      prog);
+   exit(1);
+}
+
+int
+main(int argc, const char **argv)
+{
+   int i, err = 0, maxshift = 0, firstsrc = 1, code = 0, denominator = 0;
+   unsigned long target, shift = 0;
+
+   while (argc > 1)
+   {
+      if (strcmp(argv[firstsrc], "--maxshift") == 0)
+      {
+         maxshift = 1;
+         ++firstsrc;
+      }
+
+      else if (strcmp(argv[firstsrc], "--code") == 0)
+      {
+         code = 1;
+         ++firstsrc;
+      }
+
+      else if (strcmp(argv[firstsrc], "--denominator") == 0)
+      {
+         denominator = 1;
+         ++firstsrc;
+      }
+
+      else
+         break;
+   }
+
+
+   if (argc < 2+firstsrc)
+      usage(argv[0]);
+
+   target = strtoul(argv[firstsrc++], 0, 0);
+   if (target == 0) usage(argv[0]);
+
+   for (i=firstsrc; i<argc; ++i)
+   {
+      unsigned long source = strtoul(argv[i], 0, 0);
+
+      if (source == 0) usage(argv[0]);
+
+      if (!find(denominator ? source : target, denominator ? target : source,
+         maxshift ? 0 : 1+code, -1))
+         err = 1;
+
+      if (minshift > shift) shift = minshift;
+   }
+
+   if (maxshift) for (i=firstsrc; i<argc; ++i)
+   {
+      unsigned long source = strtoul(argv[i], 0, 0);
+
+      if (!find(denominator ? source : target, denominator ? target : source,
+         code ? 4 : 1, shift))
+         err = 1;
+   }
+
+   /* Just an exit code - the printout above lists the problem */
+   return err;
+}

contrib/visupng/PngFile.c

@@ -1,17 +1,19 @@
-//-------------------------------------
-//  PNGFILE.C -- Image File Functions
-//-------------------------------------
-
-// Copyright 2000, Willem van Schaik.
-//
-// This code is released under the libpng license.
-// For conditions of distribution and use, see the disclaimer
-// and license in png.h
+/*-------------------------------------
+ *  PNGFILE.C -- Image File Functions
+ *-------------------------------------
+ *
+ * Copyright 2000, Willem van Schaik.
+ *
+ * This code is released under the libpng license.
+ * For conditions of distribution and use, see the disclaimer
+ * and license in png.h
+ */
 
 #include <windows.h>
 #include <commdlg.h>
 #include <stdio.h>
 #include <stdlib.h>
+#include <zlib.h>
 
 #include "png.h"
 #include "pngfile.h"
@@ -28,7 +30,7 @@ static png_structp png_ptr = NULL;
 static png_infop info_ptr = NULL;
 
 
-// cexcept interface
+/* cexcept interface */
 
 static void
 png_cexcept_error(png_structp png_ptr, png_const_charp msg)
@@ -43,7 +45,7 @@ png_cexcept_error(png_structp png_ptr, png_const_charp msg)
    }
 }
 
-// Windows open-file functions
+/* Windows open-file functions */
 
 void PngFileInitialize (HWND hwnd)
 {
@@ -57,13 +59,13 @@ void PngFileInitialize (HWND hwnd)
     ofn.lpstrCustomFilter = NULL;
     ofn.nMaxCustFilter    = 0;
     ofn.nFilterIndex      = 0;
-    ofn.lpstrFile         = NULL;          // Set in Open and Close functions
+    ofn.lpstrFile         = NULL;          /* Set in Open and Close functions */
     ofn.nMaxFile          = MAX_PATH;
-    ofn.lpstrFileTitle    = NULL;          // Set in Open and Close functions
+    ofn.lpstrFileTitle    = NULL;          /* Set in Open and Close functions */
     ofn.nMaxFileTitle     = MAX_PATH;
     ofn.lpstrInitialDir   = NULL;
     ofn.lpstrTitle        = NULL;
-    ofn.Flags             = 0;             // Set in Open and Close functions
+    ofn.Flags             = 0;             /* Set in Open and Close functions */
     ofn.nFileOffset       = 0;
     ofn.nFileExtension    = 0;
     ofn.lpstrDefExt       = TEXT ("png");
@@ -92,7 +94,7 @@ BOOL PngFileSaveDlg (HWND hwnd, PTSTR pstrFileName, PTSTR pstrTitleName)
     return GetSaveFileName (&ofn);
 }
 
-// PNG image handler functions
+/* PNG image handler functions */
 
 BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
                    int *piWidth, int *piHeight, int *piChannels, png_color *pBkgColor)
@@ -109,7 +111,7 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
     static png_byte   **ppbRowPointers = NULL;
     int                 i;
 
-    // open the PNG input file
+    /* open the PNG input file */
 
     if (!pstrFileName)
     {
@@ -123,7 +125,7 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
         return FALSE;
     }
 
-    // first check the eight byte PNG signature
+    /* first check the eight byte PNG signature */
 
     fread(pbSig, 1, 8, pfFile);
     if (png_sig_cmp(pbSig, 0, 8))
@@ -132,7 +134,7 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
         return FALSE;
     }
 
-    // create the two png(-info) structures
+    /* create the two png(-info) structures */
 
     png_ptr = png_create_read_struct(PNG_LIBPNG_VER_STRING, NULL,
       (png_error_ptr)png_cexcept_error, (png_error_ptr)NULL);
@@ -152,31 +154,37 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
 
     Try
     {
-        
-        // initialize the png structure
-        
+
+        /* initialize the png structure */
+
 #ifdef PNG_STDIO_SUPPORTED
         png_init_io(png_ptr, pfFile);
 #else
         png_set_read_fn(png_ptr, (png_voidp)pfFile, png_read_data);
 #endif
-        
+
         png_set_sig_bytes(png_ptr, 8);
-        
-        // read all PNG info up to image data
-        
+
+        /* read all PNG info up to image data */
+
         png_read_info(png_ptr, info_ptr);
-        
-        // get width, height, bit-depth and color-type
-        
+
+        /* get width, height, bit-depth and color-type */
+
         png_get_IHDR(png_ptr, info_ptr, piWidth, piHeight, &iBitDepth,
             &iColorType, NULL, NULL, NULL);
-        
-        // expand images of all color-type and bit-depth to 3x8 bit RGB images
-        // let the library process things like alpha, transparency, background
-        
-        if (iBitDepth == 16)
-            png_set_strip_16(png_ptr);
+
+        /* expand images of all color-type and bit-depth to 3x8-bit RGB */
+        /* let the library process alpha, transparency, background, etc. */
+
+#ifdef PNG_READ_16_TO_8_SUPPORTED
+    if (iBitDepth == 16)
+#  ifdef PNG_READ_SCALE_16_TO_8_SUPPORTED
+        png_set_scale_16(png_ptr);
+#  else
+        png_set_strip_16(png_ptr);
+#  endif
+#endif
         if (iColorType == PNG_COLOR_TYPE_PALETTE)
             png_set_expand(png_ptr);
         if (iBitDepth < 8)
@@ -186,8 +194,8 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
         if (iColorType == PNG_COLOR_TYPE_GRAY ||
             iColorType == PNG_COLOR_TYPE_GRAY_ALPHA)
             png_set_gray_to_rgb(png_ptr);
-        
-        // set the background color to draw transparent and alpha images over.
+
+        /* set the background color to draw transparent and alpha images over */
         if (png_get_bKGD(png_ptr, info_ptr, &pBackground))
         {
             png_set_background(png_ptr, pBackground, PNG_BACKGROUND_GAMMA_FILE, 1, 1.0);
@@ -199,30 +207,30 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
         {
             pBkgColor = NULL;
         }
-        
-        // if required set gamma conversion
+
+        /* if required set gamma conversion */
         if (png_get_gAMA(png_ptr, info_ptr, &dGamma))
             png_set_gamma(png_ptr, (double) 2.2, dGamma);
-        
-        // after the transformations have been registered update info_ptr data
-        
+
+        /* after the transformations are registered, update info_ptr data */
+
         png_read_update_info(png_ptr, info_ptr);
-        
-        // get again width, height and the new bit-depth and color-type
-        
+
+        /* get again width, height and the new bit-depth and color-type */
+
         png_get_IHDR(png_ptr, info_ptr, piWidth, piHeight, &iBitDepth,
             &iColorType, NULL, NULL, NULL);
-        
-        
-        // row_bytes is the width x number of channels
-        
+
+
+        /* row_bytes is the width x number of channels */
+
         ulRowBytes = png_get_rowbytes(png_ptr, info_ptr);
         ulChannels = png_get_channels(png_ptr, info_ptr);
-        
+
         *piChannels = ulChannels;
-        
-        // now we can allocate memory to store the image
-        
+
+        /* now we can allocate memory to store the image */
+
         if (pbImageData)
         {
             free (pbImageData);
@@ -234,34 +242,34 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
             png_error(png_ptr, "Visual PNG: out of memory");
         }
         *ppbImageData = pbImageData;
-        
-        // and allocate memory for an array of row-pointers
-        
+
+        /* and allocate memory for an array of row-pointers */
+
         if ((ppbRowPointers = (png_bytepp) malloc((*piHeight)
                             * sizeof(png_bytep))) == NULL)
         {
             png_error(png_ptr, "Visual PNG: out of memory");
         }
-        
-        // set the individual row-pointers to point at the correct offsets
-        
+
+        /* set the individual row-pointers to point at the correct offsets */
+
         for (i = 0; i < (*piHeight); i++)
             ppbRowPointers[i] = pbImageData + i * ulRowBytes;
-        
-        // now we can go ahead and just read the whole image
-        
+
+        /* now we can go ahead and just read the whole image */
+
         png_read_image(png_ptr, ppbRowPointers);
-        
-        // read the additional chunks in the PNG file (not really needed)
-        
+
+        /* read the additional chunks in the PNG file (not really needed) */
+
         png_read_end(png_ptr, NULL);
-        
-        // and we're done
-        
+
+        /* and we're done */
+
         free (ppbRowPointers);
         ppbRowPointers = NULL;
-        
-        // yepp, done
+
+        /* yepp, done */
     }
 
     Catch (msg)
@@ -269,7 +277,7 @@ BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
         png_destroy_read_struct(&png_ptr, &info_ptr, NULL);
 
         *ppbImageData = pbImageData = NULL;
-        
+
         if(ppbRowPointers)
             free (ppbRowPointers);
 
@@ -295,7 +303,7 @@ BOOL PngSaveImage (PTSTR pstrFileName, png_byte *pDiData,
     static png_byte   **ppbRowPointers = NULL;
     int                 i;
 
-    // open the PNG output file
+    /* open the PNG output file */
 
     if (!pstrFileName)
         return FALSE;
@@ -303,7 +311,7 @@ BOOL PngSaveImage (PTSTR pstrFileName, png_byte *pDiData,
     if (!(pfFile = fopen(pstrFileName, "wb")))
         return FALSE;
 
-    // prepare the standard PNG structures
+    /* prepare the standard PNG structures */
 
     png_ptr = png_create_write_struct(PNG_LIBPNG_VER_STRING, NULL,
       (png_error_ptr)png_cexcept_error, (png_error_ptr)NULL);
@@ -322,60 +330,60 @@ BOOL PngSaveImage (PTSTR pstrFileName, png_byte *pDiData,
 
     Try
     {
-        // initialize the png structure
-        
+        /* initialize the png structure */
+
 #ifdef PNG_STDIO_SUPPORTED
         png_init_io(png_ptr, pfFile);
 #else
         png_set_write_fn(png_ptr, (png_voidp)pfFile, png_write_data, png_flush);
 #endif
-        
-        // we're going to write a very simple 3x8 bit RGB image
-        
+
+        /* we're going to write a very simple 3x8-bit RGB image */
+
         png_set_IHDR(png_ptr, info_ptr, iWidth, iHeight, ciBitDepth,
             PNG_COLOR_TYPE_RGB, PNG_INTERLACE_NONE, PNG_COMPRESSION_TYPE_BASE,
             PNG_FILTER_TYPE_BASE);
-        
-        // write the file header information
-        
+
+        /* write the file header information */
+
         png_write_info(png_ptr, info_ptr);
-        
-        // swap the BGR pixels in the DiData structure to RGB
-        
+
+        /* swap the BGR pixels in the DiData structure to RGB */
+
         png_set_bgr(png_ptr);
-        
-        // row_bytes is the width x number of channels
-        
+
+        /* row_bytes is the width x number of channels */
+
         ulRowBytes = iWidth * ciChannels;
-        
-        // we can allocate memory for an array of row-pointers
-        
+
+        /* we can allocate memory for an array of row-pointers */
+
         if ((ppbRowPointers = (png_bytepp) malloc(iHeight * sizeof(png_bytep))) == NULL)
             Throw "Visualpng: Out of memory";
-        
-        // set the individual row-pointers to point at the correct offsets
-        
+
+        /* set the individual row-pointers to point at the correct offsets */
+
         for (i = 0; i < iHeight; i++)
             ppbRowPointers[i] = pDiData + i * (((ulRowBytes + 3) >> 2) << 2);
-        
-        // write out the entire image data in one call
-        
+
+        /* write out the entire image data in one call */
+
         png_write_image (png_ptr, ppbRowPointers);
-        
-        // write the additional chunks to the PNG file (not really needed)
-        
+
+        /* write the additional chunks to the PNG file (not really needed) */
+
         png_write_end(png_ptr, info_ptr);
-        
-        // and we're done
-        
+
+        /* and we're done */
+
         free (ppbRowPointers);
         ppbRowPointers = NULL;
-        
-        // clean up after the write, and free any memory allocated
-        
+
+        /* clean up after the write, and free any memory allocated */
+
         png_destroy_write_struct(&png_ptr, (png_infopp) NULL);
-        
-        // yepp, done
+
+        /* yepp, done */
     }
 
     Catch (msg)
@@ -389,9 +397,9 @@ BOOL PngSaveImage (PTSTR pstrFileName, png_byte *pDiData,
 
         return FALSE;
     }
-    
+
     fclose (pfFile);
-    
+
     return TRUE;
 }
 
@@ -437,6 +445,7 @@ png_flush(png_structp png_ptr)
 
 #endif
 
-//-----------------
-//  end of source
-//-----------------
+/*-----------------
+ *  end of source
+ *-----------------
+ */

contrib/visupng/PngFile.h

@@ -1,12 +1,12 @@
-//------------------------------------------
-//  PNGFILE.H -- Header File for pngfile.c
-//------------------------------------------
+/*------------------------------------------*/
+/*  PNGFILE.H -- Header File for pngfile.c*/
+/*------------------------------------------*/
 
-// Copyright 2000, Willem van Schaik.
+/* Copyright 2000, Willem van Schaik.*/
 
-// This code is released under the libpng license.
-// For conditions of distribution and use, see the disclaimer
-// and license in png.h
+/* This code is released under the libpng license.*/
+/* For conditions of distribution and use, see the disclaimer*/
+/* and license in png.h*/
 
 #include <stdio.h>
 #include <stdlib.h>
@@ -17,7 +17,7 @@ void PngFileInitialize (HWND hwnd) ;
 BOOL PngFileOpenDlg (HWND hwnd, PTSTR pstrFileName, PTSTR pstrTitleName) ;
 BOOL PngFileSaveDlg (HWND hwnd, PTSTR pstrFileName, PTSTR pstrTitleName) ;
 
-BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData, 
+BOOL PngLoadImage (PTSTR pstrFileName, png_byte **ppbImageData,
                    int *piWidth, int *piHeight, int *piChannels, png_color *pBkgColor);
 BOOL PngSaveImage (PTSTR pstrFileName, png_byte *pDiData,
                    int iWidth, int iHeight, png_color BkgColor);

contrib/visupng/README.txt

@@ -8,8 +8,8 @@ For conditions of distribution and use, see the disclaimer
 and license in png.h
 
 As a PNG .dll demo VisualPng is finished. More features would only hinder
-the program's objective. However, further extensions (like support for other 
-graphics formats) are in development. To get these, or for pre-compiled 
+the program's objective. However, further extensions (like support for other
+graphics formats) are in development. To get these, or for pre-compiled
 binaries, go to "http://www.schaik.com/png/visualpng.html".
 
 ------------------------------------------------------------------------